1983 CADICAM INDUSTRY CONFERENCE

1983 CADICAM INDUSTRY CONFERENCE
1983 CADICAM
INDUSTRY CONFERENCE
The Technology Challenge:
Linking CADfCAM with
Integrated Manufacturing
September 2 6 - 2 8
The Newporter
Newport Beach, California
The Technology Challenge:
\
Linking CADICAM with
Integrated Manufacturing
\
\
\
\
•
\
\
\
\
\
SUNDAY, SEPTEMBER 25
\
\
%
\
\
•^
"ife
\
'\.
\ .
%
\
^ .
\
\\
\
\
%
\
\
\
\
8:00 p.m.
Cocktails
\
\
•%.
\
\
.
Registration
\ :
\
^t
5:00 p.m. to
8:00 p.m.
\
-%
V
\
\
\ %
^l'.
\ \
\
%
\
^ \
\
\
\
MONDAY, SEPTEMBER 26
-V„
\
^
%
\ %
\
\
8:00 a.m.
^^.
%
\ \
\
m.
•
\
8:00 a.m.
\ %.
\
Buffet Breakfast
\
%.
\
9:00 a.m.
\
. • • • • • "
Registration Continues
.^iS^^^^"
WeIcome
t P S S I O N 1—INDUSTRY OVERVIEW
SJ'
\
•H.
\
^
\
\
9:15 a.m.
.•0-
^
>-.(
•••:.•
9:45 a.m.
10:15 a.m.
. ^ • • ' '
^
11:00 a.m.
Manufacturing Is Not a Button on the
Design Data Base
Michael Sterling
Vice President Research & Development
CADLINC
11:45 a.m.
The Role of Automated Test Equipment in
the Automated Factory
Harold McAleer
Senior Vice President and General Manager
GenRad Incorporated
12:30 p.m.
Luncheon
-tB-. ^
.
^
SESSION II—PRODUCTION AND CAD
Introduction
^.gt:jji-'
•
Coffee Break
10:45 a.m.
.•Ji-
:*s*'"
CAD Market Review and Forecast
Belh W. Tucker, Research Analyst
CAD/CAM industry Service
Data que si Incorporated
' : • : • • •
rf^-
CAD/CAM Industry Overview
^"'
Timothy 0. Gauhan, Vice President and OirectCKT-',^iSP^
CAD/CAM Industry Service
Oataquest Incorporated
^
SESSION IV—ELECTRONICS ISSUES
2:00 p.m.
1:30 p.m.
Trends i n Distributed Worlcstations in C A D
Impact of Computer-Integrated
Manufacturing on a Company
Joseph F. Gloudeman, President
Maclsieal-Schwendler Corporation
2:45 p.m.
Wilfred J . Corrigan, President
LSI Logic Corporation
R o t m t i c s — T o d a y a n d Tomorrow
2:15 p.m.
Steve VoIm, Western Region Manager
Industrial Automation
IBM Corporation
6:00 p.m.
7:00 p.m.
:08:30 p.m.
Harvey Jones, Vice President, Marketing
Daisy Systems Corporation
3:00 p . m .
Cocktails
Dinner
,„,.
ProductIvIty a n d E d u c a t i d h
Dr. RobeII Noyce, Vice Chairman
Intel Corporation
M a k i n g Engineers More Productive
,P^
^jy¥0^'
'•"•'•
#
#
The Role of a Modern C u s t o m S u p p l i e r in
j u t F s t h e Field of VLSI Design
Graham Shenton
Vice President of Marketing and Engineering
international Microeiectronics Products Inc.
4:30 p.m.
Regatta
6:00 p.m.
Cocktails
7:00 p.m.
Dinner
TUESDAY, S ^ ^ E M B E R 27
,
ll--k]XI'
r^iX^^'
•'
0"7\30Mm
Buffet Breal<fast
a-if-
SESSION III—PRODUCT STRATEGIES
8:30 a.m.
CADICAIVI W o r k s t a t i o n Trends
WEDNESDAY, SEPTEMBER 28
8:00 a.m.
Edward L. Busick, President
Spectragraphics Corporation
9:15 a.m.
Computer-Integrated Manufacturing —
T e c h n o l o g i c a l Imperative
Philips W. Smith
Corporate Vice President and General Manager
Marketing Division
Computervision Corporation
10:00 a.m.
C o f f e e Breai<
10:30 a.m.
C o m p u t e r C o m p a n i e s and I n t e g r a t i o n
John W. Poduska, President
Apollo Computer
11:15 a.m.
12:00 noon
SESSION V—REALITY OF THE
FACTORY FUOOR
9:00 a.m.
Introduction
9:15 a.m.
C A M o n t h e M a n u f a c t u r i n g Floor:
Products There Today
Ai Llbbey
Vice President, New Business Development
Modicon Programmable Control Division
Gould Incorporated
10:00 a.m.
Integrated C A D I C A M : A n A c h i e v a b l e Goal
Richard L. Justice, Director
Engineering and Manufacturing
Computer Coordination
General Motors Corporation
The Impact of Engineering W o r k s t a t i o n
Networks on CAD Productivity
David E. Weisberg, Manager, Market Analysis
Auto-troi Corporation
Buffet Breakfast
10:45 a.m.
C o f f e e Break
11:00 a.m.
Role of Start-up C o m p a n i e s in
Computer-Integrated M a n u f a c t u r i n g
Luncheon
Laura Conigiiaro, Vice President
Prudentiai-Bache Securities
12:00 n o o n
Luncheon
CONFERENCE INFORMATION
WHO SHOULD ATTEND
FOOD AND LODGING
The conference is aimed at CAD/CAM industry
executives, market research and product planning
managers, and financial analysts/portfolio managers.
PLEASE NOTE THAT ALL HOTEL RESERVATIONS MUST
BE MADE THROUGH DATAQUEST.
Room rates at The Newporter for the nights of
September 26 and 27 are $135.00 single occupancy and
$207.00 double occupancy (per room, per night). Rates
include all meals taken with the group, beginning with
breakfast on Monday and ending with lunch on
Wednesday. For those arriving on Sunday, September 25,
the rate will be $75.00 single occupancy and $87.00
double occupancy. In addition, charges of 21 percent on
food and beverage (gratuity and tax) and 8 percent tax on
rooms will be added to your bill.
LOCATION
The site of DATAQUEST's CAD/CAM Industry
Conference will be The Newporter, Newport Beach,
California.
Address:
The Newporter
1107 Jamboree Road
Newport Beach, California 92660
Telephone: (714)644-1700
CONFERENCE FEE
TRANSPORTATION
The fee for the CAD/CAM Industry Service (CCIS)
conference is:
• Subscriber (or designee) from CCIS
client company
No Charge
• Each additional attendee from CCIS
client company
$550
• Each attendee who is a client of any
other DATAQUEST Industry Service
$650
• Each attendee from a nonclient
company
$795
The Newporter is 10 minutes from the Orange County
Airport and approximately 60 minutes from Los Angeles
International Airport. Complimentary limousine service is
available from the Orange County Airport to and from
The Newporter, and coach service is available from Los
Angeles International. No reservations are necessary
from Orange County Airport. Simply call upon arrival
from the direct telephone in the baggage area of the
airport. Taxis and rental cars are also available.
REGISTRATION
REGISTRATION DEADLINE IS FRIDAY,
SEPTEIVIBER 16, 1983.
To register, please complete and return the registration
form. We cannot register you until we have your arrival
and departure dates. You will receive a confirmation
packet with additional information shortly before the
conference begins.
CANCELLATION POLICY
All cancellations received by DATAQUEST after
September 16, 1983, are subject to a $100 service charge
unless the registrant sends a replacement. Registrants
who do not cancel and who fail to attend the conference
will automatically be assessed a $100 service charge.
Notice of cancellation should be made to Gail van Tubergen,
Conference Coordinator, telephone: (408) 971-9000.
RECREATION
For tennis buffs, California's number-one-rated tennis
club, the John Wayne Tennis Club, is located on the
property. This excellent facility has sixteen lighted tennis
courts, a Jacuzzi, and a sauna. There are swimming pools
and a 9-hole golf course. The Irvine Coast Country Club
is also available for golf. The beach Is only minutes
away, where you may sunbathe, swim, or surf.
[IK^Tga Dataquest
A Subsidiary of
A.C. Nielsen Company
DATAQUESTINCORPORATED
1290 Ridder Park Drive
San Jose, California 95131
Phone:(408)971-9000
Teiex 171973
kiduilry S t n k i
Dat*
i.«utlafl
Hoial rnt«Tcont|fMI^HJ
QtfittB^ SuFlZflrUVi
D Cdfrvlng tfid Cuplhcifting
January 27-28
D Sm^I CcimjuiHr
February 16-18
9<iv>f4do OitiPtiv C1U&
NaiML Calirdmift
D ElBdlranLc Pnr7l«r
March 9-11
^iiLWVdg Cvnv\fTt CJUB
NaiH, CalllAfflii
D TalflGDmFnijnltPl^i<>n4i
April 27-29
TriB knAisi1r44 Inn
t^c^nttrvi. CalLrofJIi*
Q EurvpHH SaTilcDrrilucJv
May 18-20
Hcriei de Pu^a
MofilD -Corlit, hgnvco
D
May 24
TiAffi Amttitin dvl}
b^iDOKhtclar
To*if*. J H « n
a
CepvFi^ and ]lupri[7«li^
June 8-1O
RincFio ^riMFdo I r i
S v i Difigci, CAlltwriFH
n
Elftfitrcntc Piinier
June 16-17
N 4 ^ Htlign Hg|il
CvrevB-, SMitSBf^&Tid
•
CdAjIng i n d Dlip|l[>ai<>riD
June 22
Tpkyi) A r w r t M n Cluh
TDhyD, Japtn
•
OUpHy Twrnlnal
June 22-24
Hyatt t M Mpnct
Mgnt^my, C4nfpmla
D CAD^OAM
September 26-28
TtiS NShMptlrtcr
tC»wpgrt ^Mcli. CnKTC^'a
D SMniomnoijclw
October 17-19
narcho L u ^aUnou
h i m EpnngF, Caii'prnia
D e i H i n n i c Pr^ntar
October
Tok)\» .i^nHn^flr C^Ub
T4)hvD, Jamn
C CQnifitiUr i^wnery
November 2-4
^\^vvtH(^ CdindiY Cliib
Nafd, GallTumki
•
November 16-18
H^BlT Qsl M«<IT«
Uon\armif. t^\mynnt.
M n t ProcBMlng
Dataquest
KESSs
10ft^ DATAOUEST CONFERENCE SCHEDULE
Please check box(es) If you would like to receive
information regarding any of our conferences.
D
1290 Ridder Park Drive
San Jose, California 95131
(408) 971-9000
Telex 171973
We need the following information to
register your reservation.
Please register me for the 1983 CAD/CAM Industry Conference
(please attach business card or print)
_Flrst
Last Name-
_M.I.
Title
My spouse will accompany me. D Yes D No
Name of Spouse.
Company
Please reserve:
Street
.State
City
• Zip
Telephone (
Arrival Date.
Departure Date .
For invoicing purposes, please check the appropriate box. (If you are attending in the place of a
binderholder, please fill In their name or you may be Invoiced for a conference fee.)
n
GCIS Client
•
^^ir^
_!.]...•
.
..
J
Binderholder's name
CCIS additional attendee
n Client —Other DATAQUEST service
D Client —Prudential Bache
D Client-FSP
n Nonclient company attendee
D I cannot attend, but I would like more information on DATAQUEST's CAD/CAM Industry Service.
n
n
single room
double room
_Time_
Time.
Tl
NO POSTAGE
NECESSARY IF
rWAILED IN THE
UNITED STATES
Business Reply Mail
FIRST CLASS
PERMIT NO. 7279
POSTAGE WILL BE PAID BY ADDRESSEE:
Attention: Gail van Tubergen
Dataquest Incorporated
1290 Bidder Park Drive
San Jose, CA 95131
SAN JOSE
CAD/CAM Industry Service
Conference
September 26-28, 1983
The Newporter
Newport Beach, California
Li-^. )«Ka Dataquest
1290 Ridder Park Drive
San Jose, California 95131
(408) 971-9000
Telex: 171973
Sales/Service offices:
UNITED KINGDOM
DATAQUEST UK Limited
144/146 New Bond Street
London WIY 9FD
United Kingdom
(1) 409-1427
Telex: 266195
GERMANY
DATAQUEST GmbH
In der Schneithohl 17
D-6242 Kronberg 2
West Germany
(061) 736921
Telex: 410939
FRANCE
DATAQUEST SARL
41 rue Ybry
92522 Neuilly Sur Seine Cedex
France
(1) 758-1240
Telex: 630842
JAPAN
DATAQUEST Japan, Ltd.
Azabu Heights, Suite 410
1-5-10, Roppongi, Minato-ku
Tokyo 106, Japan
(3) 582-1441
Telex: J32768
The content of this report represents our interpretation and analysis of information generally available to the public or released by responsible individuals in
the subject companies, but is not guaranteed as to accuracy or completeness. It
does not contain material provided to us in confidence by our clients.
This information is not furnished in connection with a sale or offer to sell
securities, or in connection with the solicitation of an offer to buy securities. This
firm and its parent and/or their officers, stockholders, or members of their families
may, from time to time, have a long or short position in the securities mentioned
and may sell or buy such securities.
Printed in the United States of America. All rights reserved. No part of this
publication may be reproduced, stored in retrieval systems, or transmitted, in any
form or by any means—mechanical, electronic photocopying, duplicating,
microfilming, videotape, or otherwise—without the prior written permission of
the publisher.
Copyright © 1983 by Dataquest Incorporated
Dataquest
CAD/CAM Industry Conference
September 26-28, 1983
Newport Beach, California
Final List of Attendees
Adage
Len Keene, Manager, Market
Planning
American Microsystems, Inc.
James R. Tobias, Director,
Technical Information
Systems
Analog Design Tools, Inc.
Martin Walker, President
Apollo Computer
Richard Gimbel, Marketing
Manager
Apollo Computer
John Robotham, Engineer
Apple Computer
Matt Slavik, Program Manager
Applicon, Inc.
Russell Doty, Manager, Market
Analysis
Arthur Young & Coit?>any
Michael Dulion
Auto-Trol Technology Corporation
David P. Hanna, Director of
Marketing
Robert Stevenson, Director of
Product Marketing
David Weisberg, Manager,
Market Analysis
BA Investment Management Corporation
Rupert Grimm, Vice President
BULL SEMS
M. Jean Paul Carcel
Francois DeBelenet
Michel N. Sauvage, Manager,
Marketing and Plans
Bank of America
Indre M. Bauza, Assistant
Vice President
Bank of America
Donald Cvietusa, Vice
President - Group Head
- 1 Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
Bank of Boston
Mark MacLennan, Vice President
Daphne Strong, Banking
Associate
Jeffrey Wellington, Loan
Officer
Peter White, Loan Officer
Bank of California
Bob Bourke, Vice President
Bell Laboratories
J. E. Iwersen, Director
Joe Sitarik, Supervisor
Burroughs Corporation
Morrie Binder, Regional
Marketing Specialist
Ron Budacz, Senior Marketing
Manager, Int'l Group
Stan Eaton, Director of LOB
Manufactur ing/D istr ibution
Robert D. Merrell, Vice
President, Industrial
Products
David Rumsey, Director, Coml.
LOB Marketing Planning
CADAM Inc.
Howard Wilczynski, Member of
Technical Staff
Rhonda Lindsey, Product
Planning Administrator
CADLINC
Michael Sterling, Vice
President, Research &
Development
John West, President
CADTRAK Corporation
Jim Callan, Director of
Marketing
CAE Systems, Inc.
Keonran Elahian,
Director/Founder
Cad/Cam Specialists, Inc.
Daniel Garms, Vice President
Michael McDonald, President
CadimaticHi, Inc.
Chad Alber, Senior Staff
Scientist
Cadtec Corporation
Steve Schopbach, Vice
President, Marketing & Sales
- 2Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
California Automated Design, Inc.
Randy Smith, Vice President
Kayo Smith
California Computer Products, Inc.
Richard A. Peters, Vice
President
Calma Company
John Allum-Poon, Manager of
Customer Education
Robert Kuhling, Director,
Domestic Business
Development
Suki Nakamoto, Market
Research Analyst
Donald C. Stewart, Vice
President, Corporate
Development
Thomas R. Teeters, Director,
Marketing Information
Planning
Cascade Graphics Development
Steve Ball, Executive Vice
President
Cericore, Inc.
Ed Allred, Vice President
David Bailey, President
Conputervision Corporation
Henry Kellogg, Manager,
Competitive Analysis
Susan B. Ray, Market Research
Philips W. Smith, Corporate
Vice President & General
Manager
Consultant
Frank Gianattasio
Continental Illinois Bank
Leon McMillan, Senior
Investment Analyst, Equity
Research
Control Data Corporation
Michael E. Sokolski, Manager,
Engineering
Daisy Systems Corporation
Harvey Jones, Vice President,
Marketing
Lucio Lanza, Customer
Marketing Manager
Data General
Sherman L. Rutherford,
Director, Industrial
Automation
Dave Rome, Marketing Manager
Of CAE/CADCAM
Corporation
- 3 Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
Dataquest Incorporated
Steve Cottrell, Manager,
Consulting Activities
0. Ralph Finley, Senior Vice
President
Frank Florence, Marketing
Manager
Timothy Gauhan, Vice
President & Director
Howard Hagen, Director, Small
Computer Industry Svc.
Sally Hasson, Conference
Assistant
John Jackson, Marketing
Manager
David Jorgensen, Chairman of
the Board
Wendy Ledamun, Research
Associate
Cathy Marvin, Conference
Assistant
James R. Newcomb, Associate
Director
Laura Nichols, Industry
Analyst
Gayle Phillips, Conference
Assistant
Donna Plath, Graphics
Supervisor
John Randall, Marketing
Manager
James Riley, Senior Vice
President
Frank Sammann, Vice
President, Marketing
Robert Speicher, Marketing
Manager
Beth W. Tucker, Research
Analyst
Gail Van Tubergen, Conference
Coordinator
Ed Washington, Senior
Industry Analyst
Digital Equipment Corporation
Bob Barnes, Engineering Mgr.
Storage Technical Ops
Wes Brovm, Technical
Operations Manager
Richard Cook, Engineering
Services Manager
4 Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
Digital Equipment Corporation
Paul Kruger, Market
Development Specialist
Donald McGinnis, Senior
Corporate Planner
Bill Seaver, Manager,
Manufacturing Engineering
Sri Sriram, CAEM Marketing
Specialist
Irv Winters, Manager, Plant
Manufacturing Engineering
Douglas Wood, CAEM Electronic
Marketing Group Manager
Art Worsh, Manager,
Manufacturing MIS
Dreyfus Corporation
Tom McManus, Securities
Analyst
EDS Nuclear/Impell
Chuck Rosselle, Manager
Joan Rosselle
Evans & Sutherland Computer Corporation
Steve Artick, Marketing
Research Analyst
Diane Artick
Ginny Banerjee, Project
Manager
Fairchild Test Systems
David Birnbaum, Product
Marketing Engineer
First Pennsylvania Bank
John Maxwell, III, Securities
Analyst
Fujitsu America, Inc.
Satoru Hayashi, Senior
Engineer
Katsuhide Hirai
Fujitsu Microelectronics, Inc.
Jim Coe, Vice President,
Custom Product Operation
GenRad, Incorporated
Harold McAleer, Senior Vice
President
General Dynamics Corporation
John G. Heit, Senior
Engineering Specialist
General Electric Company
Anthony Corsi, Program
Manager, Corp. Info. Systems
-
5 -
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
General Motors C o r p o r a t i o n
Richard L. .Justice, Director
Engineering & Manufacturing
Computer Coordination
Genisco Computers C o r p o r a t i o n
Daniel R. Jones, Director of
Product Marketing
Gould D i s t r i b u t e d Systems
Kenneth Brody, Product Line
Manager CAE
Gould I n c o r p o r a t e d
Al Libbey, Vice President
Hewlett-Packard Conpany
Pat Castro, Lab Director
Calvin Cobb, Marketing
Evaluator
Tom Newsom
Mihir Parikh, Department
Manager
Hill Partnership
Paul Kirby, General Manger
Honeywell, I n c .
David P. Peters, Manager, CAE
Product Marketing
Lillian Peters
Hughes A i r c r a f t Coirpany
Frank Goebels, Portfolio
Manager
Kevin Shambrook, Technical
Director, Automation
Lawrence Schmitz, Manager,
MOS Design
IBM Corporation
David Durish, Advisory
Engineer
Richard Ma, Senior Marketing
Analyst
Bob Olmsted, CAD/CAM Market
Support
William Quigley, Program
Manager
Steve Volm, Western Regional
Manager
Judith Volm
Fred Wagner, Consultant,
General Engineering
IDS Growth Spectrum Advisors
Edward Hartmann, President
Intel Corporation
Robert Noyce, Vice Chairman
- 6Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
InterWest Partners
Scott Hedrick, General Partner
International Microelectronics Products
Bob Gardner, Manager,
Marketing
Monique Gardner
David Lucas, Manager CAD
Graham Shenton, Vice
President, Marketing
IntraWest Bank of Denver
Leo Beserra, Vice President &
Trust Officer
Jennison Associates
David T. Poiesz, Vice
President
Kemper Financial Services, Inc.
Bruce Ebel, Research Analyst
LSI Logic Corporation
Wilfred Corrigan, President
Richard Perry, Marketing
Manager
MacNeal-Schwendler Corporation
Joseph Gloudeman, President
Jan Gloudeman
Mayfield Fund
A. Grant Heidrich, Associate
Glenn Mueller, General Partner
McDonnell Douglas Automation Company
Brush Bradley, Consultant,
Product Marketing
Bob Daugherty, Principal
Specialist
Megatek Corporation
Susan Harvey, Planning Manager
Nancy Savinelli, Director of
Product Marketing
Mentor Graphics Corporation
Mike Bosworth, Director of
Marketing
Metheus Corporation
Chong C. Lee, Vice President
& General Manager, VLSI
Michael Sisavic, Vice
President
Terry Smith, Manager, Product
Marketing
James C. Towne, President
Micro Component Technology
Ward T. Bell, General Manager
- 7 Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
Mostek Corporation
Frank Hope, Graphics
Marketing Manager
NCA Corporation
Barry Goss, Vice President,
Operations
NUTECH Engineers, Inc.
George Groves, General Manager
Gary Kost, Executive
Consultant
Judy Kost
Nissho Iwai American Corporation
Edmond Murphy, Consultant,
Ccanputer Graphics
Nixdorf Computer Corporation
Philip D. Wilson, Product
Line Manager
Dora Freidman
Parallax Systems Inc.
Martin Picco
Polaroid Corporation
Phillip Mestancik, Senior
Administrator, Business
Planning
Prime Computer, Inc
Rod MacDow, Marketing Manager
Bob Sanchez, Manager, CAD/CAM
Relations
Robin Sanchez
Barry Wolman, Director,
Manufacturing Systems
Prudential Insurance Company
Dana O'Brien, Investment
Analyst
Prudential-Bache Securities
Steven Colbert, Vice President
Laura Conigliaro, Vice
President
Carol E. Muratore, Vice
President
Mitch Quain, Vice President
David Stueber, Technology
Strategist
Michael Weisberg, Vice
President
Pyramid Technology Corporation
John Hime, Director Product
Marketing
Quantel Business Computers
Milo F. Miras. Manager
Marketing Technical Support
- 8Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
Qubix Graphic Systems
Bob Harmison, Director of
Marketing
Jim Lucas, Vice P r e s i d e n t ,
Marketing
Qume Corporation
Han Park, Product Marketing
Manager/CRT Terminals
R & D Funding Corp.
Hugh McClung, President
Racal-Redac Ltd.
Richard Barrett, Technical
Director
Ian Orrock, Manager Director
Susan Orrock
Raster Technology I n c .
Stephen Coit, Director of
Marketing
Schlumberger
Thomas Schaefer, Marketing
Manager, Manufacturing
Systems
S e a t t l e F i r s t National Bank
Joseph Suty, Vice President
Seiko Instruments USA Inc.
Robert Coats, Director of
Sales
Tom Hutchins, Marketing
Analyst
Siemens AG
E. Baltin, Diploma Engineer
Silicon Compilers
Ann Doerr, Marketing Sales
S i l i c o n Graphics, I n c .
Vernon R. Anderson, President
James Clark, Chairman & Chief
Technical Officer
Silvar-Lisco
Thomas O. Binder, Vice
President, Marketing & Sales
Skantek Corporation
Roger Paradis, President
Spectragraphias Corporation
Robert Blumberg, Chairman of
the Board
Edward L. Busick, President
Alan Kaechele, Applications
Engineer
John Moreland, Director of
Marketing
- 9Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
Tektronix, Inc.
Norman B. Acker, Ccxnpetitive
Assessment Manager
Cathy Cramer, Product
Marketing Manager, Logic
Design Systems
John Crist, Information
Manager, Design Automation
Division
Robert Martin, Manager,
Engineering
Jerry Sullivan, Director,
Computer Science Group
Joe Yesenofski, Program
Manager, Graphic Design
System
Teradyne, Inc.
Jim Fogle, Engineer
Texas Instruments, Inc.
Bob Myers, Strategic Planner
The Putnam Mangement Company, Inc.
Richard Jodka, Vice President
Toyo Denki Seize K.K.
Kazuhiko Yoshinaga, Manager,
Software Dept.
Triad Systems Corporation
Steve LaComroare, Vice
President, Engineering &
Manufactur ing
Ed Morgan, Manager,
Manufacturing
Ray Osofsky, Manager,
Engineering
Sheldon Toso, Manager,
Quality Services
Trust Company of the West
James Jeffs, Research
Associate
VLSI Technologies Inc.
Henri Jarrat, President
Douglas G. Fairbairn, Vice
President, User-Designed
Technology
Valid Logic Systems
Frank Lynch, Director of
Marketing
Vectron Graphics
Charles Simon, President
Cathy Simon
Versatec
Edwin Saunders, Manager,
Market Research
- 10 Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Parl< Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
WICAT Systems
Glenn Stewart, Industry
Marketing Manager
Kay Stewart
Weitek Corporation
Timothy Neuland, Western
Regional Mgr, Graphics
Products
Western Asset Management Company
Lisa Hammersly, Research
Analyst
Wisconsin State Investment Board
Jean Ledford, Investment
Officer
Xerox Corporation
Michael Nagel, Product Manager
ZyMOS Corporation
Bob Lorentzen, CAE Tools
Marketing Manager
- 11 Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
Dataquest
CAD/CAM I n d u s t r y Conference
September 26-28, 1983
Newport Beach,
California
F i n a l L i s t of Attendees
Norman B. Acker
Tektronix, Inc.
Chad Alber
Cadimation, Inc.
Ed A l l r e d
Cericore, Inc.
John Allum-Poon
Calma Company
Vernon R. Anderson
Silicon Graphics, Inc.
Steve Artick
Diane Artick
Evans & Sutherland Computer Corporation
David B a i l e y
Cericore, Inc.
Steve B a l l
Cascade Graphics Development
E Baitin
Siemens AG
Ginny Banerjee
Evans & Sutherland Computer C o r p o r a t i o n
Bcto Barnes
D i g i t a l Equipment Corporation
Richard B a r r e t t
Racal-Redac L t d .
Indre M. Bauza
Bank of America
Ward T. B e l l
Micro Component Technology
Leo Beserra
IntraWest Bank of Denver
Morrie Binder
Burroughs Corporation
Thomas O. Binder
Silvar-Lisco
David Birnbaum
Pairchild
Robert Blumberg
S p e c t r a g r a p h i c s Corporation
Mike Bosvrorth
Mentor Graphics Corporation
T e s t Systems
- 1 Dataquest Incorporated, A Subsidiary of A.G. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
Bob Bourke
Bank of California
Brush Bradley
McDonnell Douglas Automation Ccanpany
Kenneth Brody
Gould Distributed Systems
Wes Brown
Digital Equipment Corporation
Ron Budacz
Burroughs Corporation
Edward L. Busick
Spectragraphics Corporation
Jim Callan
CADTRAK Corporation
M. Jean Paul Carcel
BULL SEMS
Pat Castro
Hewlett-Packard Company
James Clark
Silicon Graphics, Inc.
Robert Coats
Seiko I n s t r u m e n t s USA I n c .
Calvin Cobb
Hewlett-Packard Ccmpany
Jim Coe
Fujitsu Microelectronics,
Stephen Coit
Raster Technology I n c .
Steven Colbert
Prudential-Bache
Laura Conigliaro
Prudential-Bache S e c u r i t i e s
Richard Cook
D i g i t a l Equipment C o r p o r a t i o n
Wilfred Corrigan
LSI Logic C o r p o r a t i o n
Anthony Corsi
General E l e c t r i c Company
Steve Cottrell
Dataquest
Cathy Cramer
Tektronix, Inc.
John Crist
Tektronix,
Donald Cvietusa
Bank of America
Bob Daugherty
McDonnell Douglas Automation Cc»npany
Francois DeBelenet
BULL SEMS
Inc.
Securities
Incorporated
Inc.
- 2 -
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
Ann Doerr
Silicon Compilers
Russell Doty
Applicon, Inc.
Michael Dulion
Arthur Young & Company
David Durish
IBM Corporation
Stan Eaton
Burroughs Corporation
Bruce Ebel
Kemper Financial Services, Inc.
Kamran Elahian
CAE Systems, Inc.
Douglas G. Fairbairn
VLSI Technology, I n c .
Phil Fernandez
Tektronix,
O. Ralph Finley
Dataquest Incorporated
Frank Florence
Dataquest Incorporated
Jim Fogle
Teradyne, Inc.
Dora Freidman
Nixdorf Computers
Bob Gardner
Monique Gardner
International Microelectronics Products
Daniel Garms
Cad/Cam Specialists, Inc.
Timothy Gauhan
Dataquest, Incorporated
Frank Gianattasio
Consultant
Richard Gimbel
Apollo Computer
Joseph Gloudeman
Jan Gloudeman
MacNeal-Schwendler Corporation
Frank Goebels
Hughes Aircraft Ccmpany
Barry Goss
NCA Corporation
Rupert Grimm
BA Investment Management Corporation
George Groves
NUTECH Engineers, Inc.
Howard Hagen
Dataquest Incorporated
Inc.
- 3 -
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
Lisa Hammersly
Western Asset Management Company
David P. Hanna
Auto-Trol Technology Corporation
Bob Harmison
Qubix Graphic Systems
Edward Hartmann
IDS Growth Spectrum Advisors
Susan Harvey
Megatek Corporation
Sally Hasson
Dataquest Incorporated
Satoru Hayashi
Fujitsu America, Inc.
Scott Hedrick
InterWest Partners
A. Grant Heidrich
Mayfield Fund
John G. Heit
General Dynamics Corporation
John Hime
Pyramid Technology Corporation
Katsuhide Hirai
Fujitsu America. Inc.
Frank Hope
Mostek Corporation
Tom Hutchins
Seiko Instruments USA Inc.
J. E. I wersen
Bell Laboratories
John Jackson
Dataquest Incorporated
Henri Jarrat
VLSI Technologies Inc.
James Jeffs
Trust Conpany of the West
Richard Jodka
The Putnam Mangement Company, Inc.
Harvey Jones
Daisy Systems Corporation
Daniel R. Jones
Genisco Computers Corporation
David Jorgensen
Dataquest Incorporated
Richard L. Justice
General Motors Corporation
Alan Kaechele
Spectragraphics Corporation
Len Keene
Adage
- 4 -
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
Henry Kellogg
Computervision Corporation
Paul Kirby
Hill Partnership
Gary Kost
Judy Kost
NUTECH Engineers, Inc.
Paul Kruger
Digital Equipment Corporation
Robert Kuhling
Calma Company
Steve LaCcmmare
Triad Systems Corporation
Lucio Lanza
Daisy Systems Corporation
Wendy Ledamun
Dataquest, Incorporated
Jean Ledford
Wisconsin State Investment Board
Chong C. Lee
Metheus Corporation
Al Libbey
Gould Incorporated
Rhonda Lindsey
CADAM Inc.
Bob Lorentzen
Zymos Corporation
David Lucas
International Microelectronics Products
Jim Lucas
Qubix Graphic Systems
Frank Lynch
Valid Logic Systems
Richard Ma
IBM Corporation
Rod MacDow
Prime Computer, Inc
Mark MacLennan
Bank of Boston
Rc*)ert Martin
Tektronix, Inc.
Cathy Marvin
Dataquest Incorporated
John Maxwell, III
First Pennsylvania Bank
Harold McAleer
GenRad, Incorporated
Hugh McClung
R & D Funding Corporation
Michael McDonald
Cad/Cam Specialists, Inc.
- 5 -
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
Donald McGinnis
Digital Equipment Corporation
Tern McManus
Dreyfus Corporation
Leon McMillan
Continental Illinois Bank
Robert D. Merrell
Burroughs Corporation
Phillip Mestancik
Polaroid Corporation
Milo F. Miras
Quantel Business Computers
John Moreland
Spectrgraphics
Ed Morgan
Triad Systems Corporation
Glenn Mueller
Mayfield Fund
Carol E. Muratore
Prudential-Bache Securities
Edmond Murphy
Nissho Iwai American Corporation
Bob Myers
Texas Instruments, Inc.
Michael Nagel
Xerox Corporation
Suki Nakamoto
Calma Company
Timothy Neuland
Weitek Corporation
James R. Newcomb
Dataquest, Incorporated
Tom Newsom
Hewlett-Packard Company
Laura Nichols
Dataquest Incorporated
Robert Noyce
Intel Corporation
Dana O'Brien
Prudential Insurance Company
Bob Olmsted
IBM Corporation
I a n Orrock
Susan Orrock
Racal-Redac Ltd.
Ray Osofsky
Triad Systems Corporation
Roger Paradis
Skantek Corporation
Mihir Parikh
Hewlett-Packard Conpany
- 6 -
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
Han Park
Qume Corporation
Richard P e r r y
LSI Logic Corporation
Richard A. P e t e r s
California Computer Products, Inc.
David P. P e t e r s
Lillian Peters
Honeywell, Inc.
Gayle P h i l l i p s
Dataquest, Incorporated
Martin Picco
Parallax Systems Inc.
Donna P l a t h
Dataquest Incorporated
David T. Poiesz
Jennison Associates
Mitch Quain
Prudential-Bache Securities
William Quigley
IBM Corporation
John Randall
Dataquest Incorporated
Susan B. Ray
Computervision Corporation
James R i l e y
Dataquest Incorporated
John Robotham
Apollo Computer
Dave Rome
Data General Corporation
Chuck Rosselle
Joan Rosselle
EDS Nuclear/Impell
David Rumsey
Burroughs Corporation
Sherman L. Rutherford
Data General
Frank Sammann
Dataquest Incorporated
Bob Sanchez
Robin Sanchez
Prime Computer, Inc.
Edwin Saunders
Versatec
Michel N. Sauvage
BULL SEMS
Nancy Savinelli
Megatek Corporation
Thomas Schaefer
Schlumberger
Corporation
7 -
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 /Telex 171973
Lawrence Schmitz
Hughes Aircraft, Inc.
Steve Schopbach
Cadtec Corporation
Bill Seaver
Digital Equipment Corporation
Kevin Shambrook
Hughes Aircraft Company
Graham Shenton
International Microelectronic Products
C h a r l e s Simon
Cathy Simon
Vectron Graphics
Michael S i s a v i c
Metheus Corporation
Joe S i t a r i k
Bell LalDoratories
Matt Slavik
Apple Compute r
Randy Smith
Kayo Smith
California Automated Design, Inc.
P h i l i p s W. Smith
CcxTiputervision Corporation
Terry Smith
Metheus Corporation
Michael E. Sokolski
Control Data Corporation
Robert Speicher
Dataquest Incorporated
Sri Sriram
Digital Equipment Corporation
Michael S t e r l i n g
CADLINC
Robert Stevenson
Auto-Trol Technology C o r p o r a t i o n
Donald C. Stewart
Calma Company
Glenn S t e w a r t
Kay Stewart
WICAT Systems
Daphne Strong
Bank of Boston
David Stueber
Prudential-Bache
Jerry Sullivan
Tektronix,
Joseph Suty
S e a t t l e F i r s t N a t i o n a l Bank
Thomas R. Teeters
Calma Company
Securities
Inc.
- 8
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
James R. Tobias
American Microsystems, Inc.
Sheldon Toso
Triad Systems Corporation
James C. Towne
Metheus Corporation
Beth W. Tucker
Dataquest, Incorporated
G a i l Van Tubergen
Dataquest, Incorporated
Steve Volnt
J u d i t h Volm
IBM Corporation
Fred Wagner
IBM Corporation
Martin Walker
Analog Design Tools, Inc.
Ed Washington
Dataquest Incorporated
David Weisberg
Auto-Trol Technology Corporation
Michael Weisberg
Prudential-Bache Securities
J e f f r e y Wellington
Bank of Boston
John West
CADLINC
P e t e r White
Bank of Boston
Howard Wilczynski
CADAM
P h i l i p D. Wilson
Nixdorf Computer Corporation
I r v Winters
Digital Equipment Corporation
B a r r y Wolman
Prime Computer, Inc.
Douglas Wood
Digital Equipment Corporation
A r t Worsh
Digital Equipment Corporation
Joe Yesenofski
Tektronix, Inc.
Kazuhiko Yoshinaga
Toyo Denki S e i z o K.K.
Inc.
- 9 Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
Dataquest
Dataquest
MANUFACTURING IS NOT A BUTTON ON
THE DESIGN DATA BASE
Michael W. Sterling
Vice President
CADLINC, INC.
Mr. Sterling is Vice President of Engineering for CADLINC, INC., a
CAD/CAM company. Prior to the company's acquisition by CADLINC, he was
President of System Associates, Inc.
Previously, he was a Senior
Mathematician
and
Senior Development Engineer
for
General
Motors
Corporation, specializiing in CAD/CAM.
Mr. Sterling received B.A. and
M.A. degrees in Mathematics from the University of Detroit.
Dataquest Incorporated
CAD/CAM INDUSTRY CONFERENCE
September 26-28, 1983
Newport Beach, California
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
MANUFACTURING IS NOT A BUTTON ON THE DESIGN DATA BASE
INTRODUCTION
It is a shock to buyers of $100,000.00 per seat CAD/CAM systems when they realize
that they are not saving money. The reasons for this are complex and our industry
must be blamed, not so much for over-sell as for lack of understanding of the
mechanical CAD/GAM problem set itself.
We hope to shed some light on the problems and point out some areas of development
that will lead to solutions.
The evidence that we will bring out will support the thesis that a certain CIM
architecture and a methodology can be gently placed on current manual methods in
a fashion that will accelerate productivity. Along the way we will have to
illustrate solutions by pointing out problems.
(See Figure 1)
The architecture we will illustrate will be the now emerging Local Area Network
scheme. The idea here is to put personal engineering workstations on a very
high-speed communication channel. Each node will be a powerful self-sufficient
computer with no degradation in performance.
Each network will communicate to brother and sister networks via telecommunication
gateways or other high-speed internet gateways. We will show that this architecture
is the proper one because it has good characteristics:
1.
Simplicity
2.
Low Cost
3.
Functionality
4.
Mimics the well-established conventional world.
HISTORICAL BACKGROUND
The mechanical engineering portion of the CAD/CAM industry is complicated. Most
successes and economic advantages have taken place elsewhere, such as electrical
engineering.
The reasons for this are:
New technology is easiest to apply in a new industry.
Lack of past manual methods help make new ideas more acceptable.
If the industriy is new, then organizations grow with technology:
Technology is not imposed on the organization.
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
GATEWAY To
Other Systems
o
g.
CADLINC SHOP FUOOR SYSTEM
B>
n
IBM
VAX
etc....
c
(S
--IL
a>
—-.
5"
oo
CD
•3
o
a
TOOL CRIB
TOOL DESIGN
FILE MANAGER
03
c
CT
(/>
gj
>2
o_
>
b
(D
3
w
7^
^
0)
o
o
3
GENERATES!
CATALOGS:
LAYOUTS
FIXTURES
DRAWINGS
TOOLS
FIXTURES
CLAMPS
PART PGMS & TOOL PATHS
DRAWINGS
TOOL FILES
PROCESS SHEETS
PLANT LAYOUTS
PLOTTER
><
Q.
a.
<a
•X
t
ETHERNET
3
lOMILLION BIT/SEC
MACHINES
162MB DISK
STORES:
•a
lU
3
<o
O
MAG TAPE
FOREIW'S OFFICE
NC PROGRAMMING
O
^-J
w
HI
o
CO
o
>
.b.
o
00
CO
o
o
o
GEf^ERATES:
ACCEPTS:
PART PGMS &
TOOL PATHS
DRAWINGS
TOOL FILES
PROCESS SHEETS
SCHEMATICS
H/c l b
DAy
MachiIrS~^
n
Figure 1
CO
w
• ^
PART PGMS & TOOL PATHS
DRAWINGS
TOOL FILES
PROCESS SHEETS
MACHINE SCHEMATICS
Mechanical Engineering and Manufacturing are ancient disciplines. They
have their roots in antiquity. Therefore, as chaotic as they may seem at
times, these manual methods are surprisingly efficient. One only has to
spend a few weeks with a die or mold maker to be impressed with this fact.
Toolmakers are smart people !!
The field is very broad. A designer cannot be expected to grasp the manufacturing technology of the spectrum of parts he or she designs.
It takes about ten (10) years to develop a good mechanical designer. It
also takes ten (10) years to mold a manufacturing engineer. To combine these
two into one individual is asking too much over an entire industry.
The CAM portion of electrical engineering is so new that it has grown up
with the design technology. For example, VLSI designers are thinking about
most of the details of the manufacturing process as they design. As this
field matures and the manufacturing techniques become more exotic, then they,
too, will have trouble with the CM)/CAM connection in its entirety.
*
Since most computer science people know little about manufacturing engineering, the systems they produce place little emphasis on CAM.
FUNCTIONS OF DESIGN
We should reflect on the purpose of design in order to clarify our thoughts about
Computer Integrated Manufacturing. The first fact that is apparent is:
"DESIGN IS NOT AN END IN ITSELF"
This simple idea is often not understood.
This is especially true with draftsmen.
Designers are sometimes frustrated artists.
The functions of design should be:
Designs make ideas clear to others.
time.
Designs are ideas that function in
Designs are great vehicles to work out problems. The problems are usually
spatial in nature like clearances. Many times engineering properties need
to be analyzed by the designer.
Designs might be clear to other designers, but not to manufacturing. In
order to make them clear to manufacturing, one needs to add notes. So designs
are not all geometry.
Designs represent a product that can be manufactured.
TYPES OF DESIGN
There are two basic kinds of design - product and tool. Tool design is much richer
since it includes automation. A tree of design shows far more branches on the manufacturing side than on the product side. (See Figure 2)
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
Design
Tool
Product
Mold
Die
Fixture
Automation
Figure 2
For every product designer, there are five designers on the tooling end of the spectrum. They think of themselves as manufacturing engineers.
We should notice that a product design may remain quite constant over a long duration,
while the manufacturing process changes many times underneath it.
.
Most people have thought of CAM in terms of numerical control. This is only
part of a vast CAM process. In order to appreciate it we must think in terms
of CIM - Computer Integrated Manufacturing.
In major companies product and tool design activities are organizationally distinct.
The information network they encompass reaches across the main arteries of the organizational charts of the companies. It also extends to satellite companies that
serve the industrial giants.
The conventional world is a Local Area Network and we should recognize it as such.
DATA BASES
We need to look at a conventional data base that is maintained on paper with the
human being as the relational integrator.
The CIM equation will be similar, with the man being the initial integrator of
technology.
Therefore, we should try to define the conventional data base.
data base consists of a number of items:
The product design
Geometric Indicators
Notes and Dimensions
History of Development
.
Relation to Other Products
Dataquest incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
GEOMETRIC INDICATORS
Drawings of products contain geometry. This is apparant to everyone.
indicates the entire geometry and is very precise about part of it.
For example, no draftsman will draw every tooth on a gear hub.
about one and indicate the others.
The drawing
He will be precise
No draftsman will draw a full thread. He will only be precise about it if it is the
first of its kind. Otherwise, the geometry will not be shown at all.
In many complicated parts, the non-essential geometry is indicated only with notes
such as "blend from .5R to .25R" over a given distance. This gives a toolmaker
latitude without hindering functionality. In this case, the designer is using good
sense.
Good product designers are very stingy with precision. They know that precision
must only be used when it is necessary. Over use precision and it will dramatically
increase cost because manufacturing tries to produce what is drawn.
NOTES AND DIMENSIONS
Notes are great refiners of geometry, yet they in fact do much more. A note
indicating "heat treat" can greatly change the total geometry configuration of a
part and its functionality. The designer relies totally on the manufacturing
engineer in this case.
For example, many turned parts are heat treated. Parts that are turned often serve
in mechanisms that have mating rolling surfaces. The designer who says "heat treat"
in a note should be cognizant of the manufacturing process, but most times is not.
Heat treating, dimensional tolerancing, possible grinding operations, numerical
control, metallurgy and tolerance charting may be influenced by that one note.
The manufacturing process highly influences functionality of mechanisms. In molded
or formed parts a note that says "maintain sharp corners on decorative trim design"
can have the effect of greatly increasing tooling cost.
The manufacturing engineer must now consider a whole new set of problems because
of such a note. If the volume is modest, then he may be able to get away with low
cost aluminum tooling. If the volume is medium, say 20,000 parts, then that one
note can add weeks and $100,000.00 and more to tooling costs.
The reason is that soft tools may not be able to do the job and tool steel may be
necessary to maintain sharp comers on a decorative trim. Tool wear becomes
important at 20,000 parts.
Going back to the geometric side for a moment, hundreds of thousands of dollars are
added to tooling cost by small product geometry differences.
The simplest die is the one that has a neat, clean mating between male and female
pieces with little or no binder development. Minute differences in the product
cause dramatic changes in the geometry of the die.
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Parl< Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
In plastic parts this is also true. The simplest mold is the best one. The biggest
cost increment often comes from a small change in geometry. Consider the terminal
housing shown in Figure 3 (Black Housing).
The geometry for the bezel is OK, since a simple two-piece mold is sufficient.
The housing itself is made with two molds - one for the top and one for the bottom.
In the top mold, there is a back draft of just a few millimeters. This no longer is
a simple two piece mold. Two courses of action are open to the manufacturing engineer:
1.
Build a multiple pull mold with the backdraft area opening via separate
cylinder. This will add weeks to mold production and almost double its
cost.
2.
Go back to the designer and ask "Is this backdraft essential?"
What we always have to keep in mind is that this type of expensive tooling is the
essence of high production American manufacturing. It used to be that volume overcame the cost in captive markets. This is no longer true, since we must respond to
change that much faster.
Before we leave this terminal design, let's examine it from the standpoint of the
impact of dimensions on economic production.
The product designer took about six weeks using CAD techniques to design the
enclosure. The design included:
Esthetics - The surfaces are all free-form and pleasing.
Fit - The bezel fits the tube, and the housing, the base.
The production of a full scale wood model for evaluation and possible
external tooling.
The completion of this phase would constitute about 1/5 of the total time to production and at least five other designer/manufacturing engineers would have to become
involved before final tooling would be done.
Only about 1/5 of the CAD methods are complete. The other 4/5 are consumed in the
design of the tool. Since the original design culminated in the production of a
fully NC generated model, the geometric information is available. If the design
had not gone to a model, it would have required much more work.
If the conventional world, the human being fills in the gaps. Lines drawn on
paper jump out to form parts and tools in the manufacturing engineer's mind set.
The CIM equation has to do this also. This is not done easily by most system.
In order to drive cutters around objects, the geometry must make digital as well
as visual sense. We will call this topological integrity.
Dataquest Incorporated, A Subsidiai7 of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
<u
u
3
60
•H
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
Very few designs hang together topologically.
creating an oblong in CAD.
Step 1 —
Construction Lines
Step 2 —
Fillets
Step 3 —
Trim
Let me give a simple example of
No Integrity Here
1
Step 4 —
^X
Profile Cut
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
I have kept the line directions to show you how the design is visually complete,
but is topological nonsense. In order to profile cut this shape with a tool, we
must have one geometrically complete entity.
**MOST DESIGN SYSTEMS WORK FOR VISUAL COMPLETENESS**
**ALL MANUFACTURING SYSTEMS REQUIRE TOPOLOGICAL INTEGRITY**
Herein lies the dilemma. More integrity means more man hours in the design.
add integrity to the design means adding more to the design budget.
To
Remember, we are trying to lay a new process gently on an old discipline that is
quite efficient. The CIM local area net must not produce gross budgetary bumps on
the conventional network.
As soon as chief draftsmen are asked to do more work, they rebel. They are responsible for drawings (visuals). Stitching things together for data integrity is not
their job. Organizationally they never worried about it and they want it to remain
so.
Notes and dimensions can be very confusing to manufacturing. For example, we have
seen many aerospace parts with notes that say "spline" and then give a table of
(x,y) values. On the same drawing, we might see a table of five (x,y) values and
a note that says "conic".
Spline is imprecise since splining techniques are as varied as the wind. For
example, did the designer mean cubic spline, rotated cubic spline (Fowler-Wilson),
quintic spline, etc.?
Now five point conies are exact, since they define a general conic:
is somewhat obscure for most manufacturing engineers.
but the note
Here is a case where conventional and CIM techniques have come together to produce
nonsense. The writer of such a note was a west coast aerospace firm while the
receiver was a midwest tooling firm trying to use CIM tooling methods.
HISTORY OF DEVELOPMENT
Normal drafting room practice dictates that a drawing have a change record or block
on it. The purpose of this is to track a design record from its creation to the
present. This serves as an audit trail so that related parts can be checked to see
how they have been influenced by changes.
The assembly drawing is usually the court of last resort because it will often show
a "bubble" of the latest change of a piece part.
All this is well known. What is not apparent is that changes that occur once
manufacturing has started literally influence an average of five (5) drawings.
Once the manufacturing process has started, change is the big money loser. This is
one of the reasons that manufacturing engineers try to wait as long as possible to
begin tooling that which is close to the product surface itself.
What they do is get an idea of the product and start on the auxiliary areas, knowing
that if they move too close to the product structure too early they will be ruined
by change.
Dataquest incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Pari< Drive / San Jose, CA 95131 / (408) 971-9000 / Teiex 171973
Change has become such a strong Influence In American Industry that it has created
an entire sub-industry. This is known as the model and prototype tooling industry.
Models and prototype tooling are absolutely essential to high volume manufacturing.
The computer Industry has helped to create prototypes, but has done little to
eliminate them.
Two things are apparent:
Our computer data bases are not rich enough today to support going to
hard tools without intermediate steps for complicated parts.
Models and pcototype tooling are valuable as communicators of information
and ways to work out fit and form, fixture, automation and all sorts of
manufacturing problems.
Ue have been fooled by going directly to product via NC. This is sensational when
it is possible. We should seek out every instance where it is possible and automate
the process to as high a degree as possible. There is, however, an area, a large
area, where major tooling is required that is very hard to get to directly. This is
the heavy metal cutting business-, so vital to our world industrial economic power.
This area is the vital one for the computer industry to look at because it involves
the most potential for savings.
WHAT HAS HAPPENED TO THE DATA BASE IN THE SKY ?
In the early I&70*s, I thought that the next decade would yield useful geometric
data bases running from initial design to the shop floor. This has not happened.
Corporate commitment has taken the blame. "They did not understand", "They were not
ready", "They stuck to the old ways", "They" meaning the Universal Corporate ogre.
They did slow things down, but the reasons are a lot deeper than superficial slow
stepping. These are the reasons, I think, that the design manufacturing data base
has not taken hold:
Manufacturing has been left out. Software and hardware has been allocated
to a small cadre of superior designers in the engineering office.
Manufacturing problems are too little understood.
Display technology has been too expensive for high quality pictures for
the masses.
Too much emphasis has been placed on large host machines because this was
thought to be the way to capture and channel data (corporate data base control).
In short, we have been working with the wrong basic architecture because the corporate
data base centered around huge machines cuts across the bias of the conventional world.
We must mimic the conventional world and bring the Innate abilities of the man back
into the CXM equation.
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
We really did not know enough about data to channel or capture it. We certainly did
not know that our machines would not support relational data bases in a meaningful
way. Our present architecture is set up to compute, not store and rapidly retrieve
masses of pointer-based data structures.
Every manager who is a non-professional computer person should ask his computer
staff to tell him how their present CAD/CAM system can support a manufacturing data
base in full. Ask them to compute the number of characters required to store the
data sequentially. Do not have them bother with the relational aspects.
Ask them to have a report ready two weeks from Tuesday. The CAD/CAM manager is in
for a surprise! His staff is unlikely to return two years from Tuesday with any good
answers if the company is a Fortune 1000 company with any manufacturing capability
at all. No wonder manufacturing data bases have been so hard to grab onto. The
organizations that are charged with building it do not have any manufacturing people
in them. Where they do, they are picked because they know how to program the
company's latest NC equipment, not because of their expertise in the manufacturing
process.
How, then, does one go about building a decent data base.
We have to use two primary management tools:
1.
Put the right people on the problem.
2.
Give them some tools that fit the problem set.
Here is an outline:
1.
Put in the proper computer architecture. This means one that mimics the
current manual process. This will be Local Area Networks.
2.
Get someone on your staff who knows the organization from design through
manufacturing. This might be hard to get because manufacturing data base
design is not the key to the executive suite.
3.
Establish a large block diagram of organizational working units.
4.
Find out 10 inputs and outputs of these groups.
5.
Order the I/O in priority in each block of the diagram.
6.
Create sequential ASCII files to represent each part of the organization.
7.
Let the individual groups use their own local area networks and ask them
to enter the data as they see fit so it can be looked at by people in their
local area, as well as remotely using communication techniques that are
standard.
8.
Allow communication to take place. People will find ways to use the data
effectively if they can look at it on their terminal.
9.
Rather than force everyone to one standard before we know what the standard
is, establish some guidelines for the data that are easy to achieve. Later,
ASCII data stored in an orderly fashion is easy to transfer to a "proper"
data base if it warrants it. IGES is a good start here.
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
The reason we should do all this is that it is an orderly approach to the problem
that saves money. It does so by getting the whole organization working the quickest.
This is the CIM approach.
THE MYSTERY OF THE 3D DATA BASE
In over 20 years of working with designers and skilled tradesmen, it has become
apparent that most designers design in multiple 2D views while tool-makers work in
3D. Maybe the analogy of the painter vs. the sculptor is proper. The difference
in thought process can cost us big dollars if we do not have a system that can deal
with it.
Toolmakers get to be journeymen with two skills:
1.
The ability to visualize 2D drawings as 3D complements. Most tools are
in a sense the geometric complemenet of the product to be made.
2.
Manual dexterity.
Exceptions to this are loftsmen, who are natural 3D thinkers.
Mechanical designers are "flat-land people". The design problem that might help you
understand this is the design of a copying machine. This is a problem of a mass of
mating and inter-related mechanical parts. The design is primarily a two-view
design —
Side view for lever movement.
Plan view for interference and clarity.
The parts are primarily swept shapes.
axis.
That is, 2D profiles pushed through a third
I make this claim today in September 1983 — "To use a CAD system to design a copier
in full 3D with a set of good mechanical designers will add to the cost of the design.
This is a testimony to the capability of current CAD systems and the state of skills
that mechanical designers have.
Another statement of fact — "The tool designer would not benefit by the '3D-ness'
of the data base because the product data base is not the tooling data base." If
it were simpler to design in 3D and it cost less, then great. If it benefitted
tooling, then also great.
There are some items in the design that would benefit by the use of full 3D. These
are items that may undergo some sort of engineering analysis or those for which
special tooling may be required.
You can get more benefit out of an early tooling input to the design than you can
the 3D aspects of it. The tooling engineer is capable of saving great sums of money
if they are available for early design review.
Let us look at another part of the mechanical industry that has been touted as a
source of savings — Die Design.
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
The product itself may be as 3D as
but die designers primarily design
of their 3D-ness, once the product
female die, is to add bumps on the
place.
can be with all sorts of sculptured surfaces,
in plan view looking down at the die. The extent
complement surface is positioned in the male or
binder surface so that proper forming will take
About 65% of all machining of these surfaces is 2-1/2D. This means the tool-maker
needs profiles and hole locations at Z depths to build the tool.
One of the largest industrial companies in the world worked long and hard at 3D
die design. Their findings were:
1.
The cost of design went up dramatically.
2.
The best designers said it was not needed.
3.
The system response slowed dramatically with a topologically complete
data base.
4.
The tooling engineer had more work undoing the ornate topology to do 65%
of their work.
So what they have opted to do is revert to 2-1/2D design and use 3D only when it
will directly save money as dictated by tooling.
FUTURE OF 3D
Of course, 3D is here to stay. It has so many benefits whether it be wire frame or
solid, but I am amazed that two facts are not realized about where we are today.
1.
3D costs more.
2.
3D has trouble competing with conventional methods in any way across the
broad spectrum of applications.
So what can we do.
I propose the industry do two things:
1.
Recognize the problem and work on it so we can afford full 3D. This
means data base research and architectures to run them on suitably.
2.
Companies should adopt structured design techniques.
STRUCTURED DESIGN TECHNIQUES
The software industry has found that top down design is the way to go.
means is:
All this
1.
Generalize the problem into gross block diagram form.
2.
Only break down the blocks when the whole process is understood.
3.
Have regular and consistent design reviews.
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
Mechanical CAD/CAM c o u l d use this approach, involving a tooling engineer from
design i n c e p t i o n .
It usually works the other way. A contact p e r s o n from design works with tooling
after the design is done. Usually, contact people are not top quality. They are
put in this job either when too Inexperienced or when they are not competent in a
more useful role. They serve as couriers of bad news — "We're late, you're late,
there is a problem they say" etc...
If the tooling engineer were present at early design reviews, then all sorts of
problems would disappear. Early consensus is more vital in the CIM process than it
is in the conventional world because the data base in the sky is obscured by clouds
much longer.
OUT SOURCING
Forty to fifty percent of tooling is outsourced by our industrial giants. Since
bids for tooling are done late in the design process and are competitive, how do
we involve manufacturing at early stages ?
The Japanese have solved this problem. They tacitly recognize that dominant industries create captive satellite shops. They buy shares in them, sit on their
boards and try to keep them profitable. These satellites participate in the famous
Japanese consensus.
U.S. companies cannot do this, but they can alter their purchasing procedures in a
way that recognizes what is apparent to everyone:
1.
The satellites are cleverly run by skilled people,
2.
The head count and, therefore, the cost profiles of these companies are
much more attuned to boom r- recession economy.
3.
These companies need to make a profit.
Purchasing agents of high quality recognize all these while others exercise the
disposable vendor philosophy, which says:
1.
Hake them invest in CIM equipment to get the work.
2.
Work them against others for price.
3.
Dispose of them if their price rises above conventional processes.
I am sure that major tooling programs can be thought out better for the long haul.
Now if this is solved, then we will have to extend the Local Area Network concept
to small shops. This has been done in the Detroit area. We call this CIM VENDOR.
The idea is to be able to send data from a dominant industrial giant to a tooling
shop.
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
Ideally, it works like this:
1.
The shop is called in early to bid on a portion of a major tooling
program.
2.
The manufacturing engineer knows what kind and type of data he will
receive.
3.
The bid is accepted by the dominant company.
4.
The data is electronically sent via a gateway utility from the large
corporate network to the small shop.
5.
The shop manufacturing engineer processes the data to produce the tools,
CONCLUSION
My experience tells me that we can really step up productivity by using simple
strategies that are at the essence of the CIM philosophy.
Put in Local Area Networks.
Do not overdo the data base too early.
Allow data to flow.
Use the innate computing and visualization power of the human beings
in the organization.
The key idea is to allow the technology to settle on top of the existing organizations. Maybe we should call this technology CIM-M — Computer Integrated Manufacturing with man as the integrator.
We need to get this technology into American heavy industry in order to revitalize
it and allow us to compete in world markets where other countries compete with lower
labor rates and the benefit of new technologies growing up with new organizations.
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Parl< Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
(OKiDataquest
::;>a Dataquest
Dataquest
CAD MARKET REVIEW AND FORECAST
Beth W. Tucker
Research Analyst
CAD/CAM Industry Service
Dataquest Incorporated
Ms. Tucker is a Research Analyst for DATAQUEST's CAD/CAM Industry
Service.
Prior to joining DATAQUEST, Ms. Tucker was Market Planning
Manager of Calma Company's Microelectronics Division, where she was
responsible for identifying market opportunities and developing marketing
strategies for the company's electronic CAD product line.
She later
contributed to the development of marketing information systems at
Calma's corporate level. Prior to that, she was a Senior Market Research
Analyst with Monroe Systems for Business, where she was responsible for
market research and development of small business computer products. Ms.
Tucker holds an A.A. degree in Computer Technology from Morris County
College in New Jersey and is doing course work in Business Administration
at San Jose State University.
Dataquest Incorporated
CAD/CAM INDUSTRY CONFERENCE
September 26-28, 1983
Newport Beach, California
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
INTRODUCTION
Computer-aided design and computer-aided manufacturing play key roles
in contributing toward productivity gains in most major U.S. industries.
The complexity of automating the engineering, manufacturing, and testing
of products of all types is being addressed by CAD/CAM vendors and their
products.
Users of such products are demanding, and their level of
sophistication is increasing almost exponentially.
MARKET SIZE
DATAQUEST estimates that CAD/CAM industry worldwide revenue
reach $9 billion by 1987, increasing from $2 billion in 1983.
represents a compound annual growth rate of 43 percent.
will
This
The
industry
comprises
five
major
segments:
mechanical;
architecture, engineering, and construction (AEC); printed circuit (PC) ;
integrated circuit (IC); and mapping and other. The mechanical segment
is the largest, with 1983 revenues expected to reach $1.1 billion,
growing at 44 percent per year to almost $5 billion in 1987. The
fastest-growing segments are IC and PC, with compound annual growth rates
of 47 and 46 percent and estimated
1987 worldwide revenues of
$863 million and $1,551 million, respectively.
The AEC segment is
growing at 39 percent per year, and represents 15 percent of the CAD/CAM
market with 1983 revenues of approximately $326 million.
Analyzing industry revenue is a traditional and significant approach
to determine market size and market share.
However, because system
selling prices vary considerably, it is perhaps more meaningful to
determine market size by analyzing the number of "seats," or workstations
sold. A workstation, in this case, is defined as a single-user station
and may be standalone or host dependent.
DATAQUEST
estimates
that
during
1983,
approximately
17,000
workstations will be shipped and that this number will increase by
63 percent per year to 119,000 workstations shipped in 1987. Workstation
shipments are increasing significantly faster than our revenue estimates,
indicating that the average cost per workstation is decreasing. Sales of
workstations for designing integrated circuits and printed circuit boards
will experience the fastest growth rate, followed closely by workstations
for mechanical engineering.
- 1 Copyright © 2 6 September 1983 Dataquest Incorporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
For this analysis, the average cost per workstation or seat includes
the cost of the CPU, peripherals, stations, and all software spread
between the average number of workstations per system. The average cost
also
takes into account
standalone
and
host-dependent
stations.
DATAQUEST predicts that the average cost per workstation will decrease a
total of 40 percent, from $126,000 in 1983 to $75,000 in 1987. The IC
segment will experierxre the largest cost decline, at 16 percent per
year. This decline is largely due to the electronic design automation
(EDA) subsegment's wide use of low-cost 32-bit standalone systems.
SYSTEM USE
DATAQUEST's CAD/CAM Industry Service recently conducted a survey of
4,000 CAD/CAM users.
Among other things, we were interested
in
determining what " percentage of system time is spent on drafting, on
design/modeling, on analysis, and on manufacturing applications.
A
preliminary analysis indicates that the percentage of system time spent
on each of these applications is 60 percent for drafting, 27 percent for
design/modeling, 5 percent for analysis, and 8 percent for manufacturing
applications for all market segments. As expected, the percentage of
system use varies by segment. For example, AEC systems are not used for
manufacturing applications and so the largest percentage of system time
is spent on drafting.
In contrast, manufacturing applications for
mechanical systems comprise 11 percent of system time. PC systems are
used mainly for drafting, due to the wide use of systems for to creating
artwork.
As better and easier routers become available, DATAQUEST
anticipates a shift from straight drafting toward more system time spent
on design/modeling. By 1987, DATAQUEST predicts a shift in system use,
toward more design/modeling, analysis, and manufacturing, with less
emphasis on drafting.
CAD/CAM TRENDS
These changes will occur for many reasons. First, and perhaps most
importantly, the market demands it. CAD systems originally filled the
need to automate drafting and the associated changes and documentation.
This need has been successfully filled by many vendors in each of the
CAD/CAM market segments. As users and vendors become more sophisticated
in using and designing CAD systems, a higher level of sophisticated
application programs is being developed with a shift toward more
analytical programs and more integration of manufacturing applications.
- 2 Copyright © 2 6 September 1983 Dataquest Incorporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
Second, significant trends in hardware design, software development,
and systems integration philosophies are helping to make the goal of
providing a true end-to-end systems approach a reality.
Distributed processing provides each seat with local intelligence,
enabling the user to perform interactive, creative, and analytical
processes locally.
Relatively young companies are dominating the
distributed processing market with low-cost 32-bit workstation products.
The more well-established computer companies that have traditionally
supplied CAD/CAM hardware are also developing or introducing distributed
processing systems. The user can now purchase a CAD/CAM system at a
relatively low initial purchase price and expand the system without
spending hundreds of thousands of dollars.
Distributed processing
provides a consistent level of performance, regardless of the number of
seats, and if one workstation crashes, only that user is affected.
Improved communications capabilities and data base management systems
go hand in hand with distributed processing. The size and complexity of
CAD/CAM data bases is increasing to the point where local area networks
(LANs) and data base management systems (DBMSs) are needed to manage the
engineering and design processes. LANs and DBMSs support clusters of
local systems so that data and system resources can be shared among
users.
Manufacturing design functions, i.e., numerically controlled
machines and factory robots, tooling, and group technology data bases,
are dependent on communicating with the same design data base created in
engineering.
The innovative designs and applications of VLSI technology have made
it possible to increase CPU performance by 200 percent every two to three
years. Circuits are being designed with over 400,000 transistors, and
smaller is not necessarily better, according to the start-up company
Trilogy. At the same time performance is dramatically increasing, the
cost of hardware is decreasing at 15 percent per year.
Companies are developing microprocessor-based graphics stations,
with 1,000-line, 60 Hz, noninterlaced technology, offering users the
resolution and speed needed for displaying complex designs.
Fast
graphics processors that can display complex solid models in real time
are available. End-user pricing of graphics stations is also declining
at approximately 15 percent per year.
- 3 Copyright © 2 6 September 1983 Dataquest Incorporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
The CAD/CAM industry is software intensive, with development time
running into thousands of man-years for a single applications program.
Third-party software agreements are prevelant in this highly competitive
environment. The race to provide full-system design capability requires
vendors to provide as much applications software in as little time as
possible.
Virtually
all major
vendors have joint marketing
or
development agreements with software houses, have merged or acquired such
companies, or provide interfaces to existing software.
Third-partydeveloped software is typically not the core of a CAD vendor's product,
however, but is usually large, number crunching, analytic programs, such
as solids modeling, finite element modeling and analysis, or simulation
software.
Companies using an end-to-end CAD/CAM systems solution are able to
automate their product development cycle, from concept and engineering,
to manufacturing and testing, thus enabling them to compete effectively
and profitably in worldwide markets.
EDA—AN INTEGRATED SOLUTION
The electronic design automation (EDA) subsegraent of the IC CAD
market is one example of the successful implementation of the CAD/CAM
trends discussed above. DATAQUEST believes EDA is the fastest-growing
market segment, and that sales will reach $500 million by 1987, growing
at 65 percent compounded annually. To date, the majority of companies in
this market (Daisy Systems, Mentor Graphics, and Valid Logic, to name
just a few) are less than three years old; yet, together, they have
invested millions of man-years worth of state-of-the-art
product
development.
Distributed processing is prevelant in the EDA marketplace, with
low-cost 32-bit workstations being used by electrical engineers. The
traditional 16-bit IC CAD physical layout system's average cost per
workstation is declining, from $121,000 in 1983 to $78,000 in 1987.
However, the electrical engineer is not able to justify this cost; it is
too high and does not offer the necessary 32-bit performance.
An EDA system must handle massive amounts of electrical and graphics
data, so that the engineering design functions are integrated among
themselves, as well as with the physical layout and manufacturing
functions.
Most EDA vendors have formed many third-party software
agreements, ranging from interfaces to circuit simulators, to mergers
with software companies, to providing a front end to a PCB router.
- 4 Copyright © 2 6 September 1983 Dataquest Incorporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
Manufacturing integration is achieved by providing a front-end design
system to a silicon foundry's routing and placement programs for standard
cells or gate arrays. In this manner, companies developing electronic
systems can afford to develop customized and proprietary circuits. The
design
data
is
electronically
transmitted
to
the • foundry
for
fabrication. Although the electronic design autcanation market is young
and not yet fully developed, clearly the EDA vendor's goal is to provide
electronic components and systems manufacturers an integrated, end-to-end
systems solution.
COWCLDSION
The EDA market is just one example of integrating the design,
engineering, and manufacturing phases of the product development cycle to
provide CAD/CAM users an integrated solution. As the total available
market is penetrated with
low-cost, powerful systems capable of
integrating design with manufacturing, CAD/CAM users will have the tools
to increase productivity, and the reasons for spending nearly $25 on
CAD/CAM products during the next four years.
- 5 Copyright © 2 6 September 1983 Dataquest Incorporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
WORLDWIDE CAD/CAM REVENUE
(Millions of Dollars)
1983 TOTAL $2,175
MECHANICAL
$1,164
y
/
/
AEG
$326
MAPPING & OTHER
$143
INTEGRATED CIRCUITS
$185
PRINTED CIRCUITS
$357
Source: DATAQUEST
- 6 Copyright ©26 September 1983 Dataquest Incorporated-Reproduction
Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
WORLDWIDE CAD/CAM REVENUE
(Millions of Dollars)
1987 TOTAL $9,045
MECHANICAL
$4,955
MAPPING & OTHER
$467
INTEGRATED CIRCUITS
$863
AEC
$1,209
PRINTED CIRCUITS
$1,551
Source: DATAQUEST
- 7 Copyright © 26 September 1983 Dataquest Incorporated-Reproduction
Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
DEFINITION
WORKSTATION — A SINGLE USER STATION, EITHER
STAND-ALONE OR HOST-DEPENDENT
AVERAGE COST — AVERAGE TOTAL SYSTEM PRICE,
INCLUDING CPU, PERIPHERALS,
STATIONS AND SOFTWARE, DIVIDED
BY AVERAGE NUMBER OF STATIONS
- 8 Copyright © 2 6 September 1983 Dataquest Incorporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
WORKSTATION SHIPMENT FORECAST
1983
UNITS
1987
UNITS
CAG
MECHANICAL
8,317
59,822
64%
AEG
3,023
18,091
56%
PC
2,853
21,594
66%
IC
1,558
12,959
76%
MAPPING & OTHER
1,197
6,370
44%
16,948
118,836
63%
APPLICATION
TOTAL
Source: DATAQUEST
- 9 Copyright © 26 September 1983 Dataquest Incorporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
AVERAGE COST PER WORKSTATION
1983
COST
1987
COST
PERCENT
CHANGE
MECHANICAL
140,000
83,000
-41%
AEC
108,000
67,000
-38%
PC
119,000
70,000
-41%
IC
127,000
62,000
-51%
97,000
73,000
-25%
126,000
75,000
-40%
APPLICATION
MAPPING & OTHER
TOTAL
Source: DATAQUEST
- 10 Copyright ©26 September 1983 Dataquest Incorporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000/ Telex 171973
DATAQUEST USER SURVEY
PERCENTAGE OF SYSTEM TIME SPENT O N . . .
DRAFTING
DESIGN/MODELING
ANALYSIS
MANUFACTURING
- 11 Copyright © 2 6 September 1983 Dataquest Incorporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
1983 CAD/CAM SYSTEM USAGE
DRAFTING 60%
ANALYSIS 5%
MANUFACTURING 8%
DESIGN/MODELING 27%
Source: DATAQUEST
- 12 Copyright © 26 September 1983 Dataquest Incorporated-Reproduction P r o h i b i t e d
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Parl< Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
MECHANICAL CAD/CAM SYSTEM USAGE
DRAFTING 61%
ANALYSIS 6%
MANUFACTURING II0/0
DESIGN/MODELING 22%
Source: DATAQUEST
- 13 Copyright ©26 September 1983 Dataquest Incorporated-Reproduction P r o h i b i t e d
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
AEC CAD/CAM SYSTEM USAGE
DRAFTING 81%
ANALYSIS 40/0
DESIGN/MODELING 15%
Source: DATAQUEST
- 14 Copyright © 26 September 1983 Dataquest Incorporated-Reproduction Prohibited
Dataquest incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
PC CAD/CAM SYSTEM USAGE
DRAFTING 53%
ANALYSIS 2%
MANUFACTURING 7o/o
DESIGN/MODELING 38%
Source: DATAQUEST
- 15 Copyright © 26 September 1983 Dataquest Incorporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
IC CAD/CAM SYSTEM USAGE
DRAFTING 44%
ANALYSIS 70/0
MANUFACTURING 6%
DESIGN/MODELING 43%
Source: DATAQUEST
- 16 Copyright © 2 6 September 1983 Dataquest Incorporated-Reproduction P r o h i b i t e d
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Parl< Drive / San Jose, CA 95131 / (408) 971-9000 / Teiex 171973
..r
1987 ESTIMATED CAD/CAM USAGE
DESIGN/MODELING 40%
DRAFTING 25%
ANALYSIS 20%
MANUFACTURING 15%
Source: DATAQUEST
- 17 Copyright ©26 September 1983 Dataquest Incorporated-Reproduction P r o h i b i t e d
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
WHAT WILL MAKE THESE
CHANGES POSSIBLE?
- 18 Copyright ©26 September 1983 Dataquest Incorporated-Reproduction Prohibited
Jataquest Incorporated, A Subsidiary of A.C. Nielsen Company/1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
CAD/CAM TRENDS
DISTRIBUTED PROCESSING
IMPROVED COMMUNICATIONS/DATA BASE MANAGEMENT
IMPROVED PRICE/PERFORMANCE
THIRD PARTY SOFTWARE
INTEGRATED SYSTEMS
- 19 Copyright ©26 September 1983 Dataquest Incorporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Parl< Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
DISTRIBUTED PROCESSING
LOCAL INTELLIGENCE
CONSISTENT LEVEL OF PERFORMANCE
SYSTEM DOWNTIME AFFECTS SINGLE USER
LOWER ENTRY PRICE
INCREMENTAL GROWTH
SHARED RESOURCES
- 20 Copyright © 2 6 September 1983 Dataquest Incorporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
IMPROVED COMMUNICATIONS/DATA BASE
MANAGEMENT
CLUSTERS OF LOCAL SYSTEMS
SHARED DATA AND RESOURCES
EFFICIENT WORK FLOW
INTEGRATE AND MANAGE DATA BASE
PROJECT MANAGEMENT AND CONTROL
- 21 Copyright © 2 6 September 1983 Dataquest Incorporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Pari< Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
IMPROVED PRICE/PERFORMANCE
•
SEMICUSTOM VLSI
•
CPU PERFORMANCE INCREASING 200o/o EVERY 2-3 YEARS
•
HARDWARE DECREASING 15% EVERY YEAR
•
INTELLIGENT, POWERFUL GRAPHICS STATIONS
•
REAL-TIME DISPLAYS
- 22 Copyright © 2 6 September 1983 Dataquest Incorporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
#
THIRD-PARTY SOFTWARE
NO NEED TO REINVENT THE WHEEL
MAXIMIZE RESOURCES
— PEOPLE
— MONEY
— TIME
ACQUISITIONS, MERGERS, JOINT AGREEMENTS
- 23 Copyright © 26 September 1983 Dataquest Incorporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
INTEGRATED SOLUTIONS
•
COMPUTER-AIDED ENGINEERING
•
DESIGN AUTOMATION
•
COMPUTER-AIDED DESIGN
•
COMPUTER-AIDED MANUFACTURE
•
COMPUTER-AIDED TESTING
COMPUTER INTEGRATED MANUFACTURING
- 24 Copyright © 26 September 1983 Dataquest Incorporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
EDA —
AN INTEGRATED SOLUTION
- 25 Copyright © 26 September 1983 Dataquest Incorporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
WHAT MADE EDA POSSIBLE?
PHYSICAL LAYOUT, SATURATION/PRICE
NEED: ELECTRICAL ENGINEERING, DESIGN AUTOMATION
TECHNOLOGY
- 26 Copyright © 2 6 September 1983 Dataquest Incorporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
AVERAGE COST PER WORKSTATION
IC CAD AND EDA APPLICATIONS
ACTUAL DOLLARS
150
100
^
^
—
^
* ^
50
78
79
80
81
82
83
84
85
^
• ^ ^ • .
86
87
YEAR
- "
ICCAD
»
EDA
Source: DATAQUEST
- 27 Copyright © 2 6 September 1983 Dataquest Incorporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
EDA REVENUE FORECAST
MILLIONS OF DOLLARS
500
Source: DATAQUEST
- 28 Copyright © 26 September 1983 Dataquest Incorporated-Reproduction
Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
CLOSING THE GAP
INTEGRATION
ENGINEERING
AND DESIGN
MANUFACTURING
- 29 Copyright © 2 6 September 1983 Dataquest Incorporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 l^idder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
E>ataquest
•
Dataquest
CAD/CAM INDUSTRY OVERVIEW
Timothy 0. Gauhan
Vice President and Director
CAD/CAM Industry Service
Dataquest Incorporated
Dr. Gauhan is Vice President and Director of DATAQOEST's CAD/CAM Industry
Service. He was formerly with Calma Conpany, where he served as Manager
of Marketing for the Mechanical Products Division and as Director of
Marketing and Technical Publications.
Earlier, he served on overseas
research projects sponsored by the National Science Foundation and the
U.S. Agency for International Development. Or. Gauhan has taught courses
in Research Methodology at Rice University auid Arizona State University,
and courses in International Business at the American Graduate School of
International Management.
He received a B.A. degree in International
Relations fron San Jose State University and M.A. and Ph.D. degrees from
Rice University.
Dataquest Incorporated
CAD/CAM INDUSTRY CONFERENCE
September 26-28, 1983
Newport Beach, California
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
As all of us who work in the CAD/CAM industry fully appreciate, this
is a very complex business, full of interesting but oftentimes perplexing
problems. Being successful in this business often requires equal parts
of hard work, luck, and aspirin. Consider for a mcanent this list of
technical buzzwords that confront CAD/CAM marketing people on a daily
basis.
3-D Modeling
Minicomputers
Cc^mnun ica t ion
NC/FEM
Data base management
Networking
Design rule checks
Plotters
Graphics software
Routing and placement
Microprocessors
Solid modeling
Having mastered tihese little tidbits of technology (spanning the
discj,plines of computer science, mathematics, electronics, and heaven
knows how many branches of engineering) the CAD/CAM marketeer must then
turn to this list of issues in order to make sure his or her products are
successful in the marketplace.
Advertising
Product literature
Benchmarks
Product management
Corporate agreements
Sales proposals
Demos
Technical centers
Marketing agreements
Tradeshows
PR
User's groups
I won't bother you with t:he list of manufacturing, financial,
accounting, and personnel woes that confront CAD/CAM companies, but
suffice it to say, while tihe rewards can be great, so can the bills for
Bufferin.
In February of 1981, DATAQUEST began its efforts to help CAD/CAM
professionals make sense out of this oftentimes confusing world.
We
spent the following months meeting the necessary contacts throughout the
industry, collecting and analyzing what data tJiere were available, and on
August 31 of that year we sent our Marketing Analysts' Source Book to our
first clients, seme of whom are in this room this morning.
We are often accused of being compulsive keepers of statistics (and
rightly so, I suppose). Beth Tucker will report a number of statistics
on the industry in a few minutes, but I would like to take just a brief
time to share some statistics that DATAQUEST's CAD/CAM Industry Service
has collected on itself after two years of serving the industry.
- 1 Copyright © 26 September 1983 Dataquest Incorporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
While I thought there were only six companies in the business when we
first started, I am pleased to note that our client list now consists of
some 95 companies.
In terms of types of clients that we serve,
67 percent of our clients are in the manufacturing and marketing of
CAD/CAM equipment, 32 percent are in the financial community, and 11
percent are end users of CAD/CAM equipment or are other types of
companies.
Obviously the vendors of CAD/CAM equipment constitute the largest
user group of DATAQUEST's CCIS. I thought it would be interesting to see
what kind of companies these are. The following breakdown will give you
scane idea of the types of vendor companies that DATAQUEST provides
information to.
Turnkey—34 percent
Computer—20 percent
Software—8 percent
Graphics—21 percent
Other—17 percent
In order to keep track of the information IJiese companies require
from us, we currently have files on over 350 companies.
In our two years of existence we have published over 2,000 pages of
information for our clients, the number of statistics that can be found
in our volumes is in the tens of thousands.
As those of you who use CCIS are aware, a very important aspect of
our business is answering inquiries from our clients. No matter how much
information we publish in our books, we could not possibly publish enough
information to suit our clients' needs. Consequently, we have responded
to about 7,000 inquiries from our clients since the inception of the
service; and as I have said in the past, in this business, every answer
seems to raise five more questions. While we may not have been right 100
percent of the time, I think our staff has provided the industry with a
very valuable service, which we plan to improve year after year.
- 2 Copyright © 26 September 1983 Dataquest Incorporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
Gathering these numbers is a difficult enough task in itself, but
because of the dynamics of the industry over the past few years, I have
found that figuring out where to put the numbers is even more difficult.
When we first published our numbers two years ago, we estimated that the
minicomputer-based turnkey companies constituted well over 90 percent of
the total CAD/CAM market. So we published our estimates and forecasts of
the revenues cuid shipnents of the six major turnkey companies and
discussed the fact that ccnnputer companies were becoming a more and more
important factor in the business. Our most recent estimates indicate
that computer cc»npanies constitute over 30 percent of the CAD/CAM market,
so that we have now ei^anded our coverage to include detailed information
on such companies as IBM, Prime, Digital, and Sperry.
As usual, the industry has not stood still long enough for our
categories
to be
the
"final
word,"
and new products
like
microprocessor-based workstations and even CAD systems running on
personal computers have now ccme on the scene. As if that weren't
confusing
enough,
companies
like
Daisy,
Mentor,
Valid
Logic,
Silvar Lisco, Metheus, and others, have virtually created a new market,
which we have dubbed Electronic I}esign Automation (EDA) , that addresses a
whole set of application issues not previously addressed by the CAO/CAM
industry.
We are busy reorganizing our data base to accommodate these new
products and approaches, but it is not the easiest job in the world.
Here is a brief list of categories that have been suggested thus far:
•
CAD/CAM
•
Engineering workstations
%
CAE
•
Software
«
CIM
•
Terminal
•
Computer
•
Turnkey
•
EDA
After we have developed the categories that are useful in helping us
present our data in a meaningful way to the industry, that's really just
where the fun begins. I have had what I thought were semiconductor
companies tell me they were more accurately described as CAD companies,
software companies tell me they were turnkey companies, and turnkey
c(»Dpanies say they were actually into CIM.
If you are the "scorekeepers of the industry," which is a part of
DATAQUEST's role, it is a lot easier if you know whether the players
think they are playing racquetball or rugby!
- 3 Copyright ® 26 September 1983 Dataquest Incorporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
In the years to follow, though, we promise to do our best to provide
accurate information on the industry by product type, company type,
application area, and region of the world.
Before turning to a discussion of the major issues that confront the
CAD/CAM industry today, I thought I would give you just a brief glance at
some of the work that we have in progress that DATAQUEST subscribers can
expect to see from us in the coming year.
For example, we have just ccMnpleted a major survey of users' groups
which will reveal a good deal of information on their patterns of usage
of CAD/CAM.
Beth Tucker, who will give you a summary view of the
industry in a few moments, has just begun to work on a survey of 27
systems for printed circuit board design. This survey will provide us
with much needed data in what has turned out to be a very strong
marketplace for PCB design.
On the international scene, DATAQUEST's staff has taken two trips to
DATAQUEST Japan this year. Preliminary analysis of the work we did there
reveals some very interesting facts, which we will report on in the
Cloning year.
We were able to unearth 27 firms that are in the CAD/CAM business in
Japan, although some of them have sold few or no systems. The installed
base in Japan is much smaller than the base in the United States, both in
a relative and an absolute sense. As we noted in a newsletter this year,
given its industrial base, Japan is effectively five to six years behind
the United States in implementing CAD.
Looking at the spread
that t:he Japanese seem to
mechanical and electronic
emphasis in the Japanese
system electronics.
in application use, however, we are finding
be more intensive in their use of systems for
circuit design.
This probably reflects the
econcniy on heavy industry and consumer and
While tihe Japanese have gained a reputation in the area of
manufacturing autcanation, robotics, and the like, it appears as if Japan
is lagging in the area of computer-aided design.
These are but a few of the areas that we will be researching for our
clients over the next year. As we look at the changes that have taken
place in the industry since the DATAQUEST CAD/CAM Industry Service was
begun just over two years ago, we cannot help but foresee that we will be
kept very busy keeping track of the events that take place in the CAD/CAM
Industry next year and in the years to follow.
- 4 Copyright © 2 6 September 1983 Dataquest Incorporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
TECHNICAL HEADACHES
- 5 Copyright © 26 September 1983 Dataquest Incorporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
MARKETING HEADACHES
- 6 Copyright © 26 September 1983 Dataquest Incorporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
CAD/CAM INDUSTRY SERVICE
CLIENT LIST
Agence de I'lnformatique
Alpha Fund
American Microsystems Inc.
Arthur Andersen & Company
Apollo Computer
Applicon
Assigraph
Auto-trol Technology
Bank of America
Bank of California
Bank of the West
Banque Nationale de Paris
Bell Laboratories
BULL-SEMS
Burroughs Corporation
CADAM, Incorporated
Cadlinc
Cadnetix
Cadtec Corporation
Cadtrak Corporation
CAE Systems, Incorporated
CALCOMP
California Automated Design Inc.
Calma Company
CGX Corporation
Churchill Group
Holley Carburetor
Computervision
Crocker Bank
Daisy Systems
- 7 Copyright © 2 6 September 1983 Dataquest Incor po rated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
CAD/CAM INDUSTRY SERVICE
CLIENT LIST
GenRaD
GTE Automatic
Electric Labs
Hewlett-Packard
IBM Corporation
IBM World Trade Asia
Corporation
International Microelectronics
Products
InterWest Partners
ITT Engineering Support
Centre
Landmarks Group
LSI Logic Corporation
Matrix Partners
Data General
Dept. of Industry
Digital Equipment
EDS Nuclear/lmpell
Evans & Sutherland
Fairchild
First Interstate Capital
First National Bank
of Boston
Fuji Xerox
Fujitsu America
General Electric
GE Industrial
Automation Ltd.
Genisco
-
8 -
Copyright © 2 6 September 1983 Dataquest Incorpotated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
CAD/CAM INDUSTRY SERVICE
CUENT LIST
Mayfield Fund
MCAUTO
Measuronics
Megatek
Membrane Ltd.
Metheus
Moody's Investor Service
Motorola Semiconductor
Murray Electronics PLC
Nutech
Oak Management Corp.
OCE Nederland B.V
Oxford Venture
Phoenix Leasing
Prime Capital Management
Prime Computer
Qublx
Racal Ltd.
Racal-Redac
Ramtek
Rank Xerox Ltd.
Sanders Associates, Inc.
Scientific Calculations
Seiko Instruments USA
Siemens AG
Silicon Compilers
Silicon Graphics
Silvar-Lisco
Spectragraphics
Tektronix
- 9 Copyright © 2 6 September 1983 Dataquest Incorporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
CAD/CAM INDUSTRY SERVICE
CLIENT LIST
Telemecanique Electrique
Telesis
Toyo Corporation
Valid Logic Systems
Vector Scientific
Vectron Graphics
VLSI Technology, Inc.
Union Bank
Xerox
Arthur Young & Company
- 10 Copyright © 2 6 September 1983 Dataquest Incorporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Parl< Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
DISTRIBUTION OF CClS CLIENTS
VENDORS 67%
ENDERS & OTHERS 11%
SPENDERS 22%
Source: DATAQUEST
- 11 Copyright © 2 6 September 1983 Dataquest Incorporated-Reproduction P r o h i b i t e d
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
CCIS VENDOR COMPANIES
TURNKEY
COMPUTER
35%
20%
SOFTWARE
OTHER 17%
8%
GRAPHICS 20%
Source: DATAQUEST
- 12 Copyright © 2 6 September 1983 Dataquest Incorporated-Reproduction P r o h i b i t e d
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
CAD/CAM INDUSTRY
SERVICE COMPANY COVERAGE LIST
Advanced Electronics
Andromeda Systems
ASK Computer
Bausch & Lomb
Bridgeport Machines
Boeing Aerospace
California Automated Design
Control Data
Cray Research
Datapoint
Digital Engineering
Display Interface
Efficient Engineering
Evans & Sutherland
Everett-Charles
Fairchild
Florida Computer Graphics
Fujitsu
General Motors
Genrad
Graphic Construction
Hewlett-Packard
Hitachi
Honeywell
IBM
Integrated Computer Systems
Intersil, Inc.
Jacobs Engineering Group
James River Graphics
Jupiter
- 13 Copyright © 2 6 September 1983 Dataquest Incocporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
CAD/CAM INDUSTRY
SERVICE COMPANY COVERAGE LIST
Kanematsu
Kinegraphics
KOH-I-NOOR
Lear-Siegler
Lockheed
LSI Logic
MacNeal-Schwendler
Morgan Fairchild Graphics
Metrex Management
National Computer Systems
Northrup
Numeritronix
Oce Industries
Omnitech Graphics
Orcatech
Phoenix Computer Graphics
Precision Visuals
Pyramid Technology
Quality Micro Systems
Quad rex
Raster Technology
RCA
ROLM
Scientific Calculations
Seiko Intruments
Sperry-Univac
Tandem
Teradyne
Texas Instruments
Trilog
- 14 Copyright © 2 6 September 1983 Dataquest Incorporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
CAD/CAM INDUSTRY
SERVICE COMPANY COVERAGE LIST
University Computing Company
Unimation
United States Robots
Varian
Vector Automation
VLSI Teclinology
Wang Labs
Copyright
Western Electric
Westingliouse
Xerox
Xiphias
Zilog
Zonic Corporation
Zytronix
- 15 26 September 1983 Dataquest Incorporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Parl< Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
DATAQUEST CCIS DATA SUMMARY
2,000 PAGES PUBLISHED
7,000 INQUIRIES
10,000 CAD RELATED DOCUMENTS
95 CLIENTS
346 COMPANIES TRACKED
- 16 Copyright © 2 6 September 1983 Dataquest Incorporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
SUGGESTED CATEGORY NAMES
TURNKEY
COMPUTER
ENGINEERING WORKSTATION
CLUSTER NETWORK
GRAPHIC TERMINALS
SOFTWARE
CAE
EDA
- 17 Copyright © 2 6 September 1983 Dataquest Incorporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
SUGGESTED CATEGORY NAMES
HOST DEPENDENT
MAINFRAME
MINI
MICRO
PERSONNEL
YELLOW
ORANGE
BLUE
- 18 Copyright © 2 6 September 1983 Dataquest Incorporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
ESTIMATED
SYSTEM INSTALLATIONS
1982
UNITS
•»,wvw
2,000
1
0
JAPAN
U.S.
Source: DATAQUEST
- 19 Copyright © 2 6 September 1983 Dataquest Incorporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
ESTIMATED
TOTAL INSTALLED BASE OF SYSTEMS
1982
UNITS
8,000
4,000
JAPAN
U.S.
Soures: DATAQUEST
- 20 Copyright © 2 6 September 1983 Dataquest Incorporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
APPUCATION DISTRIBUTION
OF CAD/CAM SYSTEMS IN
U 3 . AND JAPAN
1982
PERCENT
100
AEC & MAPPING 6%
PC 10%
AEC & MAPPING 25%
IC 20%
PC 14%
IC 10%
MECHANICAL 64%
JAPAN
MECHANICAL 51%
U.S.
Source: DATAQUEST
- 21 Copyright © 26 September 1983 Dataquest Incorporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
CAD/CAM
THE BIG ISSUES
OF THE '80s
- 22 Copyright © 2 6 September 1983 Dataquest Incorporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 RIdder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
TURNKEY COMPANIES
LEADING EDGE IN:
— GRAPHICS INTEGRATION
— GEOMETRIC MODELING INTEGRATION (SOLID MODELING)
— APPLICATIONS SOFTWARE DEVELOPMENT
•
NEW EMPHASIS ON THE "CAM" IN CAD/CAM
•
A BROADER LINE OF PRODUCTS
- 23 Copyright © 2 6 September 1983 Dataquest Incorporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
COMPUTER COMPANIES
•
PROVIDING ENGINEERING TOOLS FOR INSTALLED BASE
•
PROVIDING A
•
PROVIDING GOOD GRAPHICS HARDWARE SOLUTIONS
•
PROVIDING COMPANY-WIDE INTEGRATED SOLUTIONS
WIDE
VARIETY OF THIRD PARTY SOFTWARE
- 24 Copyright © 26 September 1983 Dataquest Incorporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
SOFTWARE COMPANIES
TRACKING WITH ADVANCES IN HARDWARE TECHNOLOGY
PROVIDING "HOOKS" TO EXISTING TURNKEY AND SOFTWARE
SOLUTIONS
RESPONDING TO THE NEED FOR PORTABILITY
- 25 Copyright © 2 6 September 1983 Dataquest Incorporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
ENGINEERING WORKSTATIONS
DEFINITION
— CPU DEDICATED TO SINGLE USER
— OPERATING AS A STAND ALONE, IN A NETWORK,
AND/OR INTERACTING WITH A HOST
— SPECIALIZED HARDWARE CAPABILITIES FOR ENGINEERING
TASTS (E.G., GRAPHICS)
— COMPATIBILITY WITH ENGINEERING SOFTWARE
- 26 Copyright © 2 6 September 1983 Dataquest Incorporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
ENGINEERING WORKSTATIONS
TRENDS
— CONTINUING PRICE/PERFORMANCE IMPROVEMENT
— DISTRIBUTED PROCESSING AS A WAY OF LIFE IN THE
ENGINEERING ENVIRONMENT
— MORE THAN JUST "HIGH TECH ETCH-A-SKETCH"
STANDARDIZATION?
- 27 Copyright © 26 September 1983 Dataquest Incorporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
NEW APPLICATIONS
•
ELECTRONIC DESIGN AUTOMATION (EDA)
I
ARCHITECTURE-^LOGIC-^CIRCUIT*^^^'*!^^
I
•
CAM
'
•
DOCUMENTATION AND PUBLICATION
•
DRAWING STORAGE AND RETRIEVAL
- 28 Copyright © 2 6 September 1983 Dataquest Incorporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen CoITipany / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
NEW SOFTWARE TECHNOLOGY
STRUCTURED DESIGN OF ELECTRONIC CIRCUITS AND SYSTEMS
SOLID MODELING
CIM
- 29 Copyright © 2 6 September 1983 Dataquest Incorporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
NEW HARDWARE TECHNOLOGY
THE "SUPERMARKET APPROACH TO TECHNOLOGY"
THE "CAROUSEL OF CAD/CAM"
- 30 Copyright © 26 September 1983 Dataquest Incorporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
START-UPS —
HAS THE FEVER BROKEN?
- 31 Copyright © 26 September 1983 Dataquest Incorporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
CAD/CAM AND WALL STREET
OLD FRIENDS
SOME NEW KIDS ON THE BLOCK
- 32 Copyright © 2 6 September 1983 Dataquest Incorporated-Reproduction Prohibited
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
DATAQUEST
CADICAM INDUSTRY SERVICE
SECOND ANNUAL CONFERENCE
T H E TECHNOLOGY CHALLENGE:
LINKING CAD/CAM WITH
INTEGRATED MANUFACTURING
•
•
•
•
PRODUCTION AND CAD
PRODUCT STRATEGIES
ELECTRONICS ISSUES
REALITY OF THE FACTORY FLOOR
- 33 Copyright © 2 6 September 1983 Dataquest Incorporated-Reproduction Prohibited
Dataquest incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
Dataquest
THE ROLE OF ADtOMATED TEST EQUIPMENT
IN THE ADTOMATED FACTORY
Harold McAleer
Senior Vice President
GenRad Incorporated
Mr. McAleer is a Senior Vice President of GenRad and General Manager of
its Canponent Test and Quality Management Divisions. A 30-year veteran
at GenRad, he has had a variety of engineering and management positions,
such as Manager of Custom Products, Vice President of Engineering,
General Manager of the Electronic Instruments Division in Massachusetts,
and General Manager of the Vibration Analysis Division in California. He
received B.S. and M.S. degrees in Electrical Engineering from MIT and is
a member of IEEE and Tau Beta Pi.
Dataquest Incorporated
CAD/CAM INDUSTRY CONFERENCE
September 26-28, 1983
Newport Beach, California
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
THE SOLE OF TESTING
IN THE AimXfATBD FACTORY
The industrial revolution is still raging all around us as we
struggle to satisfy the insatiable demand of society for the products of
technology—vehicles and machines and computers and video games. Several
waves of this revolution are taking clearer shape—the thrust to factory
automation to improve productivity and, of ever-increasing importance, to
improve quality. As John Naisbitt says in his book "Megatrends" the
industrial-machine
nature of
the
revolution
has
changed
to
an
information-processing revolution—and that, I believe is the key to
factory automation.
I hope to give you a view of factory automation f rom . the perspective
of the ATE industry—Automatic Test Equipment for manufacturing of
electronic equipment—and describe a few points of thought and language
to expand your own viewpoint, I hope to convince you that a network of
aut(^natic test equipment linked together with information processors and
a^lications
software can
provide
the missing
link
in
factory
automation—an integrated quality management system to join with the cost
and schedule systems to produce true improvements in quality and
productivity. But first a few words about ATE.
This slide shows the ATE market as a Subset of Factory Automation,
along with CAE and CAD/CAM.
The acronym CAT, Computer-Aided Testing,
would probably fit the jargon better.
The commercial ATE industry is an outgrowth of the older Test and
Measuronent industry which included laboratory instruments for electronic
engineers. These instruments found their way to the factory floor and
the industry boomed. In the late 60*s and early 70's the miniccnnputer
combined with electronic instruments to form automatic test systems and a
new ATE industry was born.
This figure shows past and projected growth rate for the ATE
industry. It doubled every two years through the seventies, slowed down
a bit during the recent recession, and is predicted to pick up again to a
rate of doubling every three years through the eighties.
The next slide shows the rough magnitude of ATE sales last year. The
slide also gives a view into the process for electronic manufacturers
which is not so different from the process of any manufacturer.
Ccxnponents are produced by ccmiponent manufacturers—in this case the
basic components are integrated circuits from the magic kingdoms of the
West, and empty printed circuit boards. These ccsnponents are combined
into assemblies and these into systems to be shipped to end users.
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Parl< Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
ATE is used at just about every step in this process in an attempt to
weed out defects before more value is added and to diagnose and repair
failures. Component testers are used by semiconductor manufacturers and
also by electronic manufacturers at incoming inspection, testers are used
for both bare and stuffed printed-circuit boards and also for finished
assemblies. After products are in use, testers are used to repair them
when they fail.
The fundamental motivation for the use of ATE has been relatively
simple economic payback. The yield in electronic processes has been
notoriously poor. Even after careful screening, burn in, and testing by
semiconductor manufacturers drop-out rates of 1 to 5 percent are not
uncommon at incoming inspection—depending on the maturity of the
component and its own manufacturing process.
If you stuff a PC board with 100 IC's from a population with 1
percent defects, statistics tell us that 64 percent of the boards will be
bad from that cause alone not to mention bent pins and solder blobs. So
manufacturers attanpt to cull out bad units as early in the process as
possible. The theme being "don't add value to bad parts." This is also
suminarized in the "rule of ten". It cost you about ten times as much to
find a defect as it moves from step to step in your process. Thus it may
cost 50 cents to find a bad component at incc»ning, $5 at board test, $50
at final test, and, heaven forbid, $500 to send a service technician to
your custc»ners' site—not to mention the embarassment and loss of good
will.
Because of the poor yield, ATE is used not so imich as an off-line
sampling-inspection tool to gather data on the statistics of the process,
but rather as an essential part of the process to get the darn things
working. Testing therefore is a major element of manufacturing costs and
a critical factor in throughput. I suspect that that is true in areas of
manufacturing other than electronics.
Changes are occurring, however, in the role of ATE (especially
regarding factory automation) caused by competitive necessity and new
converts to the religious wave of Quality. As I mentioned, the classical
role of ATE has been to find and fix bad product-defect detection. That
role is now being expanded to include defect prevention and we can look
forward to an even greater role in an integrated quality management
system. Here's how it works:
As I said, the automatic test equipment is used at several phases of
the production process. In keeping with the new language we refer to
these as workstations and we call the activity. Level 1 where the tester
finds the faults and separates the bad from the good. It turns out that
the tester gathers a lot of data that is often not used; the tester knows
more than we ask it. In fact, in a relatively low-yield operation, like
electronic manufacturing, the tester is the only thing that really knows
what's hai^ening.
Ot:her management systems—the cost system, the MRP
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
system, etc., are predicated on high and uniform yields. They assume a
"standard" cost, a "standard" lead time, a "standard" processing time,
etc.
These standards are corrected very slowly.
The result can be
significant pileup of excess inventory-components, faulty boards. But
the tester really knows the yield and the time, if we can only figure out
how to ask it the right questions and how to use the answers.
It has been said that most electronic manufacturers are running two
factories—one to make the good stuff, and one to make the bad stuff.
The cost to run the bad factory is wasted, since notihing comes out. If
the bad factory can be shrunk, tremendous savings can be realized. Enter
a new concept—Level 2, wherein data acquired by the tester is turned
into information about the process so that action can be taken to improve
the yield. The role of the tester (and an associated data processor} is
augmented beyond just finding faults—fault detection—to stopping the
faults—fault prevention. We call this activity Level 2, an autonated
work
center—a
cc»nbination of
automatic workstations,
a central
processor, and the proper mix of peripherals and, of course, software, to
provide information to control and improve the process.
The automated ATE work centers—islands of automation—are beginning
to appear more frequently and are producing effective results. We see
th«n in inconing-inspection departments where several cc»nponent testers
are networked to a common host computer to produce reports of inspection
yield, vendor performance, etc. We also see them in the printed circuit
test and repair center where board testers are linked to a host computer
and video terminals display instructions to repair technicians while, all
the while, 1:he host gathers information on the activity of the center.
We will likewise see similar automation in the depot and field-service
areas even as we see it already in the CAE/CAD centers of engineering
departments.
The trick now is to tie all these automated work centers together
into yet cmother level of the hierarchy that we call Level 3, Integrated
Quality Management Systems. In this new world, the ATE networks join
their fellow islands of automation and tie to the big-brother systems of
MIS to provide something new that wasn't there before—Essentially, we're
talking about moving information throughout all parts of an organization
in four possible directions. Backward, forward, up and down.
First, Feedback. In an automated factory, the quality monitoring and
management
system
is
an
essential
element
in
a
real-time
servomechanism-sensing and feeding back information on quality that is
vital to the productive operation of the entire manufacturing process and
the essence of automation. The whole purpose for automating the factory
is to achieve greater productivity, greater control, and greater
manufacturing econranies. Real-time quality management makes these goals
achievable.
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 RIdder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
Second,
feedforward.
Feedforward
means
to
move
information
automatically from one stage in the process to a later stage.
This
capability can avoid duplication of effort, avert errors, reduce time to
market, and increase the return on investment. A good example here is
what happens today in most operations before a new pc-board can be
tested. A test program must first be developed. Before that can happen,
the circuit description must be laboriously entered into the board
tester. With a CAD station linked directly to the tester, the circuit
description, already housed in the CAD station, can be fed directly to
the tester. It's faster, more accurate, and no manual intervention is
involved.
Let's look at the other two directions feed up and feed down. Here
information flows automatically up from production and down from
management. The result? Better overall business control. For example a
link between a higher-level local-area network and the test network can
provide a more direct and more timely connection between "plan" and
"actual" results:
you feed plans down, feed actual results up.
Information is more timely; results are more productive. This is where
the real link between quality, cost, and schedule happens.
Automatically feeding information forward/back/up/down
throughout
this organization is a capability that has begun to emerge, but clearly,
we're not there yet. Much more remains to be done before comprehensive
quality management tools can assume this central role in the automated
factory. Much depends on the growth of the automated factory itself, the
larger universe of which quality management is only a part.
George
Gagliardi of Arthur D. Little has cited two obstacles that have slowed
down implanentation to date.
First, most corporate managers tend to view short-term payback as the
only justification of capital
investment,
despite
the
long-term
productivity improvements that factory automation promises. We all need
to look farther down tJie road than next quarter's or next year's earnings
statement. Our successful counterparts in Japan are doing just that,
with spectacular results in several industries, including semiconductors
and consumer electronics.
And
second,
there
are
no
general
standards
for
integrated
computer-aided manufacturing.
This ccxnplicates the task of building
bridges between the many islands of factory automation that already exist
on the market.
As a result, it is both tempting and convenient to
wait-and-see attitude toward implanenting the autianated factory.
take
a
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
But another set of forces is gradually overcoming these obstacles,
and presenting opportunities that are pushing the entire electronics
industry toward greater automation:
•
As more companies adopt automated manufacturing, we all move
further down the learning curve.
The risks of obsolescence
diminish and the rewards for automation grow.
•
The trend toward shorter product lifetimes requires flexible
manufacturing systems to permit quicker implementation of new
designs and smoother transitions from prototype to production.
•
Increasing production costs, for energy, materials, and labor,
require better productivity to maintain profit margins. Better
automated quality management means higher product yields, which
in turn means lower unit costs.
The conclusion is: We can't affort to wait. Companies that are
movir^ toward automation now are the companies that will get quality
products to market more cost-effectively amd ahead of the competition to
establish and maintain leadership positions.
And this is not an
unprecedented development—it fits the historical pattern perfectly.
The automation of electronics manufacturing has taken giant strides
over the past 25 to 30 years, and developments in test instrumentation
have played an increasingly important role. As the level of integration
of electronic devices has increased, so has the level of integration of
the test equipment used to manufacture both those devices and products
that incorporate them.
As a result, the testing function has also
provided better and better information.
When the basic electronic device was the discrete component the basic
testing tool was the dicrete instrument.
It tested specific device
parameters, but it required manual sequencing of tests and manual
interpretation
of
results.
With
the advent
of
the
integrated
circuit—which is a system of ccmiponents—came the integrated test
system, which is a system of instruments. These systems tested many
parameters much more quickly than discrete instruments; they featured
automatic
sequencing
of
tests
and,
in
most
cases,
automatic
interpretation of results.
Today, we not only have components integrated into circuits and
circuits integrated into systems, we now have systems integrated into
networks of systems, distributed systems. The parallel phase in ATE's
evolution is the network of test systems—distributed testing. And the
scope of ATE's role has grown accordingly.
Where once the test
instrument was called upon only to check parameters in the development
laboratory, the role of the test network is no less than to monitor and
manage quality over the product's entire life cycle—from design to
production to field service.
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Parl< Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
It's a change in testing's role—and in your perception of that
role. Testing can provide a powerful lever for improvement—or it can
define the limit of how good the operation can ever hope to be. The
testers themselves will create the information that will fuel the process
to
turn
out
high-quality
products.
Strategic
decisions
will
determine—and
are
today
determining—how
to
leverage
that
information—how good that process can become.
And there's nothing futuristic about it. That's what we mean when we
say "the factory of the future ain't what it used to be."
It's here
today: conceptually, strategically, substantively.
I realize that many of rry remarks are like preaching to the
choir—you already believe. My purpose will have been served, however,
if, when you run into some brash members of the ATE fraternity thrashing
around in what you may have viewed as your exclusive territory you don't
think, "what are t±ose turkeys doing here—this town ain't big enough for
both of us", but rather, "oh, wow, welcome guys. What can we do to help
each other to bring the fruits of technology to this wonderful
marketplace of ours?"
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company/ 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
Dataquest
Dataquest
TRENDS IN DISTRIBUTED WORKSTATIONS IN CAD
Joseph F. Gloudeman
President
The MacNeal-Schwendler Corporation
Dr. Gloudeman is President and Chief Operating Officer of The
MacNeal-Schwendler Corporation (MSC). In addition to a bachelor's degree
from Marquette University and a master's degree from the University of
Southern California, he holds a Ph.D. degree in Aerospace Engineering
from the University of Stuttgart (West Germany).
Dataquest Incorporated
CAD/CAM INDUSTRY CONFERENCE
September 26-28, 1983
Newport Beach, California
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
TRENDS IN DISTRIBUTEU
WORKSTATIONS IN CAE
Joseph F. Gloudeman
The MacNeal-Schwendler Corporation
815 Colorado Boulevard
Los Angeles, CA 90041
INTRODUCTION
This
paper
addresses
the
expected
trends
in
engineering
workstations. This is done with some trepidation because in today's
highly dynamic computing environment the price, performance and technology of computing hardware ire changing so rapidly that even last
month's forecasts are somewhat obsolete.
Contribjting to the confusion of this highly dynamic environment is
the proliferation of acronyms and terminology.
In the interest of
improved communications, the first part of this presentation will deal
with some definitions of the terminology used, as well as a discussion
of the scope that the author is attempting to cover.
Next, a brief history of the technology related to engineering workstations is presented in order to place the recent sequence of events
into proper prospective. Emphasis is, of course, placed on events of
the past few years leading to the current status of engineering workstations. This background makes it possible to offer some guesses as
to the future directions.
For the sake of sanity and selfpreservation, some of the challenges and open issues that require
attention and correction will be presented in the event that hard
copies of this presentation come back to haunt the author more than
six months from now.
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
DEFINITIONS AND SCOPE
This presentation places emphasis on engineering workstations used
primarily to conduct engineering analysis.
This type of activity
generally falls within the category of Computer-Aided-Engineering
(CAE).
The diagram shows the relationship of CAE to the CAD/CAM
environment.
At the risk of over-simplification, let's assume the
primary thrust of CAD is in the area of Computer-Aided-Design
Drafting; it is not necessary to further discuss the scope of CAO/CAM
for this audience.
The Spectacular development of the microprocessor in the early 1980's
now provides the engineering community with a most attractive opportunity to use low-cost workstations for analysis including the further
integration of various engineering analyses. While representing one
of the most dramatic breakthroughs of the past 30 years of engineering
computing, this opportunity also presents a major challenge due to
both the large number of diverse computers on which today's analysis
programs operate and the differences in these systems and their operating system and applications software programs. The new engineering
workstations provide capability that was only provided by large processors in the mid-1960's and exceeds the power of the computers
available in the 1 9 5 0 ' s .
This explosive change in microprocessor technology comes at a most
opportune time for the engineering community.
For engineers, whose
main activity is analysis, the promise of a personal-level, highpowered computer seems to offer the best of all worlds - most of the
advantages of the large-scale processor without suffering poor turnaround because of the overloading of resources usually encountered in
the lailti-user environment. Microprocessors are also well received by
engineering management who continually look for innovative ways to
improve engineering productivity, especially in those situations in
which the computational process directly affects the engineering
activity.
However, because engineers tend to work as a team,
especially on large projects, there is a need to have efficient means
of communicating technical information from one engineer to another frequently outside of the local engineering organization. Significant
improvements in telecommunications show signs of resolving this difficulty.
The recent advances in local area networks (LAN), 'ire best exemplified
by those provided through coaxial cable connections or fiber optics
such as the Xerox Ethernet and the Apollo Domain network. These networks provide a network transfer rate in the 10 to 12 megabit per
second range.
The file transfer time is not noticeably different
whether the engineer is using a disk attached directly to his own
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
workstation or one somewhere else in the network.
This capability
allows the local workstation to have very rapid access to other
servers in the network which specialize in services that are too
expensive or too infrequently used to justify locally. These servers
are usually 1) A high speed computer server with a speed of ten to one
hundred times the workstation; 2) A nigh quality laser print server
with both higher speed and quality than is available on the local low
cost dot matrix printer; 3) A large disk server with capacity that is
one hundred to several thousand times the local disk; 4) A plotting
server that provides both larger sizes and higher resolution. Clearly
the disk server will also provide various data bases of engineering,
financial, and procedural information which includes computer programs
for general usage.
Software improvements have tended to lag the explosive advances of
hardware.
^ery little has been done to significantly improve programmer productivity via improved Higher-Order-Languages in real life
environments.
FORTRAN is still FORTRAN. Rather than dwell on other
shortcomings, which usually stem from the labor-intensive nature of
software development, it now seems appropriate to try to define the
basic software needed for both individual workstations and distributed
systems in general.
FUNDAMENTAL SOFTWARE NEEDS
Every workstation has a need for a fundamental set of generic software
that can be classified in seven classes.
These can be most easily
remembered by the use of the neumonic 'WCGOLDS which stands for:
W
C
G
0
L
0
S
-
Word Processing
Communications
Graphics
Operating System
Languages
Data Base System
Spread Sheet Systems
Each class is discussed b r i e f l y below:
W Word Processing.
Most engineers spend a l o t of time w r i t i n g
r e p o r t s , memos, l e t t e r s , e t c , both at home and in the o f f i c e . Many
c u r r e n t l y a v a i l a b l e word processing programs provide outstanding capabilities
for
both short
memos and long complicated
technical
reports.
The e f f e c t i v e use of these tools w i l l be a mandatory s k i l l
f o r e^ery f u t u r e engineering graduate.
Even the Apple offers such
extras as l e f t - and right-hand j u s t i f i c a t i o n , page f o r m a t t i n g , hyphena t i o n , and checking of s p e l l i n g .
The future o f f e r s syntax and gram-
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
matical correction. Also to be looked for is the ready incorporation
of graphics into text.
C - Communications. A fundamental job of most engineers is to communicate his designs and analysis to other people. In order to do these
tasks much information is required from the customer or other
engineers. A workstation with the appropriate communications hardware
and software provides the ideal tool for receiving and transmitting
technical information.
Graphical information still provides some
difficulties which will be discussed in the next paragraph.
G - Graphics. Low to medium resolution graphics is now available on
CRT displays and dot matrix printers at very low cost.
These very
recent developments have jumped ahead of both the two and three dimensional geometric model theory and the software to support this lowcost new hardware. Rapid progress in the software is being made and
many very useful programs for both engineering and business graphics
are now beginning to appear.
The new proposed CBEMA standard for
videotext called NAPLPS offers a very powerful tool for elementary
graphics displays.
Although the recently adopted GKS graphics standard may be useful in
providing an interim set of tools, it needs major revisions and will
probably be totally replaced.
0
- Operating System.
The operating system provides the primary
user interface and the utility programs for management of files and
programs.
This then makes software most important as it is the
foundation on which all other programs depend.
Most workstations
designed for the engineering market are providing UNIX-like operating
systems. The recently delivered PC/DOS 2.0 has most of the key UNIX
concepts and assures, by its popularity, that these will become the
basis for most operating systems in the late 19«0's.
L - Languages. FORTRAN has been the principal programming language
for engineers and scientists. Neither BASIC nor Pascal will replace
FORTRAN.
Ada remains an unknown quantity, but early iraplenentdtions
show some promise.
C is an interesting combination of structured
concepts and bit iiianipulation capability which provides both good
portability and efficiency. All future engineers should know FORTRAN
and a second language still to be determined.
D - Data Base System. The critical nature of data base systems in
engineering has long been recognized but such systems are still
primarily designed for business use. As a result, most systems do not
adequately support various engineering needs such as floating point,
graphics, vectors, matrices, e t c
Recent systems such as RIM and
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
INGRES have attempted to overcome some of these deficiencies. Most
engineers will develop their own data bases which will become as
important to them as their collection of technical books and papers.
This accumulation will start in college and continue throughout their
careers.
S
Spread Sheet System.
Manpower and financial planning is
something that most engineers do from the day they go to work until
they retire. Initially the planning may be very simple and only
involve individual efforts, but it usually grows quickly to a more
complicated activity.
Spread sheet systems provide a most effective
tools for handling this other>>ise tedious function. Recent programs
such as VisiCalc and Lotus 1-2-3 provide excellent tools for
performing this key activity.
The above discussion summarizes the seven pieces of software that Tiost
engineers should know and use frequently regardless of the nature of
their technical duties which are discussed below.
ENGINEERING DESIGN AND ANALYSIS
Engineers are usually engaged in either design or analysis.
Let's
consider the mechanical engineer since that activity is more closely
related to this conference. The design engineer typically receives a
conceptual input in written or graphic form depending on the status of
the project in the design cycle. This concept or preliminary design
should either be available on a disk to which the engineer has ready
access with the other engineers on the same project or transmitted to
a local disk storage if the project is small. Further detailed design
or analysis is performed with the results transmitted or made
available to others on the project. The critical concepts then are 1)
Who has the need for and access to information?; 2) What analysis
programs are available and should be used? A complex, multi-faceted
program like MSC/NASTRAN has great capability in the hands of a
Skilled Structural analyst. However, it must be available and understood by the involved engineers.
Further, the geometric model describing a part needs to be augmented with instructions telling the
engineers what analysis needs to be done prior to that ,jart being
accepted and released.
The three key concepts discussed in the above paragraph are 1) Tne
management of design information; 2) The management of the analysis
programs to be used and; (3) The management of the resulting output
information from the analysis programs.
Currently, much of this
information is available in various computers throughout a typical
high-tech company. Unfortunately, the incompatibility of the various
data formats, the lack of high-speed local area networks, and the lack
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
and use of good engineering data base systems make the i n t e g r a t i o n of
the information d i f f i c u l t .
Also few companies have yet to consider an
engineering workstation and associated network as a mandatory tool f o r
every engineer.
A BRIEF HISTORICAL REVIEW
The i n t r o d u c t i o n of the stored program computer in the early 19bU's
enabled engineers to b e t t e r perform t h e i r a n a l y s i s , especially in high
technology areas.
IBM's f i r s t stored program computer (the IBM 7U1)
was o r i g i n a l l y called the Defense Calculator since i t was intended
p r i m a r i l y f o r use in aerospace and related engineering indust:ries.
I n i t i a l l y these engineering computing e f f o r t s were highly fragmented
and offered v i r t u a l l y no computer intercommunications of engineering
analysis or i n t e g r a t i o n of the various analyses except through human
communications.
While the majority of computers used in support of engineering
retained t h e i r fundamental batch o r i e n t a t i o n , improvements were made
over the years to make the central computing f a c i l i t y more accessible
to the engineer.
For example, IBM's Time Sharing Option (TSO) came
i n t o being in the l a t e 196U's and did o f f e r improved a c c e s s i b i l i t y ,
but was not without i t s shortcomings.
I t was often expensive because
of the premium prices for on-line users set by the central computing
organizations.
The l i m i t a t i o n of the number of simultaneous users
affected both turnaround and a c c e s s i b i l i t y , especially during peak
usage hours.
This was f u r t h e r aggravated by s e l f - s e r v i n g abuses of
the log-on procedures.
The communication l i n e s , especially public
telephone l i n e s , were not always the most r e l i a b l e - and were overpriced for what the user actually received.
The i n t r o d u c t i o n of minicomputers for handling d i s t r i b u t e d timesharing environments and microcomputer-based systems as stand-alone
workstations represented an attempt to resolve the extant problems.
These systems, which s t a r t e d to appear in the early 197U's, suffered a
common problem of inadequate processing power, word length and
addressing, data handling l i m i t a t i o n s , weaknesses in t h e i r operating
systems and programming languages, and lack of low-cost passive and
i n t e r a c t i v e graphics. Many of these deficiencies were overcome in the
l a t e I970's as 32-bit minicomputers made great s t r i d e s .
T^nis brings us to today's environment in which such technologies as
the Motorola 68000 chip have made possible very powerful microcomputers at \/ery reasonable costs, as discussed e a r l i e r .
FUTURE ENGINEERING WORKSTATIONS
First,
let's
discuss the hardware required - CPU, memory, disic, CRT,
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
keyboard, and printer. Thirty-two bit microprocessors with a speed of
one million operations per second will be available in 1984. This
then provides each engineer with the capability of today's best
selling superminicomputer at his desk. A one to two megabyte memory
should provide more than enough capability for the individual workstation. Winchester disks of forty to several hundred megabytes provide adequate local read/write storage when combined with removable
cartridges of similar capacity.
Such units are just becoming
available at costs of one to five thousand dollars for the smaller
units with larger, more cost effective units to be available next
year. CRT's with low to medium resolution are available now at under
one thousand dollars with high resolution (1024 x 1U24) at two or
three times that cost. The costs for the high resolution units should
drop in 1984 and 1985 as higher resolution units become available from
commercial television.
Keyboards will be supplemented with mouses
this year and will make the input of graphics and the general interfacing by engineers easier.
Low-cost dot matrix printers with
graphics are currently adequate for most engineering work and improvements in the next year or two will make quality better with no significant increase in cost.
In addition to the above hardware, communications interface hardware
is required.
This hardware is now available and the cost will be
significantly reduced as better microprocessors become available.
By as early as 1985 an engineering workstation with all of the above
key features should be available for a cost of around five thousand
dollars. Most future engineers will expect one, and engineering companies will tend to base their activities around the workstation and
associated network.
Engineering software costs will not change as rapidly due to the
smaller market. The more general software as described above will be
very low in cost as some of it is today; but, the more limited
specialized engineering analysis programs will not significantly
decrease in price because of the more limited distribution and great
complexity of the code.
However, compatibility between various
engineering software will be greatly enhanced by the establishment of
various standards such as IGES.
In finite element modeling, the semi-automatic generation of neshes
from geometric models will begin to make some inroads. This process
will require good human interfacing to allow the adjustment of the
automatically generated meshes in a rapid and effective manner. Such
code can be written but will require niich experimentation to make it
cost effective.
Recent
advances
in geometric modeling
research
give hope that solid
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
modeling is the wave of the future for defining mechanical parts and
assemblies.
As industry begins to design parts using geometrical
modeling
systems,
the semi-automation
of finite element
mesh
generation becomes more desirable. Recent research in octrees provide
hope that this type of process is economical and technically feasible.
CURRENT EXPERIENCE IN USING WORKSTATIONS
During the past several years, both Hughes Aircraft Company and The
MacNeal-Schwendler Corporation have been studying and evaluating
various workstations.
Included in this evaluation have been the DEC
350, DEC Rainbow 100, Dynalogic Hyperion, Altos 68000, Apollo ON 420
and ON 300. Apple Lisa, NEC ARC, Compaq, IBM PC, Onyx 3002, HP 9816,
HP 9000, Columbia PC, SUN, and Victor.
To date two units have stood out as candidates for finite element
use:
the Apollo Domain and the IBM PC. The Apollo ON 420 is a
computer in the one quarter megaop class.
It has four outstanding
features: 800 x 1024 hi-speed bit mapped mapped graphics, 12 megabit
per second interprocessor communications, virtual memory, and floating
point hardware.
Its disadvantages were only noise, size, and cost.
The Apollo DN 300 remedies significantly
improves all three
deficiencies and should be an outstanding engineering workstation when
the floating point hardware becomes available in the fourth quarter of
1983 at a cost in the $20,000 range.
At a significantly lower cost the IBM PC and the various IBM lookalikes offer outstanding value with lower capability.
The lack of
virtual memory, the low resolution graphics and the the lack of good
support for the recently announced 8087 floating point chip are major
defects in an otherwise most-promising workstation.
These defects
should all be overcome by late 1983 and this will then allow the
engineer to have a truly outstanding low-cost workstation. The speed
of the unit will probably be improved by the new Intel 286 microprocessor in 1984. The price of the improved unit is not expected to
increase since the cost of the microprocessor and associated chips are
less.
CONCLUSIONS
In the next five years, many engineers will have a workstation in
their office and one in their home. These workstations will be of the
one million instructions per second (mips) class with a one-tenth to
one-fifth of a million floating point instructions per second
(megaflops) capability.
They will have color graphics with a
resolution of approximately 1,000 by 1,000 pixels on a 15 or 17 inch
CRT.
The primary memory will be one or two megabytes with a
Winchester-type disk providing another 40 to 100 megabytes of
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
auxiliary storage. This workstation will include a color dot matrix
printer with a resolution of better than 250 dots per inch. There is
some Strong indication that this could become available sooner than
earlier anticipated.
The communication capability will include a local area network with a
Speed of at least several megabits per second for distances of up to
several thousand feet and a global network in the range of more than
100 kilobits per second. Included in this local area network will be
computer engines with speeds that are 10 to 10,000 times the speed of
the local workstation. Also available in the LAN will be large disk
farms with large data bases which will include graphic images and
geometric models. High quality, high-speed laser printers will also
be available for volume printing.
The current heterogeneous unfriendly interface problem will hopefully
have been eliminated by integrating interface programs and use of
mouses and menus will make the use of the workstation easy and fun.
In the late 1980's we will begin to see the use of techniques now
being developed in Artificial Intelligence research. The most significant of these seems to be those concerned with machine learning. By
this is meant the modification or construction by program of stored
information structures, so that the machine-deliverable information is
more accurate, larger in amount, cheaper to obtain, or some combination of these.
The implications that the computer will ijrovide
some part of the engineering analysis in a more accurate, rapid, and
cost-effective manner are staggering. The current research leaves no
doubt of this ultimate capability, the question is can we learn to
integrate this with effective use of human engineers and to benefit
mankind through its use.
The challenges of both using workstations effectively in the near-term
and then to incorporate A.I. techniques into these provides a
tremendous technical and management challenge for the rest of this
century.
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
o
CO
W
J2
c
CD
W
TRENDS IN DISTRIBUTED
o
g
•3
WORKSTATIONS IN CAE
o
u
(D
'•
a.
>
c/)
c
CT
c»
a
lu'
•<
o^
^
z
2!
CO
(D
o
o
3
I
1983 CAD/CAM INDUSTRY CONFERENCE
•
Newport Beach, California
September 26 - 28, 1983
CO
o
a.
Q.
CD
-a
0)
I
I
o
o
§
o
c»
CO
-J
(O
o
8
2.
CD
CO
Joseph F. Qloudeman
President
The MacNeai-Schwendler Corporation
815 Colorado Boulevard
Los Angeles. California 90041
•
o
u
c
• I .1 |I.»P ...BU I • .11 IIJ..L. ....I.I..IHI1JJI.1 Jllljll ..i||IH>Bprpi|W|pw|WU..I I..!Ill .1L. II.. J i . • Iijl • I I
I Flgi.pi I H.I1 .UKII
(J
tn
o
o
•D
O
CD
Q.
w
c
CT
en
a!
u'
><
o^
>
b
The Role of the Company *s Products in
Computer^Aided Enghieering
REVISION
<L
,-,„.-_.
U
<j>
CD
3
O
o
3
•D
0)
DESIGN
^ ^ ^pwr^
ro
(O
o
I
f
(CAD)
V
\
/
ENGINEERING
ANALYSIS
• : V-'
••
./,
*•*.
•
^.V
—-MK
(CAE)
/
':
MANUFACTURING
(CAIVI)
"X^
\,
O
(f>
0)
3
C_
o
u>
to
O
Company Products
>
o
00
to
MSGMESH IMSC/NASTRAN MSC/GRASP
MSC/GRASP
Output
Data
--I
(O
o
o
o
2.
Input
Preparation
Display
Processing
(D
u
" * ' ^ i . ' . . ! ' " i . r » : " i ' ' . ' ' r ' " * ' . ' ^ . ^ j " ^ « " •»:*.•••••*." \ 'M
'
c
(/D
E £
O
C/)
TO
*-
D O
C/) O
'S
H
Jr. CL
PTO-^ E
ro o ro o
- I c/) S O
CL C
f^
CJ
i£
. =J
3
GL
Q) £ •
C/3
en
»-
2E
°- ^
3
-.-/i
.y o
O
L
O) O
CO
04
00
o>
\
4-^
o
c/D
CO
O)
Q
•
•
rCO
CJ
o
fs.
o>
CO
r*.
K,.
O)
CM
O)
'«w
X
o
-^
o
CNi
in
CN
1 -
o
o
to
o
o
rvi
^
o
o
o
in
O)
SISOO 3av/v\aavH wgiSAS
<
o
U
HQI
wi^
LU
I
o
-h
o
• o
?<
O H-
il)^
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
o
B
U
n
c
3
(D
O
O
•5
o
(D
Q.
>
w
c
CT
(0
d
0)'
•<
o_
TRENDS IN COMPUTING SYSTEMS
>
b
m
(S
o
o
3
•o
D)
3
8
d
Q.
tD
T)
D)
jr
O
CO
0)
o
>
CO
Ol
.b
o
3
(O
(D
O
O
O
2.
<x>
X
<o
-.J
w
CPU SPEED
MEMORY SIZE
STORAGE DEVICES
TRANSMISSION AND DISPLAY
COMPUTER HARDWARE COST TRENDS
SUPPLY AND DEMAND TRENDS FOR
PROGRAMMERS/ANALYSTS
to
2 ^ 9 b * l to
O)
cnO
UJt-
o
o
V
A
- eo
X - . T"
-JO
J O
c*
go
Q-Z^
Ho
<o»
Q.O
<^ z
O
o
u
^ am
00
O)
^
P-4
UJ
o
U
5*
1^
o
.00
oc
^ •ie
c
o
o
(O
b.
CD
"
o
Q
u
n9
n
o
Q
O
tn o
o
to
o
o
to
t«» O
to
U
S
Q
U
eg
1
!• ! •
z
o o
o
o
• = o^
eo r<e in
-
<
lU
>
N.
s
1
o
m
00
(O
O)
• sOD
0)
eo*
CM
CM
1
o
n
1
o
«
1
o
N.
o>
($) ..aor QUVONVISM V JO I S O O
Dataquest Incorporated, A Subsidiary of A C . Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
o
£.
E'
c
T3
O
(D
a
>
w
c
WORK STATION CONCEPT
IT
a.
Bl'
•<
>
b
ECONOMICS OF MICROPROCESSORS
2.
w
POWERFUL COMMAND LANGUAGE (USER ORIENTED)
ID
o
O
3
•D
a>
COMPILERS FOR HIGHER LEVEL LANGUAGES
•<
N>
(O
O
TEXT PROCESSING (MECHANIZED REPORT GENERATION)
Q.
a
a>
TJ
0)
TECHNICALLY ORIENTED RELATIONAL DATA BASE
MANAGEMENT SYSTEM
CO
01
GRAPHIC FUNCTION
CO
U)
_<D
O
>
WINCHESTER DISK TECHNOLOGY
CO-AXIAL CABLES
.tv
00
o
to
U3
O
O
O
o
OJ
o
g
•o
o
(D
a
>
CO
Q.
9J'
ROLE OF ANALYSIS IN DESIGN
>
b
REQUIREMENTS,
GUIDELINES
<*
w
n>
O
o
3
i
•o
B>
DESIGN
8
I
a
a.
(D
ANALYSIS,
TEST
TJ
Bl
C/5
S>
C-
o
NO
GOOD
1
EVALUATE
(/>
(D
O
>
U3
cn
o
CB
§
o
o
!5.
(0
ID
•~i
u
I
GOOD
OK FOR
PRODUCTION
o
u
ST
13
C
(D
(A
5"
o
o
•3
o
(D
Q.
>
w
c
cr
9^
a>'
•<
o_
<A
EXISTING TOOLS FOR ANALYSIS
>
b
o
I
a POWERFUL COMPUTER HARDWARE
— HIGH SPEED PROCESSORS
— LARGE MEMORY, STORAGE DEVICES
— FILE MANAGEMENT
T.
0)
— REMOTE TERMINALS
* WIDE SELECTION OF ANALYTICAL PROGRAMS
CO
D)
c_
o
CA
O
>
(C
01
.».
o
00
<o
•>l
<o
o
o
o
5.
n
X
->l
<o
OJ
— FEM POPULARITY, GENERALITY
D
to
n
c
<ti
(A
A MAJOR CHALLENGE IN ANALYSIS
^
<0
a
>
w
c
cr
(A
a
0)'
TOOLING, ETC.
INHIBITS DESIGN
CHANGE
>
b
o
o
3
•o
0)
to
8
J]
DESIGN
DATA
'FROZEN'
SOME DESIGN DATA AVAILABLE
.-uofSs.-*^
a
Q.
(D
• ^
T3
01
-1
?r
O
w
0)
3
o
>
to
Ol
o
00
CO
o
o
o
(D
START
DESIGN
V
START
PRODUCTION
o
ST
c
<a
DYNAMIC ANALYSIS GROWTH TRENDS
(0
o
o
•5
o
(D
NO. D.O.F.
1,000.000
a.
w
c
CA
cL
0)'
100,000 -
>
b
10,000 - (D
3
o
o
3
•o
lU
1,000 - -
M
<o
O
a
Q.
(D
100 ••
T)
L^_
Q)
CO
B>
3
SUPER
»ER _
COMPUTER
SUPER
MINICOMPUTER
IP : :
DIGITAL COMPUTER
o
>
<o
OI
1 'I
1940
o
CO
(O
o
o
o
2.
ta
•<
u
;[-«^ANALOG COMPUTER
CALCULATOR-H
I---J
1
1950
1900
d
—PUBLIC PROGRAMS-
\
1
1970
1960
1990
D
ST
.o
c
n>
(A
•o
o
(D
a
w
c
cr
CO
a.
lu'
>
b
o
HORIZONTAL INTEGRATION
FINITE ELEMENT METHOD USED IN MANY
APPLICATION AREAS
o
3
•o
I
N)
(O
O
I
I
I
I
I
'I
o
00
CO
8
o
'
-
STRUCTURAL ANALYSIS
HEAT TRANSFER
FLUID MECHANICS
ACOUSTICS
ELECTROMAGNETIC FIELDS
AEROELASTICITY
FLUID - STRUCTURE - INTERACTION
NON-LINEAR MECHANICS
o
ffi
c
(D
0)
O
O
-^
•o
o
(»
a
>
c
CT
Cfl
Q!
0)'
-^
•<
o^
>
b
VERTICAL INTEGRATION OF APPLICATIONS
(t>
o
o
3
•D
0)
ro
(O
O
I
I
O
Cfl
ffi
J
>
Ol
J^
o
00
(O
o
o
o
(D
X
(O
00
• BETTER INTERFACES BETWEEN DESIGN AND ANALYSIS
ORGANIZATIONS
— DIGITIZED GENERATION OF GEOMETRY AND TOPOLOGY
— GEOMETRY MODELING (SOLIDS)
• FINITE ELEMENT METHOD OPPORTUNITIES: PRE- AND
POST-PROCESSORS
— ACCURACY, PRODUCTIVITY
o
(U
(t>
0)
o
o
•5
o
AN INTEGRATED SYSTEM
(B
a
>
w
a.
(U
>
b
PRE-PROCESSING
ACTIVITIES
POST-PROCESSING
ACTIVITIES
GEOMETRY
DESCRIPTION
RESULTS
VALIDATION
MESH
GENERATION
STRESS
DISPLAY
LOAD
GENERATION
DEFORMED
MODEL
re
3
o
o
3
•D
0)
to
(O
o
ao.
(B
T3
EU
W
0)
C_
O
(A
<D
O
>
(£)
Ui
FEM ANALYSIS
MECH. MAT'L.
PROPERTIES
i
CONTOURS
i
NODE
NUMBERING
SECTIONAL
CUTS
MODEL
EDITING
REPORT
GENERATOR
?
1
}
'
DATA
MANAGEMENT
,
PROJECT X
I
<o
o
o
o
2.
(D
-xl
CD
\
o
£.
»
£1
C
CD
CO
S
o
o
•o
o
ID
_a
>
w
c
CT
0)
a
5'
>
b
Z
I
o
3
»
><
•o
l>J
CO
o
I
i
Q
cn
o>
3
C-
O
TRENDS IN INTERFACING CAD AND
ANALYSIS METHODS
PHASE I: STANDALONE
• EACH APPLICATION "INDEPENDENT*
• SUBSTANTIAL MANUAL INTERVENTION + ERRORS + DELAYS
PHASE II: INTERFACING KEY SYSTEMS
• SOME COMMON DATA
• MECHANICAL TRANSLATION OF MUCH DATA
• MANUAL INTERVENTION TO ACTIVATE COMPUTER PROGRAMS
PHASE ill: THE SEEDS OF INTEGRATION
• MUCH COMMONDATA
• SOME MECHANICAL TRANSLATION
• SEMI-AUTOMATED PROGRAM ACTIVATION
CO
CD
O
>
CO
en
I
CO
o
o
o
a_
CD
X
-J
CO
CO
PHASE IV: INTEGRATED SYSTEMS
• COMMON DATA BASE (DBMS)
— CONFIGURATION MANAGEMENT
— CONSISTENT DATA AND RESULTS
— EASE OF ACCESS FOR QUALIFIED
USERS. PROGRAMS
o
SJ.
rt
c
(D
Cfl
o
O
-1
•o
o
C/3
c
cr
tn
a
&>'
>
b
(D
o
Distributed Engineering Work Stations
TELEPHONE
SATELLITE
NETWORK
PRINT
SERVER
(DISKS)
WORK
STATION
O
3
DISK
SERVER
•D
fi)
c
(
GATEWAY
d
c
(B
GATEWAY
WORK
STATION
(DISKS)
TJ
01
HI-SPEED
COMPUTE
ENGINE
Cfl
Q)
ETHERNETS
o
>
<o
Ul
o
00
ID
CO
o
o
o
CD
(O
u
son
DISK
SERVER
PRINT
SERVER
I
I
WORK
STATION
WORK
STATION
WORK
STATION
o
£.
c
CO
O
O
•5
o
(t>
Q.
>
Cfl
c
CT
(/)
9!
0)'
>
P
Z
2.
CHALLENGES IN DATA INTERCHANGES
(/>
CD
3
O
o
3
•o
0)
3
l<0
8
U
s.
a.
• LARGE VOLUMES
~
BILLIONS OF DATA ELEMENTS
• TIME LAG BETWEEN DESIGN AND ANALYSIS
(C
TJ
a>
V>
0)
O
>
a>
tn
o
00
to
o
o
o
(D
(O
--J
u
• DESIGN CHANGE ACTIVITY
— CONFIGURATION MANAGEMENT
• PROJECT LEADER VISIBILITY AND CONTROL
— RECORDKEEPING
— ARCHIVING
a
2J.
BT
n
c
<I)
CA
3"
o
o
•5
o
S
<B
_a
>
CO
c
cr
w
a
D>'
•<
>
in
z
<p
0)
(D
i
•o
0)
3
ro
<o
o
a.
a.
(I
TJ
0)
SOME OF THE CHALLENGES IN
COMPUTER SYSTEMS
DATA BASE MANAGEMENT SYSTEM
—
—
—
—
RELATIONAL/NETWORK DATA BASE
GRAPHIC AND GEOMETRIC REPRESENTATION
EFFICIENCY, ACCESSIBILITY
CONTROL
• DISTRIBUTED ENVIRONMENT
— DIS-SIMILAR PROCESSORS (E.G. WORD SIZE. ARCHITECTURE)
— MICROPROCESSORS
^ SUPERCOMPUTERS
W
0>
:)
c
o
w
p
O
>
U3
Ui
O
00
(O
CO
O
O
O
2.
0)
CJ
WORK STATIONS!
o
ffi
ST
n
c
CD
O
O
-^
•o
o
Q.
AREAS REQUIRING DEFINITION,
DEVELOPMENT
w
c
CT
U)
a
tu'
>
b
« TECHNICALLY-ORIENTED DBMS
~ LOGICAL SCHEMA DEFINITION AN OPEN ISSUE
(D
o
o
3
••
0)
• ''STANDARD" INTERFACES
— MULTI-VENDOR COMPUTERS
— GRAPHICS DEVICES
N>
(O
O
a.
a.
CD
-0
%
p
i
I
g
§
o
00
<o
o
o
o
(D
CD
•
UNIVERSAL" NETWORKINO, TELECOMMyNICATIONS
«
• USER-ORIENTED LANGUAGESi STANQARDB
— DATA HANDLING
— GRAPHICS
— GEOMETRY
o
£.
fo
£>
c
CD
3
O
o
•3
o
(D
a
>
w
c
cr
o>
a
SOME ALTERNATE PATHS
0)'
o_
>
b
I
• INTERMEDIATE GRAPHICS EXCHANGE STANDARD
o
i
— EXTENSION TO NEUTRAL DATA BASE
0)
3
I
— TRANSLATION/TRANSMISSION OF DATA
<£>
O
a
a
— NEED PEOPLE, MONEY, SPONSOR
(D
TJ
D)
NASA'S I PAD
W
B>
C_
o
DOD'S ICAM
W
O
>
o
00
CD
O
O
O
2.
(D
(O
OJ
-
RFPFORPIK^DUCTC^^INmONDAm^YSTEM
#
a
to
n
c
a
CO
5"
o
o
•5
g
(D
p.
>
0)
c
AN ARGUMENT FOR INTEGRATION
CT
(A
a
m'
•<
o^
(PRE- AND POST-PROCESSORS)
>
p
Z
2!
CA
CD
I
1
•<
K)
CO
O
a.
a.
CD
Tl
B>
05
0)
3
c-
o
en
CO
O
>
CO
01
o
(»
CO
o
o
o
CD_
(0
CO
~g
u
% OF
COST
^
^
% OF
SCHEDULE
.
PREPARE INPUT
50
60
COMPUTER PROCESSING
30
20
INTERPRET,
DOCUMENT RESULTS
20
20
::aDataquest
Dataquest
ROBOTICS—TODAY AND TOMORROW
Stephen P. Volm
Western Region Manager
Advanced Manufacturing Systems
IBM Corporation
Mr. Volm is Western Region Manager for IBM's Advanced Manufacturing
Systems
business
unit.
AMS
has
responsibility
for
developing,
manufacturing, and marketing IBM's robotic systems. During his 14 years
with IBM he has held various marketing and management positions. He has
concentrated
in
marketing
manufacturing
control
systems,
plant
automation, and CAD systems.
Prior to joining IBM, he held several
design engineering and factory management positions with International
Harvester, Mr. Volm received his degree in Mechanical Engineering from
Marquette University.
Dataquest Incorporated
CAD/CAM INDUSTRY CONFERENCE
September 26-28, 1983
Newport Beach, California
Dataquest Incorporated, A Subsidiary of A.G. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
ROBOTICS TODAY & TOMORROW
STEPHEN P. VOLM
I B M CORPORATION
ADVANCED MANUFACTURING SYSTEMS
SAN DIEGO, CALIFORNIA 92108
In February 1982 IBM formed a new organization called Advanced Manufacturing Systems to develop, manufacture and market IBM's robotic
products.
The organization was set up as an independent business unit which operates independently of any existing IBM marketing organization. Our
organization consists of a General Manager who reports to a Board of Directors
made up of IBM group executives.
The business unit is still apart of IBf, but
its independency gives it an opportunity to react to changing market conditions with this new technology.
IBM's internal automation group
(SEDAB-Special Engineering Design And Build) is also part of our AMS organization.
This group has responsibility for both hard and flexible auto-
mation projects at IBM plant sites worldwide.
This adds additional skill to
our new organization that is \/ery helpful when working with prospects on
flexible automation requirements.
The AMS organization is headquartered at
the IBM Boca Raton, Florida facility.
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
This site was chosen because it is the location of our manufacturing organization that builds the Personal Computer and the IBM Series 1 Computer
which drive the robotic products.
In addition, our SEDAB organization is also
located at this facility.
The interest and use of flexible automation is accelerating rapidly in
the United States.
However, when compared to the robotic usage in Japan, the
United States is very far behind.
Recent surveys have estimated that Japan
has three-fourths of the world's robots installed. They have been implementing
robots for the past ten years or longer, and are now implementing computer
integrated factories.Increasing interest in this country can best be described
by recent attendance at the National Robot Shows.
people attended the National Robot Show.
Two years ago, about 4,000
The Robot VII Show this year in
Chicago had over 20,000 attendees.
There are many reasons why companies must begin to make widespread use of
robotics and flexible automation.
Competition has increased significantly en
a worldwide basis over the past ten years.
If you examine some of the causes
of increased competition particularly from overseas businesses, you v/ill find
that the expenditures for plant modernization and flexible automation have far
exceeded those in the United States.
This has increased productivity at a
rapid rate in other countries and created a condition where our productivity
improvement and output per manhour in the United States now lags all major industrial nations along with the United Kingdom.
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408)971-9000 / Telex 171973
In absolute terms, the United States still leads the rest of the industrialized world in productivity, but in productivity growth, we are last. This is
a trend that must be reversed in the near term if we are to regain our industrial leadership.
been about 2J%.
Our rate of productivity growth over the past decade has
Japan's has been 7.3%.
Much has been written about large
corporations in the United States beginning to take an aggressive attitude
towards automation.
General Electric, for example, is building one
of the world's most advanced locomotive factories in Erie, Pennsylvania.
Using new manufacturing technology, robots and computer control, the new
factory will bring a dramatic gain in productivity.
Where it now takes 68
skilled machine operators 16 days to build a locomotive motor frame, the new
factory will turn them out in one day with one-tenth of the operators.
Japan's Yamazaki machine works is one that has been written up in a number of
publications.
They are building a new 15 million dollar machine tool parts
plant in Florence, Kentucky.
This factory will be one of the world's first
entirely automated factories, and when they complete it in 1983, it will
require 6 workers to tend to the master controller and the robots in
the factory.
Continued inflation creates additional justification for robotic opportunities into the decade of the '80's.
Labor costs have risen about 48
over
the last 6 years in the United States, and that trend is expected to continue.
Robots can help stabilize these costs.
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
In many businesses we have seen a growing requirement for improved
product quality.
low quality.
Items made in Japan used to be ^ery
inexpensive and of
Today that trend has been reversed where their products are now
the Standard of quality.
The demand for quality will be achieved with flexi-
ble automation and robotic systems that can produce with repeatability and
impact on the overall product quality.
We have also found that many of the assembly operations in the United
States are being switched to small batch lot assembly.
With the recent rising
cost of carrying inventory and the volatile market conditions, it is ^ery
difficult to predict volume requirements of any one product in a time period.
As a result, manufacturers are shifting their operations to produce in
smaller lots to reduce the amount of inventory and be more responsive tc
market fluctuations.
Programmable robots give companies a tremendous
opportunity to change lines and manufacture multiple products over a single
flexible automation line.
The changeover of these lines that are designed for
flexibility is minimal and permits manufacturers to produce in smaller lots.
The evolution of the robotic industry can be traced back to the initial
implementation in the early '60's of mechanical arms.
computer controlled robots.
sensory feedback robots.
an automated factory.
By rrid 70's there were
In the early '80's we saw the introduction of
By 1985 the robotic systems will be integrated into
IBM has focused its efforts on the intelligent robots
with sensory feedback and integrated systems for assembly and test.
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
This is a very
new area for robotic automation because it requires sensory
capability in the robot and requires the computer that is driving the robot
to handle many other tasks besides controlling the manipulator.
What are we trying to achieve with robotic automation, and why is this an
advantage?
If you examine the opportunities for automating assembly prior to
intelligent robots, there were basically two options: 1) manual labor; and 2)
fixed or hard automation.
taught, very
Manual labor was the most flexible being easily
adaptable to change, required little tooling, and would auto-
matically inspect its work.
The negatives associated with manual labor were
fatigue, quality could vary, speed could vary, and costs were increasing.
With fixed automation, it was s/ery repeatable, reliable and was usually
fast.
very
However, in many cases, it was built for a single purpose; you could
not vary the product; it was not flexible; and it was expensive compared to
flexible automation.
The key question that we are asking is how do we achieve
breakthroughs in productivity where we can get the best of both manual labor
and fixed automation.
Robotics fits in this middle ground between manual
labor and fixed automation.
Intelligent robots offer an opportunity to
realize the advantages of manual labor and hard automation while eliminating
the negatives associated with each.
IBM has addressed the assembly market with two types of intelligent
robots.
The 7535 is an intelligent robot which is quick and precise.
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
The 7565 is a wery
intelligent robot with tactile sensing that can adapt to
its work environment.
The key to intelligent robots is the software language.
IBM developed the language used to drive its robotic products called AML
in the early '70's.
This language was put to use with robotic systems in the
IBM plants in the mid '70's and has been continually enhanced since that time.
An intelligent robotic language must have several capabilities which AML has,
such as: robot control, data processing capabilities, calculational capabilities, debugging, program development, error handling.
language must be extendable.
In addition, the
AML has the capability to design your own
interface; you can develop a higher level language within the AML language to
customize subroutines for your own use; and IBM has added a high level
language for manufacturing engineers called PROBE which stands for Programming
Robot By Example.
This easy-to-use language allows people who are not prc-
gramming professionals to gain access to the system and do productive
work.
AML is a single step interpretive language which allows you to make
program changes line by line and execute them.
This interpretive language
capability is \/ery important in making it easier to put robotic applicatiors
in production.
The 7565 and 7535 basic systems were intended to cover the spectrum of
assembly operations.
The 7535 is used for simple assembly operations and can
be used to service many hard automation devices such as loading and unloadir^g
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive/ San Jose, CA 95131 / (408)971-9000 / Telex 171973
auto insertion equipment, test stands, etc
The 7565, on the other hand, is
intended to address the more complex assembly operations where tactile sensing
is required to duplicate the assembly process.
The types of applications per-
formed by these robots are stacking, component insertion, fitting,
palletizing, testing, load and unload, parts handling, kitting, and light
metal removing and fabrication such as drilling, tapping, deburring.
IBM developed its robot technology and the AML language in 1972 and 1973
and installed the first prototype system at Lexington and Poughkeepsie in 1976
and 1977.
This first application was used to test the back panels of the IBM
large scale processors.
There were about 30,000 pins on these back panels
that had to be tested for continuity.
each system.
It took about 92 hours to test
We put 2 arms in the 7565 frame and rotated it 90 degrees and
did the probe testing on those panels in about 14 hours with the robot.
Since
the first system was installed in 1976, we have continued to expand this testing system.
IBM's latest large computers are 100% tested in a cell that now
consists of 16 2-armed 7565 robotic systems.
Many other tests are now per-
formed on circuit quality in addition to circuit continuity.
The major
benefit from this system was improved testing which made a major contribution
to improved product quality. It has also produced substantial labor savings.
The new modules are tested in about 3 hours and are performing many more test?
that would have been uneconomical to do manually.
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
The IBM 7565 adaptable robot system has a workspace about the size of a
workbench.
The longest axis is about 58 inches in length with an 18 inch
X-axis and a 17 inch Y-axis.
It is driven by the IBM Series 1 Computer that
has the capability of storing up to nine million bytes of data.
printer are available for programming and operator control.
A CRT and
This system
is a hydraulic unit with a speed of 40 inches per second and has 6 degrees of
freedom (X, Y, Z, roll, pitch and yawl) in addition to the gripper. The
linear hydraulic motors are a special design and operate along the parabolic
cam for each linear axis.
They have proven to be very reliable in all the
installations both in IBM plants and in customer locations.
gives us a very
The configuration
high force to wait ratio and a very reliable system that can
adapt to its surroundings.
The system's adaptiveness is due to the strain
gauges that are built into the fingers of the gripper that detect X, Y and
Z forces.
These highly sensitive gauges that can detect force in 10 gram
increments along with the software that monitors the pulse beat of the robot
every
20 miliseconds are the basic ingredients that allow the 7565 to adapt to
a variety of changing conditions that occur in its working environment. For
example, if a drift builds up in the tooling of the fixtures, the robot can
sense this and take corrective action and then re-register all the points of
the new location to accommodate for the drift.
In addition to the tactile
sensing in the gripper, there is also a light emitting diode that allows
presence sensing to occur between the fingers.
This feature is very useful in
calibration.
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 /<408) 971-9000 / Telex 171973
The operator merely has to move the gripper near the calibration post and
execute a subroutine called "Find Post", and the 7565 through an AML software routine will calibrate itself on the alignment post.
This makes setting
up a job \/ery easy for the operator.
The other major robotic system being offered by IBM is the 7535 SCARA
robot.
This system was designed for precision assembly and is specified with
a repeatability of ±.002 inches across the entire workspace with up to a 13
lb. payload.
It has 4 basic degrees of freedom (X, Y, Z, and roll) in addi-
tion to the gripper.
This system does not have tactile sensing but does have
14 points of digital input and digital output sensing that are useful in
moving conveyor systems or detecting part presence.
It is programmed with the
IBM Personal Computer, and one Personal Computer can service many 7535's by
downloading the programs into the 7535's controller.
The system is \/ery
flexible and can be located over existing conveyor linesor workbench areas.
Management's focus has been on payback when justifying robots.
This
thinking is changing as it becomes obvious we do not understand all the
benefits of this new technology prior to installation.
The major components of robotic justification are:
labor displacement,
quality improvement, reusable tooling, cost stability, reduced tool-ing costs,
manufacturing flexibility and shorten the design-to-build cycle.
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
Users that have installed intelligent robotic systems have learned the
following:
it requires an upfront investment; it takes dedicated people; you
should start with simple appl i'cations; intelligent robots can reduce tooling;
they learned how to design the product for automated assembly; and they had to
learn by doing.
The steps necessary to expand into the computer integrated factory are:
the first step could be a robot assembly cell; the second step could be moving
the work flow to several of the automated assembly cells; the third step is
the integration of both the work flow and the data flow, achieving a computer
integrated manufacturing system.
The technology is there today to implement
computer integrated manufacturing. It requires management commitment and an
initial investment of resources and capital to begin the implementation
process.
Computer integrated manufacturing and flexible automation will help
manufacturers
become ."low cost producers" and improve their level of product
quality.
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
HISTORY OF IBM ROBOTICS
1972
RESEARCH DIVISION (YORKTOWN.NY)
CHARTERED TO
DEVELOP ROBOT TECHNOLOGY
1973
FIRST ROBOT BUILT
1976-1977
PROTOTYPE SYSTEMS INSTALLED AT LEXINGTON AND
POUGHKEEPSIE PLANTS
1978
ROBOTICS GROUP FORMED IN BOCA RATON
1979
ROBOT SYSTEM INSTALLED AT JOINT STUDY SITE
1980
IBM ROBOTS INSTALLED THROUGHOUT IBM PLANTS AND
LABORATORIES
1981
FIRST TEST MARKETING ROBOT CUSTOMER SHIP
1982
FEBRUARY - IBM ROBOTIC ANNOUNCEMENT
MARCH - ROBOTS VI EXPOSITION
1983
JANUARY - IBM 7565 MANUFACTURING SYSTEM
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
KEYS TO ROBOTIC JUSTIFICATION
INTELLIGENT
VERY INTELLIGENT
LABOR DISPLACEMENT
QUALITY mPROVEMENTS
UNIFORM PERFORMANCE
TEST WHILE BUILD
REUSE
RETOOL
REPROGRAM
COST STABILITY
STEADY RATE
MANUFACTURING FLEXIBILITY
ASSEMBLE TO "."lEED"
PRODUCT MIX CHANGES
REDUCED TOOLING COST
ADAPT WITH SENSIiJG/SOFTWARE
SHORTENED DESIGN/BUILD CYCLE
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive /San Jose, CA 95131 / (408) 971-9000 / Telex 171973
7535 APPLICATION POTENTIAL
A
P
P
L
I
C
A
T
I
0
N
C
0
M
P
L
E
X
I
T
Y
CABLE MFG.
CARD MFG.
BATCH ASSEMBLY
PILOT HOLE DRILLING
HIGH VOLUME ASSEMBLY
7535
FUitCTIOi'l/CCST
INTELLIGENT ROBOT REQUIREMENTS
°PRICE PERFCRiMANCE
°SPEED
'PAYLOAD
^PROGRAMMING SIMPLICITY
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
ROBOTICS EVOLUTION
21
c
O
O
>
w
(O
^
S.
a
ta
-<i
u
1962
1975
1980
1985
.MECHANICAL
ARMS
COMPUTER
CONTROL
SENSORY
FEEDBACK
' INTEGRATED '
ROBOTIC SYSTEMS
Autom^ion Migration
o
S.
c
(D
t/>
O
O
-1
•o
Computer
o
(n
c
cn
Q!
Computer Aided Design
0)'
p Capability
5'
Integrated Materials
Systems
O
I
«
|lntegration
•o
fi)
/
Automated Material
Handling Systems
M
(O
o
a
Q.
(D
TJ
tu
/
Cell Or Island
Of Automation
CO
D)
c_
o
/
(A
_<D
o
>
CO
01
O
00
(O
<o
o
o
o
(D
<o
u
Multiple Robot
application
/
Single Robot
Application
Time
Data
f
E
O)
p.
C
CJ
<0
9!
u'
•<
0_
>
b
(A
(B
o
O
3
•o
a>
•3
2
a
TJ
U
W
0
(/>
(D
O
>
<D
!J>
i
s
Mgt
Syatam
Computer^
: on t r o l l e d
tfare h o u s e
Parts
Mfg
V
\Z_
J
c
Computer Integrated Manufacturing
9
(0
3
O
o
•o
o
-1
Engineering
CAD
(D
p.
>
0)
c
CT
t»
a
Operations
Accounting
Inv. Ctl.
Purchasing
Raw Material
S'
><
o^
• ^
>
b
(/)
(V
o
o
3
T3
B)
3
8
a
a
(B
Inspection
Packaging
Shipping
•a
a>
3
<_
O
o
>
i
(p
Parts
Manufacturing
CNC
Machine Cells
Assembly
Robotic Cells
ROBOT ASSEMBLY CELL
o
ST
.o
Floor
Sub
Host
(D
E
Assembly
Cell
(C
a
W
c
cr
C o m p u t er
Controlled
Warehouse
(A
a
lu'
-<
-1
./
Material
o^
>
o
Handling
!L
0)
(B
o
o
3
•o
B>
3
a
a
a
—\
Manual
Material
Interface
Handling
-a
01
CO
X
O
>
CO
Ul
Material
Assembly
Cell
Handling
o
00
Computer
Controlled
CO
o
o
o
Warehouse
:sr
5.
(D
CO
-NI
u
Shipping
Dataquest
PRODUCTIVITY AND EDUCATION
Robert N. Noyce
Vice Chairman
Intel Corporation
Santa Clara, California
A cofounder, with Gordon Moore, of Intel Corporation in 1968, Dr. Noyce
was President until 1975 and Chairman of the Board from 1975 to 1979.
Dr. Noyce is co-inventor of the integrated circuit with Jack Kilby. They
jointly received the Ballantine medal of the Franklin Institute, and the
Cledo Brunetti Award of the IEEE for this work. He has also received,
with Gordon Moore, the AFIPS Harry Goode award for leadership in computer
science. Dr. Noyce was awarded the National Medal of Science and the
I.E.E. Faraday Medal in 1979, and the IEEE Medal of Honor in 1978. He is
a member of the National Academy of Science, the National Academy of
Engineering, the American Academy of Arts and Sciences, and is a Fellow
of the IEEE. Dr. Noyce holds 16 patents for semiconductor devices,
methods, and structures. He did research at Philco Corporation before
joining Shockley Semiconductor Laboratory, Palo Alto, California. In
1957, Dr. Noyce cofounded Fairchild Semiconductor Corporation, Mountain
View, California. He was Research Director until early 1959, when he
became Vice President and General Manager. Dr. Noyce received a B.A.
degree from Grinnell College (Iowa), where he was elected to Phi Beta
Kappa.
He received a Ph.D. in Physical Electronics from the
Massachusetts Institute of Technology.
Dataquest Incorporated
CAD/CAM INDUSTRY CONFERENCE
September 26-28, 1983
Newport Beach, California
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
PRODUCTIVITY and EDUCATION
R. N. Novce
There is probably no better statement of our national goals than the
preamble to the American Constitution. I'd like to read it to you as a
reminder, and use it as a point of departure in talking about productivity
and education.
"We, the people of the United States, in order to form a more
perfect union, establish justice, insure domestic tranquility,
provide for the common defense, promote the general welfare,
and secure the blessings of liberty to ourselves and our
posterity, do ordain and establish this Constitution for the
United States of America."
Our history, with some periods of backsliding, has indeed been in
pursuit of these lofty goals. Most of us would agree that we have not met
them yet, and that we have a lot of work left to do. It is the sum total of our
efforts to meet these goals which I would like to define as productivity in
the
broadest
sense.
There
are
many
facets
to
the
issue
of
productivity—the provision of the best quality of life possible—and for
the most people in a democracy.
Clearly, there are many conflicts to be
resolved in striving toward these g o a l s . Effort to resolve these conflicts
has been at the center of the political process since politics began.
Essential to reaching these goals, however, if we have not yet reached
the optimum, is change. Although this simple fact should be self evident,
many of the voices raised in our society wouldseem to say that although we
must improve the quality of life, change is deleterious.
As an example,
some of our most liberal thinkers seem to say that we must preserve the jobs
in the older industries which are no longer neede. They decry change in
these traditional industries, or in the us© of our national resources. Yet
they advocate raising the standard of living for Americans, which clearly
cannot be done without changing what we are doing. Even the labels we apply
to these viewpoints are exchanged as liberals work for the conservation of
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
the environment, and those we call conservatives would see expanded use of
the national forests and primitive areas.
These apparently conflicting positions arise from wishful thinking,
from the refusal to admit that the objectives of our society call for
continuous compromise., hoping we can have our cake and eat it too. There is
a need to think about
accommodating
the change necessary to approach our
goals more closely.
It is indeed through change that we have been able to achieve the
highest standard of living humankind has ever known, and it will be through
further change that we can improve more. A century and a half ago, more than
half our population was engaged in agriculture, in the provision of the food
necessary for survival.
I'm sure that if, at that time, a politician
projected that most of the farmers would lose their jobs, and that less than
three percent of the population would provide the food for all Americans,, he
would have been ridden out of town after being tarred and feathered, and
furthermore would have lost the next election. It is precicely because that
farm labor was available to provide for other needs or wants of the society
that the American productivity increased so rapidly.
It was that labor
which built the highways and the hospitals, built the automobiles and the
refrigerators, built the jet airplanes and computers which we have come to
e.xpect.
This massive improvement in the productivity of agriculture was
undertaken as America's first, and perhaps only, successful inijustrial
policy, if we put it in the language of today's debate. That policy was
carried out through the establishment of the Land Grant colleges, with the
consequent
massive upgrading of the skills of our farmers and farm
management,
and
the subsequent
research centers.
establishment
of major
agricultural
Those centers were directed toward finding better
varieties of grain to grow, or better stock to raise. It may be useful for a
minute to speculate about the debate that might have been carried 0L4t at
that time if those engaged in the debate could have forseen some of the
effects of their policy.
"My constituents are farmers—I can't support a policy which will
eliminate their
jobs."
"How can we ever hope to retrain the farm population to make tractors or
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
develop hybrid corn?"
"This policy will cause massive technological unemployment."
Ths policy would clearly have failed if only that part of its effect
could have been envisioned. Only with a better vision of the future impact
could it have been enacted. I suspect that none of the displacement effects
were f orseen, but that the improvement in knowledge was seen as benefitting
the more than half the population who were farming.
The makeup of our worl<force has, as you know, gone through a massive
shift since that time. Now more than half of the population is involved with
the creation, handling, and dissemination of information. Is it possible
today to implement a policy similar to that implemented for agriculture so
long ago? That might involve the establishment of schools directed at the
improvement of the performance of, and the tools available to the knowledge
worker.
We could establish research centers which would examine the
methods used by those workers.
New equipment might be developed in
concept,
as is being done in the Japanese fifth generation computer
project.
We might even establish incentives for the adoption of new
reethodologies,
such
as
computer-aided
design
or
computer-aided
manufacturing.
A rational man could easily think that such a program which would
enhance the productivity of more than half the workforce would be quickly
adopted.
Unless, and until we understand our other national goals and make
due accommodation for them, they won't be. Clearly there are many conflicts
which must be resolved. At the same time, there are many forces which are
driving us to accommodate the technology available to us. Primary among
those are the aspirations we all have to lead a better life.
Advancing
technology is simultaneously the cause of, and the response to many of these
aspirations.
We can easilyidentify some of the changes which are changing our
historical response to these aspirations:
We have seen the disappearance of our western frontier. Those
seeking new opportunity and challenge must seek it not on a
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
homestead, but in a new activity
Our population is aging as a result of improved public health
and medical intervention, and the drop in the birth rate, fln
ever increasing fraction of our population will be dependent on
the working population either as they are being trained for
productive activity, or after their retirement.
Improvements in communication and transportation have made
competition global, whether it be for markets, or for adoption
of ideas. Wars in Ireland or in Lebanon are concerns for all,
and the establishment of the fifth generation computer project
is seen as a threat to our own industry.
America has raised the priority of establishing justice for
those of all races or economic background, sacrificing short
term benefit for longer range and higher objectives.
In the broadest sense, increasing productivity is our attemt to more
nearly optomize the solution to the broadly stated problem: "How can we tiest
meet our conflicting national goals?"
The answer clearly includes the
improvement in the hourly output of everyone in the workforce, but it
includes many other things as well, which enhance our quality of life. This
leads me back to the subject of this talk, and the relationship between
education and productivity. The basic role of education is the preparation
of our people to optomize the solution to this problem.
The current debate on education is many-faceted.
The flmerican
Electronics Association is concerned with the short term availability of
scientists and engineers necessary for the rapidly growing electronics
industry, a concern all of us share.
The recent report of the U. S.
Commission of Higher Education decries the deterioration of the quality of
our institutions of higher learning, particularly in the sciences and
mathematics, but in communication skills as well. Congress is asking what
role education must, or should play in the solution to problems of
international
competitiveness,
or
retraining
the
technologically
unemployed. At the same time, articles are being written suggesting that
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
the increasing use of computers and automation of the production process
will require lower skills from our workforce rather than higher s k i l l s .
As an example of this line of thinking, let me quote from a recent
article by Henry Levin and Russell Rumberger of Stanford:
"Just
as ominous is the possibility
that high tech will
eliminate more jobs than it creates.
Researchers at the
Robotics Institute at Carnegie-Mellon University estimate that
in the next twenty years, robotics could replace up to three
million
manufacturing
positions
involving
operating
machinery, and potentially eliminate all eight million of these
positions
by
the
year
2025.
The
widespread
use
of
computer-aided design may virtually eliminate the occupation
Of drafter in the not-too-distant future, a potential loss of
300,000 skilled positions, ft recent study from the Upjohn
Institute estimated that robots could eliminate three times as
many jobs as they create, and the Director of Advanced Products
and Manufacturing at General Motors predicts the 'factory of
the future' will employ 30% fewer workers per car because of
robotics.
Even if laid-off production workers are retrained
for high tech positions they may not be able to achieve a
comparable income level."
That is a powerful call to arms to the draftsmen and the auto workers.
We can speculate
unemployment
that
roles.
eliminated
while
resistance
to change
the farmers of 1850 should still be on the
That it is seen as "ominous" that jobs can be
achieveng
the
same
output
indicates
and lack of understanding
of how
the enormous
our society
progresses. That resistance to change is institutionalized in the power
structure in our society.
Just as the congressman resists change in the
rules which let him be elected, the industrialist seeks to preserve the
competitive environment in which he built hiscompany, and the engineer
would like to continue to work in the field in which he has enjoyed his
success. Clearly we have our work cut out for us.
Education and basic research occupy a unique position in our society.
Both are very long-term investments.
In niether case is the benefit of
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
their support very likely to accrue to the individual or group supporting
them. To Illustrate, consider an industry with two equal competitors, and
their considerations about performing basic research. Upon reflection,
both will conclude that the winning strategy is to let the competitor pay
for the basic research , the results of which are widely disseminated, and
exploit those results when appropriate with the funds saved by not paying
the bill for the original research. Thus, left to the free market., support
for education and research will be at a lower than optimum level for the
society. The odds become better when supported by a broader coalition of
the society, but railing the power to force everyone to participate, the
individual would still decide to let his neighbor support the schools and
use those savings to buy a new car which he sees of more direct benefit to
himself. How many smog devices would be installed on American automobiles
if they were only a consumer option?
The feedback loop around the process of education is very long. The
determination of whether or not the process produced the desired result rnay
tal<e twenty years or more. As any control theorist can explain to you, it is
very difficult to stabilize such a process, and the system is likely to go
into wild oscillations.
concerned
Educators, as a result, are properly deeply
about making changes rapidly for fear of destabilizing the
system.
Traditionally, education has sought to prepare the student for the
rest of his lifetime of work. Perhaps when the rate of change was slower,
that concept was valid.
Today, in the technological fields, the concept
that training in a narrow field can prepare for his lifetime is clearly
erroneous.
If that career lasts forty years, (and that time is likely to
lengthen with the increasing vitality of the older population) we need only
to think of the technology of forty years ago to see how obsolete that
education will be. In 1943, the transistor was yet to be invented, the first
of the military jets had just begun flying, computers made with relays were
calculating
shell trajectories, sulfa drugs had been discovered but
penicillin was yet to come, and nuclear weapons and space travel existed
only in science fiction magazines.
Clearly, specific training in the
applied technology of the day would not be useful today. The objective of
education must be to provide a background such that the student can
understand our society, has the basic skills which may be used in a variety
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Parl< Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
of different activities, and can accommodate the inevitable changes which
will occur, and, hopefully, contribute to constructive change in his
lifetime. We must not sacrifice this broader objective to the short term
success of specialized training, even though that is the basis upon which
the job applicant
is most
likely
chosen.
Such a course would
be
short-sighted in the extreme.
As we seek to improve the productivity of our society profound change
is inevitable. Education is a powerful facilitator for that change. To
fulfill that role, whether in the classroom or out, many things need to be
accomplished:
First, and perhaps most important, we must establish a common
ground for resolving
the conflicts which will
arise with
Change. Failing in this task will leave us paralyzed, unable to
do that necessary for continued progress toward our goeils.
Secondly,
basic skills in communication
and
quantitative
thinking must be imparted to the student,in order that he can
understand
and
analyze
the
potential
solutions
offered.
Language and mathematical skills are essential.
Third, since many, if not most, of the changes which will occur
will be based in the advancement of scientific knowledge the
student needs to have a basic understanding of the method by
which new knowledge is developed, an understanding of how the
scientific method distinguishes between truth and falsehood,
and
an understanding
of the limits of todays scientific
knowledge.
Fourth, in order to live a satisfying and productive life, he
must have an understanding of the culture in which he is to
function, which may be inparted through its history, art, or
philosophical foundation.
Finally, he should be trained to make a specific contribution to
the society, whether
as a dental
technician or dcctc?r,. a
repairman or engineer. This "vocational" training may be done
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 RIdder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
within our institutions of learning or outside, but it surely
must be done if the student is to be gainfully employed. This: is
the most transient of the education, likely to be repeated many
times in a career.
For the society, education is a highly leveraged investment,
providing more than half the inprovement in productivity which we have
achieved historically. Yet, because of its long-term nature, it can easily
be sacrificed to shorter term goals. Consequently, it needs to be nurtured
and supported by all of us if we are to "promote the general welfare" as
envisioned by our f o r e f a t h e r s .
Oataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
Dataquest
CAD/CAM WORKSTATION TRENDS
Edward L. Busick
President
Spectragraphics Corporation
Mr. Busick is founder cuid President of Spectragraphics Corporation in San
Diego, California.
Previously, he was in management with Parson's
Engineering and Construction and with other international companies on
major construction projects in Australia, Alaska, and Saudi Arabia. Mr.
Busick was also with Adage, Inc., a computer graphics workstation
manufacturer in Billerica, Massachusetts. He received his B.S. degree
from the Sloan School of Industrial Management, Massachusetts Institute
of Technology.
Dataquest Incorporated
CAD/CAM INDUSTRY CONFERENCE
September 26-28, 1983
Newport Beach, California
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
o
c
CD
E
a
a.
C/5
c
g
u
a
0)'
>
b
CD
O
O
3
•o
B)
ro
(O
o
Q.
Q.
CD
T3
0)
05
tu
3
o
>
<o
Ul
.Cv
o
00
<o
o
o
o
2.
to
X
~j
(O
u
CAD/CAM
WORKSTATION
TRENDS
o
(u
a>
(A
•o
o
(S
a
>
CO
o.
>
b
2.
u>
n>3
o
o
3
•o
0)
ro
(O
o
• HISTORICAL PERSPECTIVE
• TECHNOLOGICAL AND COST FACTORS
a
a
(D
T3
EU
• CURRENT TRENDS
C/J
B>
• FUTURE TRENDS
3
o
>
(O
U1
.b.
o
S
to
(O
o
o
o
!L
(B
X
•>l
(O
• FINAL CONSIDERATIONS
HISTORICAL PERSPECTIVE
This year marks the AOth anniversary of the era of electronic
computing.
In 1943 at the University of Pennsylvania the effort was
begun to develop ENIAC, the first vacuum tube computer.
In 1951, the
Whirlwind computer at M.I.T. utilized the first core memory. VJhirlwind
filled a large room, had 4K of memory, and less computing power than
today's home computer.
One of the first CRT computer systems was the SAGE Air Defense
System used to detect and display the location of aircraft throughout the
U.S.
SAGE became operational in 1953.
In 1957, the TX-0 was built at
M.I.T. and was the first transistorized computer. The TX-0 had a point
plot graphics console with light pen. and could do convincing interactive
graphics.
Two events occurred in the early I960's that can be considered to
mark the beginning of the CAD/CAM era.
At General Motors, the DAC-1
console was installed by IBM on a 7094 computer.
The DAC-1 was used to
test man-machine interaction for the first time and showed a 1/3 increase
in productivity.
PDP-I.
The same year, DEC produced its first computer, the
These events marked a major point of departure in solutions to
the CAD/CAM challenge.
The CAD/CAM Tree (Figure
1) traces the history of major developments
over the past four decades.
Following the main branch on the left, after
the PDP-1, we next encounter the PDP-9 in 1966. This machine had
optional storage tube as well as stroke-vector CRT. The PDP-9 and its
successor the PDP-I5 laid the groundwork for DEC's reputation in
engineering graphics support computers.
The PDP-11 opened the door to
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company/ 1290Ridder Park Drive / San Jose, CA 95131 / (408)971-9000 / Telex 171973
0)
n
c
(D
CAD/CAM TREE
E
Q.
inJ5i?^
w
c
g
(A
a!
S'
APPLICON
AUTOTROL
>
b
1980
.. T'.
i ' ' ' \ ^..i!..
^
••A %•ji-!^H
to
APOLLO
APU
HMfr^^^'^A^U*^^
-^
•<
o^
1
GRAPH
MC AUTO
I DEC
I CALMA DG
COMPUTERVISION
GERBER
UNIVAC I
"I
.„„
CDC PRIME I
IBM
, MV
P^68060 M
'V:/m ^ - 5 i^?^^^>M^^
M*
XX
'i^j.
I
- -
^
^
1^
'••••• •
• ;•
3
•o
"
'
i
.
'^^.
•'-'
•Mi^.ieCLIPSE 'l,;;...^
1--•
1,
11'.,
I^
I,
••,', i i
lU
"nm
rv)
a
a.
a
/ . #
370
-i':';'::':-:-:-:--T^fffS^S^
"^iiii^.
-0
u
pDp-d!'
O
w
o t ^ ^ ^ ^
PDP-1t
33
J M
/i30XX
t .1
qVBER
^OVA..
1970
8
'V77
4300
r^
'
3
o
o
/.y
H
HP-1000
360
1960
3
pDP-tf:?-':pAC-i
o
>
TX-O^i'"
(D
Ul
S
SAGE
1950
^H^w&im
<£>
O
O
O
••••.raiiASs...
.,^m^
(O
figure 1
to
# '
>X*>My.v
COLOR RASTER
STORAGE TUBE
STROKE
CAD/CAM for DEC. Applicon
introduced its first system using the II in
the early 1970's and was followed a few years later by Intergraph and
McAuto.
As you can see, all these companies have upgraded to the VAX
series and were joined by Autotrol and Calma as DEC-based turnkey
suppliers.
The right branch of the CAD/CAM Tree shows the introduction of the
IBM 360 in 1964. Lockheed Corporation along with General Motors and
McDonnell Douglas were early pioneers in IBM's CAD/CAM effort. Today
Lockheed is one of the largest users of CAD/CAM workstations and has many
turnkey IBM systems, some with as many as 100 workstations on a single
mainframe. Lockheed's subsidiary, CADAM, Inc., is a major third party
software vendor.
CADAM software is sold under license by IBM.
The 370,
30XX and 4300 series round out the past 20 years of progress for IBM in
CAD/CAM computer support.
IBM is now the second largest turnkey systems
supplier.
Computervision is the No. 1 vendor in the industry.
With total
installations approaching 10,000 workstations, Computervision has set the
pace for the industry.
Recent product announcements indicate further
Computeirvision innovations.
Agreements with IBM and Sun Microsystems
epitomize the new trends that CAD/CAM is establishing.
The 32-bit
mainframe and the 32-bit microprocessor are taking their position along
with the 32-bit supermini as major factors in turnkey systems.
Apollo, another manufacturer of networked microprocessor-based
systems, has made a significant impact on CAD/CAM.
Started in 1979,
Apollo supplies workstation hardware to Autotrol, Calma, as well as
several of the new Computer-Aided-Engineering (CAE) turnkey electrical
vendors.
Data General, Univac, CDC, Prime, and Hewlett Packard are also
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
shown as major factors in CAD/CAM.
These companies represent a major
portion of the industry, but are by no means a complete list. The
purpose of the CAD/CAM Tree is to show general trends for vendors over
the past few decades.
A second set of information is depicted in the CAD/CAM Tree. The
history of the workstation CRT technology used by each vendor is
shown.
As many of you know, there are three major types of workstation
hardware.
These include stroke, storage tube, and color raster CRT's.
IBM has been and still is the major proponent of stroke technology.
Stroke systems offer significant speed-of-response advantages for fast
picture update, and are well-matched to the high speed I/O channel and
CPU performance of the IBM and Univac mainframes.
Stroke systems have
limited color capability and the tendency to flicker as the picture
content becomes more complex.
The storage tube has been, for the most part, a Tektronix product.
The storage tube was the main workstation offering for most turnkey
vendors throughout the 1970's. The storage tube offers the advantage of
no flicker. Once the picture is written, it remains stored in the screen
phosphor without refresh.
The disadvantages include slow update rate and
limited color capability.
The third workstation alternative is raster scan.
Raster resolves
both the fast update problem as well as the potential flicker problem.
The picture can be refreshed at constant rates exceeding normal television quality.
Raster also offers the benefit of full color capability.
Raster workstations were first introduced into CAD/CAM in 1980, and today
most of the turnkey vendors offer color.
Because of the advantages and disadvantages of the workstation CRT
types, certain modes of operation were adopted by various vendors to
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
o
U
n
c
<D
01
O
O
•3
O
CO
a.
CO
c
CT
(A
a
fl)'
WORKSTATION TECHNOLOGIES
>
b
o
o
3
•o
01
3
STROKE
STORAGE TUBE
RASTER
FAST RESPONSE
SLOW RESPONSE
FAST RESPONSE
FLICKER
NO FLICKER
NO FLICKER
LIMITED COLOR
LIMITED COLOR
FULL COLOR
i>o
(O
o
Q.
Q.
01
•D
0)
w
O)
3
O
>
{
\
\
2D DRAFTING
2D DRAFTING
2D + 3D MODELING
SCREEN MENU
TABLET MENU
TABLET OR SCREEN MENU
CO
Ul
I
(O
o
o
o
Figure 2
(O
CO
match workstation performance.
Stroke systems take advantage of fast
update and utilize a screen menu for selection of Interactive function,
in order to circumvent the slow erase and update for storage tubes, a
tablet menu with a large set of function alternatives was employed.
The raster workstation off ers the flexibility for either screen or
tablet update.
Secondly, complex Images In full color can be presented
without flicker.
This flexibility has encouraged turnkey vendors and
third party software vendors to rapidly expand the number of features
available In various applications areas. Three-dimensional solids
modeling, robotic kinematics, work cell simulation, and 3-D piping are
but a few of the new areas opening up to CAD/CAM with the advent of
raster workstation technology.
These new opportunities raise questions
that should be addressed if the CAD/CAM industry is to take effective
advantage of the new technology.
Should the CAD/CAM system use
specialized workstation features now available in order to reduce the
host computer workload?
Off-loading the host computer means better
productivity and more users per system.
easier to cost justify.
This makes a CADfCAM system
How should intelligent workstations with local
32-bit microprocessor power be used in the system resource allocation?
(Figure 3).
The cost benefits available with intelligent workstations
must be balanced with the overall corporate CAD/CAM needs and the needs
of the user.
If these issues are not addressed by turnkey vendors and
users, the host computer may become overburdened handling the workstation
graphics functions.
The historical summary for CADfCAM (Figure 4) shows a range of
computing alternatives in the 32-bit arena — mainframes, superminls and
microprocessors —
that can provide growth potential and flexible,
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
cost-effective solutions tailored to user needs.
Raster workstations can
provide similar growth potential and a migration path to more
functionality.
The technology and cost factors must be understood to
balance CAD/CAM system resources. We will return to the system resource
question after looking at these factors.
WORKSTATION TECHNOLOGY AND COST FACTORS
Turning now from the historical to the present, let's look at some
of the technological factors of workstation design.
Figure 5 enumerates
some of the typical graphics functions that can be applied to the
Interactive geometric data base.
I have Included a brief glossary of
terms in the Appendix to help you remember these graphics functions.
Let's examine the task allocation implications of using the host computer
VS. the workstation for these functions.
Each of the graphics functions
has a required level of processing power for execution.
This graph shows
that the range of computing power varies by a factor of nearly 1,000 from
the simplest graphics function to the most complex.
Intended to show order-of-magnltude relationships.
such as shading are dependent on the algorithm used.
The figure is
Specific functions
The simplest
functions,, pan and zoom, are becoming a standard feature of most
workstations being offered today.
The medium performance functions are
being offered by many workstation manufacturers.
The more complex
functions are being addressed by custom VLSI chip manufacturers and a few
workstation manufacturers.
The highest performance function shown, point
light source shading, can require significant computing power for
execution and is only being effectively addressed with mainframe
computing.
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
K
OJ
n
c
(B
1
<B
a
>
t/3
a.
THE CAD/CAM SYSTEM
RESOURCE QUESTION
01
>
b
n>
3
o
o
3
•a
01
WHERE ?
~
^
CAD/CAM
ANALYSIS
DATA
BASE
MANAGEMENT
a.
a
SYMBOLIC
DRAWING
DEFINITION
DETAILED
PARTS
DRAWING
DEFINITION
3D
MODEL
MANIPULATION
INTERACTIVE
GRAPHIC
DATA
BASE
(D
HOST COMPUTER OR INTELLIGENT WORKSTATION
COMPUTER
WORKSTATION
w
0)
3
o
>
o
00
CO
o
o
o
ID
->J
I ' i g u r t i 3^
o
SJ.
c
<s
<l>
3"
o
o
-^
•o
o
0)
Q.
>
cn
c
HISTORICAL SUMMARY
CT
(0
ai
S'
•<
o^
>
b
32 BIT
COMPUTER
(»
(D
o
o
)
COLOR
RASTER
WORKSTATION
3
•o
0>
•<
MAIN
FRAME
ro
(O
O
SUPER
MINI MICRO
30
S.
a.
(D
TJ
B)
1980
So
CO
0)
D
C_
1970
o
cn
(0
O
>
o
00
STORAGE
1990
^juBy
(£)
<o
o
o
o
Figure 4
CJ
STROKEV
o
w
oT
42
C
(D
O
O
•3
o
CD
p.
>
V>
C
CT
C/)
TYPICAL GRAPHICS FUNCTIONS VS ESTIMATED NUMBER
REQUIRED MACHINE INSTRUCTIONS TO EXECUTE FUNCTION
d
0)'
•<
FUNCTION
o
>
b
LIGHT SOURCE SHADING
U)
CD
o
o
3
•o
0)
20000+
HIDDEN SURFACE REMOVAL
] 10000
POLYGON SHADING
]8000
]4000
ANTI-ALIASING
to
§
da.
<i>
•o
EU
]3200
DEPTH QUEUEING
llOOO
CLIPPING
]150
PERSPECTIVE
]72
3D ROTATION
CO
D)
o
>
BAfiK SURFACE TEST
;]4o
PAN & ZOOM
]40
PAN (TRANSLATE)
]24
(SCALE)
]24
(O
u>
ZOOM
i
.c^
§
10
100
1000
ESTIMATED NO. DEC ASSEMBLY INSTRUCTIONS
n>
X
-si
U3
->l
U
Eigtrre 5
J
10000
Turnkey vendors are moving toward the 3-D data base as the standard.
Parts drawings and symbolic drawings are 2-D projections of the 3-D
model.
The drawings are closely related to the model data but with added
Information for the drafting or analysis requirement.
The 3-D model or
the 2-D drawings can be changed and the changes reflected in either data
base.
The Computervision CADDS-4 and IBM-CADAM 3-D Wire Frame are
examples of these new directions.
Let's take a look at 3-D graphics manipulation as a specific example
of the cost/performance trade-off available in workstation design.
Figure 5 tells us that in order to execute an assembly language 3-D
rotation routine in a DEC VAX will require 72 instructions. This routine
will be executed once for each line in the 3-D model. A typcial 3-D
image is composed of approximately 5,000 lines. A simple calculation
tells us that in order to do a single rotation of this image will require
72 X 5,000 = 360,000 instructions. Now, let's assume that the host
CAD/CAM computer supports 16 users and half of them need this 3-D
capability at any given time. We further assume that these 8 users will
utilize the 3-D function an average of once every 10 seconds. The
calculation in Figure 6 gives the result that 14% of the computing
capacity of a 2 MIPS host computer will be used in supporting the 3-D
rotation graphics function.
hardware in the workstation.
The alternative is special purpose graphics
Available workstation technology can
off-load this graphics function from the host computer.
The 3-D rotation
graphics function can be executed by special purpose hardware, at a
performance throughput of 72 MIPS.
The cost per user in current systems
is as low as $2,000/workstation or $16,000 for our assumed eight-user
requirement.
A 2 MIPS host machine is generally priced in the $400,000
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
o
ST
J2
c
(D
(/>
3
O
o
-t
•o
o
SAMPLE CALCULATION
HOST COMPUTER 3D SUPPORT
n>
_a
>
w
c
CT
W
Q.'
0)'
>
72
z
CD
CD
(V
o
O
3
INSTRUCTIONS PER LINE IN 3D-MODEL
X 5,000
LINES IN 3D MODEL
360,000
INSTRUCTIONS TO ROTATE MODEL
X
USERS
•o
0)
s
?
d
a
(D
•0
01
w
O
>
8
2,880,000
-T-
10
288,000
INSTRUCTIONS
SECONDS AVERAGE INTERACTION TIME
INSTRUCTIONS/SECOND
(O
Ul
00
288,000
o
2,000,000
o
(O
=
1 4 ^ OF 2 mips HOST COMPUTER CAPACITY
FlgJl 6
range and up. The final cost calculation is shown in Figure 7.
Off-loading this graphics function will allow an increase in the number
of users of the CAD/CAM host machine, making for easier system cost
justification.
The reason that special purpose graphics workstation hardware is
more suited to the problem solution is that the graphics functions can be
executed most effectively with parallel-pipelined processing. Mainframes, superminis and microprocessors are not designed for this type of
parallel processing.
The polygon shading feature is a much more dramatic example of the
potential cost/performance advantages of special purpose workstation
hardware.
If this "solids" feature is required by the same number of
users as in the example above, and utilized once every two minutes, 67%
of the 2 MIPS machine capacity would be required (Figure 8).
If these
features are to become truly useful in the future of CAD/CAM, most of
the graphics functions will become workstation hardware requirements.
The next technological issue is local processing power in the
workstation.
industry.
The Motorola 68000 has started a major trend in this
The Computer-Aided-Engineering (CAR) workstation is expected
to rapidly increase its market share in CAD/CAM over the next few years.
New microprocessors from National Semiconductor, NCR, Intel, and others
will add to this lower cost computing alternative.
There are some who
see the microprocessor as the ultimate solution to the CAD/CAM problem.
Every engineer will have his own workstation and the need for the
mainframe will disappear.
this dream.
The reality of the situation does not match
The intelligent workstation can solve a sizable piece of the
CAD/CAM problem but many problems are beyond its reach.
For instance,
some of the large aerospace companies have on-line access to nearly one
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
o
Si
ST
c
(C
3
o
o
-^
•c
o
(D
COST COMPARISON
HOST COMPUTER VS. WORKSTATION
_a
>
w
c
cr
u
a
BJ'
>
b
HOST COMPUTER TOTAL COST:
o
CAPACITY REQUIRED FOR 3D-ROTATE
FUNCTION (14%):
o
3
•D
01
•<
$400,000
-14
N3
8
5!
a.
(D
T3
0)
w
a
o
u>
?
TOTAL HOST COMPUTER RE«>URCE ALLOCATION
COST (8 USERS):
$ 56,000
WORKSTATION 3D-ROTATE FUNCTION HARDWARE
COST (8 USERS):
$ 16,000
COST RATIO 3D-R0TATE FUNCTION
O
>
<a
<ji
HOST COMPUTER : WORKSTATION
3.5 : 1
o
00
CO
o
o
o
(O
^
^^
Fl
I
o
o
•D
O
(S
CL
>
Q.
0)'
SAMPLE CALCULATION
HOST COMPUTER POLYGON SHADING SUPPORT
>
b
8,000
INSTRUCTIONS PER POLYGON IN 3D-MODEL
{ 1 / 4 " X 1 / 4 " POLYGONS ON 1024 x 1024
RESOLUTION CRT)
2,500
POLYGONS IN 3D-MODEL
(D
o
o
3
•o
a>
•<
IV}
<o
o
I
i
o
(y5
0)
20,000,000
X
8
160,000,000
INSTRUCTIONS TO SHADE POLYGON 3D-M0DEL
USERS
INSTRUCTIONS
o
u>
o
>
(D
Ol
O
00
(O
-T-
120
1,333,333
1,333,333
(O
o
o
o
2,000,000
SECONDS AVERAGE INTERACTION TIME
INSTRUCTIONS PER SECOND
= 6 7 ^ OF 2 mips HOST COMPUTER CAPACITY
(B
(O
3
,
Figure 8
million CAD/CAM drawings. The mainframes that manage this data base are
located in different parts of the country.
It is with this major data
base management challenge that the local area network (LAN) falls short.
Secondly, the analysis portion of the CAD/CAM problem can be compute
intensive and mainframes will continue to outperform micros. Mainframe
power will maintain its role in these areas.
Let's address the range of three possible performance levels for a
CAD/CAM workstation.
The basic workstation has 2-D capability with a
complement of interactive devices.
It is attached directly to the host
computer over serial or high-speed parallel interface.
Workstation
Number 1 needs full host support. Only the interactive graphic data base
is processed by the basic workstation.
A second workstation with
expanded hardware to handle the 3-D or Solids graphics function can solve
more of the problem and off-load the host computer. Workstation Number 2
performs 3-D model manipulation without host support.
Finally, a
workstation with 32-bit microprocessor can function as a stand-alone
intelligent system. Workstation Number 3 can be used in symbolic and
detailed parts drawing definition.
There is an Important point to
understand in this migration path.
Going from Step 1 to Step 3 is not
possible without Step 2.
The microprocessor system without the option
for expanded workstation hardware functions will provide limited growth
potential for 3-D modeling requirements in the coming years.
Step 3
without step 2 does not have a "solid" future.
CURRENT TRENDS
In covering the historical and current technology, we have mentioned
several of the more obvious CAD/CAM workstation trends.
These include
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
D
»
DT
.a
c
(D
(/>
5'
o
o
•5
o
(t>
a.
w
c
THE CAD/CAM SYSTEM RESOURCE QUESTION
CT
!2
Q!
oj'
•<
o^
>
O
\
o
\
ANALYSIS
(0
•
o
3
X
\
\
\
"
^
SYMBOLIC
DETAILED
3D
.
!
^ DRAWING ^ PARTS
^ MODEL
\
DEFINITION DRAWING
MANIPULATIONDEFINITION
X
INTERACTIVE
GRAPHIC
DATA
BASE
•a
0)
to
(O
o
WORKSTATION
COMPUTER
a
Q.
(D
•D
(U
w
0)
o
>
§
CO
o
o
o
BASIC WORK STATION
NEEDS FULL HOST COMPUTER SUPPORT
(D
(O
OJ
I'lj^iirc
9
o
lU
CD
1/1
O
O
•3
o
THE CAD/CAM SYSTEM RESOURCE QUESTION
C/5
c
IT
(A
a.
01
>
o
z
""
ii=;
~
,^
a>
V)
a>
z>
O
3
\^.\.Y
\
MBO
—s .
/
INTERACTIVE
CAD/CAM
ANALYSIS
SYMBOLIC
DETAILED > j 3D
X
GRAPHIC
DATA
\
V
N D R A W I N G ^ PARTS
^ MODEL
MANIPULATION >^ DATA
BASE
\ ^
\ ^
DEFINITION DRAWING
y
X BASE
MANAGEMENT \
\
\
\DEFINITION /
TJ
3
WORKSTATION
*<
N>
(O
O
COMPUTER
a
o.
(D
T3
0)
D
O)
01
3
o
>
U3
Ol
.Ck
o
CO
(O
o
o
o
(O
-si
U
WORKSTATION WITH EXPANDED GRAPHICS FUNCTIONS
OFF-LOADS HOST COMPUTER
PERFORMS WORKSTATION 3D-M0DEL MANIPULATION
KlniirL- ID
o
u
ST
#
c
(D
a>
o
o
•5
o
THE CAD/CAM SYSTEM RESOURCE QUESirON
(D
Q.
>
CO
Q.
•<
>
o
z
CD
(A
:)
O
o
3
•o
0)
l>0
CO
o
' ^ ^
-
.»
~ .—
b
•
^
1
=
-
SYMBOLIC
DETAILED
DRAWING .^^PARTS
DEFINITION DRAWING'
y
y
DEFINITION
„
Y
X . '
- ,i<r , •
3D
MODEL'
y
MANIPULATION.
INTERACTIVE
QRAPHIC
DATA
BASE
•••
f
I
WORKSTATION
COMPUTER
a
Q.
CD
"D
a>
(f)
0)
O
>
.b.
3o
(O
(O
o
o
o
32-BIT MICROPROCESSOR-BASED WORKSTATION
STAND ALONE CAPABILITY
3D MANIPULATION + DRAWING DEFINITION
<D
(D
~.l
U
F igiirf
1 I
o
c
a>
01
•o
o
n>
Q.
>
w
c
cr
GRAPHICS WORKSTATION MIGRATION
CA
•:
Bj'
•<
>
b
n>
3
o
O
3
•a
(U
3
ro
(O
o
a
CL
<D
-o
0)
BASIC
FUNCTIONALITY
w
01
O
>
EXPANDED
FUNCTIONALITY
3D MODELING
CO
en
.c^
o
00
CO
->j
CO
o
o
o
2.
n
CO
->J
w
I' I g i i r t '
t
\1
STAND
ALONE
LOCAL
INTELLIGENCE
3-D MODELING
UJ
oc
u.
9
o
(0
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
dominance of color raster in CRT technology, 32-bit tnicroprocessor-based
intelligent workstations, local area networks, and expanded workstation
graphics functions.
An important new trend is workstation compatibility
with the proposed Graphics Kernal System (GKS).
GKS is being evaluated
by the American National Standards Institute as the standard for graphics
communication between the host computer and the workstation.
eight companies have announced GKS compatibility.
At least
Recent articles and
technical presentations seem to indicate that Tektronix and IBM plan to
ioin this group.
GKS is the result of a world-wide cooperative effort to
establish a method for transportability of graphics applications from one
workstation to another.
Commitment to this standard will have to be
addressed by the major turnkey vendors. GKS presents an opportunity for
the turnkey vendors to minimize future development costs. Once an
application is GKS compatible, the migration costs to another type of
workstation will be minimized-
For the user, GKS also offers benefits.
The user will be able to take advantage of GKS compatible applications
from multiple vendors and independent third-party software companies.
Another recent trend has been the introduction of several low-cost
color hard copy printers in the $10,000 range.
These devices will be
suitable for the LAN microprocessor-based systems, and will have some
limited use with superminis and mainframe host computers.
The larger
host systems will continue to rely on higher performance hard copy such
as the Versatec, Benson, and Calcomp devices. Many system developers
believe that black and white hard copy will continue to play an important
role in predominantly color systems. This line of reasoning is based on
the fact that manuals and supporting documentation derived from CAD/CAM
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
o
u
nS
c
(D
W
3"
o
o
•5
o
(D
Q.
W
c
CURRENT TRENDS
IT
—
a
S'
>
o
5.
en
CD
o
o
3
• COLOR RASTER CRT
• 32 BIT MICROPROCESSOR BASED
INTELLIGENT WORKSTATION
•o
0)
IV)
to
o
Q.
a.
CD
•D
EU
(/>
0)
3
• LOCAL AREA NETWORK
• EXPANDED WORKSTATION
GRAPHICS FUNCTIONS-3D, ETC.
• GRAPHICS KERNAL SYSTEM
(GKS) COMPATIBILITY
o
>
• LOCAL COLOR HARD COPY
o
00
(O
o
o
o
• LOW-COST PASSIVE REVIEW
(PR) TERMINAL
5.
CD
X
(O
->J
CJ
Figure 14
are printed in black and white.
The costs are major for converting this
support documentation infrastructure to color.
The final item on our list of current trends is a low-cost graphics
terminal that can be used for passive viewing of drawings by the CAD
documentation support groups, process planning group, and the
manufacturing organization.
This concept is being addressed by IBM with
the 3279 "softcopy" terminal.
This Passive Review (P/R)
type terminal
will be a necessary part of an Integrated Manufacturing scheme. Target
prices under $5,000 will be necessary in order to fulfill this
requirement.
The number of terminals to be sold in this market may
exceed the number of interactive CAD/CAM workstations by a factor of two
or three. The market is potentially large.
FUTURE TRENDS
Many of the major Fortune 500 users have progressed to the stage of
setting corporate CAD/CAM goals for future growth.
The goals often
include a requirement for communications among divisions throughout the
U.S. and, in some cases, even include European operations.
Remote Area
Networks will become more commonplace for major users. These requirements coupled with the emerging trend toward process planning and group
technology will necessitate expanded communications support from turnkev
vendors.
Since the turnkey suppliers use a wide variety of CPU's and
workstation configurations, a hierarchical network of mainframes, superminis, and 32-bit microprocessors will likely emerge.
include Passive Review (P/R)
The network will
Terminals and Interactive Workstations (W/S)
(Figure 16). The Initial Graphics Exchange Specification (IGES)
represents part of the solution to this network requirement.
IGES allows
the exchange of CAD/CAM drawings between otherwise non-ccmpatibie turnkey
svstems.
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
o
u
ST
n
c
01
o
o
•3
o
(t>
CL
>
v>
c
cr
(A
a
u'
>
b
2.
0>
(S
O
o
3
FUTURE TRENDS
•D
01
N)
tf>
O
a.
a
• REMOTE AREA NETWORK (RAN)
CD
•D
01
w
01
D
• INITIAL GRAPHICS EXCHANGE SPECIFICATION (IGES)
• ERGONOMICS AND AESTHETICS
o
>
o
to
(O
o
o
o
(D
(t
3
Figure- 15
o
Si
DT
.o
c
(A
3"
o
g
•o
o
(C
IGES/GKS INTEGRATION
_a
>
w
c
CT
(A
a
oj'
•<
o
>
b
P/R
MAIN
FRAME
^^T'X^I^Mi
P/R
o
o
3
•o
0)
W/S
^^^SfliJ«?H***='**•
V n.'X'W:^'^*^
fl;«*^*:^«^^:-^:<:^.^^^-v^.^.:^^^^^4^.x^«'^:'^^+yowo(:l;1^«•^:4*!i^i
'•^J^i^KSjJJ.^.^^.^.
W/S
**>:-^«-.>:.x*„,,
™->««»:.„
P/R
W/S
ro
(O
o
I
i
P/R
v:^Tr:,t.Toy4j«>j«.-.Mr:
CO
fil
P/R[
32 BIT
MICRO
SUPER
MINI
V
/;%',^^S-;^*ftfi^>&^rr^.^.-''
W/S
>
32 BIT
MICRO
itt^<W'>.ljt'tWT*!>
j-iiswes*
SUPER
MINI
^•WibWiV^
o
>
(O
Ol
o
00
IGES
U9
o
o
o
GKS
P/R
'r->:.:4:4ie4di)£i:S^^
Fi
16
32 BIT
MICRO
z
32 BIT
MICRO
W/S
W/S
W/S
Recent developments by Autotrol and Applicon indicate that unbundling of software may be an emerging trend.
unbundled software on DEC-VAX.
Both these companies offer
If unbundling becomes a trend and GKS is -
accepted as the workstation interface standard, CAD/CAM software may
eventually become independent of both CPU and workstation hardware. The
software and workstation could be upward compatible, not only to other
CPU's in a single manufacturer's line, but also across lines to CPU's
from other major manufacturers.
The corporation would be able to choose
CPU, workstation, and software based on the merits of each, or choose-a
packaged turnkey system which includes all three components.
The final item on the list of future workstation trends is
ergonomics and aesthetics. The major turnkey vendors and workstation
manufacturers will continue to pay increasing attention to workstation
ergonomics.
Various international standards will make this a necessity.
The user will benefit with an improved work environment.
Many CAD/CAM
users spend a full 8-hour workday at the workstation and deserve maximum
consideration in workstation feature design.
Aesthetics of the work-
station will gain in importance in future years.
FINAL CONSIDERATIONS
Our summary of trends for CAD/CAM workstations in the 1980's focuses on
three important areas. The first area is the effective use of emerginp
new standards such as ICES and GKS.
These standards will allow more
freedom in moving applications software and data from one CPU to another,
and between the workstation and the CPU.
Each branch in the CAD/CA>! Tree
has had major R & D cost implications for turnkey vendors and software
companies.
There have also been transition cost for the users in taking
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
o
0)
(D
cn
3
O
O
•5
o
<t>
a.
CO
c
CT
CA
g!
—,
><
o_
>
p
z
5^
(A
(D
SUMMARY
o
o
3
•o
B>
3
N)
O
?
d
• TRANSPORTABLE APPLICATIONS SOFTWARE
Q.
•D
0)
• MAXIMUM USE OF COMPUTER POWER
.1
• WORKSTATION HARDWARE MIGRATION
5
o
00
(O
-vl
(O
o
o
o
2.
(D
Pijiure 17
advantage of available new technology.
Product line continuity and user
satisfaction will be maximized if future R & D efforts give full
consideration to these new standards.
The second area is the utilization of an effective blend of
mainframe, supermini, and microprocessor power.
Adding to the existing
base of CAD design softw.are is only a portion of the challenge to be
addressed in this decade.
Group technology and process planning will
alter the balance of computer resources required for CAD/CAM in the
coming years. A flexible network of CPU resources will be an advantage
in meeting this challenge.
The third area of importance is workstation hardware migration.
Graphics functions require parallel processing which is a different type
of problem than the serial processing required for the CAD/CAII data base.
Parallel processing is most easily handled within the workstation
hardware framework.
The workstation hardware migration path leads to
full realization of system growth potential.
Consideration to all three
of these areas is important to final success.
Dataquesf Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
GLOSSARY
ANTI-ALIASING: The process of "smoothing" a line on a raster CRT by
providing a more accurate pixel (dot) approximation to the line than
provided by the normal CRT pixel addressing.
BACK SURFACE TEST: Removal of any surface (plane) which is located on
a rear surface (farthest from viewer) of a displayed image. Gives
an approximation to Hidden Surface Removal for convex objects.
CLIPPING: Defining a boundary for a screen image. The clip boundary is
often used to eliminate the portion of the image falling outside the
boundary, similar to clipping a newspaper article.
DEPTH CUEING: Providing a visual indication of the third dimension by
varying the line color or intensity.
*HIDDEN SURFACE REMOVAL: Representation of solid 3-dimensional objects on
the screen in which all surfaces not normally viewable are eliminated.
*LIGHT SOURCE SHADING: Representation of a solid 3-dimensional object on
the screen depicted as if a light emanating from a single point
source were being shone on the object.
PAN:
Translation of an object across the screen to a new position.
PERSPECTIVE: Simulation of depth and distance by representing parallel
lines merging at a vanishing point.
*POLYGON SHADING: Providing a unique color or intensity for a planar
polygon (constant
shading). The polygons often form the surface
boundary of a 3-D solid object. Polygon shading can be the final
step in light source shading.
3-D ROTATION: Alteration of the angle of view. Generally applied to a
3-D "wire frame" model composed of line segments defining the
surface boundary.
ZOOM:
*NOTE:
To scale up or down the size of an image.
The calculation of the number of instructions required for these
functions in Figure 5 is based on .25" x .25" square polygons,
102A X 1024 pixel addressability, and a 19" CRT. Other functions
are generally applicable to lines and points.
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
Dataquest
Dataquest
COMPUTER-INTEGRATED MANUFACTURINGTECHNOLOGICAL IMPERATIVE
Phillips W. Smith
Corporate Vice President
Computervision Corporation
Dr. Smith is a Corporate Vice President of the Computervision Corporation
and is also a Corporate Officer. His responsibilities include corporate
strategic planning, product line management for hardware and software
products,
pricing,
and
competitive
analysis.
He
has
direct
responsibility for the seven vertical industry segments, aerospace,
autanotive, AEC, mechanical machinery, electrical machinery, fabricated
metals, and manufacturing/construction prcxJucts. Prior to his current
assignment, he was President of Applied Research Laboratories, a division
of Bausch and Lomb.
Previously, he was Division President of Lear
Siegler's Environmental Technology Division. His experience includes the
general
management
of
high-technology,
internationally
oriented
corporations. Dr. Smith studied Engineering at the U.S. Military Academy
(West Point) and received his M.B.A. from Michigan State University and
his Ph.D. in Business Administration from St. Louis University.
Dataquest Incorporated
CAD/CAM INDUSTRY CONFERENCE
September 26-28, 1983
Newport Beach, California
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
COMPUTER-IIMTEGRflTED MflNUFflCTURING—THE TECHNOLOGICflL IMPERATIUE
Intense competition is forcing manufacturers all ooer the world
to find new ways to increase productioity. One method that has prooen
particurlarly effective is the application of computer equipment to
oarious manufacturing operations, But to reap maximum benefits, all
these computerized operations need to be integrated into one
centralized system, The term computer-integrated manufacturing, or
CIM, has eooloed to describe this integration.
But calling it computer-integrated manufacturing can be
slightly misleading. CIM is a global concept that includes far more
than the traditional manufacturing processes found in the factory such
as drilling, punching or stamping. It integrates all data processing
functions within the company, including financial accounting,
inventory control and payroll, as well as those traditional
manufacturing operations.
In the next half hour or so, I would like to define why I
belieoe CIM is a technological imperative. I'll trace some of CIM's
technological roots, talk about some of the future directions I think
the technology is heading, use Computeroision as an example to show
how the marketing strategy of a CAD/CAM company has shifted to
accomodate these sweeping changes, and discuss a few of the obstacles
that I think are slowing the implementation of CIM technology.
Dataquest Incorporated, A Subsiidiary of A.C. Nielsen Connpany /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
Since the mid-1960s, industrial countries in Europe and Asia
haoe been gaining market share in such industries as steel,
automobiles, electronic products and machine tools,
These countries
haoe captured market share by offering higher quality products that
provide more performance at less cost.
While these Industrial
countries often pay lower labor costs, they have also been very
agiiressive at installing factory automation systems to boost overall
productivity.
By Investing in and effectively implementing these automated
systems in their plants, Asian and European manufacturers Are creating
superior products that cost less to produce, providing significant
market advantages.
One of the most poignant examples of this advantage can be
found in the automotive Industry.
U.S. automakers, for example, are
simply unable to compete effectively against foreign manufacturers in
the subcompact market.
In a recent WALL STREET JOURIMAL article
entitled "U.S. Car Industry Has Full-Sized Problems in Subcompact
Market", author Amal Nag points out that high production costs are
probably Detroit's single biggest problem in the production of small
cars.
U.S. auto workers are now paid an average of $8.00 more an hour
in wages and benefits than their Japanese counterparts.
And according
to James Harbour, an automotive consultant in Detroit, better factory
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
layouts, more flexible use of workers and automated manufacturing
equipment allow Japanese automakers to assemble a small car in
approximately 18 man-hours compared with as many as 30 hours for
American automakers.
(Figure 1) When all the cost advantages are
considered, the Japanese haoe an approximate $2500 cost advantage ooer
U.S. manufacturers on each subcompact they produce.
THE EUOLUTION OF CIM
With the increasing sophistication and decreasing price of
microprocessors, factory automation equipment such as computer-aided
design and manufacturing (CAO/CAM) systems, robots, numerical control
mci'liines, material requirement planning systems and coordinate
measurement machines are permeating manufacturing facilities.
Rapid
advances in a number of areas are pushing the development of
increasingly sophisticated manufacturing equipment, including:
1)
Computers—from micros up to mainframes
2)
Inexpensive electronic data storaye
3)
Data communications systems
4)
Sensors and actuators
5)
Machine recognition of visual patterns
6)
Uoice input and output for machines
7)
Electronic graphic displays
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
CO
z>
c
CO
CL
cd
-D
(0
CO O
1
1
1
1
o
o
o
T—
o
CO
21
CM
<0
0)
•AB
ss
C 3
(0"a
Q.O
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
Just as the steam engine seroed as the catalyst for the Industrial
Reoolutlon, the microprocessor is playing an analagous role '• ^ I he
automated factory.
Until recently, these oarious automation technologies haue been
deueloping somewhat independently, creating "islands of automation" in
both the office and on the shop floor.
But the key to inc ^ ised
productioity in the future will be the successful linking of these
independent islands into one centralized computer integrated
manufacturing system.
(Figure 2)
Two fundamental aspects of manufacturing, originally defined by
Dr. Joseph Harrington, Jr. of Arthur D. Little, help to explain why
CIM is SO important to manufacturing.
The first point is the fact
that manufacturing—from product design through production to service
support—is a monolithic, indioisible function.
Because all the
oarious components are interrelated, no single part can be considered
singularly, but only by its relationship to the whole.
Second, the
common denominator for all manufacturing operations is the processing
of data.
The creation, storing, analysis, transmi
• n ,ind
modification of data is key to all the operations of a manufacturing
company.
The importance of computer-integrated manufacturing is that it
directly addresses these two fundamental aspects.
By succe^s. ( n lly
linking together and integrating all operations through a common data
link, CIM maximizes and streamlines the manufacturing process.
The ultimate goal is to link all management, engineering and
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
production functions, and integrate these into a heirarchical network
of computers.
This CII^I heirarchy prooides four major benefits:
electronic communications, decision support analysis, automated
equipment and database management.
Electronic communication provides decision makers with the ability
to access up-to-date information on a wide range of actiuities,
including shop floor operations, new product design.s, current orders,
process plans and inventory. Decisions based on this information can
be rapidly returned to the appropriate area for immediate and timely
execution.
With decision support analysis capabilities, managers can easily
access and manipulate large amounts of data to make more meaningful
decisions based on costs and benefits.
Automated equipment such as numerical control machines and CAD/CAM
systems can substantially improve quality and reliability, streamline
material flow, and improoe inventory control.
Database management is perhaps one of the most important benefits
that CIM systems can offer. It helps to store, track, update and
retrieve data generated from and required by each area of the
automated manufacturing operation.
The heart of database management is the centralized product
database.
This product database is the central reservoir for all the
information about a particular product—the design information needed
to actually create it, cost variables and production flow.
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
Information created in one application area is stored and then easily
accessed and used by other applicatons.
As each application adds its
specific information, the product database becomes more useful to all
areas.
Computer-integrated manufacturing technology offers progressioely
greater benefits as it integrates more and more sectors of a plant's
operations.
Ualue added comes from the productioity benefits gained
at each step of the operation as a result of sharing and capita] :/ing
on information from the same centralized database.
Since data for
each area of automation does not haoe to be re-entered each time,
costly redefinition and reformatting errors are eliminated.
And
because the information is more accurate, product quality and
reliability are enhanced.
Another important benefit is that
communication betiAjeen design, manufacturing and the other major groups
of a company is oastly improoed as many of the traditional barriers
between groups are broken down.
(Figure 3)
COMPUTERUISION'S CIM STRATEGY
Computeroision has recently announced a major change in strategy
that is designed to offer companies a wider range of productioe CIM
solutions.
I would like to use my company as an example to show how the
marketing strategy of a CAD/CAM company has changed to maximize the
benefits of integrated manufacturing operations.
Computeroision has traditionally supplied turnkey CAD/CAM systems
for product design and manufacturing processes, including
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
o
ffi
S
SI
c
(D
3
O
O
•o
o
(D
Q.
>
OT
c
CT
U
£
»'
o_
>
p
z
•2
5.
CO
»
o
O
3
•D
IB
Centralized
Product
Data Base
lO
(O
o
a.
OL
o
—\
-o
D>
CO
O
CO
O
>
Engineering
§
O
o
2.
(V
(O
-.1
CO
computer-aided engineering, computer-aided design and computer-aided
manufacturing. These products and seroices haue established the
company as market share leader in these segments of the industry
automation market.
(Fifjure 4)
But Computeroision has recently chosen to redefine and expand the
range of markets it serves from the traditional core of
graphics-oriented turnkey CAD/CAM applications to a broader range of
engineering and manufacturing actioities encompassing total product
data management and control.
This strategy was deoeloped in response ho market demands for
expanded capabilities in several major areas: improved functionality
and performance in the core software application areas; en' MM i?d
abilities to manage and control expanding product design and
manufacturing databases; and greater capabilities to shar • 'he
benefits of this technology throughout engineering, manufacturing and
all Other information management functions within a company.
Computeroision's new strategy is to prooide customers with
capabilities in three areas. First, the traditional
single-user/focused application enoiornment will incorporate
intelligent workstations, standard operating systems, and local area
networking, and will spread the auailability of CAD/CAM technology
throughout engineering and manufacturing operations.
Engineers will
be able to easily access drawings and information from small,
intelligent workstations that will be linked together in large
networks.
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
0)
^
o
cgSjo
o^B-
'"DQ-S
m < CO
I^OS
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
Secorid, the multiuser/multiapplication enoironment, the core of
our traditional offerings, will continue to be enhanced in both
functionality and performance. Our new CDS 4000 product line, for
example, is the first member of this new product family.
And third, a large scale system that will be linked to large main
frame computers and provide total database management for an • ni Ire
company. Huge databases will be supported on-line, so users can
integrate all data processing applications into a single i.ilL for
comprehensioe manufacturing automation.
This three tier strategy is aimed at providing total CIM
capabilities to manufacturing companies of almost euery size and
need—ranging from standalone engineering workstations to large
comp<uiy uiide database management systems .
Computervision has traditionally developed the majority of its
products internally, but a part of this new strategy is to address the
needs of the expanded marketplace not only through internal pi .i.lIict
development, but also by joining forces with other industry leaders in
CAD/CAM related fields. The company has acquired a number of software
firms that provide key application capabilities. Our recent purchase
of the Organization for Industrial Research, for example, provides
group technology and process planning capabilities—key elements of a
CIM configuration.
To begin the implementation of this strategy, Computervision has
also entered into a joint development agreement with Sun Mic
^ '
LMMS
for the development of engineering workstations, and recently signed
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
an c!;j: jment with IBM Corporation for the purchase of IBM products for
integration in Computeroision syiems..
This new agreement with IBM will prouide a means to link together
all the information resources within a company to create what mig!.! be
called a configuration management system, and the Sun agreement will
prouide a series of powerful intelligent workstations.
FUTURE DIRECTIONS
But where will all this lead?
When engineers create a product
with CAD/CAM today, they typically use a series of part drawings.
While all information reuoloes around this part drawing, there is
currently no effectioe way to relate part drawings with one another.
When an engineering change order requires a change in one coi .
riL of
a subassembly, for example, there is no way to determine how that
change will affect all the other components in the same assembly.
There can often be as many as 20,000 or more part drawings for a
particular assembly, creating a logistics nightmare to control and
access all the relevant information.
The configuration management system incorporates an ex"
; • Lye
on-line storage capability that will allow all drawings for
subassemblies and other components to be integrated and associ.-'i-'d.
When a change is made on one component of an assembly, the system will
quickly identify all part drawings affected by that change.
Eoen with
as iiinny as 20,000 drawings on file, the configuration management
system will be able to quickly search the entire file.
"' ' ' ige
Dataquest Incorporated, A Subsidiary of A.G. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
on-l1ri;'
ory uiill also make group technology procedures eoen more
effectiue as the system can search for similar designs, process plans
and tooling that may haue been preuiously created, minimizing design
redundancy,
The traditional data processing functions of a company haoe
eoolued independently from the graphic CAD/CA(«I functions, but it will
be essential that the two databases be linked together if an effective
company-wide computer-integrated manufacturing system i
'''•^>ed.
When the traditional data processing functions such as financial
accounting, inventory control and payroll are integrated wilih
manufacturing functions such as numerical control, process planning
and materials requirement planning, the resultant integrated di ' •'•a^ze
can provide a company with an extremely powerful pool of information.
This information can be used for a variety of "what if" scenariir^ to
determine optimum economies of scale, and allow companies to
accurately project total costs for a specific project.
I would even go so far as to project that the size of the
manufacturing database will ultimately be anywhere from 20 to 50 times
larger bhan most current data processing databases.
This extensive
database will be composed of process plans, fixture and t M"! Lng
designs, NC programs, quality control and inspection programs, robot
programs and process programs.
With as many as 48 billion !ii/les
devoliM: to manufacturing data compared to perhaps two billion bytes
for data processing, it is quite conceivable that data processing will
ultimately become a subset of the manufacturing operation once a truly
integrated CIM operation has been put i> " ;Mace.
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
Intelligent engineering workstations will also become increasingly
important to a wide oariety of job functions.- Data is already being
generated much faster than it can be consumed.
Ultimately, the
paperless office and factory will become a reality with information
accessed on a terminal rather than a piece of paper.
Eoen more
documentation and other supporting data will accompany specifi
drawings and information.
i-^art
Working at an engineering workstation, an
indiuidual will be able to combine documentation, word processing,
business graphics, reports and a variety of additional information and
haoe it all tied directly to a specific part driudng.
For example, a part design might be sent to a finite element
modeling specialist for analysis.
The specialist would .^
.iii? a
mesh on the engineering workstation, and then access a larger
processor to run an analysis on the part using Strudl or some similar
andlv/sis package.
He would then write a summary of the analysis,
perhaps defining a structural problem in a specific area of the part.
When he sends his report back, he would include a file showing the
mesh and the actual dynamic deformation of the part to supper i his
summary and conclusions,
l*1any networking and communication problems still must be re^wlued
before the integration of the alphanumeric data from data processing
and the graphics CftD/CAM data can take place.
But significant
prugre;i;s is being made in this area, and it is likely that these
problems will soon be resolued.
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
OBSTACLES TO CIM IMPLEMENTATION
With all of the tremendous aduantages CIM systems can offer, !Aihy
aren't more of them currently in place?
wide-ranging.
The reasons are complex and
But an understanding of the problems is usually one of
the best ways to begin to solue them.
Most manufacturers will agree
that CIM offers significant advantages in terms of inci'
^ing
productiuity, lowering costs for materials, labor and energy, and
improving product quality.
But because of its integrative nature, f:[|wi
technology is haoing and will continue to haoe a profound impact on a
company's industrial organization, occupational structure and labor
relations.
One of the major roadblocks to the implementation of CIM Sj,' ' .>ins
is the apparent lack of long-term planning in American industry.
Up
to and during World War II, the movers and shakers of American
industry were individuals from manufacturing and production.
Production specialists showed their genius during World War ?", nnd
were largely responsible for quickly and radically stepping up
production during the war and in the economic boom years that follcmed,
With the proliferation of new products during the 19BOs and 1960s,
companies shifted their central focus to marketing as a means to gain
atlt^-ntion for their products.
A few years later, with the advent of conglomerates and mergers,
the central focus changed again to lawyers and financial experts.
Manufacturing companies soon began to give in to the pressure of
steadily increasing revenues each quarter to insure glowing financial
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
rr-ijiM hs .
Long range planning was largely forgotten in the scramble
for consistent quarterly increases.
Today, when companies consider large capital investment decisions,
the financial people are usually there asking how soon a return O" Lhe
investment can be expected, what effect the inuestment will haoe on
earnings, and how they can justify it to stock holders.
This s'lurt
term iliinking is peroasioe in American industry, and while still
important, it fails to take into consideration the essential long term
quesl. ion of whether the company will eoen be able to compete in the
international marketplace in fiue to ten years.
The f±\<'<>< ial
representatives typically are unfamiliar with manufacturing processes
and naturally gravitate toward options with the lowest perc ^jed
risk- a classic case of not being able to see the forest for the trees,
A recent article in the HARUARD BUSINESS REUIEW by Belo ".-.Id,
entitled "CAM Sets New Rules for Production", states several more
specific reasons why American managers have been slow to realize the
significance of CII*1:
1)
They don't realize that CIM is a special kind of technology,
whose adoption requires a broader level of analysis than is
normally applied to the purchase of equipment and f • "^ii ies.
2)
They rely on bottom-up impetus for the generation of new
equipment i- i"v<i^ls.
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
3)
They still depend on traditional capital budgeting techniques
to evaluate proposals.
CIM provides a systematic capability by integrating adjacent and
interrelated operations, so neither the purchase or the performain e of
CIM systems should be evaluated in the traditional way a single piece
of equipment such as a machine tool is evaluated.
When managers
insisi on evaluating CIM systems with these narrow attitudes, it
becomes very difficult to make meaningful decisions about a new
technology that has implications for the entire company.
The existing manufacturing organization often presents signifiiant
barriers to the implementation of CIM systems.
The traditional
heirarchical structure firmly in place in most plants has these
inherent problems:
a)
Policies, systems and procedures are outlihed.
b)
Work tasks are often polarized.
c)
Top-down chain of command allows little feedback fi
' ^ Lorn
up.
d)
Fixed logic management decision systems are not equipped to
handl<» the variable decisions needed to justify and maintain
CIM sys ! lems .
Some experts are calling for a complete revamping of the manufacturing
organization from a heirachical structure to more of a matrix
structure where information is easily accessed by individuals at every
Dataquest Incorporated, A Subsidiary of A.G. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
leoel.
The fixed logic so prevalent in the industrial era is giuing
way to a more variable logic made possible by the information
reoolution.
When individuals at all levels can quickly and easily
access information, communication is vastly improved and more
intelligent decisions can be made.
The lack of properly trained personnel also contributes to the
problem.
CIM technologies are complex and changing rapidly.
Many
have only been developed in the last five to ten years, so it can
often be difficult to find people uiith the necessary background and
experience.
With the trend toward more and more specialization, it
can also be difficult to find people that understand how all the
various technologies integrate into one centralized system.
This problem is only temporary, however, because as more and more
systems are put into place, a larger pool of experienced people will
be available who understand the wide-ranging implications of CIM
systems.
Another problem is that, with CIM technologies evolving so
quickly, it becomes very difficult to keep up with new developments.
(Figure 5)
Trade magazines and industrial trade shows and seminars
can help, but the technology is changing so rapidly that manufacturers
still have a general lack of understanding of CIM systems and their
implications on the manufacturing process.
CIM has a significant
impact on the structure, competitive positioning, financial
performance, employment patterns and labor relations of a company, so
in some respects it is understandable why managers have been reluctant
to enbrace CIM.
Computer-integrated manufacturing technology is still
Dataquest Incorporated, A Subsidiary of A.G. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Parl< Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
in the somewhat early stages of deoelopment, and there are some gaps
in the technology.
Users must still bear the burden of identifying
and developing many specific applications.
They must also choose
among a wide range of potential systems, and are still largely on
their own in integrating the entire CIM structure in their
manufacturing operations.
But their failure to at least learn more
about the advantages of CII<1 will haue serious long-term implications.
THE TECHNOLOGICflL IMPERflTIUE
Computer-integrated manufacturing prooides engineers and
manufacturers with a powerful new set of tools.
But these individuals
must educate themselves if they are to successfully capitalize on
CIM's advantages.
Because CIM drastically changes the traditional way
a new product is designed and manufactured, massive revamping of
current design and manufacturing techniques must take place.
of this magnitude can only come from one place:
Change
top management.
But this goes against the grain of the typical manner in which a
company with a hierarchical structure purchases capital manufacturing
equipment.
Change in this type of company typically starts at the
bottom and works its way up.
I^lhen purchasing a machine tool, for
example, the bottom up approach is fine because the person or group
actually using the system knows their needs best.
But because Clfi
integrates and automates every phase of the manufacturing process, it
must be the responsibility of top management to insure effective
implementation in the face of almost certain reactionary resistance at
all lower levels.
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
If CIM Is to be successfully implemented, the old philosophy of
designing the system to meet user requirements must be abandoned.
Instead, manufacturers must carefully evaluate and understand user
requirements relative to the total manufacturing picture rather than
for indloldual elements of it.
Manufacturers face a dramatic change in the methods and operations
as fundamental as the changes that set the stage for the Industrial
Reuolution.
Until now, relatively few companies have risen to the
challenge, and many are not eoen aware that it exists.
But most are
starting to see the effects.
The free
market system, with all of its pluses and minuses, does
Insure one thing:
only the strongest competitors suruioe.
CIM truly
is a technological imperative that can help to insure economic
survival in an Increasingly competitive markelp"!.ice.
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
(::)a Dataquest
Dataquest
COMPUTER CC»4PANIES AND INTEGRATION
Richard L. Gimbel
Manager, End-User Marketing
Apollo Computer, Inc.
Mr. Gimbel is Manager of End-User Marketing for Apollo Computer, Inc. He
is responsible for developing and implementing marketing support programs
for
established
markets
and
customers, as well
as
identifying,
developing, and implementing business plans for new end-user markets.
Prior to joining Apollo, Mr. Gimbel worked for Digital Equipment
Corporation, where he held a number of positions within Digital's
Engineering Systems Group, most recently as Manager of Marketing
Programs. Before joining Digital, Mr. Gimbel was a Software Engineer for
Control Data Corporation.
Mr. Gimbel holds a B.S. degree from the
University of Oregon and an M.S. degree in Computer Science from Purdue
University.
Dataquest Incorporated
CAD/CAM INDUSTRY CONFERENCE
September 26-28, 1983
Newport Beach, California
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
D
m
ST
r>
c
(D
CO
cpolb
computer h a
w
c
CT
(A
gl
B)'
>
p
z
!5.
0)
(D
o
ro
CO
o
XI
a.
a.
CD
T)
0)
3r
O
0)
ni
o
u>
P
O
>
CO
o
QO
CO
o
o
o
CO
-J
CO
COMPUTER COMPANIES
AND
INTEGRATION
o
lu
E
apollo
computer Ina
>
w
c
C7
m
a
0)'
-1
•<
>
b
COMPUTER COMPANIES
AND INTEGRATION
o
o
3
• INTEGRAnON: DEFINITION AND REQUIREMENTS
CO
o
a
a.
a
• ROLE OF COMPUTER VENDOR
• "THE INDUSTRr
w
(u
3
o
>
o
00
(D
O
o
O
iL
UD
-J
CJ
D
01
•o
o
^
apolb
computerlna
en
c
CJ
CI
» •
>
b
iL
0)
(D
o
o
COMPUTER INTEGRATED
MANUFACTURING
3
•o
01
3
lO
• TASK AUTOMATION
s
ci
Q.
(D
T)
01
• UNITE INTERRELATED TASKS
• VIEWED AS AUTOMATING THE ffiOOEM
CO
01
O
>
•tv
O
00
CO
o
o
o
(O
-NI
CO
o
c
•D
O
cpob
computer ha,
>
w
Ia
PROCESS M •
>
o
o
o
3
•a
01
N)
to
O
a.
a.
a>
TJ
o>
w
3
o
>
•fc.
o
00
TECHNICAL TASKS AUTOMATED
CO
o
o
o
--J
CO
~J
CJ
INFORMATION SHARED/COMMUNICATED
i
D
a
X)
c
(D
U>
O
O
-1
•o
o
(D
cpob
computer ho,
a
>
CD
a.
a
EXTENDED PROCESS
>
OFFICE AUTOMATION
DEaSION SUPPORT
PROJECT MANAGEMENT TRAINING
b
CD
o
o
3
•o
lu
3
l>0
CO
o
Q!
Q.
a>
T3
(U
a
(/>
01
c_
O
0>
O
>
PROCESS IS MANAGED
o
OD
<o
O
o
o
to
• i.
EXISTS WITHIN AN ORGANIZATION
o
o
q
"O
o
cpdlo
computer IriK
p.
>
a.
a'
>
p
z
o
o
INTEGRATION
UNITE TASI^ INTO MORE EFFICIENT PROCESS
3
INJ
CO
o
CONSIDER ALL ASPECTS OF WORK PROCESS
a.
a.
(D
T3
lU
5-
BENEFITS
w
0)
o
u>
P
O
>
ELIMINATES INFORMATION TRANSFER STEPS/ERRORS
<D
CJ1
O
00
<D
O
O
O
CO
-si
CJ
• MORE PREDICTABLE/MANAGEABLE ACTIVITY
o
01
o
o
"O
o
(D
p.
cpollo
i
computerlna
—
'— 1
>
(f)
c
IT
Q.'
B)'
>
b
CIM IMPLEMENTATION CONSIDERATIONS
2.
(fl
(D
3
o
o
3
• WIDE RANGE OF APPLICATIONS
5
• RANGE OF COMPUTING ENGINES
•o
A)
3
d
Q.
(D
•D
0)
• NETWORK AND DATA MANAGEMENT
CO
0)
• HUMAN INTERFACE
o
>
CO
• SUPPORT
i
<D
O
O
O
5.
(D
X
~g
(O
->i
w
o
[u
1
cpdo
computer ITKX
>
C/J
c
cr
(0
Q.'
0)'
>
O
o
M
(O
O
3)
ROLE OF A COMPUTER VENDOR
CONSIDERATIONS
• MANY MARKET SEGMENTS
CL
a.
(B
TJ
0)
CO
0)
• PRODUCT LIFE CYCLES
• INTERLOCKING TECHNOUXIIES
o
(A
(D
o
>
CO
CJl
i
CO
o
o
o
• INERTIA OF INSTALLED SOLUTIONS
o
tu
3
O
o
o
•o
(D
Q.
cpob
computer IfKL
>
g.
0)'
>
o
o
o
ROLE OF A COMPUTER VENDOR
APPLICATIONS
3
•o
0)
lo
<o
o
• SOFTWARE DEVELOPMENT TOOLS
a.
Q.
(D
-^
•D
01
0)
o
tn
o
>
<o
o
o
o
CO
-~i
CO
• 3RD PARTY PROGRAM
• "GENERIC" OFFERINGS
o
lu
3
O
o
(D
cpollo
cofuputerha
a.
W
c
CT
U
g!
0)'
•<
>
O
ROLE OF A COMPUTER VENDOR
COMPUTING ENGINES
o
o
3
•a
01
3
lO
CO
o
• MANAGE TECHNOLOGY
A. Architecture
Q!
Q.
m
-a
0)
B. Standards
C. Alternatives
(/>
lu
o
01
(D
• GENERAL PURPOSE WITH MARKET SPECIFIC OPTIONS
O
>
(O
OI
i
(O
o
o
o
(0
• PROTECT SOFT INVESTMENTS
o
2.
oT
n
c
(B
E
(D
Q.
cpollo
computer inc.
>
0)
Q.
B)'
>
b
COMMUNICATIONS
GATEWAY SERVER
HLEANO
PERIPHERAL SERVER
n
:J
O
o
3
T3
01
3
lU
to
o
3}
a
a.
<»
-^
TJ
lU
W
D)
3
o
>
•u
o
CO
to
oo
o
•vi
USER INrEnFM;ED PERIPHERAiS
o
SI
o
o
•5
o
cpolb
I
con«3ulefln&
^j-«n>liB#l
>
CO
c
IT
(J>
a!
a)'
-s
•<
o
>
p
ROLE OF A COMPUTER VENDOR
2
CD
0)
ID
i
NETWORK/DATA MANAGEMENT
D>
•<
i
• BASIC FILE TRANSFER
3J
CL
Q.
i
DATA BASE & TRANSUTOR TECHNIQUES
o
01
o
u>
(D
o
>
.t>.
o
00
CO
o
o
o
5.
(D
DOMAIN NETWORK DATA MANAGEMENT
a
Si.
w
n
c
(D
E
(D
Q.
W
c
CT
CO
9!
B)'
>
b
o
o
3
"O
0)
3
l\3
CO
o
3)
•.'
Q.
(D
—%
•D
Di
W
01
o
>
CO
tn
j^
o
00
CO
o
o
o
CO
->J
CO
cpollO
compuferhCL
DOMAIN CAPABILITIES
o
01
^
cpollo
coinputorhc.
w
c
cr
to
g!
0)'
>
O
ROLE OF A COMPUTER VENDOR
(D
o
O
3
•o
(U
8
I
5-
w
(u
O
>
.(»
o
00
<o
o
o
o
5.
tD
CO
•vl
w
CONSISTENT HUMAN INTERR^^E
B MULTI'TASKED WINDOWS AS FRONT END INTEGRATOR'
• EASE OF USE FEATURES
o
O
O
•5
o
(D
Q.
03
C
CT
«
a
cipollo
computer he
B)'
>
o
ROLE OF A COMPUTER VENDOR
O
o
3
•o
0)
-<
I
S?
W
D)
3
o
>
.b.
o
00
to
o
o
o
CD.
CD
X
-.1
CO
-vl
u
• MANAGED TECHNOLOGY
• INDUSTRY RELATIONSHIPS
(0
3
a
LU
X
CO
flC
o
o
z
u
>
R
u
o
o
<
o
E
<
oc
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Pafl< Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
a
01
o
o
—.
•o
o
(D
a.
cpdo
computorlne*
>
w
I
THE INDUSTRY
2,
2.
CO
CD
O
o
3
•o
ID
3
ro
to
o
Q.
a
(D
TJ
o>
CO
g>
c_
O
u>
_eo
O
>
.c
o
00
(O
o
o
o
(D
(O
•CO
^^
ARCHITECTURAL EVOLUTION
o
(U
o
o
-1
•o
o
(D
a.
w
c
cpob
computer knet
CT
(A
THE INDUSTRY
a!
0)'
>
o
o
o
3
"HIGH PERFORMANCE PERSONAL
NETWORK-BASED COMPUTER SYSTEMS'
T3
0)
•<
ro
(O
o
HIGH PERFORMANCE
SUPERMINI-CUSS WITH PREDICTABLE
RESPONSE PER USER
PERSONAL
COMPUTER WITH INTERACTIVE 6lUI>HlCS
PERUSER
NETWORK-BASED
INTEGRATED COMPUTING ENVIRONMENT WITH
ECONOMICAL RESOURCE SHARING AND
MODUUR GROWTH
a
a
CD
-1
TJ
a>
w
O
>
I
CO
o
o
o
£5.
CD
o
fi)
^
o
o
—1
•o
o
9-
cpob
' compul8rh&
>
w
Ia
>
o
CO
zs
O
o
3
•a
lu
-<
lO
§
DOMAIN PROCESSING SYSTEM
LARGE MACHINE PERFORMANCE AND FUNCTlONALHY
DEDICATED TO COMPUTER PROFESSIONALS
-32-Brr HIGH PERFORMANCE PROCESSOR
-HIGH RESOLUTION, BIT-MAPPED GRAPHICS
31
a
a
CD
T)
B>
03
a>
3
o
>
§
<o
o
o
o
CO
•>l
CO
NETWORKED COMMUNITY OF USERS
-HIGH-SPEED LOCAL AREA NETWORK
-NETWORK-WIDE VIRTUAL MEMORY OPERATING SYSTEM
-DISTRIBUTED DATA BASE MANAGEMENT SYSTEM
I
o
o
•3
o
(D
a.
^
c
cr
(fl
9!
apdb
' compLrterha
0)"
•<
o^
>
O
THE INDUSTRY
o
o
3
•o
0)
3
STANDARDS?!
CO
o
a
a
(D
-1
•o
0>
w
0)
3
C_
o
(A
(B
O
>
CO
i
CO
o
o
o
iE.
CD
CO
CO
SOFTWARE
A. SUPPORT
B. LICENSING
C. TRAINING
D. CUSTOMIZATION
o
u
u
X)
c
(D
3
O
o
-1
•o
o
<D
p.
>
w
c
cr
Cfl
cpdb
compiler kK;
9:
-^
-<
tu'
o^
>
b
(D
o
COMPUTER COMPANIES AND INTEGRATION
O
3
•o
lU
Q.
a
(D
"INTEGRATE OR PERISH"
w
01
o
>
(O
01
.1^
3o
(O
-Ni
<o
o
o
o
iS.
CD
~J
o
(u
o
o
•5
o
(D
a.
>
w
o
o
apoio
computer inc.
o
CO
0)
3
o
>
o
o
o
™.
CD
X
-^J
to
-vl
CO
Dataquest
Dataquest
THE IMPACT OF ENGINEERING WORKSTATION
NETWORKS ON CAD PRODUCTIVITY
David E. Weisberg
Manager, Market Analysis
Auto-trol Technology Corporation
Mr. Weisberg is Manager, Market Analysis, at Auto-trol Technology
Corporation, where he is responsible for the evaluation of industry
trends within the ccsnputer graphics and CAD/CAM industries. He has also
worked in management positions involving Product Planning, Product
Marketing,
and
Applications
Software
Development
at
Auto-trol.
Previously, he was Field Marketing Manager for Tektronix's Mechanical
Engineering Graphics activity.
He has been engaged in the computer
graphics field in both the development and marketing of software/hardware
systems since 1961. His work has been in applications areas including
cartography, process
control,
IC
design, mechanical
design,
and
architecture and engineering design. Mr. Weisberg received his B.S. and
M.S, degrees in Civil Engineering
from Massachusetts
Institute of
Technology.
Dataquest Incorporated
CAD/CAM INDUSTRY CONFERENCE
September 26-28, 1983
Newport Beach, California
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
PERFORMANCE
A N D P R O D U C T I V I T Y IN
COMPUTER AIDED DESIGN
SYSTEMS
W i t h C o m p u t e r Aided D e s i g n < C A D ) s y s t e m s becoming a common e n g i n e e r i n g
tool, it is n e c e s s a r y that there be a solid u n d e r s t a n d i n g of the i m p o r t a n c e
of o v e r a l l system p e r f o r m a n c e and its impact on p r o d u c t i v i t y . This paper
p r e s e n t s r e a s o n s why a p r e d i c t a b l e level of p e r f o r m a n c e can e n h a n c e the
p r o d u c t i v i t y of system u s e r s and an a n a l y s i s of the f a c t o r s that can impact
p e r f o r m a n c e . P e r h a p s this will also e x p l a i n why s y s t e m s can look very good
u n d e r c e r t a i n c i r c u m s t a n c e s and fail to live up to e x p e c t a t i o n s when ins t a l l e d in a real w o r l d e n v i r o n m e n t .
We are now e n t e r i n g the f o u r t h d e c a d e of the s e r i o u s use of c o m p u t e r s
in e n g i n e e r i n g design. D u r i n g this p e r i o d a l m o s t u n b e l i e v a b l e changes have
taken p l a c e in t h e tools that are a v a i l a b l e to the user. With these c h a n g e s
in t e c h n o l o g y have come c h a n g e s in the level of e x p e c t a t i o n s . T h e c o m p u t e r
m y s t i q u e is g o n e when you go home from a hard day at the o f f i c e and find
your p r e t e e n son or d a u g h t e r h u n c h e d over the family A P P L E II or T R S - 8 0 .
T e c h n i c a l m a n a g e m e n t today will not a c c e p t the tools of y e s t e r d a y .
One good e x a m p l e of this is in the area of s t r u c t u r a l d e s i g n using
FINITE
E L E M E N T A N A L Y S I S ( F E A ) . In the early I S S O ' s , the first p r o g r a m s using
this t e c h n i q u e a p p e a r e d on the scene. In o r d e r to use these a n a l y s i s p r o grams, the e n g i n e e r had to m a n u a l l y c r e a t e his node and element m o d e l ,
convert the m o d e l to input on a coding form, have it k e y p u n c h e d , take it to
the c o m p u t e r site, run it, and then r e c e i v e back a huge stack of n u m e r i c a l
listings. T h e e n t i r e p r o c e s s for a m o d e r a t e size problem could take 4 to 6
weeks.
B e c a u s e of the lack of g r a p h i c c a p a b i l i t i e s , e r r o r s were common
This
r e s u l t e d in f r e q u e n t r e r u n s and large a m o u n t s of e n g i n e e r i n g effort spent
r e v i e w i n g coding f o r m s and listings. Today, this sounds like an a g o n i z i n g
p r o c e s s . W e must r e m e m b e r , however, that w i t h o u t the F E A p r o g r a m s , these
a n a l y s e s just could not be done. E f f o r t then began to a d d r e s s s p e e d i n g up
the p r o c e s s .
In the last 2 0 y e a r s we have seen many d e v e l o p m e n t s that have improved
the p r o d u c t i v i t y of u s i n g F E A t e c h n i q u e s . R e m o t e b a t c h e n t r y of p r o b l e m s
e l i m i n a t e d the need to p h y s i c a l l y take card decks to the c o m p u t e r . P l o t t e r s
w e r e used to r e d u c e the huge v o l u m e of p r i n t e d output to a few easy to
u n d e r s t a n d plots. With the advent of i n t e r a c t i v e g r a p h i c s in the e a r l y
1970's a huge s t e p f o r w a r d was made. It was p o s s i b l e to create a model
and submit it for a n a l y s i s from a g r a p h i c s w o r k s t a t i o n . U n f o r t u n a t e l y ,
t i m e - s h a r i n g g r a p h i c s w e r e slow and costly. But It did r e d u c e the t u r n a Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
round time to a few days o r less f o r f a i r l y large p r o b l e m s . The next step
was to use s t a n d - a l o n e m o d e l i n g s y s t e m s or e x t e n s i o n s of C A O systems to
build the model and then transmit the model in a b a t c h m o d e to the c o m p u t e r
doing the a n a l y s i s . This w a s a m o r e e f f i c i e n t use of the e n g i n e e r ' s time
The t u r n a r o u n d time w a s s t i l l m e a s u r e d 'in many hours, if not days.
R e c e n t l y , C A D s y s t e m s have been i m p l e m e n t e d on " s u p e r m i n i s " that a r e
a l s o c a p a b l e of p e r f o r m i n g the actual a n a l y s i s as well. N o w it is p o s s i b l e
to d e s i g n a part, c o n v e r t it to a f i n i t e element m o d e l , add the n e c e s s a r y
loading data, p e r f o r m the a n a l y s i s , view the results and d e c i d e what
Changes need to be m a d e to the m o d e l ; all in one c o n t i n u o u s o p e r a t i o n
For
a m o d e r a t e l y c o m p l e x part, an e n t i r e I t e r a t i o n may only take a few hours.
As the a b o v e d e s c r i b e d p r o c e s s has gone from taking many w e e k s to only
a few hours, the u s e of the t e c h n o l o g y has s k y r o c k e t e d . E n g i n e e r s today
will u s e the t e c h n i q u e b e c a u s e the cost to do s o is a c c e p t a b l e , the
process is e f f e c t i v e , and it is a s a t i s f y i n g way to a c c o m p l i s h the work.
O t h e r e x a m p l e s of d e s i g n time c o m p r e s s i o n can easily be found in areas
such as I n t e g r a t e d C i r c u i t design. P r i n t e d Circuit Board layout, N u m e r i c a l
C o n t r o l , A r c h i t e c t u r e , and P r o c e s s P l a n t design. The key point is that
d e s i g n e r s and e n g i n e e r s today r e q u i r e and need systems that can s u p p o r t
the e x p e c t e d c o m p r e s s i o n of the d e s i g n and d e v e l o p m e n t time cycle. CAO
systems p e r f o r m a n c e is v e r y important to m e e t i n g these o b j e c t i v e s .
As a b a s i c p r i n c i p l e let us take as our thesis that " An e n g i n e e r or
d e s i g n e r is most e f f e c t i v e when his thought p r o c e s s is not d e l a y e d by
a r t i f i c i a l c o n s t r a i n t s " . A second
p r i n c i p l e is that " A n e n g i n e e r or
d e s i g n e r w o r k s m o r e e f f e c t i v e l y with a c o m p u t e r system that has p r e d i c t a b l e
p e r f o r m a n c e than w i t h one w h e r e the p e r f o r m a n c e is u n p r e d i c t a b l e moment to
moment". A third f a c t o r to be c o n s i d e r e d is that a level of system p e r f o r m a n c e that w a s a c c e p t a b l e s e v e r a l y e a r s a g o is no l o n g e r felt to be
a c c e p t a b l e by many u s e r s .
What we a r e d e a l i n g with is the s u b j e c t of the e m o t i o n a l needs of a
t e c h n i c a l l y o r i e n t e d s y s t e m s user. Me a r e not digging d i t c h e s or u n l o a d i n g
boxcars. The d e s i g n e r is w o r k i n g with the system to c r e a t e something m e a n ingful. The thought p r o c e s s e s a r e o f t e n complex and e x t e n s i v e . IF the s y s tem is s l o w to r e s p o n d to an o p e r a t o r action, it is very easy to lose one's
train of thought. P e r h a p s this is o n e r e a s o n why e n g i n e e r s with s l m i l i a r
i n t e l l e c t u a l c a p a b i l i t i e s turn out such d r a m a t i c a l l y d i f f e r e n t v o l u m e s of
completed work. T h e high p e r f o r m e r is the one w h o can best keep himselF
focused on the task at hand. What we want to do with a w e l l d e s i g n e d C A D
system is not c r e a t e any a r t i f i c i a l b a r r i e r s to this c o n c e n t r a t i o n . We want
the system and the u s e r to w o r k t o g e t h e r as one in s o l v i n g the problem. The
easiest way to d e s t r o y this r e l a t i o n s h i p is to f r u s t r a t e the u s e r to the
point that he loses his c o n c e n t r a t i o n .
A very key point is that we a r e not talking Just about blazing speed.
In fact a system that can p e r f o r m e x t r e m e l y fast on o c c a s i o n and on o t h e r
o c c a s i o n s is very slow b e c a u s e of c o n t e n t i o n from o t h e r u s e r s may a c t u a l l y
be c o u n t e r p r o d u c t i v e . What the u s e r needs is a r e a s o n a b l y high level of
p e r f o r m a n c e that he can count on. When the r e s p o n s e of a system is smooth,
the user gets into a rhythm. iHe knows h o w long an a c t i o n is going to take
to e x e c u t e and is p r e p a r e d f o r the next step. A system that d e m o n s t r a t e s
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, GA 95131 / (408) 971-9000 / Telex 171973
A n o t h e r r e a s o n why p r e d i c t a b l e p e r f o r m a n c e is important is to avoid
the f r u s t r a t i o n of d e l a y s h a p p e n i n g that a r e o u t s i d e of the u s e r ' s c o n t r o l .
If a d e s i g n e r r e q u e s t s t h e system to c a l c u l a t e the i n t e r s e c t i o n b e t w e e n
two v e r y complex s u r f a c e s , he knows that this Involves a s u b s t a n t i a l amount
Of c a l c u l a t i o n s but the s y s t e m Is doing t h e s e c a l c u l a t i o n s f o r him. When
s o m e o n e in a n o t h e r room or even a n o t h e r d e p a r t m e n t , r e q u e s t s a complex
a n a l y s i s be p e r f o r m e d It is going to d e g r a d e the p e r f o r m a n c e of the total
system. O u r u s e r does not k n o w why p e r f o r m a n c e has just gone t h r o u g h the
f l o o r , he Just knows that It h a s . "If that is the case, why not Just go
down to the company c o f f e e s h o p and get a cup and m a y b e the s y s t e m will be
cooking when I get back". W h e r e did his train of thought and c o n c e n t r a t i o n
on the p r o b l e m g o ?
U n t i l recently, the p r i m a r y m e a n s of i m p l e m e n t i n g f o r a C A O system was
to u s e a m e d i u m to large m i n i c o m p u t e r w i t h a capacity of 4 to 16 w o r k s t a tions. T h e a r c h i t e c t u r e of these s y s t e m s is v e r y c o n s i s t e n t and it is not
d i f f i c u l t to u n d e r s t a n d h o w p e r f o r m a n c e can s i g n i f i c a n t l y d e g r a d e u n d e r
many c i r c u m s t a n c e s . O n e way to a p p r e c i a t e this is to look at a system as
a series of g a t e s with all the u s e r s trying to get t h r o u g h at once. Some of
these g a t e s as shown in F i g u r e 1 are:
1. A c c e s s to disk f i l e s .
Most C A D tasks a r e v e r y disk file I n t e n s i v e . This i n v o l v e s the r e t r i e v al Of both p r o g r a m m o d u l e s and data f i l e s . T h e typical system has a single
disk c h a n n e l with one or m o r e disk d r i v e s c o n n e c t e d t h r o u g h a s i n g l e cont r o l l e r to that c h a n n e l . The channel speed In most s y s t e m s is a b o u t 1.2M
bytes p e r second. Data f r e q u e n t l y can not be t r a n s f e r r e d o v e r this channel
w h i l e the r e a d / w r i t e head is being p o s i t i o n e d so the e f f e c t i v e t h r o u g h p u t
is Often a lot lower. It is not hard to see that with m o r e u s e r s a c t i v e l y
using the system, the m o r e likely that a c c e s s to the disk will become a
c o n t r o l l i n g f a c t o r on p e r f o r m a n c e .
T h e r e a r e h a r d w a r e and s o f t w a r e t e c h n i q u e s that can be used to r e d u c e
the n e g a t i v e Impact of this problem. From a h a r d w a r e point of view, we can
add m u l t i p l e data c h a n n e l s , c o n t r o l l e r s , and disk drives so that there can
be s e v e r a l data t r a n s f e r s being p e r f o r m e d at the same time. T h e s e d e v i c e s
can add s u b s t a n t i a l l y to the cost of a s y s t e m and a r e not n e c e s s a r i l y
a v a i l a b l e on all m i n i c o m p u t e r s .
The s t r u c t u r e of data files is also a c h a r a c t e r i s t i c of data t r a n s f e r
p e r f o r m a n c e . T h e m o r e s e l e c t i v e the a c c e s s process is and the amount of
data r e q u i r e d to d e s c r i b e a g e o m e t r i c e l e m e n t also have an Impact on p e r f ormance.
2. C e n t r a l p r o c e s s o r p e r f o r m a n c e
T h e r e a r e a large n u m b e r of f a c t o r s that impact p r o c e s s o r p e r f o r m a n c e
in a m u l t i - u s e r system. S o m e of these a r e h a r d w a r e r e l a t e d such as the
basic speed of the unit in e x e c u t i n g i n s t r u c t i o n s , the m e m o r y speed, and
extent to w h i c h o p e r a t i o n s can be o v e r l a p p e d . The size of m e m o r y and how
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
it is o r g a n i z e d can be very I m p o r t a n t . L a r g e C A D s y s t e m s can involve l i t e r ally m i l l i o n s of bytes of p r o g r a m . If only a small p o r t i o n of m e m o r y is
a v a i l a b l e to each user, the computer will f r e q u e n t l y be going back to the
disIc to r e t r i e v e a d d i t i o n a l program m o d u l e s . The bottom line is that if a
computer can e x e c u t e 1 m i l l i o n i n s t r u c t i o n s per second and there are 10
users on the system, then e a c h user has in effect a 1 0 0 , 0 0 0 i n s t r u c t i o n per
second computer at his d i s p o s a l . G r a n t e d , when the s y s t e m is lightly loaded, he has a lot m o r e power a v a i l a b l e , but what h a p p e n s when two or three
users decide to do complex a n a l y s i s tasks. Off to the c o f f e e shop!
3. O p e r a t i n g s y s t e m s
S i n c e there are many users doing a v a r i e t y of tasks on the system, it
is n e c e s s a r y to use a complex o p e r a t i n g system to a l l o c a t e r e s o u r c e s
The
s c i e n c e of d e v e l o p i n g these o p e r a t i n g s y s t e m s has r e a c h e d a very a d v a n c e d
state. The m a j o r problem is they have been o p t i m i z e d to handle a wide
range of needs and are not n e c e s s a r i l y the best for m u l t i p l e CAO users
Ease of use and f l e x i b i l i t y have been e m p h a s i z e d over p e r f o r m a n c e . This
g e n e r a l i z a t i o n a l l o w s the user to h a v e a c c e s s to a w i d e r a n g e of c a p a b i l ities, but at a s i g n i f i c a n t cost of m a c h i n e r e s o u r c e s .
Some CAO s y s t e m s attempt to s o l v e this problem by u t i l i z i n g special
p u r p o s e o p e r a t i n g s y s t e m s . This a p p r o a c h trades off i n c r e a s e d p e r f o r m a n c e
against g e n e r a l i z e d support. It also m a k e s it very hard to support s o f t w a r e
from a v a r i e t y of s o u r c e s .
4. Output to w o r k s t a t i o n
The next step in the process is to output i n f o r m a t i o n to the user
w o r k s t a t i o n . T h « most important f a c t o r s here are the s p e e d of the transfer
and the amount of data that has to be t r a n s f e r r e d . A large drawing may
r e q u i r e the output of 2 0 , 0 0 0 v e c t o r s . At 3 2 bits per v e c t o r <a fairly
common s i t u a t i o n ) , this m e a n s t r a n s f e r r i n g 6 4 0 , 0 0 0 bits of data. If a
terminal is o p e r a t i n g at 9 6 0 0 bits per second this t r a n s f e r w o u l d take
67 s e c o n d s . T h i s is typical of many n o n - t u r n k e y s y s t e m s . S p e c i a l i z e d
systems c u r r e n t l y on the market tend to use a lot higher transfer rates.
A n o t h e r m e a s u r e of- output capacity is how many w o r k s t a t i o n s are linked
to each output channel. A system may have a 5 0 0 , 0 0 0 bits per second c h a n n e l ,
but it is shared by 16 w o r k s t a t i o n s . When only one or two w o r k s t a t i o n s are
active, this will not d e g r a d e p e r f o r m a n c e . If e v e r y o n e is busy, however,
each user is r e d u c e d to a 3 1 , 2 5 0 bps channel. In actual practice, the
e f f e c t i v e t h r o u g h p u t will be a lot lower b e c a u s e of the system overhead
in m a n a g i n g the c h a n n e l . S o m e c o m m u n i c a t i o n p r o t o c o l s can r e d u c e t h r o u g h p u t
to 60'/. or less of the t h e o r e t i c a l data link capacity.
The v o l u m e of data that must be sent to the w o r k s t a t i o n also has a big
impact on r e s p o n s i v e n e s s . S o m e w o r k s t a t i o n s have very limited local c o m p u tational c a p a b i l i t y and a very large v o l u m e of data must be t r a n s f e r r e d .
Other systems a l l o w data to be sent in a very compact form.
5. D i s p l a y g e n e r a t i o n
Most of the w o r k s t a t i o n s being used in C A D systems today use raster
technology. T h i s r e q u i r e s that v e c t o r s be c o n v e r t e d to i n d i v i d u a l points or
pixels on the display. T h e r e are many different s o f t w a r e and h a r d w a r e
t e c h n i q u e s for doing this that can result in varying levels of p e r f o r m a n c e
Some systems u s e fairly fast h a r d w a r e , but the h a r d w a r e is shared by s e v e r al w o r k s t a t i o n s . O t h e r d i s p l a y f u n c t i o n s such as fitting a portion of the
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Parl< Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
d r a w i n g onto a section of the screen can e i t h e r be done by the host c o m p u ter or by a p r o c e s s o r in the w o r k s t a t i o n . In systems w h e r e this h a r d w a r e is
shared by several worlcstat ions, p e r f o r m a n c e is a d v e r s e l y affected.
6. U s e r i n t e r f a c e
How the system h a n d l e s the Interface b e t w e e n the u s e r and the w o r k s t a tion can have a m a j o r impact on p e r f o r m a n c e . If these tasks a r e h a n d l e d
by logic in the w o r k s t a t i o n , one u s e r will have m i n i m a l impact on o t h e r
u s e r s w h e n he is e n t e r i n g commands, picking e l e m e n t s on the screen, or
e n t e r i n g textual i n f o r m a t i o n . If the central p r o c e s s o r is used, each
a c t i v i t y will have an impact on o t h e r u s e r s .
The a b o v e d i s c u s s i o n is a c t u a l l y a f a i r l y brief d i s c u s s i o n of what
goes on in typical m u l t i - u s e r e n v i r o n m e n t s . It is easy to see that an e f f e cient system r e q u i r e s the very careful b a l a n c e of a large n u m b e r of conf l i c t i n g f a c t o r s . S o m e p e o p l e have done this well, w h i l e many o t h e r s
have ended up with s y s t e m s that did not live up to e x p e c t a t i o n s or w h e r e
p e r f o r m a n c e could d e g r a d e badly u n d e r c e r t a i n c i r c u m s t a n c e s .
T h e r e a r e two p r i m a r y ways in w h i c h o u r d e s i r e d level of p r e d i c t a b l e
p e r f o r m a n c e can be a c h i e v e d , O n e is to use a very large or s u p e r - c o m p u t e r
that may be in the r a n g e of 10 to 3 0 m i p s (million of i n s t r u c t i o n s per
s e c o n d ) . T h e s e c o m p u t e r s a r e e x p e n s i v e w i t h p r i c e tags in the m i l l i o n s
of d o l l a r s . E v e n a v e r y large system can s u f f e r n o t i c e a b l e d e g r a d a t i o n when
a complex a n a l y s i s task is e x e c u t e d u n l e s s the number of w o r k s t a t i o n s is
kept v e r y low. U n f o r t u n a t e l y , this is a s o l u t i o n that few can afford. It
also s u f f e r s from the p r o b l e m that when it goes down, your w h o l e e n g i n e e r ing d e s i g n a c t i v i t y is out of b u s i n e s s . E v e r y o n e to the coffee shop!
A second a l t e r n a t i v e is to e l i m i n a t e the central host c o m p u t e r and p r o vide e a c h w o r k s t a t i o n w i t h a c o m p l e t e set of computer, file storage, and
g r a p h i c s c a p a b i l i t i e s . T h e c o n t e n t i o n f o r a c c e s s to a c o m m u n i t y disk file
is e l i m i n a t e d b e c a u s e e a c h w o r k s t a t i o n can have its own disk. A s i m i l i a r
s i t u a t i o n e x i s t s with d i s p l a y output s i n c e the w o r k s t a t i o n s display p r o c e s s o r is i n t e g r a t e d into the w o r k s t a t i o n ' s computer. This c o m p u t e r needs
to have s u f f i c i e n t p o w e r to p r o v i d e a high level of p e r f o r m a n c e with a
w i d e range of e n g i n e e r i n g design, a n a l y s i s , and drafting s o f t w a r e . S i n c e
no system v e n d o r can p r o v i d e all the s p e c i f i c a p p l i c a t i o n s that e v e r y o n e
needs, the u s e r must be able to c r e a t e his own a p p l i c a t i o n software. These
c o m p u t a t i o n r e q u i r e m e n t s can be met by p r o v i d i n g e a c h u s e r with a s t a t e - o f art 32-blt c o m p u t e r ,a 1 to 16 M B y t e v i r t u a l memory, and an industry
standard o p e r a t i n g s y s t e m that s u p p o r t s programming languages such as
F O R T R A N , P A S C A L , a n d C. S e e F i g u r e 2 and F i g u r e 3.
On first i m p r e s s i o n , it seems that this will result in a c o l l e c t i o n of
p o w e r f u l , s t a n d a l o n e s y s t e m s . F o r some a p p l i c a t i o n s , that might be a c c e p t able. Most u s e r s , h o w e v e r , need to have a s u b s t a n t i a l amount of i n t e r a c t i o n
b e t w e e n u s e r s or w i s h to have a h i g h e r d e g r e e of control o v e r the s i t u a t i o n
than can be a c h i e v e d w i t h independent w o r k s t a t i o n s . The s o l u t i o n is to tie
the w o r k s t a t i o n s t o g e t h e r via a Local Area c o m m u n i c a t i o n N e t w o r k (LAN) and
p r o v i d e an o p e r a t i n g s y s t e m that permits t r a n s p a r e n t sharing of p r o g r a m s ,
data, and d r a w i n g s as well as c e n t r a l i z e d control over common p e r i p h e r a l
d e v i c e s such as p l o t t e r s . This m e a n s that a u s e r at one w o r k s t a t i o n can,
with a p p r o p r i a t e a c c e s s codes, obtain a drawing or symbol file from a n o t h e r
w o r k s t a t i o n as easily as from his own.
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
The best of both w o r l d s ; a very high level of p r e d i c t a b l e p e r f o r m a n c e
as well as a tightly linked network of users sharing data and r e s o u r c e s .
Such a system is a v a i l a b l e today from A u t o - t r o l . It c o n s i s t s of the
Advanced G r a p h i c s W o r k s t a t i o n and the a s s o c i a t e d A u t o - n e t c o m m u n i c a t i o n
n e t w o r k . P e r f o r m a n c e of a single w o r k s t a t i o n is c o m p a r a b l e to any a v a i l able in the I n d u s t r y . T h e key fact is that this p e r f o r m a n c e is a v a i l a b l e
whether there is 1 w o r k s t a t i o n or 100 in the n e t w o r k ! S e e F i g u r e 4.
The AGW o v e r c o m e s the p e r f o r m a n c e limiting gates d e s c r i b e d above by
1) providing a s e p a r a t e disk for each w o r k s t a t i o n , 2 ) c o n t a i n i n g a high
p e r f o r m a n c e 32-bit virtual memory o r i e n t e d processor in each w o r k s t a t i o n ,
with 1 M bytes or m o r e of physical memory, 3 ) an industry standard o p e r a ting system, 4 ) an integrated c o m p u t a t i o n a l and g r a p h i c s system with no
transfer of data to an e x t e r n a l display, 5 ) high p e r f o r m a n c e graphic d i s play c a p a b i l i t i e s i n t e g r a t e d into the system, 6 ) and very efficient user
interface t e c h n i q u e s . O v e r a l l , a well b a l a n c e d system d e s i g n e d for a high
level of p r e d i c t a b l e p e r f o r m a n c e .
Dataquest Incorporated, A Subsidiary of A C . Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
o
u
iu
r>
c
(D
U>
O
O
•3
O
>
DISK
oim
CO
c
cr
u
a!
» •
>
p
z
(A
(D
3
o
o
3
CENTRAL
PROCESSOR
X3
D)
3
DATA
CHANNEL
§
a
DATA
Q.
(D
TJ
0)
CHANNEL
w
D>
3
WORKSTATION
o
>
CONTROLLER
o
CO
(O
o
o
o
DISPLAY
DISPLAY
DISPLAY
-vl
(O
•vl
CO
TRADITIONAL CAD SYSTEM
FIGURE 1
DISK
CONTROLLER
o
B
ST
r>
c
(D
CO
O
o
•5
o
>
CO
c
CT
cn
a
5"
-^
-<
O^
>
b
(A
(D
3
o
o
3
•a
fi>
3
•<
^
s o
3)
a
a.
CD
to
i^
to
cC
o
lU
O)
z
LU
t—f
LU
1—
UJ
_l
0.
-^
z
o
o
CO
Ijj
TJ
(U
*iPb
>
o
^
,^t
xS*^
i^
-iS^^'
c^^
^^^:
INTELLIGENT CAD WORKSTATION
n
O
>
s
to
(O
T
T
T
T
5
NUMBER OF WORKStftTIONiS
7
8
9
T
^
o
o
o
H
(D
PERFORMftNCE DEGRAOfiTIdN
FIGURE 2
n
10
o
01
3
O
o
•3
o
DISK
>
w
DISPLAY
c
CT
(A
d
u
•<
o_
>
b
•
CO
(D
3
CENTRAL
PROCESSOR
o
o
3
DISK
CONTROLLER
•o
»
3
•<
s
d
Q.
(D
-1
TJ
0)
DISK
DISK
CO
fil
o
(/>
(»
o
>
DISK
CONTROLLER
NETI^ORK
INTERFACE
DISK
CONTROLLER
.b.
o
oa
DISPLAY
to
CENTRAL
PROCESSOR
NETWORK
INTERFACE
NETWORK
INTERFACE
o
o
o
(D
CO
u
INTELLIGENT WORKSTATION NETWORK
FIGURE 3
CENTRAL
PROCESSOR
DISPLAY
z
o
CO
Of
o
en
u
S^
" 3
oS2
^LT
o"z
eC
>
a
oa;
a:
hI
o
h-
a:
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
Dataquest
Dataquest
IMPACT OF COMPUTER-INTEGRATED MANUFACTURING ON A COMPANY
Wilfred J. Corrigan
President
LSI Logic Corporation
Mr. Corrigan is President, Chairman, and Chief Executive Officer of LSI
Logic Corporation. Formerly President, Chairman, and Chief Executive
Officer of Fairchild Camera and Instrument Corporation in Mt. View,
California, Mr. Corrigan also held a series of management positions at
Fairchild.
Before joining Fairchild, Mr. Corrigan was Director of
Transistor Operations at Motorola, Inc.'s, Semiconductor Products
Division in Phoenix, Arizona. He graduated from the Imperial College of
Science in London, England, with a B.Sc. degree in Chemical Engineering.
Dataquest Incorporated
CAD/CAM INDUSTRY CONFERENCE
September 26-28, 1983
Newport Beach, California
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
SLIDE 1
1.
WORIJ) SEMICC»tDUCTOR MARKET
Chip sales will swing to non-standard devices as we approach end
of this decade.
We will have a 50 billion dollar per year
marlcet for semiconductors; 50 percent of
this market will
consist of devices v^ich are customized to the applications:
Custom Circuits
Semi Custom
EPROMS
ROMS
E^ Devices
Pals
etc.
Customization of these products may be done by the semiconductor
manufacturer
or by the user. Many of these products will he
reprogrammable in situ. This is a major change.
Largely, this
integration.
was
inevitable
as
we
move
to
large
scale
As we trend towards tJie vrtiole system being on a chip, either we
must move to all chips being different, or all systems being the
same.
The area that we focus on at LSI Logic is very Dense Logic
Arrays v^ich is the best way at the present time to address this
need for customized random logic in both small and large
electronic systems. What I will do today is quantify the extent
of the need, describe what a Logic Array is, and show how the
only way to address this market opportunity was to use advanced
CAD/CAE. In the process, we have found that we had to organize
into two (2) "Factories" - aj_ "Design Factory" with t:he charter
to manufacture designs, and b]_ "Silicon Factory" is virtually
100 percent conputer controlled, both in terms of technical
aspects, and scheduling.
SLIDE 2
World Wide Logic Array Market.
This is Strategic, Inc. data and
being a new field any data should be only viewed as approximate, but
the projection of growth from tJie $200 Millon+ range to $7 Billion+
between '82 and '92 gives a flavor for the size of the opportunity.
Of course, the Semicustom Products of 1992 will be much more con^lex
and probably use different technology than we currently conceive.
Also note the rapid predc»ninance of HQ4OS in the general purpose area.
- 1 Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 RIdder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
SLIDE 3
What is a Logic Array?
Basically, it is a semi finished product that is customizeable at the
metal layers to implement a user defined function. It is very close
to being a printed circuit board executed in Silicon.
SLIDE 4*
This is what a High Density CMOS Array looks like.
SLIDE 5*
The dual layer metal which is the key to high performance arrays
shows a strong similarity to the ubiquitous printed circuit board this is several hundred times magnification.
SLIDE 6*
This shows metalization at 2,000 times magnification.
Metal widths
are about 3 microns, centre to centre on metal lines about 10 microns.
SLIDE 7*
The finished, packaged
144 pin package.
array.
This
happens
to be 6,000 gates
in
SLIDE 8
Array Application
A)
B)
C)
SSI Integration
User Defined LSI
Subsystem on a Chip
300 - 1,000 Gate QKiS
700 - 2,000 Gate CMOS
2,000 - 6,000 Gate CMOS
SLIDE 9
2
Example Conversion of 27 T*L
I.e.'s to a Single Array.
- 2Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
SLIDE 10
ECONOMICS OF LOGIC ARRAYS
•
Implementing a system in LSI usually creates a dramatically
lower cost system. In most cases the part costs are less also.
Particularly as T^L is rapidly increasing in price.
•
Design cost is usually lower and more predictable, the forced
use o£ t:he computer necessitates a discipline which improves the
"Right the First Time" attitude.
It also, creates a more
measureable and controllable design cycle.
•
Field service costs with LSI are much lower.
Reliability
improves as the interconnects move on the chip. As the chips
become more complex, the system becomes simpler and more
reliable.
e.g.
Calculator
1 chip
Very Reliable
For example: To illustrate the H(34OS intact on a traditional
T^L design, we analyzed a popular mini computer implemented in
T^L MSI and DID a hypothetical redesign in HC31OS 3 Micron
Arrays.
The printed circuit boards in the system dropped frcan 35
to ±'.'.
Relative cost dropped by 60 percent J.'
SLIDE 11
So, in addressing this large and rapidly growing market the challenge
basically was that with:
•
System and chip complexity increasing
•
Development times stretching when they need to compress
•
Increased con^etion .... and<an
•
Engineer shortage.
How to meet these needs in a cost effective fashion:
-
3 -
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 RIdder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
SLIDE 12
THE SOLUTION
COMPUTER AIDED ENGINEERING
As we were postulating a design factory, we needed computer driven
manufacturing. This required a heavy duty approach, and to address
the issue on an appropriate scale — not "How do we use the Computer
to do a Design?" But
"How do we use the Computer to Produce:"
a Design per day
10 Designs per day
100 Designs per day
i.e.
A Design Factory —
Not an Engineering Department.
SLIDE 13
WORLD DESIGN RESOURCES
There are several thousand Silicon Designers feeding a user base of
several hundred thousand system designers. With appropriate computer
resources, and software tools we can convert same percentage of these
designers into Silicon Designers. Thus the bottleneck disappears and
the LSI revolution can continue.
The main software system we use at LSI Logic is the LDS-II.
SLIDE 14
This gives us the ability to train a customer's system designer in
about two (2) days. Typically, most of the work is done an a large
Amdahl, either V6 or V8, but we are able to interface with smaller
machines where all or portions of a sinuilaticxi have been performed on
a workstation.
SLIDE 15*
Milpitas Computer Center
SLIDE 16
Customer Design Support—System Designer orientation rather than
Semiconductor Designer orientation. We provide local design centers
with applications support, and direct 9600 baud access to our main
computer facility.
- 4 Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
SLIDE 17
More detail of our current LPS-II System
Using this approach we can handle a large number of designs flowing
through our design factory with a surprisingly small numtier of
applications engineers.
Ctoviously, every
customer
design is
different, but we have been able to sell the concept of a standard
methodology, with a flexible interface to the outside world. We
expect by the end of this year to have working interfaces with all
the major workstations vendors. We already are able to interface with
the more popular software simulators.
SLIDE 18
Our future thrust will be to move lihis capability up to the system
level to assure that major LSI subsystems work together prior to
implementing tihe Silicon. This will further congress system design
time and improve design accuracy. We will be describing this system
later this year.
SLIDE 19
Close.
LSI Logic
*Hardcopy of these slides unavailable.
- 5 Dataquest Incorporated, A SubsIdIary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
LSI LOGIC
CORPORATION
40
g 30
O
O
O 20
WORLD
SEMICONDUCTOR
MARKET
Chip sales
will swing to
nonstandard
devices.
W)
Z
o
m 10
1981
LATE 1980s
* Includes custom, semlcustofn, and programmable logk and memory chips.
SOURCE: BUSINESS WEEK
Data: Texas Instruments Inc.. BWestimates.
o
J2
c
a>
CO
o
o
-1
•o
LSI LOGIC
CORPORATION
o
(D
O.
CO
c
WORLDWIDE LOGIC ARRAY MARKET
IT
CO
a
BY MAX>R PROCESS TECHNOLOGY
0)'
—1
•<
o^
>
o
BIPOLAR
DK5ITAL
(D
o
o
3
TJ
lU
BIPOLAR
DIGfTAL
ru
CO
o
a
a.
CD
Tl
0)
w
Q)
3
o
>
BIPOLAR
ANAtOG
CMOS
CMOS
BIPOLAR
ANALOG
1982 - $231M
1992-S7.3B
COMPOUND A N N U A L GROWTH RATE: 4 1 %
CO
8
o
CD
CD
CO
W
SOURCE: STRATEGIC, INC.
LSI LOGIC
CORPORATION
WHAT IS A LOGIC ARRAY?
A n uncommitted array of logic
functions that are interconr>ected
at the metal level to implement
a user defined function.
POLYSILKION
GATE
2 LAYER METAL
IfS/TERCONNECT
sa^^k.
The silicon wafer is preprocessed
to the standard array cell design.
The metal Interconnect is
subsequently processed to
implement the \ner defir>ed
logic function.
• CUSTOMIZED
METAL
riSrrERCONNECT
STANDARD
SILICON
WAFER
HCMOS STRUaURE
^
o
o
•3
o
^^^^^^—^^^^f^^t^-^^m^^i^m^^ma^ami^m^mi^^^m^^^^^^m^m^m^mmK^^mmmm^i^l^^
LSI LOGIC
CORPORATION
(D
p.
>
CO
c
cr
(n
a.
0)'
-3
o^
>
b
m
u
ARRAY
ABDICATIONS
(D
3
o
o
3
••
B)
3
to
CO
o
Q.
a
SSI INTEGRATION
TYPICALLY 300-1000 &m
28-40 FiN Dip,
CMOS.
(D
-o
0>
w
01
o
>
CO
CJl
o
00
USER DEFINED LSI ELEMENTS
TYPICALLY 700-2000 G « E CMOS OR HCMOS.
600-1200 GATE ECL. W-6S PINS.
SUBSYSTEM ON A CHIP
2000-6000 GATE HCMOS,
66-120 HNS.
CO
o
o
o
5.
CD
CO
CO
mmmimmmmm»mmmitmmMmmmmm
o
B)
•o
O
LSI LOGIC
CORPORATION
(D
a.
>
V)
c
CT
(0
5!
o>'
EXAMPLE OF A CONVERSION FROM
27 TTL ICs TO A SINGLE ARRAY
>
b
EQUrV. GATES
PER IC
TOTAL EOUIV.
GATES
2 ea. 74LS91
43
86
2 ea. 74LS100
16
36
4 ea. 74155
21
84
2 ea. 74174
37
74
YO ea. 7402
4
40
2 e a . 74L5Y0
6
12
5 ea. 7430
6
30
1 ea. 74163
68
68
10 ea. Norvlnverting Output
Buffers; 2X Drive {for array)
4
40
2 ea. Non-Inverting Tri-State
Output Buffers (for array)
6
12
EXISTING SYSTEM
o
o
3
•o
01
3
to
CO
o
aa
(D
T)
0>
CO
01
3
O
>
4^
O
CD
8
CO
(D
X
--J
<o
--I
TOTAL EQUfVALEfVrr GATES
482
o
n
c
a>
o
o
•o
o
LSI LOGIC
CORPORATION
(D
a
>
Q.
OJ'
>
o
ECONOMICS OF
LOGIC ARRAYS
o
o
3
••
»
ro
<o
o
aa
(D
T3
0)
3
o
>
Silicon vs. System Costs
Design Costs
Market Window Costs
—Speed of Design
— Rapid Manufacturing Ramp
— Ease of Technology Insertion
Quaftfkaiion Costs
Field Service Cosu
o
(O
-J
§
(O
CO
imm
o
J3
c
(D
^
LSI LOGIC
CORPORATION
>
w
c
a
en
oi.
aa
>
p
z
CHALLENGE
01
(D
o
o
3
•o
B>
3
o
• SYSTEM AND CHIP COMPLEXITY INCREASING
I
I
• DEVELOPMENT TIMES STRETCHING OUT
ro
CD
O
W
Q>
3
o
>
.8
CO
8
o
5.
(D
CO
->l
CO
• INCREASED U.S AND JAPANESE COMPETITION
• ENGINEER SHORTAGE
o
mm^i^mm
w
B
.O
c
(D
O
O
•3
o
LSI LOGIC
CORPORATION
>
0)
c
cr
(0
o!
0)'
-t
-<
o^
>
b
o
o
3
•a
fi>
3
(O
o
SOLUTION:
COMPUTER-AIDED ENGINEERING
FOR EXAMPLE,
OUR LDS-n LSI DESIGN SYSTEM
Q.
Q.
(D
"D
ID
• MULTIPLIES ENGINEERING PRODUCTIVITY
CO
0)
O
(0
(D
O
>
<o
(O
• ALLONX/S SYSTEM ENGINEERS TO DESIGN ICs
• GENERATES PRODUCTION TOOUNG AND TESTS
• REDUCES DEVELOPMENT TIME
<o
o
8
(O
CO
i^MBpaWPCB
o
(u
LSI LOGIC
CORPORATION
WORLD IC DE;SIGN RESOURClES
^ ^ ^ ^ ^ ^ ^ ^
rtti~r*^g v^^^Ht^vSs^fflj^Sa ?!?
k
l3SlJFJ-i^!^te^^i^J.^p^^j^=-Cy'^g^B
ydti^r^V^^^^^^^^^^^^t^
^^Ct^-^nPtm^Sitl^B^'V^'fiSsB^^S^^W^rJ ^ T^^^^-S^^y^" i^i^^#^*- SJT c^^^r^u.'^B
lK(^^;^f|ea^^^'^^^^9
^^¥r~j^i«%#'^»!on-'v^^£tfl
^^^j^rs^a^^^ia^^^^yiy
^^^^^^^^^^m^^JT
^^i^^^^^^^^^^^S
WBf^^f^^S^sJ^sSf^^^^SS^ff^^B^S^aS^^^m
H^^r^ES£pMM^E#S^^^S^^I
oew •
1
Mg|[aL.^ ^*^tJf-£"??3^^=^,..^i*ilJ^S^Biwl^B
^^^; ^v9wi^^Mi^B(^^^9
M ^ > - " ^ S L Thin, ^^z' ^<5^?^MS^fcj'¥5t-(^^B
SYSTEM DESfGNERS 1
200,000
1
EIVKJINEERS
I
EjQEK^giKBljBS^
• 3 r w^'iin? a?^2^H
^ ^ ^ ^ ^ ^ ^ ^
s^l-S^SaSEiaBEggK^S^B
^^v^^^j^^g^M^B
^
^
iSiiCS^^jBrj^^H
BPt^'^-tyler jFlirg'iim^aw^ B
E*g^£S?^ •.ffafc^E*Sf3^ft' .^Jw^j^^hJE^B
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company
/1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
o
"^•W—^•^•^^"^'^i^^
I—IWIWIW^H——W—p^Mp^fWII*^
LSI LOGIC
CORPORATION
CUSTOMER DESIGN
SUPPORT
• System designer rather than
semiconductor designer oriented Interface
• Training centers and courses
• Applications support on har>d
• Multi Design centers
• Documentation
LSI L O G I C
CORPORATION
INDIVIDUAL ARRAY
DESIGN FLOW
LDS-in
OR
DESIGN
STATION
;
•VSTCM
SIMUUTIOM :
LDS-nr
AIWAY PHYSICAL
OKSIQN
HIfMhMilMM
-jMliiaiiaiHii
"'•,! m
ai^
FINAL
BYSTCM
ptMgLATipii^
J
TO CUSTOMER
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
LSI LOGIC
CORPORATION
SYSTEM S I M U L A T I O N
ARRAY «1
DESIQN FILE
ARRAY #N
tWSKlN PILE
ARRAY f a
oesioN nti
LINK
ARRAY MET
STANDARD
ic Mooet^
PWTTCRN
•LJ
mfmtwnon
itlMmiiiiiJi
I
.£=
;WMULATIOI|
iiM^
.toiiU
TEST
fORMATTINQ
ARRAYS A BOARD
PHYaiCAL, peSIQN
•-•:;>L......
PtMAL0YBTCM
SIMULATION
h>-.itf^tA*wi.iii»
.
.
,t,^
• •
,
^Mwiihi*
ARRAYS « BOAM>
MSRICATION
-
- • • • • - ' • ' • • "
- - ^
hUldMi
•
tf
;^
ir
BOARD '
^...TCtT/-:;,,,
^ ^ . • • L . -..
• .d---.....i....^
I
SYSTEM CHECKOUT
>
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
Dataquest
o
oT
.a
c
(D
cpdlo
cofTtputerhGL
(/)
c
CJ
!!?
Ol'
>
p
2
5!
u>
lu
CO
o
aa
CD
TJ
0)
w
B>
C-
o
(0
_<D
O
>
•fc.
o
CD
(O
O
O
O
CO
--J
CJ
COMPUTER COMPANIES
AND
INTEGRATION
o
BT
•Q
C
(D
1
(0
Q.
qpollo
computer inc
CO
c
cr
u
9!
Qj'
COMPUTER COMPANIES
AND INTEGRATION
>
6
o
o
3
•o
lU
• INTIGRATION: DEFINITION AND REQUIREMENTS
10
(O
o
a.
Q.
CD
• ROLE OF COMPUTER VENDOR
TJ
a>
•
w
0)
rj
O
>
o
00
CO
o
o
o
SE.
(D
U5
-J
CO
THEINDUSTRr
A COPY OF THIS PRESENTATION WAS NOT AVAILABLE AT PRESS TIME. IF
THE PRESENTATION IS MADE AVAILABLE AT THE CONFERENCE, COPIES
WILL BE mSTRIBUTED.
Dataquest
THE ROLE OF A MODERN CUSTOM SUPPLIER
IN THE FIELD OF VLSI DESIGN
Graham Shenton
Vice President Marketing
International Microelectronic Products
Mr. Shenton is a Vice President of IMP and is responsible for Marketing
and Engineering Design Systems. Prior to assuming this responsibility,
he was a founder and Vice President of Engineering at IMP. Previously,
he was Director of International Engineering for AMI, Inc., based in the
United Kingdom. His earlier experience includes positions at Plessey and
Amalgamated Wireless Value Company, involving product design, engineering
managenent,
marketing,
quality
assurance,
test
engineering,
and
applications. Mr. Shenton studied Electrical and Electronic Engineering
at the Universities of Sydney and NSW and holds B.S. and M.S. degrees in
Engineering.
Dataquest Incorporated
CAD/CAM INDUSTRY CONFERENCE
September 26-28, 1983
Newport Beach, California
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, GA 95131 / (408) 971-9000 / Telex 171973
THE ROLE O F A MODERN CUSTOM
INTEGRATED CIRCOIT SEMICONDOCTOR MANUFACTURER
by
Graham Shenton
Vice President, Marketing
International Microelectronic Products
San Joser California
What I plan to do today is look at the role of a modern
custom—and the emphasis is on custom—semiconductor
manufacturer from the user's point of view.
I want to
explore how the prospective user of a custom IC should select
a vendor, when he should use the resources and specialized
knowledge of the vendor, and when the decision is his alone.
A corollary to this is the use of a custom IC manufacturer to
help determine if a custom IC is even the best solution. In
fact, that should be the starting point.
By now you have probably heard all the reasons for
choosing a custom circuit:
Increased confidentiality so the
competition will have a harder time copying your design. A
unique application for which no standard products exist.
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
Lower production cost traded off against higher development
cost and maybe, but not always, a longer development cycle.
What you haven't heard from custom manufacturers (and we
may be doing you a disservice), is why and when you don't
need a custom IC to do your job. Sometimes we don't tell you
because of ego—yours. A system designer comes to us and
wants his latest system cast in silicon because it is sexy,
or because everybody else is doing it, and "who do you think
you are, Mr. Semiconductor Manufacturer, to tell me whether
or not I need a custom circuit.
Just build what I tell you
to build."
With markets changing so rapidly, some designers want
their systems in a custom IC to prevent easy reverse
engineering, before they have had a chance to recover their
engineering costs and meet their business plan's ROI.
However, that rapidly changing marketplace happens for two
reasons and the designer must decide which reason causes the
Change before he can decide whether or not to use a custom
IC.
If the market is changing because the manufacturer has
hit upon a gold mine and it is easy for his competitors to
copy his system and enter the market with little or no
development cost, then there are few alternatives. You could
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
encase your product in a metal box, weld the box shut and
have it self-destruct when forcibly opened—messy.
the parts—expensive.
Mislabel
Or switch to custom circuits, which
also saves you costs and increases your reliability while
giving you the confidentiality you sought.
The personal computer market is a textbook case. Apple
II computers are easy to copy and have been copied by many
manufacturers.
It's simple:
just pop the cover, read the
labels on the ICs, and bingo 1 The fact that Apple is using
more and more custom ICs in their machines, and that the IBM
PC Started with six semi-custom gate arrays (as has been
reported in the media), supports the use of custom ICs in
this market environment as the right move.
However, if the market is changing because products have
a very short lifecycle, then it may not be appropriate to put
your system in a custom IC, Some markets, particularly in
the consumer fad area, have such a short market window, that
to lose any of it to development time would be a disaster. A
product built of off-the-shelf components still takes some
time to copy, and by the time your competitors are to market,
there is no market left.
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
I hedged my bet on that last item because there are
times that a custom IC can be advantageous in markets with a
Short time span. One is when TTL lead times get so long it
is faster to design and build a custom circuit.
In fact,
that is the case today if your design is dependent on certain
TTL parts that have lead times of 20 weelcs or more.
A second case is when you know your marltet and you know
what some of the changes are going to be, or you are
confident of the new bells and whistles your product is going
to need in six months. Then go ahead and build your custom
IC, but include on it all the features that you can release
over a reasonable market timeframe in order to continuously
upgrade your product. The features are then on all products,
but you're not releasing them for general use until you are
ready.
Having a feature on a chip is useless unless someone
tells you it is there and how to access it.
Some watch manufacturers used that strategy and they
Still do. First, they designed one chip with all the bells
and whistles. Then they introduced a plain vanilla watch
with time and date. Their next version added day-of-theweek.
This was followed by a timepiece with a stopwatch
built in. And if they thought far enough ahead, they
included a calculator function with memory.
They did this
partly because of the economy of scale—it was cheaper to
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
build one chip and not use all its functions in one product
than to Iteep redesigning the chip—and partly because the
Chip was one of the least expensive items in the watch.
Now that you have decided that a custom IC is the only
way to produce your product, your design problem starts—how
do you picic a vendor?
If a custom manufacturer has helped
you get to this point, then the question is probably answered
already.
If not, then we start to get into the role of the
modern custom IC manufacturer.
I will assume that you
compile your vendor list from industry lists, advertisements
word-of-mouth, e t c
And that you have whittled down the list
by eliminating those vendors whose minimum requirements in
terms of volume and/or money are beyond your needs. Now, how
do you select from the dozens that are left?
First, you need to realize some things about the custom
business. You have all heard how custom vendors are looking
for Close, long-term relationships with their customers. But
I'm not sure you really understand what that means and how
profoundly it can affect your business. Most of you look at
this relationship in terms of "can I trust the custom
manufacturer to keep my confidence and not reveal any of my
secrets, even inadvertently, to my competitors?
And, can he
deliver a reasonably reliable product on time?"
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
That's not enough.
Let's start at the beginning.
You
decide you need a custom chip and have heard, £or example,
that IMP can give you a full custom circuit, using a
Standard-cell design methodology, in 14 weeits from approved
logic diagram to packaged, tested woricing parts.
What a
shoclc when you discover that before you can even get to the
actual design phase, you still have to negotiate for price,
NRE, delivery, e t c , go through the traditional credit
checJcs, and then finally get to the definition phase 1
This is where you have to carefully evaluate the custom
supplier from a long-term point of view.
In the short-term,
you are going to have to endure the same routine with every
supplier on the first circuit.
than Others.
And some can do it faster
This is where the experience of the custom
vendor can malce a tremendous difference.
The learning overhead is so high, that an experienced
custom supplier with a good reputation can minimize this
phase of the project by preventing the customer from
reinventing the wheel.
The right custom supplier knows the
customer interface well, he knows what v/ill and won't work,
and he knows how to help the customer define his own needs.
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
Furthermore, buying a custom circuit is so different
from buying a standard product that top management (as well
as engineering management) must get involved.
In fact.
Purchasing may be the last to get involved.
Remember, we're still just tallcing about the first
Circuit you are going with a particular custom supplier.
However, it is rare that a company does a single custom
circuit unless they have been burned.
If the circuit works,
the benefits are so great that they often look at any other
designs they have that can be reduced to a single silicon
chip.
This is where that long-term relationship really pays
off.
By working well with the customer, the negotiations on
each succeeding circuit can be minimized.
Passing on to the
systems designer as many design skills as possible, either
through teaching or CAD, minimizes the definition and maybe
the whole design phase. Additionally, intimate knowledge of
the custom design process allows the systems designer to
incorporate superior cost/performance in his next design.
It is extremely important that you establish that the
vendor will continue to be as responsive after the initial
contacts.
It does you no good to have the president and
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
chief design engineer put a full-court press on you for the
initial circuit and then
circuits.
never see them again on succeeding
What you are buying with this long-term
relationship is a commitment from the vendor to always give
you the best possible design and manufacturing available.
That he will help you decide what should emd should not be
put in Silicon and not just try for every circuit he can get.
But most of allf that the vendor will grow with you; that he
will continuously upgrade his processing and design
capabilities to keep you on the leading edge of technology.
Ask where the vendor is positioned in terms of process
capabilities.
He must be on the leading edge if he is to
give you the best processing there is.
This is where you are
going to get your cost/performance edge in the marketplace.
You have to be given the ability to build your products in
processes comparable to those used on standard productsf but
you also have to balance aggressive technology advances
against production requirements.
And, even if you can't make use of that state-of-the-art
technology, you need to know that it will be available when
you can.
It also shows you a commitment by the custom IC
vendor to continually hone his skills and not simply rest on
past performance.
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 17197?
The custom IC vendor who is not aware of where the
market is going and who is not reasonably close to it is
going to be left behind.
This can prove especially deadly to
a customer because the security that builds up with the
credibility of a custom vendor over time can lull a customer
into not keeping tabs on the custom market. The customer is
caught unaware and, before he knows it/ he is scrambling to
find another vendor on a rush basis.
Keeping up on process technology is only one aspect of
the custom business today. The other is design. In
particular, computer-aided design.
The high cost of custom ICs can be directly attributed
to the labor intensive aspects of design. This also
accounted for the long design cycles. The newer CAD
equipment and the methodology for using it has significantly
decreased both the cost and design cycle time, as you have
previously heard at this conference.
The task of selecting among vendors based on their
design tools, then, becomes increasingly difficult if you
Start focusing on minor variations in these factors. The
difference between an 11- and 14-week design cycle or $30,000
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
or $40,000 design cost are minuscule against the 36-week and
$200,000 cost of a fully hand-drawn design. The focus should
be on design methodology first and then CAD tools; the trick
is not to get hung up on details.
Don't get me wrong.
I'm not trying to say that $10,000
is insignificant, but other factors may be more important and
may offset that cost difference. One of those factors is
vendor reliability; i.e., does he deliver on time?
do what he says he is going to do?
Does he
Another is credibility;
i.e., what is the probability that the circuit will work the
first time?
Of increasing importance are customer-usable CAD tools
that are touted as "user-friendly".
word.
Be careful of this buzz
Find out if it means easy-to-use or easy-to-learn.
Ideally it means both but, if not, be sure that you get the
one you need.
If a few people or a single group are going to
be responsible for designing all custom chips in your
company, then an easy-to-use system is more important. It
becomes worthwhile to learn a great number of sophisticated
and complicated commands once because constant use will
ingrain them in the chip designer's mind.
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / {408) 971-9000 / Telex 171973
Ifr on the Other hand^ chip design is just one aspect of
the engineer's job, then an easy-to-learn system is more
desirable. There is going to be a certain amount of time
wasted going up the learning curve each time the systems
designer returns to the CAD system to design his next chip.
An easy-to-learn system minimizes this time.
Regardless of how the system is used, the ideal design
methodology should encompass the following:
o
Minimize the time needed to implement a system design
in a working chip in production. The emphasis should
be placed on design accuracy and on achieving
"first-time correct" performance.
o
Allow for designer controlled flexibility in the
trade-off between development time and silicon area,
i.e., the trade-off between development cost and
production cost.
o
Hierarchical system to allow the user to enter at the
level of his expertise or need.
For example, system
designer should be able to work at a functional block
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
level and to produce good designs without becoming a
semiconductor circuit design and layout expert.
o
Capable o£ designing the most complex circuits
possible with advanced technology.
This also means
the ability to easily assimilate technology
improvements into the design system.
o
Provide techniques to ensure that the chips produced
are testable. The best time to think about the
testing problem is during the system and logic design
phase.
To implement the above methodology in a system expressly
designed for the systems designer would require the following
features:
o
Design capture via schematic entry on a color graphics
terminal or via a direct textual description of the
logic design on a video display unit.
In either case,
the design must be captured once and automatically
translated to the same hierarchical network
description from which the design proceeds and to
which the design is continually compared.
Dataquest incorporated, A Subsidiary of A.G. Nielsen Company /1290 RIdder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
o
Design consistency implemented by having a single
integrated data base that contains all the various
levels o£ abstraction o£ the designr system
description, logic diagrams, transistor representation
and final physical layout in mask and wafer form.
o
Library of functional bloclcs that exist at various
levels, e.g., logic, transistor, layout, but to the
designer represent the fundamental design components
of the system.
The system designer should be able to
call on a range of commonly used TTL elements or their
CMOS equivalents from which to construct a system.
o
Construction of the functional bloclcs should be as
process independent as possible to avoid a major
limitation inherent in most libraries.
In attempting
to work with the most advanced technologies, there is
a limited amount of time to characterize new design
elements before the process advances and you have to
go back and recharacterize all elements. Process
independence allows the custom IC vendor to
concentrate on adding new cells and features to the
CAD system rather than on updating an older library.
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
Process independence allows the whole library to
"instantly" talce advantage of process advances.
o
The designer roust be allowed to the malce the complex
design decisions. The system should give him rapid
answers from which to proceed with the design process.
For example, if the requirement is to implement a TTL
design with minimal design time and cost, he may
decide to make maximum use of existing library
components.
If more engineering effort is justified
in the form of developing new functional elements or
in the planning of the layout because'^ of high
production volume, then more interactive design time
can be spent with subsequent reduction in silicon area
and production cost savings.
Features such as these malce a complete design system and
not just another development aid.
One further point:
the purpose of all custom IC CAD
systems is to help more engineers design better circuits more
easily.
And all CAD systems on the market do this to a
greater or lesser extent. However, if the system is not
universally available, then you are being locked in to a sole
source design and lose a lot of flexibility.
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
Design flexibility comes from having a company's
functional library resident on many different workstations,
CAD systems houses, to customer mainframes to the specific IC
manufacturer's design system.
Each offers a different level
of versatility for the user.
To make maximum use of a custom IC manufacturer's
capabilities requires using the manufacturer's own design
system.
It was built with their manufacturing facility in
mind and there is almost an implied guarantee that what is
designed can be built. The system will be even more powerful
if the manufacturer uses the same system as his in house
design system.
This will insure that it will be constantly
upgraded and include all the latest features.
It also means
the manufacturer has an intimate knowledge of what the design
system can and can't do and how to get the user out of the
jams he will inevitably find himself.
Purchasing just the design software and running it on
your own mainframe will accomplish almost the same thing.
The possible glitch here is that no two mainframe systems are
identical and care must be taken in bringing up the software.
And then you have the problem of maintaining the system and
incorporating updates as they are issued.
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
The most flexible position for the user is the
Standalone woricstation that can access any number of
different libraries ranging from gate arrays to full custom
Standard cells, to PC board layout, etc
Here, you must be
very careful to ensure that the design methodology is not
missing.
CAD systems houses don't design ICs, Only the
custom IC manufacturer is the source of that knowledge. So
it is important when selecting an independent CAD system to
malte sure that this training is available to you either from
the owner of the library directly or through the CAD
manufacturer.
In fact, that last point is so important, I want to
repeat it. In order for the relationship between the custom
manufacturer and the custom user to work smoothly over the
long haul, the IC manufacturer must be willing to help the
user acquire a design capability regardless of the design
system or manufacturer chosen. The custom IC vendor must
help in the selection of hardware. The custom IC vendor must
recognize the need for and help select a second source. And,
the IC manufacturer must, above all, remember that the design
belongs to the customer and all he can expect is a reasonable
production return.
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
Another point which is often overloolted because it was
not a problem during the recent recession, is the
availability of manufacturing capacity to build the parts
designed.
Many custom IC manufacturers do not have their own
wafer fabrication facilities; they contract actual production
out to silicon foundries. This worked fine during the
recession when there was a lot of excess capacity available,
even though it did require them to juggle the various process
differences of different foundries. But if a foundry loses a
process or decides to change or even cancel it, they have no
recourse but to find another foundry.
Their customer doesn't
want to hear about another once-removed vendor's problem. A
more liicely scenario is unavailability of capacity as the
business starts to boom again. Those vendors without the
ability to control their own manufacturing, those without a
factory, could well be squeezed out of the market.
A custom IC vendor with manufacturing capability must
perform on both the design and manufacturing phases to be
credible. This ensures customers that the right attention
will be paid to producing their parts with the best possible
yields.
In the final analysis, a custom IC manufacturer is
selected based on the company's reputation more than any
Other factor.
This is an easy thing for a new customer to
check via references and industry sources.
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
In this discussion I have tried to emphasize the roles
o£ a modern custom IC semiconductor manufacturer, how to
select a custom solution (and how this choice is directly
influenced by the market), how to select a vendor and when to
use his Icnowledge and resources, and how to implement design
methodologies for a flexible, all-custom IC CAD system. A
company that satisfies these roles is an extremely valuable
partner in system development.
The conclusion I hope you
have drawn is that you need a well-balanced custom IC
manufacturer if your designs are to be consistently
successful.
Dalaquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
Dataquest
Dataquest
CAM ON THE MANUFACTURING FLOOR:
PRODUCTS THERE TODAY
Albert H. Libbey
Vice President New Business Development
Programmable Controller Division
Gould Inc.
Mr. Libbey is Vice President of New Business Development at Gould Inc.'s,
Prograiranable Controller Division in Andover, Massachusetts.
Prior to
that position, he was Vice President of Engineering at Gould.
His
experience
includes management
and
design
roles
associated
with
programmable
controllers and
their
I/O products, applications
of
programmable controllers, communication systems, and instrument systems.
Previously, he was Scientific Executive at E.G.&.G., Inc., Bedford,
Massachusetts, where he worked in the Bedford Division for twelve years.
He received his B.S.E.E. degree and his M.S. degree in Electrical
Engineering from the Massachusetts Institute of Technology.
He is a
member of the Instruments Society of America and the Institute of
Electrical and Electronic Engineers.
He also holds memberships in
Etta Kappa Nu and Sigma XI.
Dataquest Incorporated
CAD/CAM INDUSTRY CONFERENCE
September 26-28, 1983
Newport Beach, California
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
o
Si.
ST
.Q
c
a>
u>
S
o
o
•D
O
(D
a
>
CO
c
(T
W
Q!
B)
>
b
o
o
3
EU
3
CAM ON THE MANUFACTURING
3)
Q!
a.
m
TJ
u
W
0)
o
>
o
00
to
o
o
o
FLOOR:
PRODUCTS THERE TODAY
o
HI
o
o
•5
o
S
-^
ABQUI
>
Cfl
c
ow
•:
0
DISCRETE MANUFACTURING
0
MAKING ENGINES
0
CANNING SOUP
01
>
b
(B
°
•o
B)
3
I
0 ASSEMBLING TV'S
§
s.
tx
n
MOT ABOUT
i
1
0
CONTINUOUS PROCESS MANUFACTURING
0
MAKING PLASTIC
o
w
2
(O
Ol
—fc
OS
i
0 DISTILLING GASOLINE
o
00
"
2.
(B
<o
-s(
CO
0
GENERATING ELECTRIC POWER
o
S
J3
C
CD
3
O
o
•5
o
0)
p.
>
CO
c
(»
g.'
Ol'
•<
o
>
p
z
<2.
^flTF THAT
Ul
CD
O
o
0)
4
0
so CALLED PROCESS INDUSTRIES HAVE DISCRETE MANUFACTURING
•<
fo
<o
o
I
f
p
CO
0>
c_
o
01
_<D
O
>
o
00
to
CO
o
o
o
(O
->l
u
AND VICE VERSA.
^
1
S%-
%
>
w
cow
a
S'
•3
J
z
s!
I
NOTE THAT (COHT'P)
0 CONTINUOUS PROCESS ALREADY AUTOMATED.
b*
3
5I
I
I
i
s
.1
2
(D
CP
•>J
(.1
THROUGHMANUFACTURING
CAD AND CAM, IS STRUGGLING TO INCREASE PRODUCTIVITY
0 DISCRETE
f
3
O
O
•o
o
(D
Q.
>
w
Q.
0)'
>
b
1 ^ STAHPftRP SOmTlOMS
EXAMPLE:
METAL REMOVAL OPTIONS
0)
<t>
3
o
o
3
n
0)
3
CHOOSE A SET AND SFtJilFliCF
FIXTURES
TAPPING
MILLING
DRILLING
REAMING
LAPPING
DEBURRING
ETCHING
CLEANING
FACING
THREADING
GRINDING
TURNING
TUMBLING
BORING
BURNING
to
(O
o
I
i
o
w
O
U>
a
O
>
.b
o
S
<o
(C
oo
o
(O
--J
CJ
PLANING
ec
0£
UJ
oc
LU
.m
o
(_>
L9
U9
<_>
U.
o
Ditaquesl Incorporated, A Subsidiary of A.C. Nielsen Company / 1290Ridder Park Drive/San Jose, GA 95131 / (408) 971-9000 / Telex 171973
o
JS.
D>
JQ
C
<D
W
O
O
•o
o
n>
a
>
W
c
cr
tn
TODAY'S
MANUFACTURING
COMPUTER HIERACHY
CORPORATE PLANNING (LARGE MAINFRAME)
AND CONTROL
>
b
(D
MANUFACTURING
CONTROL (MAINFRAMES)
o
O
T3
0)
ro
(O
o
SUPERVISORY
CONTROL (MINIS AND SUPER
MICROS)
CL
a.
n>
•D
u
SO
O)
3
EQUIPMENT CONTROL
(PCR'S, NC'S. Y ' S ,
ROBOTS. RELAYS. ETC.)
O
>
.o.
o
us
-NI
(D
o
o
o
a
(D
-Nj
u
etc.
etc.
SENSORS AND ACTUATORS
(SWITCHES. LIGHTS.
MOTORS, MAN
MACHINE INTERFACES.
ETC.)
t
3
o
O
-1
•D
o(D
p.
>
Ui
C
»g
a
CHARACtEtilSTICS OF THE MAMUrACmftlllfi CQlW>BTFB kTFIUttftnr
0)'
•<
o_
>
p
z
THE MONEY IS SPENT AT THE BOTTOM.
2!
(A
(t>
3
I
?
•<
0
THE NUMBER OF UNITS GOES DOWN BY A FACTOR OF 10 TO 100 FOR
EACH STEP UP THE HIERARCHY.
0
THE RESPONSE TIME REQUIREMENT GOES DOWN BY A FACTOR OF 10
TO 100 FOR EACH STEP UP THE HIERARCHY.
3
ro
g
a
Q.
(D
•o
THE LEVEL OF ABSTRACTION GOES UP WITH EACH STEP UP,
«:
0)
«
ou>
I
a
0
THE NUMBER OF PEOPLE WHO CAN UNDERSTAND EACH LEVEL GOES DOWN
FOR EACH STEP UP THE HIERARCHY.
>
5
0
THE COMMUNICATION SYSTEM ARCHITECTURE OFTEN FOLLOWS THE HIERARCHY
ARCHITECTURE,
s
g
0
THE BOXES AT THE BOTTOM OF THE HIERARCHY ARE ON THE FLOOR NEAR THE
EQUIPMENT BEING CONTROLLED.
^
I
(S
(O
-~)
u
a
m
S
n
c
a>
(A
5"
r>
g
•5
o
a>
a.
HAMUFArTURTMG FtQU
>
in
c
IT
<A
Q!
lu'
•<
o^
>
b
ENERGY AND
MATERIAL
INPUTS
CD
OUTPUT
o
o
3
n
0)
ro
EACH NODE, • . REPRESENTS
AN ASSEMBLY OR PROCESS
STATION.
(£>
o
Q.
a
(D
EACH B R A N C H , - , REPRESENTS
A STORAGE OR TRANSPORT
SYSTEM.
TJ
tu
V)
0)
3
o
>
(O
tn
.b.
o
00
TIME AND POSITION
(O
o
o
o
5.
(O
->J
CJ
i
cca
u>
3
o
o
• ^
•o
o
(D
,p.
>
05
CHARACTFRISTICS OF THE MANUFACTURING Fl OU
c
CT
W
a
lu'
>3
o_
>
b
0
EACH NODE AND BRANCH IS DIRECTLY CONTROLLED BY THE BOTTOM
LAYER OF THE MANUFACTURING COMPUTER HIERARCHY.
i
^
0
THE FLOW DIAGRAM HAS A ONE - ONE RELATIONSHIP WITH THE
PHYSICAL FLOW OF WORK IN PROCESS.
ro
<0
o
0
MANUFACTURING MARGINS GO UP AS THE INVENTORY IN PROCESS
GOES DOWN.
0
MANUFACTURING MARGINS GO UP AS THE PRODUCT QUALITY GOES UP.
^
0
MANUFACTURING MARGINS GO UP AS POOR QUALITY IS DETECTED AND
CORRECTED AT ITS INCEPTION.
O
0
THE PRODUCT DEFINITION IS CHARACTERIZED BY THE MATERIAL INPUTS AND
ASSEMBLY OR PROCESS STATION DEFINITION.
a
Q.
(D
TJ
Q)
<'
>
(O
tjl
O
00
-.1
8
•H
(B
o
»
DT
n
c
0)
•D
O
(D
Q.
(n
c
cr
MDIllRFUFIitS wm THF HAMUEACTIIIMHB HlERARtHY
(A
a!
a>
•<
o^
>
b
REDUCE THE INVENTORY IN PROCESS
en
<D
0
THROUGH ABILITY TO HANDLE SMALL BATCHES,
o
1
THROUGH ABILITY PROCESS BATCHES QUICKER
to
<o
o
THROUGH ABILITY TO MANUFACTURE IN RESPONSE
TO ACTUAL SALES RATHER THAN SALES FORECASTS,
»
THROUGH HIGH AVAILABILITY.
•D
0)
3r
'^
i
0
CO
01
3
INCREASE THE PRODUCT QUALITY.
THROUGH INSPECTION AT ASSEMBLY OR PROCESS STATIONS OR
AT THE NEXT STATION WITH IMMEDIATE FEEDBACK.
O
>
S
0
RESPOND QUICKLY WITH LOW COST TO PRODUCT CHANGES,
0
TOYOTA EXAMPLE.
"S
I
(O
o
o
o
5.
a>
OJ
o
SI
D)
n
c
CB
3
O
q
•s
o
CP
_a
>
w
c
aCA
a
ft)'
><
o^
FinoR rnMTROi FFATIIRFS RF<;PQNSIVE TO HANUFACTURIMG NFFDS
>
b
(D
3
o
o
3
•D
HOW PROVIDED
FEATURE
EASILY UPDATED WITH NEW PROGRAM
NEW PRODUCT
0
0
0
5
PROGRAMMABIHTY
PR06RAMMABILITY BY
MANUFACTURING PERSONNEL
COMMUNICATION TO
SUPERVISORY COMPUTER
Q!
•-»
•0
B)
o
w
RAPID C O N T R O L RESPONSE FOR RAPID
THROUGHPUT
0
0
0)
a
o
>
<o
Ol
.b.
CO
2.
CD
CO
CJ
0
SHORT RESPONSE TINE
PROGRAM EASILY TUNED BY
MANUFACTURING PERSONNEL
PEER - PEER COMMUNCIATION
o
m
m
.o
cCD
(A
3"
o
o
•a
o
- 1
>
w
c
a
U)
FLOOR EOMTROI FEJLtURFS RFSPiJIt^lVF TO imiUFJl£TlMlifi MFFUS I t M I T ' l l l
a
u'
h
o
o
3
EE&mSE
HIGH. CONTROL AND CONTROLLED. AVAILABILITY
•o
fi)
3
•<
KO
<o
o
Q.
Q.
(D
T3
lU
m
fi>
C_
O
0)
<D
O
>
(O
Ol
o
3
<o
-J
<D
O
O
O
(O
~^l
CO
PROVISION OF WORK IN PROCESS
AND PROCESS FLOW DATA FOR PLANT PERSONNEL
FOR PROCESS TUNING
HaB PRWtBgtt
0 RUGGED CONTROL
0 EASILY MAINTAINED BY
MANUFACTURING PERSONNEL
0 CONTROL HAS ABILITY TO
DIAGNOSE OUR FAULT
0 CONTROL HAS ABILITY TO
DIAGNOSE CONTROLLED
EQUIPMENT FAULT
COMMUNICATION TO
MANUFACTURING FLOOR
MAN/MACHINE INTERFACE
QUALITY MEASUREMENT DATA
GATHERING CAPABILITY
COMMUNICATE TO SUPERVISORY
COMPUTER
GLOBAL COMMUNICATION
o
0)
o
o
•n
•o
o
n>
Q.
>
w
FtnOR rONTROI FFATURFS RESPONSIVE TO MANUFACTURING NFFDS (CONT'D)
Ia
.>
p
i
o
o
3
•o
B>
><
ro
<£)
o
I
IP
m
0)
3
C
o
0)
a
o
>
<o
oi
o
00
CO
o
o
o
(O
u
IMMEDIATE FEEDBACK OF OUT OF SPECIFICATION
PERFORMANCE TO ALL CONTROL AND PERSONNEL
0 GLOBAL COMMUNICATION
0 COMMUNICATION TO
SUPERVISORY COMPUTER
0 COMMUNICATION TO
MANUFACTURING
0 FLOOR MAN/MACHINE
INTERFACE
0 INTERFACE TO INSPECTION
EQUIPMENT
o
S
c
(D
3
O
o
•3
o
<II
a.
w
c
IT
a>
S.
a>
>
b
a
3
o
HANUFAriliRINR FIOOR mMPIITATlOM DFVirFS
0
RELAYS
0
NUMERICAL CONTROLS
0
PROGRAMMABLE CONTROLS
0
MINIS AND MICROCOMPUTERS
0
ROBOT CONTROLS
O
3
T3
0)
(C
o
a
a.
a
TJ
0>
o
(C
c?
B)
C_
O
n
fl)
O
>
.u
o
00
8o
o
5.
(D
CO
^J
u
a
ST
c
(n
o>
3
o
o
•D
1
a
RELAYS
I
0
BOOLEAN LOGIC, HANDWIRED
t
0
SOFTWARE AND HARDWARE MAINTAINED BY ELECTRICIANS
0
MULTITASKING DONE BY SEPARATE LOGIC
0
COST EFFECTIVE FOR ^
0
RUGGED
0
10 MS RESPONSE TIME
0
NO COMMUNICATIONS CAPABILITY
0
VERY LIMITED DIAGNOSTICS
0
MAJOR MANUFACTURERS
>
b
»
3
o
20 RELAYS
O
3
•D
0)
I
i
t
f
i
5
.t.
ALLEN BRADLEY
§
o
SQUARE D
o
00
o
-I
5.
a>
>c
^
(O
-.J
u
3
o
o
—1
T3
O
CD
Q.
>
c
CT
0>
a.
0)'
-^
•<
PROfiRAHHARI F CONTROLLERS
o
>
J
0
LADDER DIAGRAM (BOOLEAN LOfilCJ.
0
ALSO:
s!
u
I
i
?
TIMERS, COUNTERS, ARITHMETIC
AND ASSEMBLY LANGUAGE LIKE
INSTRUCTIONS.
•3<
S
0
PROGRAMS SELF DOCUMENTING, EASILY CHANGED,
0
PROGRAMS MAINTAINED BY ELECTRICIANS AND
3
5!
Q.
I
D>
TECHNICIANS.
W
O
>
U3
Ul
.b
O
00
<o
o
o
o
H
2.
(D
0
HARDWARE MAINTAINED BY ELECTRICIANS AND
TECHNICIANS.
g"
1
c
(C
(A
3'
O
I
pPOKPAMMARI F rONTROU FRS (CONT'D)
I
?
0
DIAGNOSTIC CAPABILITY BUILT IN
0
MULTITASKING EASILY DONE
0
CAM INTERFACE TO INSPECTION EQUIPMENT
(A
a
I
>
?
V>
f
0 TYPICAL COST J5000.
t3
A)
I
0
RUGGED
I
0
DISCRETE AND ANALOG I/O
5
0
10 - 100 MS RESPONSE
!
§
- (D
o
>
0
COMMUNCIATIONS VIA RS232 ASCII. AND PROPRIETARY
BUS COMMUNICATIONS PROTOCOLS
2
0
MAJOR MANUFACTURERS
I
8
GOULD
ALLEN BRADLEY
S
(O
--J
w
o
nS
c
CD
W
5"
o
o
TD
O
<B
a.
ROBOT rnHTROl
>
Cfl
c
cr
0)
0
WALK THROUGH PROGRAMMING
0
SUPERVISORY PROGRAMMING VIA:
0
PROGRAMS CAN BE DOCUMENTED
0
PROGRAMS EASILY CHANGED
0
PROGRAMS HAINTAINED BY ROBOT CONTROL TECHNICIAN
i
0
HARDWARE MAINTAINED BY COMPUTER TECHNICIAN
s
0
SOME DIAGNOSTICS
0
SINGLE TASK
0
CAN INTERFACE TO INSPECTION EQUIPMENT
S
0
TYPICAL COST J15,000.
I
0
RUGGED
i
0
SERVO AND SOME DISCRETE I/O
0
TO MILLISECOND RESPONSE TIME
a
0
RS232 ASCII COMMUNICATIONS
Io
0
MAJOR MANUFACTURERS
a.
t>
•-<^
o_
>
Ib
I
o
I3
•o
0)
•<
VAL, RAIL. AML. ETC
•
Q.
?
3r
I
fl)
Ol
t
o
(O
-g
o
SE.
G.E.
(I)
GENERAL NUMERIC
u
o
£.
S
c
(P
(A
3
O
g
•3
o
f^iMT-Mirpn roHPUTFR
(D
a
>
3
a
>
b
5.
u>
0
ASSEMBLY AND HIGH LEVEL PROGRAMMING LANGUAGE (NAME
YOUR FAVORITE).
0
PROGRAMS CAN BE DOCUMENTED.
0
PROGRAMS MAINTAINED BY SOFTWARE/SYSTEM PERSONNEL.
0
HARDWARE MAINTAINED BY COMPUTER TECHNICIAN.
0
DIAGNOSTICS MAY BE PROGRAMMED.
0
MULTITASKING SUPPORTED BY VARIOUS OPERATING SYSTEMS.
0
CAN INTERFACE TO INSPECTION EQUIPMENT.
0
TYPICAL COST J 5 0 0 0 .
<t>
z>
o
I
a.
a.
a>
at
CO
o>
3
o
>
o
00
8o
o
(£>
--J
CJ
(WITHOUT PROGRAM).
o
K.o
c
m
v>
3
O
o
o
•3
CD
_a
>
w
c
CT
CO
s
a
MTMT MTCRO r.OHPIITFR fCONT'Dl
>
b
$
o
o
0
NOT RUGGED.
0
FULL RANGE OF I/O.
0
TO 10 MICRO-SECOND RESPONSE TIME
„
0
RS232 ASCII COMMUNICATIONS
i
0
MAJOR MANUFACTURERS
I3
^
a.
a.
(B
g.
MOTOROLA
I
INTEL
>
DEC
.u
o
00
(O
o
o
o
UJ
CO
<
ffi
o
UJ
<
co
~
a:
.l o• ^^
tn
0 ( j
1^3 ro
CO
2 0
0
-'£
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
lO
00
a>
!5
(T
I lii
<
^
CO
0|CO
OC 3 H U J
O Q.OO
I-
52SO
€0
is
iS
o
o
10
S
T
I
I
i
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
o
B)
r>
O
•3
o
(B
a.
W
c
CT
W
5!
>
b
<B
W
(B
o
o
3
RELATIVE
SALES/UNIT
TIME INTO U.S.
INDUSTRIAL
MARKET
RELAYS
•o
0)
3
J)
Q!
a
(B
•D
0)
CO
Dl
D
CO
(A
(B
O
>
CO
Ul
.b.
o
TIME
00
(O
o
o
o
2.
a
(D
-sj
u
PRODUCT POSITION ON LIFE CYCLE CURVE
I
(C
p.
>
CO
c
CT
0)
ci
u'
><
o
CQMHUHCATIQH ZOO
5
o
o
RELAYS DO NOT COMMUNICATE
1
OF THE COMMUNICATORS LESS THAN 20X ARE CONNECTED TO
SUPERVISION
>
b
M
(O
O
s
I
OF THE COMMUNICATORS SOX ARE CONNECTED TO AT LEAST
ONE PEER
TJ
fi>
-^
O
jr
VENDORS ARE OFFERING:
^3
T I WAY
I
AB DATA HIGHWAY
2.
MODBUS
oi
I
MODWAY
.b.
I
6M MAP
-g
i
DEC NET
o
^
ETHERNET
CD
5
802
S
NAME YOUR FAVORITE
I—
<
u.
en
oo
Ul
o.
Q.
n
<
^
o
^3
00
(/>
(/)
t/»
yj
uu
Dataquest Incorporaaed, A Subsidiary of A.G. Nielsen ComiMiny /1290 Bidder Park Drive / San Jose, CA 95131 / (4(») 971 -9000 / Telex 171973
Dataquest
Dataquest
INTEGRATED CAO/CAM:
AN ACHIEVABLE GOAL
Richard L. Justice
Director
Engineering Computer Coordination
Advanced Product and Manufacturing Engineering Staff
General Motors Cor]?oration
Mr. Justice is Director of Engineering Computer Coordination and is one
of the principals involved in establishing the integrated CAD/CAM system
for General Motors. He has served his entire professional career with
General Motors.
Prior to his current assignment, he was Manager of
Research Development for the Personnel Administration and Development
staff.
Previously, he was
Executive
Engineer
at
the
Chevrolet
Engineering Center and was responsible for developing the On-Line
Engineering Releasing System.
Before that he served with the General
Motors
Proving
Grounds
and
with
the
General
Motors
Research
Laboratories. Mr. Justice received his bachelor's and master's degrees
in Mathematics from Purdue University.
Dataquest Incorporated
CAD/CAM INDUSTRY CONFERENCE
September 26-28, 1983
Newport Beach, California
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
INTEGRATED CAD/CAM - AN ACHIEVABLE GOAL
Almost everything seems to be Computer Aided today.
We now place CA
in front of all of our industrial processes—design, engineering, test,
manufacturing, instruction, scheduling—the list goes on and on.
important to industry leaders today is achieving
What is
the full potential of
that computer aiding.
It has been remarkably profitable to use these aids
even
cases, they were
though, in most
strictly
used
for automating a
formerly manual task—or only represented a productivity gain.
What the
industry now desires is a path that will lead to integration of these
entire systems.
General Motors is a company that is certainly able to exhibit all of
the
issues
that
computer-aided
have
spawned
technologies.
this
set
of
non-integrated
With over 500,000 employes engaged
separate manufacturing and assembly
cars and
current
trucks each year, and
in 162
facilities producing over 6,000,000
distributing
them
to more
than
10,000
dealerships around the world—we do have an opportunity to severely test
the extent to which our systems are integrated.
It
is
my
contention,
however,
that
any
industrial
concern
can
demonstrate most of these same problems, be it another General Motors or a
cottage industry.
The objective of this presentation is to outline for
you just what are the obstacles or roadblocks to integrated CAD/CAM. Even
though
some would
suggest
that
it
is
foolhardy
to
undertake
such
a
listing, I believe it to be limited to ten broad categories.
Dataquest Incorporated, A Subsidiary of A C . Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
Product Configuration Control
Data Representative and Proliferation
Data Exchange
Enterprise Data Management
2-D to 3-D Conversion
Surface/Solid Modeling
Math/Wood/Metal Models
Tool Design/Construction
Technology Transfer
Management Discipline
The format of this presentation will be to discuss each of these areas, or
roadblocks, in a manner
that
illustrates not only how they delay
the
achievement of integration, but also to indicate how and when they may be
overcome.
Others will enumerate the cost/savings of integrated CAD/CAM, a
subject to which our proj.ect could also make contributions, however, what
is
important
at
this
point
along
the
road
to
such
a
system
is
the
recognition of and agreement concerning solutions to the major roadblocks.
Consider each of them in turn:
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408> 971^9000 / Telex 171973
PRODUCT CONFIGURATION CONTRUL
In order to produce an end product, it is essential to describe in
detail every aspect of that product—from
assembly and distribution.
conception through to fional
Hence, product configuration information will
encompass physical properties, geometric data, usage data, manufacturing
concepts, assembly fixtures and a lot more.
manual world has been less than optimum.
date
have
only
attempted
to
Control of these data in a
Unfortunately, many of us to
automate
that
cumbersome
manual
procedure—and not taken advantage of electronic methods to bring it into
full operation.
From the point of view of producing a product, CAD/CAM has not yet
achieved a significant
level of integration.
CAD/CAM has been applied
separately to many aspects of design and manufacturing, and has resulted
in significant local improvements in productivity.
This first level of
return from CAD/CAM has fueled rapid growth, but why has integration been
so
elusive?
Major
efforts
have
been
going
on
for
(Integrated Program for Aerospace Design), ICAM (Integrated
Manufacturing), and all other commercial
functions.
systems
tout
years
in
IPAD
Computer-Aided
their
integrated
It is obviously easier to recognize the need for integration
than it is to achieve it.
The key ingredient to the successful CAD/CAM applications to date has
been the graphic console and the ability to use human understanding along
with
computer
tools.
Very
little
of
CAD/CAM
has
become completely
Dataquest Incorporated, A Subsidiary of A.C, Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
automated to the point where man is no longer valuable.
integration
is that
it has been
looked upon as a
The problem with
computer-automation
process that did not require an extensive human component.
All the data
would be stored in the computer in a totally neutral form and drawn off as
needed by downstream activities. More and more this simplistic view is
proving to be unrealistic.
The successful data bank at GM's Fisher Body Division and the CIIN
data management system at Boeing are largely manual processes that use
limited
computer
aids.
Every
GM
design
activity
has
a
defacto
data administrator who has developed a mostly manual system for keeping
track
of
design
files.
The
PDDI
program
(Product
Data
Definition
Interface), which is a subprogram to ICAM, is beginning to discover that
product data bases are totally inadequate to drive manufacturing systems.
The whole concept of a computer-driven CAD/CAM process is currently being
overstated.
Industry
needs
to recognize
the data
administration
function
design in the same way as it recognizes specifications people.
(There
even the possibility that these two functions could be combined.)
event, data administration processes need to be formalized
for
is
In any
rather than
continue on an ad hoc basis.
There
data
is technology
administration
becoming
function.
available
Relational
that will greatly
aid
the
data base managers, such as
IBM's DB2, show promise of being flexible enough to evolve along with the
data
administration
function and
provide any
degree of
relationships.
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 RIdder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
Integration is inherently the management of relations. The combination of
people and these new tools could result in a long-term breakthrough in
data administration.
A general conclusion is that product configuration control is not
attainable until this data administration function is put in place with a
relational data base manager.
I will return to this point later in my
presentation.
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171972
DATA REPRESENTATION AND DATA PROLIFERATION
There are many ways to represent product design in a computer.
The
most simple schemes use the old concept of "Do it like we've always done
it."
The result
representations
is a computer-aided
of
sections
of
drawing
the product
system.
are
Two dimensional
designed
and
Transfer of the 2-D data to other 3-D systems is difficult.
plotted.
There are
many of these systems and CADAM is one example.
More complex computer systems allow the design of 3-D "wire frame"
models.
Wire frames are blended with equations to create surface "math
models."
Points
and
vectors
polynomials, arcs, spirals, etc.
the
product
themselves
being
to wire
designed.
are
connected
by
other
vectors
or
by
The most suitable technique depends upon
For
frame approach.
example,
sheet
Traditionally,
metal
parts
designs
are
lend
designed
without metal thickness shown, and hidden lines are not shown as they are
in the design of castings.
Individual point locations are connected by
Euler spirals.
Somewhat more important, than specific techniques used to design and
define the product, is the compatibility with downstream users.
Sheet
metal product design is based on aesthetic and aerodynamic requirements.
Basically, the mathematical
implementation
along the arc length of individual lines.
is the control of
curvature
Ex:
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
Used for aesthetic design
K
= Increasing
dK/ds
= Increasing
d^K/ds^
= Increasing
Requirements for tooling and manufacturing are d i f f e r e n t
in product
design.
Design
of
sheet metal parts
in each
from those
stage of
manufacture Includes draw binders, draw walls, trim ledges, beads, prongs,
and other metal control devices.
Usually the design criteria involves
length of line analysis and minimum energy curve design, or-
/K^ds =
minimum along arc length.
Obviously, computer
systems need
to be different
to accomplish
different design goals.
Transfer of data from one computer to another always involves some
loss of precision and definition.
Retaining data in one computer so that
successive operations can be performed on it is a desirable goal.
Even
so, the problem remains that down stream users.have little control over the
format of the design given them. Curves designed by spirals in one system
do not always map properly into a linear or vector environment.
Two
dimensional overlays or drawing type output can be almost useless in a 3-D
machining mold.
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
Thus, we have the problem of:
1.
Data Compatibility
Integrated
CAD/CAM
systems
must
produce
data
in
formats "usable" by all users.
The "explosion" of data that naturally emerges from the processing,
tool
engineering,
managed.
machining,
assembly
and
tryout
activities
must
be
A single sheet metal part may contain 100 lines or a total of
5,000 discreet point locations.
By the time this part data is received at
the tool room machining area, it would look something like this:
(See
next page)
Dataquest IncorporateIJ, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
PART
DRMf
OP 1
TOOL 1
I
TRIM
OP 2
TOOL 2
CASTING 2
UPPER SHOE
CASTING 1
LOWER SHOE
1
FLUIGE
OP 4
TOOL 4
PIESCE
OP 3
TOOL 3
CASTING 4
ADAPTER
(
OPERATION 1
FACE MILL
1
OPERATION 2
CONTOUR MILL
WORK ELEMENT
AREA 1
COTTER 1
SETUP 2
BOTTOM
'
^
DETAL 1
STEELS
1
CASTING 3
PAD
SETUP
TOP
CAM FLANGE
OP 5
TOOL 5
1
SETUP 3
ANGLE
1
1
OPERATION 3
PROFILE
OPERATION 4
DRILL
WORK ELEMENT
AREA 2
CUTTER 2
ETC.
I
ETC.
BORE
WORK ELEMENT N
ETC
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive/ San Jose, CA 95131 / (408) 971-9000 / Telex 171973
Thus, we have problem number two:
2.
Data Proliferation
In order to produce the parts for the dies which stamp one
panel, we may be l o o k i n g
at several thousand operations.
Some of these operations require enormous data storage and
retrieval capabilities.
For example:
Complete N/C machining of a draw die for a hood panel
requires approximately 100,000 tool centers.
z coordinate
(18 scalars).
The
Each tool center is an x, y,
18 scalars plus relational information
require 40 bytes or 20 words for a total of 2 million words.
At the rate
of one-tenth inch per character, over five miles of punched mylar tape are
required to drive the N/C mill.
for just one die!
This is approximately 40 rolls of tape—
Punched tape will not be the medium of the future.
Astronomical numbers of tool centers are not the only concerns in
data proliferation.
As part surfaces are duplicated
in tree structure
processing/manufacturing operations, it becomes increasingly difficult to
quantify
the
effects
of
engineering
changes.
Which
element
operation on which detail of which die is affected—and
Is a new stock list required?
Are new setups required?
of
which
to what extent?
And so on...
Modularized CAD/CAM systems cannot deal with the above problem.
doesn't matter whether B. Splines (Computer Vision), Polysurf
Gordon Surface
It
(Prime) or
(GM CGS) were used to define the element if you cannot
readily find every occurence on the structure.
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
Perhaps
the most
important
single
change we could make
is to
completely sever the dependence on drawings. Their past importance still
pervades current thinking.
Design systems should create computer math
models with downstream compatibility.
inaccurate and o u t - o f - d a t e .
Drawings are generally incomplete,
The math model itself could be kept up to
date and, if properly connected
to a relational data base, all the
downstream operations could be kept up to date as well.
What Is needed is to have one item known to several operations.
Example:
A more sophisticated system includes not only the associated part
data to an operation, but the attitude (spatial position) of the item by
use of a transformation matrix. This further reduces data proliferation.
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
In addition to the above structure, each operation would relate to
items not transformed
operation).
(generated directly in the attitude used in that
Item B could be transformed for some other operation.
/ Item
Also needed
\
is an associative mechanism whereby
a surface
(Item)
would be known to the tooling members which contain it.
r
- 7---'
-fOP 1
Item
'''
V
If Item A is changed, it should be possible to easily determine all
the tooling members affected and all the operations affected.
In order to achieve effective integrated CAD/CAM, the £omputer Aided
Design must be formatted so that Computer ^ided Manufacturing systems can
properly
mechanisms
utilize
for
it.
This
implementing
includes
change.
compatible
Critical
data
structures
and
information, dimensions,
etc. must be packaged according to the operations which will be performed
at a later date.
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
DATA EXCHANGE
Since the early 70's, when CAD/CAM began to be used for production
activities by non-computer related engineering p r a c t i o n e r s , there has been
a serious problem in communicating between systems.
In a large multi-
product manufacturing organization there are not only the usual problems
of
communicating
departments, but
between
also
product
between
departments themselves.
and
manufacturing
the many, varied
engineering
product
engineering
A system that would satisfy the needs of a
department responsible for designing steering components might not be the
same as for a department responsible for designing heating and air conditioning components.
The facts are that many companies do not have a
central controlling area for any one discipline, and the result is many
different systems that cannot exchange data. The problem is so acute in a
multi-divisional, multi- product organization that attempts to quantify
the value of data exchange capability often fail. The lost opportunity of
not using the computer-generated CAD/CAM layout background information
alone, costs millions of dollars. The lost opportunity to use the design
data
for
thoroughly
other
dis- ciplines
investigated.
in
Such
product
straight
engineering
forward
has
not been
applications
as
documentation, technical publications, mass calculations and engineering
analysis, would amount to additional millions of dollars.
This brief statement has not even dealt with the use of the product
engineering
CAD
data
being
used
by
the
downstream
departments
in
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
manufacturing
engineering,
engineering.
Such
activities
as process
planning,
tool
material handling, packaging, numerical control machining,
robotics programming and numerical control inspection would all benefit
immensely if the CAD and CAM data from various systems could be exchanged
easily and with a substantial level of richness.
Many
companies
faced with
the frustrations
of not being
able to
exchange CAD/CAM data among their own units have developed very specific
one-to-one converters.
The
GMDES
(General
Some have met with fairly cost effective results.
Motors
Data
Exchange
Standard)
facility
has
been
extremely productive and easy to Implement by both General Motors and the
major
CAD/CAM
suppliers.
However,
the
three-dimensional point set data only.
implemented.
GMDES
software
deals
with
No dimensions or text have been
While many applications do not require exchange of dimen-
sional data, there are many applications that do require close tolerance
dimensional data to accompany the part geometry.
In
graphics
1979, GM's
Exchange
representatives
Specification)
Sciences Auditorium
accepted
attended
meeting
in Washington, D.C.
at
the
first
the
National
In early
IGES
Academy
by
of
1980, General Motors
IGES as the long-range direction of the corporation.
1983, this direction was confirmed
(^.nitial
the Executive Committee
In May,
of
the
Corporation, when it was proposed that IGES be the major data exchange
facility
in the GM
priority continues to
integrated
CAD/CAM
plan.
However, the problem of
plague the vendors in fully implementing IGES.
The
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Parl< Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
majority of the user community, although recognizing the need for data
exchange, have not enforced their desires with performance requirements on
purchase orders.
It is one thing to verbalize support and quite another
to specify with resulting withholding of funds for lack of compliance.
General Motors units will begin specifying IGES compliance on purchase
orders and the appropriation request review process will assure this
compliance requirement.
This in itself will not cause the data exchange
problem to be solved, but major purchasers like GM, working with the major
CAD/CAM
suppliers, will
hasten
the
acceptance
of
IGES
as
a
true
international standard.
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 17197:
ENTERPRISE DATA MANAGEMENT
Amongst the many good results of current CAD/CAM efforts are some new
problems.
Specifically, many members of the industrial community
themselves with a multitude of good usable data bases.
they are generally not locatable, connectable or shareable.
have been used to describe this problem.
find
Unfortunately,
Various terms
We refer to it as the need for
an Enterprise Data Manager.
In the automobile industry we have the following examples of this
problem.
resides
How do we collect all of the data for one given car product—it
in
multiple
divisions, multiple
systems, and multiple data base managers.
suppliers, multiple
computer
Some of the data is geometric,
some is material properties, some is in part usage terminology, and some
is proprietary.
Today, we have enterprise wide managers—there are people and paper
systems.
Tomorrow, we will have relational data base managers that will
automate
today's
manual
system.
managers is not new or mysterious.
The
technology
for
developing
these
Several vendors are in the marketplace
now with such product offerings.
Why has industry not stepped forward and used these offerings?
are expensive by way of resource consumption.
facilities,
and they
are demanding in
They
They use cycles, they use
their initialization, as
well as
Oataquest Incorporated, A Subsidiary of A.C. Nielsen Connpany / 1290 Bidder Parl< Drive / San Jose, CA 95131 / <408) 971-9000 / Telex 171973
subsequent care and feeding.
They will, however, perform the tasks that
we need and all of us will be implementing them soon.
It is inevitable.
The constantly decreasing c o s t of MIPS, currently at about 25% per year,
will hasten the use of these new r e l a t i o n a l data base managers.
Current p r o j e c t i o n s are that p i l o t operations with these systems will
commence shortly, run for two to three years, and become operational in
late 1985.
It may well be another two years before the vendors have fine
tuned them (along with the new hardware) so that they will be truly cost
effective. They will be commonplace this decade.
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
2-D TO 3-D CONVERTERS
A goal of Integrated CAD/CAM is to create solid models of all 3-D
solid parts so that design analysis and validation can be performed to
Improve quality and optimize the designs.
Some of the roadblocks to this g o a l are:
(1)
Many
draftsmen
have
a difficult
time handling
the spatial
relationships of 3-D, but can quickly create 2-D computerized
layouts in a natural and intuitive manner using
sketch-like
input methods.
(2)
It takes too long, using current interactive methods, to merge
2-D orthographic views to generate 3-D solid models; and hence
only those parts that absolutely must be analyzed are modeled.
(3)
There
are
thousands
of
existing
hardcopy
2-D
engineering
drawings that need to be added to the data base so that they can
be more easily used, modified, and integrated into new designs.
Existing digitizer methods are too time consuming.
(4)
2-D engineering drawings must continue
to be generated
(and
cannot yet be eliminated) due to the fact that many suppliers
still require them.
There
are two approaches
to removing
these
roadblocks.
First,
continue to improve the 3-D solid modelling design aids and integrate
existing
3-D design packages
for those designers who can handle the
spatial relationships of 3-D.
Second, automate the process of changing
2-D computer designs and manual drawings to 3-D solid models.
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
Scanner digitizers are becoming available to automatically scan and
input the 2-D engineering drawing information into CAD/CAM data base as
2-D data (i.e., points, lines, text, etc.).
Developments, in regards to
converting 2-D data to 3-D solids have been conducted at GM, IBM, MIT,
Philips
Research
Laboratory,
and
others.
Also, algorithms
developed to automatically convert 2-D drawings
lines) into 3-D
solid volumetric
(solids formed by plane faces) .
have
been
(consisting of straight
representations
of polyhedral objects
Further developments are being conducted
to extend these algorithms to handle circular arcs.
Industry
engineering
can
process
and
by
will
continue
developing
its
better
efforts
solid
to
automate
modeling
scanner digitizers, and 2-D to 3-D conversion algorithms.
the
techniques,
Indications are
that this roadblock to Integrated CAD/CAM can thereby be overcome by 1985.
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
SURFACE/SOLID MODELING
Many of the problems with solid modeling are related to the fact that
this is an emerging technology.
Most of our current efforts lie in the
CAD area of modeling a solid part.
wireframe
technology
in the area of complete
Wireframe data is often ambiguous
objects.
One of
Solid models offer advantages over
the
advantages
and
of
definition
of
can often represent
the
solid
model
is
the part.
nonsensical
its complete,
unambiguous definition of the part.
Because
computations
detection).
this
technology
cannot
be
is
performed
still
reliably
being
or
developed,
automatically
certain
(collision
Solid modeling operations are computationally intense, which
causes a large amount of the computer's resources to be used.
This means
that some operations are very slow, thereby preventing creative designers
from achieving
their
full potential.
As hardware
speeds
increase and
algorithms improve, these speed problems will be minimized.
The ability to communicate data from other design or solid modeling
systems is one other problem.
There is research underway at a number of
universities to produce solid models from 2-D and 3-D wireframe models.
Solid modelers use different representation schemes:
*
Boundary representation
*
Constructive solid geometry
*
Sweeps
Dataquest Ineorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 1.71973
It
is often
*
Spatial enumeration
*
Cell decomposition
*
Primary instancing
impossible
to communicate
this information between
different solid modeling systems.
Most solid modeling systems started with primitive solids (block,
cone, sphere) but the world is composed of more than these primitive
solids.
Currently, most solid modelers are adding free form or sculptured
surfaces that can represent blending or filleted areas.
modeling systems provide the user with graphics (hidden
Many solid
line removal or
color shading), mass properties (volume, moments of inertia, products of
inertia, and surface area), and static interference detection.
As
the
CAD
side
of
solid
modeling
becomes
applications are beginning to be developed.
more
stable, CAM
The areas that are under
development by different solid modeling systems are:
*
Automatic finite element mesh generation
*
N/C programming and tool path verification
*
Dynamic interference detection
*
Kinematics using solids
*
Robotics
*
Simulation
*
Off-line programming
*
Path and positioning of robot
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
It is through these applications that the power of a solid model can
be
realized.
When
automatic
FEM
generation
becomes
available,
the
designer will easily be able to design and then finite element analyze the
part, determine
areas of weakness
potential problems.
and modify
the design
to
eliminate
Currently, the fact that it often takes more time to
create the finite element model than to design the part severely limits
the use of finite element analysis.
When the process of generating the
finite element model is automatic, the designer can iterate through the
design and analysis process for the optimum design.
If information is placed into the solid modeling data base while the
part is designed, such as what size hole to drill or tap, then an N/C
generation
program
could
automatically
determine where to drill the holes.
interrogate
the solid
model to
The verification of the tool path and
the use of collision detection to detect when the tool interferes with the
clamps or the machine tool is another application under development.
Robotics is another good CAM application under development on some
solid
modeling
systems.
The
ability
to
simulate
the
movement
and
constraints of the robot is underway.
Currently most robots are being
programmed manually or in a teach mode.
There are some systems that are
investigating the off-line programming of robots.
This would be much like
the way N/C machine tools paths are now generated.
There
are
other very
chamber analysis.
specialized
This program
applications
such
as
combustion
uses the solid model to define the volume
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
of the combustion chamber and calculate the flame front and flame travel
as the gas is ignited. It is through these specialized applications of
solid modeling that much of its potential application can be reached.
Clearly, the roadblocks of speed, resource requirements, data base
and
applications
today.
are all being
addressed
by
the major
solid modelers
Before 1990, solid modelling will play a significant role in the
integrated CAD/CAM system.
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
MATH, WOOD, AND METAL MODELS (DIES)
The tooling industry has, historically, depended on a conventional
(manual) set of processes and procedures to construct dies and molds. The
heart of this process is a master wood model which, until recently, was
made by hand.
control.
Many wood model shops are now beginning to use numerical
The model represents the actual full-scale male shape of the
part.
Plaster/plastic male and female construction aids are duplicated off
the master model and
then traced
(kellered)
replicate the part surfaces in the dies.
on milling machines to
The current problems in this
process include poor model/aid dimensional stability, manual errors, model
construction costs, adverse impact of updated engineering changes and poor
machining efficiency and quality resulting from the tracing operation.
Final finishing is dependent on a hand-made checking model, called a
spotting rack, which may also have dimensional errors due to the effects
of temperature and humidity.
With the advent of computers and CAD/CAM/NC Systems, mathematics—
under development for over 25 years—in conjunction with computer aided
design systems have defined complex sculptured surfaces which are referred
to as math models. The results of the design operation are stored in the
computer's master data base.
I^ataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
Sculptured surface N/C programming systems are also being developed,
particularly
in
the automotive
and
aerospace
industries, to calculate
accurate, error-free tool paths on complex shapes to N/C machine models,
dies, molds, patterns, aircraft and missile surfaces, and a variety of
other complex surfaces in various industries.
The
major
automotive
and
developments
aerospace
in
this
companies.
area
The
have
been
original
spearheaded
developments
by
were
proprietary for many years; however, companies like Renault and Dassault
in France are marketing their systems, and Boeing's B-Surf has been in the
public
domain
effective
companies
for
and
many
faster
specializing
years.
With
computer
the
hardware,
in CAD/CAM
systems
availability
commercial
of
more
turnkey
are also beginning
cost
vendor
to offer
sculptured surface capability.
Sculptured
surface
design
capability
is
becoming
quite
common.
However, most N/C programming systems to machine complex shapes are still
limited in their ability to control a tool over multiple surfaces with all
interference problems solved
cutters.
correctly, particularly
for flat end mill
Intensive development effort is underway in the larger European,
American, and Japanese automotive companies to solve this problem.
While
it is generally held that math models will not totally replace wood and
metal models
(i.e., achievement of clay-to-die design and build), huge
savings will be produced by the reduction in requirements for such hard
models.
Much of this, however, will not be achieved during this decade.
Dataquest Incorporated, A Subsidiary of A.C, Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
TOOL DESIGN AND CONSTRUCTION
Tool design, like product design, lias been a highly skilled labor
intensive activity throughout the years since the Industrial Revolution.
Manual drawings have been the Interface and communications link between
design and manufacturing.
is to develop any
The purpose of tool design, generally speaking,
additional
surfaces
(addenda)
required
to make
the
actual part as in a stamping draw die, design all the details that must be
machined and assembled into a complete tool assembly, establish correct
dimensions, spacers, etc., for the die or mold to be installed in a press,
and to design the Interface equipment required to automate part handling
between presses or molds.
The tool design data serves as the primary
input to tool construction, which at the present time, is also largely
labor
Intensive with manual-based
information are used
operations.
Product and tool design
to make wood models which are then duplicated as
construction aids for the actual die surface machining.
railing, hole
drilling
and
tapping
has
been
done
The straight line
conventionally
on
manually-operated machines.
Commercial
vendor
popular for geometric
turnkey
systems
tool design.
are
now
becoming
The sculptured
increasingly
surface design
(die
layout, e.g.) remains a major difficulty with turnkey systems and is being
attempted mostly by In-house proprietary developments.
CAD/CAM systems
that have demonstrated good geometric mechanical design capability such as
Computervision, Applicon, Calma, and CADAM are being utilized
in tool
design activities to an ever-increasing extent.
.Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
Much of the machining for tool construction is of a relatively simple
geometric
point-to-point
capability
nature.
N/C
point-to-point
2-D
programming
is becoming available on integrated graphic CAD/CAM systems.
Some limited sculptured surface N/C machining capability is also available
on these systems, but the more sophisticated systems primarily reside in
the larger company proprietary operations.
The
solution, again, calls for an acceleration
transfer, since most of
ultimate user.
it exists
of
the technology
but has not been produced
for
the
Some of this will be accomplished during the eighties, but
to become fully integrated will require another ten years.
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
TECHNOLOGY TRANSFER
Moving a manufacturing facility from a traditional manual environment
to the high technology world of CAD/CAM has many problems associated with
One of the most critical is the retraining of the workforce.
it.
Without
a well-motivated and highly-trained workforce, the factory of the future
with its sophisticated graphic systems and N/C machines will not function
properly
and
predicted.
their
will not
provide
the
productivity
gains
that
are
being
In addition, the employes will be confused and frustrated in
new
environment,
destroying
any
quality-of-work-life
gains
previously made.
The
solution
to
training program.
this
problem
is
a
well-planned,
multi-faceted
Simply providing a training class for a new graphics
design system or N/C machining program will not be adequate.
must be attacked on several levels.
The first level is an overview or
introduction to the new technology for the entire workforce.
should
explain
what
CAD/CAM
is, what
machining is and how they work.
training
is
being
handled
colleges and schools.
by
The problem
computer
graphics
The overview
is, what
N/C
At several plants today, this level of
utilizing
courses
from
local
community
The second level of training involves making sure
that the employes have the basic skills that will be required in the new
environment.
abilities
training
may
can
Personal skills such as blueprint
be
required, but
also
education programs.
be
handled
not
by
all
local
reading and basic math
employes
community
The next level requires
possess
colleges
them.
This
and
adult
retraining machine operators
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
to handle the new N/C equipment. This involves not only teaching them the
new machine and controls, but also new work habits. Years of working with
old equipment has ingrained many employes with practices and procedures
that are no longer valid.
Management and the union must work cohesively
to establish new work standards.
The
functions.
final
level
requires
retraining
employes
to
perform
new
An example of this is training new N/C equipment. Because the
function is new to the selected employes and the systems are so complex,
the training will have to be handled in two steps. First training classes
will be held to provide an understanding of the systems.
Secondly,
technical experts will work with them on a series of prototype jobs.
These jobs will help strengthen their knowledge and build their confidence
before they have to function independently.
Although training the workforce for the high technology environment
is a critical problem, it should be obvious that with adequate planning,
all the training components can be covered to provide a smooth transition.
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
MANAGEMENT DISCIPLINE
Finally, in a d d i t i o n
to all of these technical roadblocks
pervasive obvious problem of management discipline.
enumerate
the physical
resources
to overcome
selling solutions.
It is one thing to
roadblocks, it is quite another
them.
This
is
is the
true whether
to obtain
you are buying
the
or
Because of the very breadth of the design-test-build
process in industry, it is not easy for any vendor to excel in solutions
to problems in mechanical-electrical, design-build, schedule-ship, etc.
Consequently, it is most often recognized that if management feels they
can solve the integration problem by specifying only one supplier, they
are mistaken.
Indeed, all of
the other nine
roadblocks must
be recognized
and
overcome, but there still remains a management resolve to be achieved.
Whereas, we speak of user training and retraining, there also exists a
need to carry that concept throughout all levels.
Effectively, there is little to be added here that is new, since we
all have for years blamed "them" for any problems.
at the feet of management.
We can always lay it
Fortunately, or unfortunately,
there is a
solution that stands four-square in front of us and will make this one
easy.
It is survival!
There is one way to gain full attention of everyone from the sweeper
to the Chairman.
That is profit!
If American industry is to survive the
great new off-shore competition, we must lead in the implementation of new
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
technology.
Let me
assure you
recog- nizes this fact.
that
the domestic
auto
industry
We are well on the way to full implementation.
Nothing will prevent us from being the leader in this process.
regard
fully
Integrated CAD/CAM as a cornerstone
object or faint resolve can be permitted.
to this goal.
We fully
No physical
We want to, we can and we will
join hands with many partners to achieve this goal.
We know what needs to
be done, and we are setting out to do it.
Please join us.
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
ACKNOWLEDGEMENTS
Special acknowledgements and thanks are offered to the following members
of the Advanced Product and Manufacturing
Engineering Staff, General
Motors:
L. J. Berch
P. D. Burt
F. W. Demasek
C. R. Lewis
G. L. Ross
E. J. Toton
C. J. Wezner
J. F. Yevtich
C. A. Zonca
Dataquesi Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose. CA 95131 I (408) 971-9000 / Telex 171973
Dataquest
Dataquest
ROLE OF START-UP COMPANIES IN
COMPUTER-INTEGRATE} MANUFACTURING
Laura Conigliaro
Vice President
Prudential-Bache Securities Inc.
As an analyst, Ms. Conigliaro covers the field of factory automation,
including robotics, computer-based vision systems, advanced technology
machine tools, and CAD/CAM, both
from an
industry
and
company
perspective. Ms. Conigliaro's previous employment includes several years
as an intelligence analyst at the National Security Agency. She was the
contributing author to a book on investanent and, more recently, has
contributed numerous articles to various publications, including her ovm
"CIM Newsletter." Ms. Conigliaro holds an M.S.A. degree in Finance. She
is a member of RI/SME, CASA/SME, the EIC Robotics Advisory Board, and Phi
Beta Kappa.
Dataquest Incorporated
CAD/CAM INDUSTRY CONFERENCE
September 26-28, 1983
Newport Beach, California
Dataquest Incorporated, A Subsidiary of A.C. Nielsen CoInpany /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
THE R O L E OF START-UP CCMPANIES IN CCMPUTER-INTEGRATED MANUFACTURING
THE WINNERS IN THE RCBOTICS AND/OR CCMPUTER-INTEGRATED MANUFACTURING AREAS WILL
BE ALLEGHENY INTERNATIOSIAL
BENDIX
CIM:INNATI MILACRCN
CTNERAL E L E C T R I C
GENERAL MOTORS
IBM
TEXTRCN
UNITED TECHNOLOGIES
WESTINOJOUSE
OR WILL THEY? I CM'T TELL YOU HOW MANY TIMES SINCE THESE CCMPANIES ENTERED THE
RCBOTICS INDUSTRY THAT I'VE HEARD IT SAID THAT THE GIANT CCMPANIES WOULD STEP CN
THE SMALLER CNES AND END UP OWNING THE INDUSTRY.
IT IS TRUE THAT THESE GIANTS HAVE SCME IMPORTANT ADVANTAGES - SUCH AS INSTANT
NAME RBCOQIITICN AND MAJOR FINAtCIAL STREt^JTH. THEY ALSO HAVE VAST DISTRIBUTICN
NETWORKS AND MAJOR FINANCIAL STRENGTH. BUT IT'S ALSO TPUE THAT IF WE TURN THE
C O I N OVER, WE SEE THAT PCWERFUL FINANCIAL STRENGTH FOR THE CORPORATIOsI AS A
WHOLE SHOULD NOT BE INTERPRETED AS BEING EQUIVALENT TO UNLIMITED FINANCIAL
RESCURCES FOR THE RCBOT ENTITY WITHIN THE COViPANY. THE ROBOT GROUP - IT'S OFTEN
NOT EVEN A DIVISICN - IS JUST A TINY, TINY FISH IN A GREAT BIG PCM). EVEN THE
GREAT - THE TRULY GREAT - IBM IS, IN THE END, RESOURCE-LIMITED. IT'S HARD TO
CONCEIVE CF A CCMPANY THAT WANTS TO BE IN RCBOTICS, FOR EXAMPI£, SO BADLY THAT
THEY WOULD CONTINUE TO WHOLEHEARTEDLY SUPPORT THEIR ROBOT GROUP IF IT TURNS CUT
T O BE A LEADER PRIMARILY IN CNLY ONE AREA - THE ABILITY TO LOSE MO^JEY.
(HE THING THAT I AM FULLY CCNVINCED OF IS THAT A NUMBER OF THESE GIANT CCMPANIES
WISH THAT THEY HAD NEVER GOTTEN INVOLVED IN THE RCBOTICS AREA OR, IF THEY HAD,
THAT THEY HAD NOT PUBLICIZED IT QUITE AS MUCH. A QUIETER ENTRY INTO THE INDUSTRY WOULD HAVE iULLOWED 1HEM TO MOVE MORE SLOWLY, TO TEST THE WATERS BEFORE
DIVING IN AND FINDING OUT THAT THIS WAS NO SWIMMING POOL - IT WAS A SHARKINFESTED MUDHOLE. A LESS PUBLICIZED ENTRY WOUID ALSO HAVE MADE IT EASIER TO
MAKE THE ULTIMATE DECISICN MORE EASILY - PULLING OUT OF THE AREA ENTIRELY.
NCW THAT'S NOT TO SAY, IMPLY, OR MEAN - IN ANY WAY - THAT A FEW OF THESE GIANT
CCMPANIES WILL NOT BE SUCCESSFUL IN THIS AREA. I AM COtNUKZW) THAT THEY WILL BUT IT'S NOT GOING TO BE EASY. AND SUCCESS FOR THEM IS GOING TO HAVE TO BE
MEASURED ULTIMATELY THE SAME WAY IT IS FOR ANY BUSINESS - IS THIS A BUSINESS
WHERE YOU CAN MAKE ENOUGH MCNEY TO JUSTIFY BEING IN IT?
I'VE CHOSEN THE RCBOTICS AREA .OF CIM AS MY EXAMPLE HERE BUT ALOT OF WHAT I'M
SAYirrc WILL BE TRUE FOR OTHER AREAS CF THE OVERALL CCMPUTER-INTEGRATED
MANUFACTURING INDUSTRY. IT IS A BRUTAL AREA. ALOT OF THINGS THAT MAY HAVE
HEID TRUE F O R OTHER INDUSTRIES DCN'T NECESSARILY HOLD TRUE HERE.
FCR CWE THING, IT IS QUITE TRUE THAT MICROPROCESSORS ARE WOK«!ING THEIR WAY INTO
THE FACTORY FLOOR - WHEIHER IN THE FORM OF PROGRAMMABLE CONTROLLERS, PERSONAL
COyiPUTERS, MINIS, PROCESS CCMTROLLERS, CNC MACHINE TOOLS, COyiPUTER AUTCMATED
MATERIALS HANDLING DEVICES, ROBOTS, CAD/CAM, ARTIFICIAL VISION SYSTEMS, OR
WHATEVER. BUT IT'S NOT THE SAME AS MICROPROCESSORS IN THE WORLD OF OFFICE
AUTOMATION OR DATA PROCESSING OR TELBCOyiMUNICATIONS. ON THE FACTORY FLOOR,
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
:.;-•"
-UV-'"' '." ^
'-"•1.._ .
.
;•• .
THERE IS GENERALLY A BASIC DISCCMFORT WITH CCMPUTERS AND SOFTWARE DESPITE THE
GRCWING REALIZATICN THAT THIS IS ALL PRETTY IXiUCH INEVITABLE, WHAT THIS
RESULTS IN IS THAT CUSTCMERS NEED AND WANT HANDS-<»I APPLICATICNS ENGINEERING AND
SUPPORT FRCM THE VENDOR. MANY OF THE CCMPANIES THAT HAD BEEN IN RCBOTICS FOR A
NUMBER O F YEARS AS WELL AS MANY THAT ENTERED THE INDUSTRY OILY WITHIN THE LAST
FEW YEARS REALLY FELT - AND HOPED BECAUSE IT'S ALOT EASIER - THAT THE BEST WAY
T O BE SUCCESSFUL IN THE INDUSTRY WAS GENERALLY STAY AWAY AS MUCH AS POSSIBLE
FRCM THE SYSTEMS END OF THE BUSINESS. IN FACT, A N OFFICER AT CNE OF THE GIANT
CCMPANIES THAT ENTERED THIS AREA A FEW YEARS AGO WAS THEN QUOTED AS SAYING THAT
THEY WERE HOPING THAT THEY COULD SELL RCBCTS PRETTY MUCH LIKE COOKIES. THAT HAS
HARDLY TURNED OUT TO BE THE CASE. A S AN ASIDE, IT'S KIND OF INTERESTING T O SEE
THAT MANY OF THE VEMXDRS WHO HAD SWORN THAT THEY WERE GOING T O STEER CLEAR OF
THE SYSTEMS END CF THE BUSINESS ARE NOW EMBRACING SYSTEMS INTBGRATICN AUySCST A S
A RELIGION.
OUR BOTTOyi LINE RIGHT NCW ABOUT THE AREA OF CCMPUTER-INTEGRATED MANUFACTURING IS
GENERALLY THIS - WE DO NOT BELIEVE THAT ANY CCMPANY IN THE AREA REALLY KNOWS YET
EXACTLY WHERE AND HOW IT IS GOING TO BE POSITICNED OVER THE LCNG TERM. T H E
MARKET IS STILL TOO DYNAMIC AND IN TOO MUCH FLUX - KIND OF A LEARNING PHA,SE WITH VERY FEW BRQAD-SCALE TRENDS EVIDENT.
THAT GENERALLY SUGGESTS THAT CCMPANIES THAT ARE GOING TO BE RESPONSIVE TO THE
MARKET HAD BETTER BE -
. :-^ •-•
!',•
.•:: ••••
(,••- :
-•t^^ -Z-nll ~\.-- -
O DYNAMIC O R ADAPTABLE, BECAUSE THEY'RE PROBABLY GOING TO HAVE TO MAKE
ALOT OF TSCTICAL CHANGES i M ) PROBABLY A FEW STRATEGIC CNES AS WELL
OVER THE NEXT FEW YEARS;
O QUICK, BECAUSE THEY HAD BETTER MAKE THOSE CHAtKSES FAST. IN THE
ROBOT INDUSTRY ALONE THERE ARE ABOUT 98 U.S.-BASED COMPANIES, 250+
JAPANESE BASED CCMPANIES AND SEVERAL DOZEN IN EUROPE. IN ARTIFICIAL
VISION, THERE ARE ALREADY 50 O R SO U.S.-BASED VISIOSI SYSTEMS
CCMPANIES.
6 WELL-MANSGED, BECAUSE WITH ALL THESE CHAISES GOING ON AiRGUND THEM
AND 'THE CNES THAT THEY'RE GOING TO HAVE TO MAKE IN THEIR OMN
ORGANIZATICNS, YOU CAN BE SURE THAT THINGS CAN START T O GET OUT OF
HAND PRETTY DAMN QUICK IF MANSGHyiENT ISN'T REALLY ON TOP OF THEM.
O MARKETING-DRIVEN IN PART, AND HEADED B Y ONE SUPER MARKETING PERSON
WHO CAN READ THE DYNAMICS OF "nffi MARKET, MOTIVATE HIS FORCES, AND
RESPOND.
<:>••
.?
o TBCHNQLOGY-DRIViM IN PART, BECAUSE THE INDUSTRY IS SO DAMNED
COMPETITIVE THAT WHEN THEY RESPOND, THM BETTER BE RESPONDING WITH
MC»E -THAN JUST HOT AIR. THIS ALL SUGGESTS THAT WHILE A CCMPANY MAY
B E SELLING PRODUCTS THAT ARE STATE-OF-THE-MARKET, IT BETTER B E
WORKING ON PRODUCTS THAT ARE STATE-OF-'THE-ART NOW.
o WELL-BACKED FINftNCIALLY - BUT WE'VE ALREADY DISCUSSED THE DIIHyJMA OF
BEING A N ENORMOUSLY WELL-FINANCED GIANT COMPANY BUT STILL NOT HAVING
INFINITE R E S O U R C E S TO THROW OFF INTO EVERY DIRECTION. AND, IN ORDER
T O GEH! T H O S E RESOURCES, YOU OFTEN HAVE TO TAKE YOUR CASE UP THE LADDER
FROM GROUP HEAD TO DIVISION HEAD TO CORPORATE AND SO ON - THROUGH
COMMITTEES. KIND OF A CIRCULAR DILEMMA WHEN YOU'RE ALSO TRYIWS TO B E
DYNAMIC AND QUICK.
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
THERE HAS NEVER BEEN AN IMPORTANT HlOi TBCH-ORIENTED AREA YET WHICH HAS BEEN SO
DCMINATED BY GIANT CCMPANIES THAT ALL THE START-UPS HAVE ESSENTIALLY BEEN
STIFLED, CUT OFF AT THE PASS, STEPPED CN, OR STRANGLED. PM!) IT'S NOT GOING TO
HAPPM THIS TIME EITHER. WE MAY END UP WITH A FEW GIMW CCMPANIES THAT HAVE
BECOME SUCCESSFUL BUT THAT WILL PROBABLY NOT BE THE NORM.
so LET'S FLIP THE COIN OVER FOR A MINUTE AND LOOK AT SOME OF THE START-UPS - THE
ONES FINANCED BY VENTURE CAPITALISTS TO THE TUNE OF OVER $100M IN THE ROBOT AND
VISION INDUSTRIES ALONE SO FAR. ON THE PLDS SIDE, THEY CERTAINLY HAVE THE
ABILITY T O ATTRACT AND HIRE SOME OF THE MOST AGO^ESSIVE AND SMARTEST INDIVIDUALS. 1HE BEST O F THESE CCMPANIES CAN BE VERY DYNAMIC, AGILE, AND QUICK. ON THE
MINUS SIDE, THEY OFTEN HAVE VERY LITTLE NAME RECOaSIITICN IN AN INDUSTRY
INHABITED BY NAMES LIKE IBM AND GE. ALSO, BECAUSE THE INDUSTRY IS SO
CCMPETITIVE, THERE'S VERY LITTLE ROCM FOR ERROR OR FOR LEARNING FRCM MISTAKES.
LE^T ME DIGRESS FOR A MOMENT TO SHOW YOU SOMETHIMS ABOUT IHESE START-UPS. YOJ
CAN JUDGE FOR YOURSELVES. PERHAPS IT POINTS UP JUST HOW AGGRESSIVE AND AMBITIOUS THEY CAN BE, HOW HU1SK3RY THEY ARE. BUT IT'S INTERESTUSKS IN ANY CASE,
I READ ALOT OF BUSINESS PLANS FOR SMALL START-UP CCSyiPANIES, ESPECIALLY THOSE
ENTERING ROBOTICS, IT'S NO EASY MATTER TO WRITE AN 80-100 PAGE BUSINESS PLAN
ESPECIALLY WHEN THE PURPOSE OF THE BUSINESS PLAN IS TO CONVINCE POTENTIAL
BACKERS THAT YOU HAVE A SOUND STRATEGY FOR SUCCESS - TO GET MONEY, SOMETIMES
SEVERAL MILLION DOLLARS, FROM FINANCIERS WHO HAVE SEEN IT AND HEARD IT BEFORE,
EVERY BUSINESS PLAN INCLUDES FORECASTS MADE BY THE COMPANY - HOW WELL THEY THINK
THEY CAN DO USING THE STRATEGY THEY'VE LAID OUT, ODDLY ENOUGH, NEARLY EVERY
BUSINESS PLAN I'VE READ SHOWS THE COMPANY GAINING 10% OR BETTER OF THE MARKET,
UNFORTUNATELY, I'VE READ OVER 50 BUSINESS PLANS - FOR A GRAND TOTAL FROM IHESE
COMPANIES ALONE OF 500% OF THE MARKET IN A GIVEN YEAR,
NOW I'VE GONE THROCOI THIS EXERCISE NOT TO DEMCNSTRATE MERELY HOW OUTLANDISH
KJSINESS PLANS CAN BE (ALTHOUGH SOME ARE OFTEN JUST THAT) BUT RATHER TO SHOW HOW
EXTREMELY AMBITIOUS AND AGGRESSIVE SOME OF THESE SMALLER CCMPANIES TEND TO BE,
THEY HAVE TO BE - REMEMBER, THEIR SURVIVAL DEPENDS CN THEIR RAPID GROWTH,
so THERE IS SCMETTHING TO BE SAID FOR BEING A SMALL COMPANY, STARTING WITH FEW
PRBCCNCEPTICNS, AND RUNNING LIKE HELL, LEST I EVEN HINT AT SUGGESTING THAT
BEING SMALL MAKES THE INDUSTRY OBSTACLES DISAPPEAR AND MEANS THAT YOU'VE GOT IT
MADE, LET'S TAKE AN IMAGINARY LOOK AT SOME OF THE OBSTACLES FROM THE POINT OF
VIEW OF A START-UP COMPANY - EARLY 1980 - YOU ENTER THE INDUSTRY FILLED WTIM GREAT IDEAS AND A
CERTAINTY THAT YOU'RE GOING TO CCNQUER THE WORLD OF CII4, YOU'RE
GOING T O BECOME THE IBM OF THE C m V30RLD,
- YOU CALL UP A FEW VENTURE CAPITALISTS - FRED ADLER, KLEINER PERKINS,
J.H, WHITNEY, O R WHOMEVER AND SCME ASSISTANT ASKS YOU FOR A BUSINESS
PLAN,
- YOU SAY - "A BUSINESS PLAN? WHILE I'M SITTING AND WRITING THIS
100-PAGE DOCUMENT, EVERYONE ELSE WILL BE GAINING GROUND," BUT,
STILL, THEY WANT A BUSINESS PLAN,
- so YOU SIT DOWN AND WRITE A BUSINESS PLAN, THEN YOU SEND IT TO
THE VENTURE CAPITALISTS WHO, AFTER ALL THAT FUSS, TREAT IT AS IF IT
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Bidder Park Drive / San Jose, CA 95131 / (408) 971-9000/ Teles* 171973
^^
WERE A PIECE OF FICTICN AND RIP IT APAJCT WHEN THEY COXIE TO VISIT
YOU. ANYWAY, YOU FINALLY GET FINANCED,
- EARLY 1981 - HERE YOO ARE, O^EAT IDEAS AND ALL, A DAMNED GOOD
MANAGEMENT TEAM, BEGINNING TO MAKE CCNTACT WITH THE MARKETPLACE
AND THE WHOLE BCCNOyiY RUNS RIGHT INTO A BRICK WALL AND DIES. YOUR
ENTIRE CUSTCMER BASE IS STILL TALKIM3 ABOUT PRCDUCTIVITY AND KANBAN
AND ROI BUT THEIR CASH FLCW IS GOING DCWN THE TUBES. YOU REALIZE
THAT THE CYCLE TIME FRCM INITIAL CCNTACT TO PURCHASE ORDER IS GOING
T O B E ABOUT IWICE AS LCNG AS YOU FORECAST IN YOUR BUSINESS PLAN.
• •
•^' ••
- ''
- UNIMATICN, THE LEADER IN THE INDUSTRY, FILES TO GO PUBLIC AT A
$115M Vi\LUATICN. THAT SOUNDS ENCOURAGING UNTIL EVERYOffi REALIZES
HOW BAD UNIMATICN'S MARGINS HAVE GOTTEN AND VENTURE CAPITALISTS
START TO ASK - "THIS IS THE INDUSTRY THAT WE PUT ALL THIS MCNEY
INTO?" EVERYONE STARTS TO TALK ABOUT THE "PROFITLESS PROSPERITY"
THAT WILL BE THE FATE OF THIS INDUSTRY.
• u'k
- -•
"'•
'• • •
- GENERAL ELECTRIC ENTERS THE CIM INDUSTRY. THEY ARE GOING TO OFFER
FACTORIES O F THE FUTURE. THEY NOW HAVE CAD/CAM, MICRDPRCCESSORS,
CNCs, PROOyyyMABLE CONTROLLERS, ROBOTS, VISION, NETWORKING AND SO
ON. AS IF THIS WEREN'T FORMIDABLE EN0UC2I, JACK WELSH - THEIR
CHAIRMAN - DRIVES HOME THE POINT BY SAYING THAT THIS IS GOING TO BE
A MAJOR AREA OF EMPHASIS FOR GE MD FOR HIM AS WELL.
- MORE AND MORE START-UPS ARE GETTING FINANCED. THE INDUSTRY NOW HAS
FAR MORE CCMPANIES IN IT THAN THERE ARE SALES TO GO AROUND.
- NATURALLY, YOUR FORECASTS HAVE VIRTUALLY FALLEN OUT OF BED WHAT WITH
THE RECESSION, THE NUMBER OF CCMPEH^ITORS i^ND "THE NORMAL RELUCTANCE
O F E N D U S E R S T O ADAPT TO NEW "TECHNOLOGIES. STILL, IT'S TIME FOR MORE
MONEY - TIME TO RAISE THE VALUATION, GIVE AWAY MORE OF THE COMPANY
AND SPEND THE NEXT MONTH OR SO PRACTICALLY IGNORING THE BUSINESS SO
YOU CAN MEET WITH VENTURE CAPITALISTS.
"•^' *
,.';)•.
- RECESSION STILL HOT AND HEAVY. STILL, YOU'VE MADE INROADS AT
GENERAL MCTORS, THE LARC2ST BUYER OF ROBOTS AND VISION SYSTEMS
PROBABLY T H R O U G H O U T THE DECADE.
- IHyi EOTERS THE INDUSTRY. SO MUCH FOR YOUR WANTING TO BE THE IBM OF
THE C3M WQRID.
- GETERAL MOTC^S ENTERS THE INDUSTRY - IN A JOINT VENTURE WITH FANUC,
A JAPANESE COMPMY VERY WELL KNOWN IN FACTORY AUTOMATION.
I C O U L D C O N T I N U E THIS FOR QUITE SOME TIME BUT I'M SURE THAT YOU GET THE
POINT. START-UPS NEVER HAVE AN EASY TIME AND, IN THIS INDUSTRY, IT'S
i"^'•",.-.. PARTICULARLY TOUGH BECAUSE THE INDUSTRY ITSELF IS SO TOUGH. BUT THE BEST OF
•;r^p: t.F.
THEM - AND ONLY THE VERY BEST - CAN BE VERY SUCCESSFUL BECAUSE THEY ARE QUICK,
DYNAMIC, AGGRESSIVE AND SOMEWHAT LIKE PERSISTENT, OBNOXIOUS INSECTS - YOU CAN
SPRAY AT THEM (KIND OF LIKE THE PERVASIVE, GLOBAL, I'M BIG SO I'LL-BE•Zi^^
_ > .
. EVERYTHING-TO-EVERYONE APPRQ?Oi TAKEN BY GIANT COMPANIES) BUT UNLESS YOU
'•'SrSX ?^: • ACTUALLY HIT THEM WTTH THE SPRAY CAN (UNI£SS LARGE CCMPANIES ACTUALLY PICK A
^ i 2 n r •,,. FOCUSED, NICHE TYPE CF APPROACH AND CONTINUALLY GO HEAD-TO-HEAD WITH THESE
START-UPS) , THE BEST OF THE START-UPS ARE HARD TO KILL.
NOW
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company / 1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
NCW LET'S TAKE A LOOK AT TWO OF THE SUBSBGMENT INDUSTRIES WITHIN CIM AND SEE
WHAT THIS MEANS M © WHERE THEY'RE GOING.
IN THE RCBOT OR RCBOI SYSTEMS SBOyENT, SALES IN 1982 DID NOT JUMP BY THE
FORECAST 35% OR SO FOR THE INDUSTRY - IHE RBCESSICN WAS JUST TOO BAD. BUT THEY
DID GO TO $190M, UP 23% OVER 1981 - NOT BAD FOR A PRETTY TERRIBLE YEAR.
THIS YEAR WE'RE EXPECTING SALES FOR THE INDUSTRY TO BE UP A3AIN BY LESS THAN 35%
- UP CNLY ABOUT 30% TO BEIWEEN $235-265M WITH A BIG JUMP OF ABOUT 40% OR SO IN
1984. BUT LET'S LOOK UNDER THE INDUSTRY-WIDE NUMBERS.
WHEN WE DO, WE SEE THAT THE ORIGINAL CCMPANIES - NAMES LIKE UNIMATICN,
CINCINNATI MHACRCN,, PRAB RCBOTS, ASEA, DETHBISS AND COPPERWELD - CCMPANIES
THAT ESSENTIALLY WERE THE INDUSTRY IN 1979, 1980 AND 1981 HAVE BEEN GIVING UP
MARKin? SHARE TO THE NEWER CCMPANIES. BY THE EMD OF 1983, THESE NEW CCMPANIES
COLLECTIVELY WILL HAVE ABOUT 48% CF THE MARKER, AND, IN 1984, ABOUT 55%.
IF WE LOOK EVEN FURTHER, IT IS CURIOUS TO SEE THAT THE START-UPS - CCMPANIES
THAT STARTED FRCM NOTHING JUST A COUPLE OF YEARS AGO AND WERE CLEARLY IN A
POSITION T O MAKE EVERY MISTAKE IN THE BOOK AND, EVEN IF THEY DITO'T THEY WERE
GOING T O G E T STEPPED CN BY THE GIANTS WHO E3SITERED THE INDUSTRY - START-UPS HAVE
GCNE FRCM A MARKET SHARE CF AIMOST NOTHING - ABOUT 2.3% IN 1980 TO AN ESTIMATED
14.4% IN 1983 AND COULD WELL GO OVER 20% OF THE TOTAL U.S. MARKET IN 1984.
GIANTS HAVE GCNE FRCM ABOUT 5.4% IN 1982 TO AN ESTIMATED 21.3% IN 1983 AND WILL
LIKELY BE OVER 30% BY 1984.
BUT IT'S BEEN THE START-UPS THAT HAVE SEI THE TCNE AND THE DIRBCTICN AND THE
PACE. THE SAME HAS BEEN TRUE IN THE ARTIFICIAL VISICN SYSTEMS INDUSTRY. IF WE
LOCK AT CNLY THAT SEGMENT OF THE VISICN INDUSTRY WHICH IS SELLING SYSTEMS THAT
ARE RELATIVELY STANDARDIZED IN GIVEN APPLICATIONS, AND WHICH DOES NOT INCLUDE
THOSE CCMPANIES SELLING IMAffi PROCESSING SYSTEMS OR SYSTEMS SINGULARLY DEDICATED
T O A PARTICULAR TASK OR PRETTY MUCH CUSTOM-BUILT AND ALSO DOESN'T INCLUDE SALES
OF VISICN-RELATED COMPONENTS, WE'VE STILL GOT MORE THAN 30 CCMPANIES. BUT C2QLY
2 OF THHyi ARE LARGE COMPANIES. THE REST ARE TINY. THIS INDUSTRY WASN'T EVEN A
SPECK I N 1981 VEm EXTERNAL SALES FOR THE INDUSTRY WERE ESTIMATED TO BE ABOUT
$8.5M. AND IT WASN'T MUCH MORE THZMSI A SPECK IN 1982, WITH $17.5M IN SALES.
THIS YEAR, IT'S STILL JUST A DOT OF AN INDUSTRY AT ABOUT $35M AND NEXT YEAR NOT
M U C H M O R E AT ABOUT $75M. BUT THIS WAS ALL DONE BY START-UPS. ONE HUNDRED
PERCENT GROWTH PER YEAR COMING FRCM TINY CCMPANIES IN AN INDUSTRY THAT'S NOT
CONSIDERED EVEN LEGITIMATE ENOUGH TO CALL IT AN INDUSTRY. IT WAS THE LITTLE
COMPANIES W H O RECOO^IZED THE POTENTIAL AND HAD TO ESSENTIALLY USE
TRIAL-AND-ERROR I N THEIR MARKETING EFFORTS BECAUSE USERS COULDN'T UNDERSTAND
WHAT IN THE WORLD THIS WAS ALL ABOUT. IT WAS THE TINY COMPANIES WHO WERE
PREPARED T O HOLD HANDS WITH ENDUSERS, SUPPORT CUSTOMERS, AND SPEND THE TIME
BECAUSE, TO THEM, THIS 100% REVENUE GROWTH COUNTED ALOT ALTHOUGH LARGER
COMPANIES GENERALLY CCmDN'T AFFORD TO DO THIS. TO THE GIANTS, IT WAS TOTALLY
INSIGNIFICANT. IN 1983, THESE CCMPANIES WILL COLLECTIVELY SELL i^BOUT 1000
ARTIFICIAL VISION SYSTEMS FOR INSPECTION, ADAPTIVE CCNTRQL, DIMENSIONAL CHECKING
AND GUIDANCE.
T O GIANTS, THIS MAY NOT MEAN MUCH - BUT SOMEWHERE OUT THERE, THERE ARE COMPANIES
THAT ARE GAINING A TREMENDOUS jyvCUNT OF VISIBILITY IN INDUSTRIES WHERE MANT OF
THE GIANT CCMPANIES - THOUGH CERTAINLY NOT ALL - ARE STILL SCRATCHING THEIR
HEADS AND TRYDK3 TO FIGURE OUT WHERE THE MARKET DISAPPEARED TO.
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000/ Telex 171973
WHILE MANY CF THE LITTLE CNES A R E RUNNING A S FAST A S THEY CAN JUST TRYING TO
STAY A STEP AHEAD CF THE SUPPOSED THUNDERING HERD OF GIANTS BEHIND THEM, MANY OF
•THE BIG ONES ARE STILL BUSY TRYING T O FORECAST THE INDUSTRY - AND W E ALL KNOW
-WHAT A THZ^JWLESS AND HOPELESS TASK THAT IS. SO MUCH OF THE DATA W E GET TENDS T O
B E RECYCLED DATA CCMING FRCM THE SAI'ffi SOURCES. IN FACT, T O DIGRESS A MOMENT, I
REMEMBER NOT TOO LCNG AGO, HAVING A CCNVERSATICN WITH THE VP OF CNE OF THE
OPERATING GROUPS A T A VERY LARGE COMPANY IN THE FACTORY AUTOXIATICN INDUSTRY. HE
ASKED M E WHAT M Y FORECASTS WERE FOR THE INDUSTRY. SO I GAVE HIM M Y NUMBERS FOR
RGBOTS, VISICN, CNC MACHINE TOOLS AND CAD/CAM AND THAT WAS THAT. A FEW WEEKS
LATER, I WAS SPEAKING T O CNE CF THE CORPORATE VPs A T THE SAME COMPANY. W E WERE
JUST CHATTING AND I HAPPENED T O MENTION T O HIM THAT I HAD HAD THIS OTHER CONVERSATION WITH THIS OPERATING V P , A T WHICH TIME H E SIMPLY STARTED T O LAUGH. I
-ASKED HIM WHAT WAS S O FUNNY AND H E SAID, "WELL, YOU'RE GOING TO B E OUT HERE
IN A FEW WmS
ON A NORMAL ANALYST'S VISIT AND, PROBABLY, A T THE END OF YOUR
VISIT, YOU'LL ASK U S , 'SO WHAT D O YOU THINK T H E INDUSTRY'S GOING T O B E DOING
OVER THE NEXT FEW YEARS?' WE'LL GIVE YOU OUR NUMBERS. AND YOU'LL WALK OUT
. SHAKING YOUR HEAD AND SAYIITC T O YOURSELF, ' I MUST B E REAL SMART. LCOK HOW CLOSE
- M Y NUMBERS CAME T O THEIR NUMBERS.'" SO MUCH FOR RECYCLED INFORMATION!
1 MENTIONED EARLIER THAT, IN MANY CASES, IT HAS BEEN THE START-UPS THAT HAVE SET
THE TONE, THE NEW DIRECTION AND THE PACE AND I SUPPOSE THAT'S TRUE FOR MOST HIGH
TECH AREAS. LET'S LOCK A T SOME OF THE TRENDS AND DIRECTIONS IN CIM THAT HAVE
BEEN GENERATED B Y T H E START-UPS s... i
cH'ijL^
.z.s;:3yS; A
--. V , qr
•/j_^-"£v
-.Vp' "•)q
O IN C O M P U T E R P R O C E S S I N G F O R T H E ENGINEERING MARKET, START-UPS HAVE
SUCCESSFULLY INTRODUCED AND MADE IMPORTANT A WHOLE NEW GENERATION OF
COMPUTERS - COMPUTER TECHNOLOGY BASED CN MODERN MICROPROCESSORS OR
32-BIT INTELLIGENT WORKSTATICNS. COMPANIES WITH NAMES LIKE APOILO,
SUN MICROSYSTEMS, THREE RIVERS AND, MASSCOMP, HAVE BEEN DEVELOPING
AND COMMERCIALIZING THIS TECHNOLOGY FOR THE CAD/CAM AND CAE
INDUSTRIES. T H E RESULT, CF COURSE, HAS BEEN T O DISTRIBUTE COMPUTER
POWER DOWN CNE MORE NOTCH AND MAKE REAL-TIME INTERACTIVE
CAPABILITIES AVAILABLE, FOR EXAMPLE, T O MANUFACTURING AND PRODUCTION
ENGINEERS.
s-
O IN MECHANICAL CAD/CAM, COMPANIES LIKE CADLIMZ HAVE DEVELOPED
SYSTEMS OUT CF THEIR OWN 32-BIT INTELLIGENT WORKSTATIONS.
.iAvV
o IN ELECTRONIC DESIGN AUTOMATION, SINGLE USER FUNCTIONALITY COyiNG
FROM 32-BIT INTELLIGENT WORKSTATIONS HAS BEEN COMyERCIALIZED BY
COMPANIES SUCH A S DAISY SYSTEMS, MENTOR GRAPHICS, VALID LOGIC,
MBTHEUS, ETC.
O WATCHING -THESE TRENDS, A FEW CF THE LARGER CAD/CAM COMPANIES, LIKE
CCMPUTERVISION, AUTO-TROL, Mm CAD-IA A R E NOW DEVELOPING PRODUCTS IN
•THIS AREA.
J:^J1
O IN RCBOTIC:S, START-UPS WITH NAMES LIKE AUTOMATIX AND ADVANCED
R O B O T I C S RECOGNIZED RIGHT FROM "THE START "THAT THE CAM INDUSTRY WAS
SIMILAR T O T H E CAD INDUSTRY A T LEAST IN CNE RESPECT: CUSTOMERS NEED
M © WANT SUPPORT FRCM THE VENDOR. A S I MENTIONED EARLIER, OVER
TIME, NEARLY A L L THE VENDORS - LARGE AND SMALL - BEGAN T O REALIZE
THAT, WHETHER "THEY LIKED IT OR NOT, THIS WAS PRCBABLY TRUE, AND THEY
ALSO BEGAN T O PRCMCriE -niEMSELVES AS BEING IN THE TURNKEY SYSTEMS
. BUSINESS ?ND OFFERING COMPLETE CUSTOMER SUPPORT.
Dataque^yncorporatedi A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / Sjin ^ s e , CA 95131 / (408) 971-9000 / Telex 171973
o ARTIFICIAL VISICN IS PERHAPS THE MOST AMAZING AND-TELLiNiS OF ALL. I
TOJCHED CN IT BEFOi^ 'HJT, -AGAIN^'THERE SIMPLY WAS NO INDUSTRY.
THERE WAS A "TECHNOLOGY" - THE TBCHNbLOGY OF SENSING IMAGES»
DIGmZING THEM, PROCESSiMS AND ANALYZING THH4' A ^ MAKTNG''DECISIONS
BASED ON THAT ANALYSIS - BOT THERE WERE NO si>BC#iC' PRiCDtlGTS AIMED
AT DOING SPECIFIC INSPECTICN, ADAPTIVE CCNTIOLi, DIMENSI<mL
CHECKING, O R GUIDANCE TASKS, EXCEPT IN VERY LIMTTED ARE2&.
POTENTIAL ENDUSERS DIEN'T KNCW WHAT IHEY WANTED AND WEREN'T PREPARED
T O WASTE MCNTHS AND MCNTHS FIOJRIISKS IT OJT giNCi'THERE WERE
ALTERNATIVES. THE MARKET WAS ALSO EXTREMELY AMORPHOUS -'fcCND CF
LIKE GETTTING YOUR HAIOS CN A LARGE JELLYFISH - BiE^MISE-'tT WAS
POTENTIALLY EVERYWHERE AND YET IT WAS, AT THE 'SAME "EEME,- REALLY
NOWHERE,
I T WAS THE START-UPS VfflO WENT THROUOi THEIR'OWN LEARNING
CURVE TRYING TO FIGURE OUT HOT TO MARKET INTO THIS'ISOT^INDUSTRY,
WHAT COULD REALLY BE SOLD TO THIS NCN-INDUSTRY, HOT IT'SHOULD BE
PACKAGED AND SOLD, AND THEN REALLY CI^EATEDM"INDUSTRY.-' 'THUS, WE HAVE
NAMES LIKE AUTCMATIX, COGNEX, CCNTRQL AUTCMATICN, ITRAN; 'PERCEPTRCN AND
OTHERS THAT ARE MAKING A BUSINESS OUT CF THIS.- THTS^'HASN'T GONE
UNNOTICED BY THE BIG CCMPMIIES. GENERAL ELECTRIC, FOR C9SIE, HAS NOW
PICKED UP IT'S CWN PACE IN THE INDUSTRY. OTHERS LIKE'OTEJIS-ILLINOIS AND
CHESEBROUGH PCNDS, HAVE TAKEN INTERNALLY DEVELOPED ja© USED TBCBNQLOGIES,
AND HAVE BBCUN TO CCMMERCIALIZE THEM, WITH OTHER GIANTS DEVELOPING THEIR
PRODUCTS IN T H E WINGS.
-^'i"
DOES THIS MEAN THAT ALL OF THESE INDUSTRIES ARE REALLY DESTINED TO BELONG TO
START-UP COMPANIES AND THAT WE CAN ESSENTIALLY IGNORE THE GIANTS? OBVIOUSLY,
THE ANSWER IS NO. MOST OF THE START-UPS WILL NEVER MAKE IT IN ANY IMPORTANT
WAY. MEANWHILE, THAT JOINT VENTURE THAT I MENTIONED EARLIER BETWEEN GENERAL
MOTORS ^ M D FANUC, T H E COMPANY CALLED C31F, IS PROBABLY GOING TO DO ABOUT $22-25M
W O R T H O F ROBOTIC BUSINESS THIS YEAR AND OVER $60M NEXT YEAR. HARDLY A SNEEZE.
A LITTLE COMPANY CALLED IBM - THE SAME ONE WHERE AN INDEPENDENT BUSINESS UNIT
SPAWNED OFF A PERSONAL COMPUTER BUSINESS THAT'S SETTI^K5 THE PACE FOR AN ENTIRE
INDUSTRY - HAS ANOIHER INDEPENDENT BUSINESS UNIT IN THE ADVANCED AUTCMATIOJ
SYSTEMS AREA THAT JUST BOOKED A $5M+ OWER FOR A SOHJISTICATED ROBOTIC ASSEMBLY
SYSTEM FOR AN ELECTRONICS COMPANY. AGAIN, HARDLY A SNEEZE SINCE THIS IS POTENTIALLY THE MOST EXCITHTC SEXSffiNT IN THE ROBOTICS AREA. AND TEIAT COMPANY THAT'S
BRINGING G O O D THINGS TO LIFE, GE, THEY HAVE SO MANY PIECES - CAD/CAM, ROHOTICS,
VISION, CNC, SOFTWARE, NETWORKING, A VERY AGGRESSIVE, COySffllTTED CEO, AND AN
UNBELIEVABLE DISTRIBUTION NETWORK IN THE INDUSTRIAL MARKET - THAT WE'D BE FOOLS
T O IQ^ORE IT.
STILL, I REALLY BELIEVE IHAT MOST CF THE LARGE COMPANIES, LIKE
MOST OF THE SMALL COMPANIES, WILL NEVER MAKE IT IN CIM IN ANY IMPORTANT WAY.
BUT THE ONES THAT DO MAKE IT - THE LARGE ONES AND THE SMALL ONES (AND IT'S NOT
REALLY HARD TO FIGURE OUT WHICH START-UPS HAVE A GOOD SHOT AT BEIMS SUCCESS
STORIES) WILL HAVE AN INCREDIBLY ATTRACTIVE INDUSTRY TO SELL IN AND 0?OT IN
BECAUSE FACTORY AUTOyiATION CR CCMPUTESl-INTBGRATED-MANUFACTURING IS ONE OF THE
GIANT AND IMPORTANT INDUSTRIES OF THE 1980s.
FINALLY, A STORY TO DEMCNSTRATE HOT PERVASIVE THE POWER OF A LITTLE START-UP CAN
BE. ABOUT 2 1/2 YEARS AGO, I WAS TALKING TO SOMEONE AT A VERY LARGE AUTaySOTIVE
COMPANY WHICH ALSO HAPPENS TO HAVE ALREADY OPENLY AEMETTED THAT THEY WILL BE AN
ENORMOUS BUYER OF CIM TECHNOLOGY OVER THE DECADE. HE WAS ASKING ME ABOUT A
PARTICULAR START-UP WHICH HM) THEN BEEN IN BUSINESS FOR ABOUT 6 MONTHS AND T-JHICH
HAD RECEIVED ALOT OF PUBLICITY ALREADY BECAUSE CF THEIR'HIGH-Pa»?ERED MANAGEMENT
TEAM. WE'LL CALL' IT COMPANY X. HE ASKED ME, "WHO IN THE WORLD IS THIS COMPANY
Dataquest Incorporated, A Subsidiary of A.'C. Niblsen Company /1290 Bidder Park Drive / San Jose, CA 9^131/(408) 97'i-^OOO'/ telei("171^7i-
0
X? WHY IS IT THAT THEY'RE BEATING CN OUR DOQR CCNSTANTLY THINKING THEY'RE GOIN(
T O GET OUR BUSINESS? WE ALREADY HAVE A LIST OF VEMX3FS THAT WE'VE BEEN DOING
BUSINESS WITH FOR YEARS AND WHO KNOW US WELL. THE PEOPLE AT COVIPANY X PROBABLY
KNOW A BIG, FAT ZERO ABOUT THE FACTORY FLOCB.. VMHE I'LL ADMIT THAT THEY'RE
PROBABLY BRILLIANT BASED CN THEIR BACRCaiOUNDS - IN FACT, I WOULDN'T MIND IF OSIE
CF THEM C O U L D SCMEHOW TAKE THE CQLLBC2:-ENTRANCE SATs FOR MY SCN CN SATURDAY THERE'S JUST NO WAY WE'RE GOING TO TURN THEM LOOSE HERE." ABOUT 6 MONTHS AFTER
THAT CCNVERSATICN, WE TALKED AGAIN. BY THEtH, ALCT CF THE GIANT CCMPANIES HAD
BEGUN T O E N T E R THE INDUSTRY. I ASKED HIM WHICH CF THE GIANTS HE HAD ALREADY
BEGUN T O DO BUSINESS WITH. HE MENTIONED CNE CR TWO CF T H m AND ADDED THAT THEY
HAD ALSO BEGUN TO DO SOME BUSINESS WITH CCMPANY X, THIS PEANUT START-UP. HE
NOTED THAT CCtffi»ANY X MAY WJS HAVE KNOWN TCNS ABOUT THE FACTORY FLOOR CCMING IN
BUT THEY WERE WILLING TO LISTEN HARDER THAN ANY CF THE CCMPANIES HE HAD ALWAYS
BEEN DEALING WITH. ITOW, A COUPLE CF YEARS LATER, I SHOULD TELL YOU THAT
CCMPANY X IS CNE CF THE MAJOR SUPPLIERS IN ITS AREA OF BUSINESS TO THIS AUTO
CCMPiVNY. THE EXECUTIVE FROM THE AUTO COMPANY IS STILL SHAKING HIS HEAD IN
i^lAZEMENT AT THAT CNE.
AS FOR ME, I LOOK AT START-UPS ALL THE TIME. I "TRY TO EVALUATE THEM BASED ON
MANftSEMENT DEPTH, TBCHNQLOGICAL STRENGTH, MARKETING, THE STRENGTH AND POTENTIAL
CF THE INDUSTRY THEY'RE IN, FINANCIAL BACKING, AND SO ON. STILL, I THINK IT'S
m INTANGIBLE IHAT SEPARATES OUT -THE REAL WINNERS. I CALL IT "SOUL." I DON'T
REALLY EVEN KNOW WHAT IT MEANS BUT I CAN TELL YOU THAT THOSE THAT HAVE IT ARE IN
A POSITION TO POTENTIALLY, IN THEIR INDUSTRY, SHOOT FOR "THE MDCN.
Dataquest Incorporated, A Subsidiary of A.C. Nielsen Company /1290 Ridder Park Drive / San Jose, CA 95131 / (408) 971-9000 / Telex 171973
•
^UJ -.A 1
:.*^5j«J,,.
•'J^--.-^^IK
^*:^ •
(::>aDataquest
c?iv;-j.4i.>.,n:-.| t.-'T'Sup;^:
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF

advertisement