Analog Dialogue 31-1

Analog Dialogue 31-1
Editor’s Notes
http://www.year2000.com
That’s one of the sites where you can
read about a major problem we
residents of Earth will face as we
celebrate the rolling over of the digits
after 23:59:59 on December 31, 1999.
The problem is that in many cases
there aren’t enough digits to roll over.
Unless something is done about it,
many machines and programs will tell
you that your age is suddenly
–(100 – y), where y is the number of years you’ve actually lived. If
you’re 105, you may get a computer letter calling on you to register
for kindergarten.
Most of you have probably read about the inability of many computer
programs to distinguish between the years 1900 and 2000, because
the year is expressed by only 2 digits. Thus, the year after 1999 of the
common era (i.e., year 99 anno moduli—AM) becomes 1900 (i.e.,
year 00 AM), neatly compressing time into a never-ending series of
identical centuries. No doubt you’ve also read about the many real
problems it will cause, the catastrophic expense involved in fixing
whatever portion of it can be found and fixed, the general indifference
to the problem, and the shortness of time and paucity of personnel
available to work on it.
Where did it all start? Probably back in the first quarter of this century
when Hollerith’s punched cards (first used more than 100 years ago)
started to become widely used for storing sortable data. With only 80
characters available, reducing the year to two characters seemed like
a worthwhile tradeoff to obtain two additional characters for important
data. Remember, this was a time when nearly all printed forms, such
as bank checks, had a date line 19__ (they still do!?), since it was
likely to be valid throughout one’s lifetime (or at least, one’s watch
interval). Scarcely a thought was given to the next century.
It’s interesting to speculate on the role played in the origins of this
problem by the fact that the twentieth century is a pre-millenial
century, i.e., that to fairly take into account dates far into the future,
all four digits have to be allowed to roll over. What if this were a
century somewhere in the middle (say, the sixteenth or forty-third)?
Then for many, many generations, the maximum century rollover would
be a single digit. It’s conceivable that more software designers would be
willing to take the long view when only a single digit had to be advanced
at some distant time within the next 99 years.
Much of the discussion in the popular press has to do with the moreobvious implications, relating to billing and inventory, tax and interest
computations, demographics. There are also problems in routines
related to safety, such as inspection intervals. While many unfixed
problems will show up simultaneously, many more, lying dormant,
will show up sporadically, or even by the absence of some important
event. Faulty computer prog rams may lie embedded in
microcomputers and microcontrollers in transportation, medical,
military, and environmental equipment, perhaps designed by engineers
now retired, and manufactured by companies no longer in existence.
It’s a hairy problem, one that it behooves us to learn more about. A
good place to start is at the above Web address, which is co-sponsored
by Peter de Jager, probably the leading consultant on the problem.
Another well-reasoned discussion can be found at http://www.bcs.uk/
b
news/millen/millen.htm#wha.
[email protected]
2
THE AUTHORS
David Skolnick (page 3) is a
Technical Writer with ADI’s
Computer Products Division,
Norwood, MA, where he writes and
illustrates manuals and data sheets
and has written or edited a number
of DSP application notes. He holds
a Bachelor degree in Electrical
Engineering Technology and is
completing a Master of Technical
and Professional Writing program at Northeastern U. He enjoys
gardening, outdoor sports, and anything to do with his kids.
Noam Levine (page 3) is a Product
Manager in the 16-bit DSP product
line, working on new-product
definitions and fixed-point DSP
applications. He holds a BSEE from
Boston University and an MSEE in
DSP from Northeastern University.
He has authored several application
notes, technical articles, and
conference papers. When not in the
digital domain, Noam can be found playing jazz saxophone and
working on his SCCA competition driving license.
Mike Byrne (page 10) is Applications Manager with the Applications
Group in Limerick, Ireland. The
group provides customer design-in
support and are involved with newproduct definitions. He holds a BSc.
in electronic systems from the
University of Limerick. A patentholder, he is the author of several
application notes and technical
articles. His leisure-time activities include music and outdoor sports.
Ken Kavanagh (page 10) is a
Technician with the Applications
Group in Limerick, Ireland, where
he provides customer support and
is involved with the design of
evaluation boards and support
software. He holds a Diploma in
Electronics and is currently
undertaking a part-time degree
course in the University of Limerick.
His leisure-time activities include music, tennis, and badminton.
[More authors on page 23]
Cover: The cover illustration was designed and executed by
Shelley Miles, of Design Encounters, Hingham MA.
One Technology Way, P.O. Box 9106, Norwood, MA 02062-9106
Published by Analog Devices, Inc. and available at no charge to engineers and
scientists who use or think about I.C. or discrete analog, conversion, data handling
and DSP circuits and systems. Correspondence is welcome and should be addressed
to Editor, Analog Dialogue, at the above address. Analog Devices, Inc., has
representatives, sales offices, and distributors throughout the world. Our web site is
http://www.analog.com/. For information regarding our products and their
applications, you are invited to use the enclosed reply card, write to the above address,
or phone 617-937-1428, 1-800-262-5643 (U.S.A. only) or fax 617-821-4273.
ISSN 0161–3626 ©Analog Devices, Inc. 1997
Analog Dialogue 31-1 (1997)
WhyUseDSP?
response shown in Figure 1, would have the following
characteristics:
Digital Signal Processing 101—
An introductory course in DSP
system design: Part 1:
• a response within the passband that is completely flat with zero
phase shift
• infinite attenuation in the stopband.
Having heard a lot about digital signal processing (DSP)
technology, you may have wanted to find out what can be done
with DSP, investigate why DSP is preferred to analog circuitry for
many types of operations, and discover how to learn enough to
design your own DSP system. This article, the first of a series, is
an opportunity to take a substantial first step towards finding
answers to your questions. This series is an introduction to DSP
topics from the point of view of analog system designers seeking
additional tools for handling analog signals. Designers reading this
series can learn about the possibilities of DSP to deal with analog
signals and where to find additional sources of information and
assistance.
What is [a] DSP? In brief , DSPs are processors or
microcomputers whose hardware, software, and instruction sets
are optimized for high-speed numeric processing applications—
an essential for processing digital data representing analog signals
in real time. What a DSP does is straightforward. When acting as a
digital filter, for example, the DSP receives digital values based on
samples of a signal, calculates the results of a filter function
operating on these values, and provides digital values that represent
the filter output; it can also provide system control signals based
on properties of these values. The DSP’s high-speed arithmetic
and logical hardware is programmed to rapidly execute algorithms
modelling the filter transformation.
The combination of design elements—arithmetic operators,
memory handling, instruction set, parallelism, data addressing—
that provide this ability forms the key difference between DSPs
and other kinds of processors. Understanding the relationship
between real-time signals and DSP calculation speed provides some
background on just how special this combination is. The real-time
signal comes to the DSP as a train of individual samples from an
analog-to-digital converter (ADC). To do filtering in real-time,
the DSP must complete all the calculations and operations required
for processing each sample (usually updating a process involving
many previous samples) before the next sample arrives. To perform
high-order filtering of real-world signals having significant
frequency content calls for really fast processors.
WHY USE A DSP?
To get an idea of the type of calculations a DSP does and get an
idea of how an analog circuit compares with a DSP system, one
could compare the two systems in terms of a filter function. The
familiar analog filter uses resistors, capacitors, inductors, amplifiers.
It can be cheap and easy to assemble, but difficult to calibrate,
modify, and maintain—a difficulty that increases exponentially with
filter order. For many purposes, one can more easily design, modify,
and depend on filters using a DSP because the filter function on
the DSP is software-based, flexible, and repeatable. Further, to
create flexibly adjustable filters with higher-order response requires
only software modifications, with no additional hardware—unlike
purely analog circuits. An ideal bandpass filter, with the frequency
Analog Dialogue 31-1 (1997)
As Figure 1 shows, an analog approach using second-order filters
would require quite a few staggered high-Q sections; the difficulty
of tuning and adjusting it can be imagined.
PASSBAND
IDEAL
RESPONSE
IDEAL
RESPONSE
|H(f)|
by David Skolnick and Noam Levine
Useful additions would include:
• passband tuning and width control
• stopband rolloff control.
ROLLOFF
ROLLOFF
2ND
ORDER
2ND
ORDER
STOPBAND
f0
STOPBAND
f
Figure 1. An ideal bandpass filter and second-order
approximations.
With DSP software, there are two basic approaches to filter design:
finite impulse response (FIR) and infinite impulse response (IIR).
The FIR filter’s time response to an impulse is the straightforward
weighted sum of the present and a finite number of previous input
samples. Having no feedback, its response to a given sample ends
when the sample reaches the “end of the line” (Figure 2). An FIR
filter’s frequency response has no poles, only zeros. The IIR filter,
by comparison, is called infinite because it is a recursive function:
its output is a weighted sum of inputs and outputs. Since it is
recursive, its response can continue indefinitely. An IIR filter
frequency response has both poles and zeros.
IN THIS ISSUE
Volume 31, Number 1,1997,24 Pages
Editor’s Notes, Authors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
Digital signal processing 101—an introductory course in DSP system design: I . . 3
Selecting mixed-signal components for digital communications systems—III . . . . 7
Controller board system allows for easy evaluation of general-purpose converters . . 10
Build a smart analog process-instrument transmitter with
low-power converters & microcontroller . . . . . . . . . . . . . . . . . . . . . . 13
New-Product Briefs:
Amplifiers, Buffered Switches and Multiplexers . . . . . . . . . . . . . . . 16
A/D and D/A Converters, Volume Controls . . . . . . . . . . . . . . . . . . 17
Power Management, Supervisory Circuits . . . . . . . . . . . . . . . . . . . 18
Mixed bag: Communications, Video, DSP . . . . . . . . . . . . . . . . . . . 19
Ask The Applications Engineer—24: Resistance . . . . . . . . . . . . . . . . . 20
Worth Reading, More authors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
3
z –1
INPUT
(n–1)
(n–N+2)
z –1
a(0)
a(1)
z –1
a(N–2)
(n–N+1)
a(N–1)
∑
FIR
STRUCTURE
OUTPUT
SAMPLING REAL-WORLD SIGNALS
N–1
y(n) =
∑a(k)
(n–k)
k=0
a(0)
x(n)
z –1
z
y(n)
a(1)
b(1)
IIR
FILTER
–1
a(2)
y(n) =
z –1
z –1
b(2)
N–1
M
k=0
k=1
∑a(k) x (n–k) + ∑b(k) y (n–k)
Figure 2. Filter equations and delay-line representation.
The xs are the input samples, ys are the output samples, as are
input sample weightings, and bs are output sample weightings. n
is the present sample time, and M and N are the number of samples
programmed (the filter’s order). Note that the arithmetic operations
indicated for both types are simply sums and products—in
potentially great number. In fact, multiply-and-add is the case for
many DSP algorithms that represent mathematical operations of
great sophistication and complexity.
Approximating an ideal filter consists of applying a transfer function
with appropriate coefficients and a high enough order, or number
of taps (considering the train of input samples as a tapped delay
line). Figure 3 shows the response of a 90-tap FIR filter compared
with sharp-cutoff Chebyshev filters of various orders. The 90-tap
example suggests how close the filter can come to approximating
an ideal filter. Within a DSP system, programming a 90-tap FIR
filter—like the one in Figure 3—is not a difficult task. By
comparison, it would not be cost-effective to attempt this level of
approximation with a purely analog circuit. Another crucial point
in favor of using a DSP to approximate the ideal filter is long-term
stability. With an FIR (or an IIR having sufficient resolution to
avoid truncation-error buildup), the programmable DSP achieves
the same response, time after time. Purely analog filter responses
of high order are less stable with time.
0
90-TAP FIR
RELATIVE MAGNITUDE – dB
PASSBAND CUTOFF
FREQUENCY 0.5fS
–25
2
Real-world phenomena are analog—the continuously changing
energy levels of physical processes like sound, light, heat, electricity,
magnetism. A transducer converts these levels into manageable
electrical voltage and current signals, and an ADC samples and
converts these signals to digital for processing. The conversion
rate, or sampling frequency, of the ADC is critically important in
digital processing of real-world signals.
This sampling rate is determined by the amount of signal
information that is needed for processing the signals adequately
for a given application. In order for an ADC to provide enough
samples to accurately describe the real-world signal, the sampling
rate must be at least twice the highest-frequency component of
the analog signal. For example, to accurately describe an audio
signal containing frequencies up to 20 kHz, the ADC must sample
the signal at a minimum of 40 kHz. Since arriving signals can easily
contain component frequencies above 20 kHz (including noise),
they must be removed before sampling by feeding the signal
through a low-pass filter ahead of the ADC. This filter, known as
an anti-aliasing filter, is intended to remove the frequencies above
20 kHz that could corrupt the converted signal.
However, the anti-aliasing filter has a finite frequency rolloff, so
additional bandwidth must be provided for the filter’s transition
band. For example, with an input signal bandwidth of 20 kHz,
one might allow 2 to 4 kHz of extra bandwidth.
1/2 LSB
NOISE
20 24
FREQUENCY – kHz
–50
48
4
6
–75
–100
–125
0
f
0.1
0.2
0.3
0.4
0.5 fS
NORMALIZED FREQUENCY =
(ACTUAL FREQ) / (SAMPLING FREQ)
Figure 3. 90-tap FIR filter response compared with those of
sharp cutoff Chebyshev filters.
Mathematical transform theory and practice are the core
requirement for creating DSP applications and understanding their
4
limits. This article series walks through a few signal-analysis and processing examples to introduce DSP concepts. The series also
provides references to texts for further study and identifies software
tools that ease the development of signal-processing software.
Figure 4. Antialiasing filter ideal response.
Figure 4 depicts the filter needed to reject any signals with
frequencies above half of a 48-kHz sampling rate. Rejection means
attenuation to less than 1/2 least-significant bit (LSB) of the ADC’s
resolution. One way to achieve this level of rejection without a
highly sophisticated analog filter is to use an oversampling converter,
such as a sigma-delta ADC. It typically obtains low-resolution (e.g.,
1-bit) samples at megahertz rates—much faster than twice the
highest frequency component—greatly easing the requirement for
the analog filter ahead of the converter. An internal digital filter
(DSP at work!) restores the required resolution and frequency
response. For many applications, oversampling converters reduce
system design effort and cost.
Analog Dialogue 31-1 (1997)
PROCESSING REAL-WORLD SIGNALS
The ADC sampling rate depends on the bandwidth of the analog
signal being sampled. This sampling rate sets the pace at which
samples are available for processing. Once the system bandwidth
requirements have established the A/D converter sampling rate,
the designer can begin to explore the speed requirements of the
DSP processor.
Processing speed at a required sample rate is influenced by
algorithm complexity. As a rule, the DSP needs to finish all
operations relating to the first sample before receiving the second
sample. The time between samples is the time budget for the DSP
to perform all processing tasks. For the audio example, a 48-kHz
sampling rate corresponds to a 20.833-µs sampling interval. Figure
5 relates the analog signal and digital sampling rate.
ANALOG
t
t
LOWPASS FILTER
X(N)
H(Z)
Y(N)
ADC
DAC
Figure 6. Example of continuous processing of samples
in digital filter.
Frame-based systems, like a spectrum analyzer, which determines
the frequency components of a time-varying waveform, acquire a
frame (or block of samples). Processing occurs on the entire frame of
data and results in a frame of transformed data, as shown in Figure 7.
SIGNAL
ANALYSIS
DSP
SYSTEM
SAMPLING
INTERVAL
DIGITAL SAMPLE TIMES
Figure 5. Sampling train and processing time.
Next consider the relation between the speed of the DSP and
complexity of the algorithm (the software containing the transform
or other set of numeric operations). Complex algorithms require
more processing tasks. Because the time between samples is fixed,
the higher complexity calls for faster processing.
For example, suppose that the algorithm requires 50 processing
operations to be performed between samples. Using the previous
example’s 48-kHz sampling rate (20.833-µs sampling interval),
one can calculate the minimum required DSP processor speed, in
millions of operations per second (MOPS) as follows:
DSP Speed =
Operations
50
=
= 2.4 MOPS
Sampling Interval 20.833 µs
Thus if all of the time between samples is available for operations
to implement the algorithm, a processor with a performance level
of 2.4 MOPS is required. Note that the two common ratings for
DSPs, based on operations per second (MOPS) and instructions
per second (MIPS), are not the same. A processor with a 10-MIPS
rating that can perform 8 operations per instruction has basically
the same performance as a faster processor with a 40 MIPS rating
that can only perform 2 operations per instruction.
SAMPLING VARIOUS REAL-WORLD SIGNALS
There are two basic ways to acquire data, either one sample at a
time or one frame at a time (continuous processing vs. batch
processing). Sample-based systems, like a digital filter, acquire data
one sample at a time. As shown in Figure 6, at each tick of the
clock, a sample comes into the system and a processed sample is
output. The output waveform develops continuously.
Analog Dialogue 31-1 (1997)
TIME
FREQUENCY
Figure 7. Example of batch processing of a block of data.
For an audio sampling rate of 48 kHz, a processor working on a
frame of 1024 samples has a frame acquisition interval of 21.33 ms
(i.e., 1024 × 20.833 µs = 21.33 ms). Here the DSP has 21.33 ms
to complete all the required processing tasks for that frame of data.
If the system handles signals in real time, it must not lose any
data; so while the DSP is processing the first frame, it must also
be acquiring the second frame. Acquiring the data is one area where
special architectural features of DSPs come into play: Seamless
data acquisition is facilitated by a processor’s flexible dataaddressing capabilities in conjunction with its direct memoryaccessing (DMA) channels.
RESPONDING TO REAL-WORLD SIGNALS
One cannot assume that all the time between samples is available
for the execution of processing instructions. In reality, time must
be budgeted for the processor to respond to external devices,
controlling the flow of data in and out. Typically, an external device
(such as an ADC) signals the processor using an interrupt. The
DSP’s response time to that interrupt, or interrupt latency, directly
influences how much time remains for actual signal processing.
Interrupt latency (response delay) depends on several factors; the
most dominant is the DSP architecture’s instruction pipelining.
An instruction pipeline consists of the number of instruction cycles
that occur between the time an interrupt is received and the time
that program execution resumes. More pipeline levels in a DSP
result in longer interrupt latency. For example, if a processor has a
20-ns cycle time and requires 10 cycles to respond to an interrupt,
200 ns elapse before it executes any signal-processing instructions.
When data is acquired one sample at a time, this 200-ns overhead
will not hurt if the DSP finishes the processing of each sample
before the next arrives. When data is acquired sample-by-sample
while processing a frame at a time, however, an interrupted system
wastes processor instruction cycles. For example, a system with a
5
200-ns interrupt response time running a frame-based algorithm,
such as the FFT, with a frame size of 1024 samples, would require
204.8 µs of overhead. That amounts to more than 10,000
instruction cycles wasted to latency—productive time when the
DSP could be performing signal processing. This waste is easy to
avoid in DSPs having architectural features such as DMA and
dual memory access; they let the DSP receive and store data
without interrupting the processor.
DEVELOPING A DSP SYSTEM
Having discussed the role of the processor, the ADC, the antialiasing filter, and the timing relationships between these
components, it is time to look at a complete DSP system. Figure 8
shows the building blocks of a typical DSP system that could be
used for data acquisition and control.
ANALOG INPUT
ANTI-ALIASING
FILTER
ADC
DSP
DAC
HOST
INTERFACE
ANTI-IMAGING
FILTER
DIGITAL I/O
ANALOG OUTPUT
Figure 8. Putting together elements of a DSP system.
Note how few components make up the DSP system, because so
much of the system’s functionality comes from the programmable
DSP. Converters funnel data into and out of the DSP; the ADC
timing is controlled by a precise sampling clock.To simplify system
design, many converter devices available today combine some or
all of the following: an A/D converter, a D/A converter, a sampling
clock, and filters for anti-aliasing and anti-imaging. The clock
oscillator in these types of I/O components is separately controlled
by an external crystal. Here are some important points about the
data flow in this sort of DSP system:
Analog Input: The analog signal is appropriately band-limited
by the anti-aliasing filter and applied to the input of the ADC. At
the selected sampling time, the converter interrupts the DSP
processor and makes the digital sample available. The choice
between serial and parallel interfacing between the ADC and DSP
depends on the amount of data, design complexity trade-offs, space,
power, and price.
Digital Signal Processing: The incoming data is handled by the
DSP’s algorithm software. When the processor completes the
required calculations, it sends the result to the DAC. Because the
signal processing is programmable, considerable flexibility is
available in handling the data and improving system performance
with incremental programming adjustments.
Analog Output: The DAC converts the DSP’s output into the
desired analog output at the next sample clock. The converter’s
output is smoothed by a low-pass, anti-imaging filter (also called a
reconstruction filter), to produce the reconstructed analog signal.
Host Interface: An optional host interface lets the DSP
communicate with external systems, sending and receiving data
and control information.
6
REVIEW AND PREVIEW
The goal of this article has been to provide an overview of major
DSP design concepts and explain some of the reasons why a DSP
is better suited that analog circuitry for some applications. The
issues introduced in this article include:
• DSP overview
• Real-time DSP operation
• Real-world signals
• Sampling rates and anti-alias filtering
• DSP algorithm time budget
• Sample driven versus frame driven data acquisition
Because these issues involve many valuable levels of detail that we
could not do justice to in this brief article, you should consider
reading Richard Higgins’s text, Digital Signal Processing in VLSI
(see References below). This text provides a complete overview of
DSP theory, implementation issues, and reduction to practice (with
devices available at the time it was published), plus exercises and
examples. The Reference section below also contains other sources
that further amplify this article’s issues. To prepare for the next
articles in this series, you might want to get free copies of the
ADSP-2100 Family User’s Manual* and the ADSP-2106x SHARC
User’s Manual.* These texts provide information on Analog
Devices’s fixed- and floating-point DSP architectures, a major topic
in these articles. The next article will cover the following territory:
• Mathematical overview of signal processing: It will present
the mathematics for the transform functions (frequency domain)
and convolution functions (time domain) that appear throughout
the series. While the mathematical treatment is necessarily
incomplete (no derivations), there will be sufficient detail for
considering how to program the operations.
• DSP architecture: The article will discuss the nature and
functioning of the DSP’s arithmetic-logic unit (ALU), multiplyaccumulator (MAC), barrel-shifter, and memory busses—and
describe the numeric operations that support DSP functions.
• DSP programming concepts: A discussion of programming
will bring together theory and practice (math and architecture).
Finally, it will lay out the main parameters for a series-length DSP
design project, provided as an example.
b
References
Higgins, R. J. Digital Signal Processing inVLSI, Englewood Cliffs, NJ: Prentice
Hall, 1990. DSP basics. Includes a wide-ranging bibliography. Available for
purchase from ADI. See the book purchase card.
Mar, A., ed. Digital Signal Processing Applications Using the ADSP-2100 Family—
Volume 1, Englewood Cliffs, NJ: Prentice Hall, 1992. Available for purchase
from ADI. See the book purchase card.
Mar, A., Babst, J., eds. Digital Signal Processing Applications Using the ADSP2100 Family—Volume 2, Englewood Cliffs, NJ: Prentice Hall, 1994. Available
for purchase from ADI. See the book purchase card.
Dearborn, G., ed. Digital Signal Processing Applications Using the ADSP-21000
Family—Volume 1, Norwood, MA: Analog Devices, Inc., 1994. Available for
purchase from ADI.0 See the book purchase card.
*Mar, A., Rempel, H., eds. ADSP-2100 Family User’s Manual, Norwood, MA:
Analog Devices, Inc., 1995. Free. Circle 1
Mar, A., Rempel, H., eds. ADSP-21020 Family User’s Manual, Norwood, MA:
Analog Devices, Inc., 1995. Free. Circle 2
*Rempel, H., ed. ADSP-21060/62 SHARC User’s Manual, Norwood, MA:
Analog Devices, Inc., 1995. Free. Circle 3
Analog Dialogue 31-1 (1997)
SelectingMixedSignal
Componentsfor
DigitalCommunication
Systems–III:Sharing
theChannel
from a given transmitter is not continuous, but in bursts. To
represent a continuous conversation (say in a cellular phone call),
the digitized information acquired during the period between
transmissions must be time compressed, transmitted in a short
burst, then expanded in the receiver to form a transparently
continuous message.
by Dave Robertson
Figure 1. Illustration of time multiplexing, showing guard
bands.
Part I provided an introduction to the concept of channel capacity—
and its dependence on bandwidth and SNR; Part II gave a brief
summary of different types of modulation schemes.This segment discusses
communications signal-processing issues that arise when multiple users
share the same transmission medium .
SHARING THE CHANNEL
Selection of an appropriate modulation scheme is only part of the
problem of defining a communications network. In most cases,
the transmission medium must accommodate signals from more
than a single transmitter. The most obvious case of such multiple
use is the airwaves; they must carry a variety of wireless traffic,
from broadcast radio and television, to cellular telephony, to CB
and short-wave radio. Even a simple twisted-pair telephone wire,
which represents a dedicated line between the phone company
central office and a user, must carry both incoming and outgoing
voice and data during a call.
In most cases, the key to effective multiplexing of independent
transmissions is proper observance of “live and let live” protocols,
enabling the effective transmission of the desired message without
undue interference to other transmissions. There are a variety of
approaches towards sharing a communications medium among
multiple users; each has its own requirements affecting component
selection. Most of these schemes are usable for both analog and
digital communications; but the flexibility of time compression,
and other features available in digital communications, opens up
more options.
TDMA—time-division multiple access: perhaps the most obvious way
of sharing the communications channel is to “take turns”: only
one transmitter at a time is allocated the channel. There must of
course be some sort of protocol to establish who has the
transmission privilege, when, how often, and for how long. A simple
example is the walkie-talkie user’s employment of the word “over”
to indicate the termination of a transmission stream and freeing
up the communications channel for other users to transmit.
A more formal arrangement is usually desirable, especially when
each user is to be allotted a very brief—but repetitive—
participation. An overall time period can be divided into designated
“slots”, with each of the transmitters assigned a different time slot
for transmission (Figure 1). This kind of scheme requires
synchronization of all the transmitters, plus a “supervisor” to assign
time slots as new transmitters want to enter the channel—and to
keep track of slots vacated. Some “overhead” space must be
provided to allow for transitions between transmitter time slots;
the better the synchronization, the less time lost to these transition
periods. Time multiplexing also means that the stream of data
Analog Dialogue 31-1 (1997)
GUARD
BAND
B
USER
TIME
SLOT
A
C
B
A
USER
USER
USER
USER
TIME
The analogy of a panel discussion is sometimes used to illustrate
the nature of TDMA. A participant who interrupts out of turn
or rattles on and on endlessly commits “violations of the TDMA
protocol.” The European GSM digital cellular telephony
standard makes use of TDMA; each channel carries eight phone
calls simultaneously in a repeated transmitting sequence of eight
time slots.
Component selection for TDMA systems must involve careful
consideration of bandwidths and settling times; long time constants
of components with insufficient bandwidth will tend to cause
signals to “bleed into” an adjacent user’s time slot.
FDMA—frequency-division multiple access: anyone who receives TV
or radio broadcasts at home is familiar with an example of
frequency-division multiple access. In this case, multiple
transmitters can simultaneously transmit without interference (at
a given power level in a given geographical area) by keeping each
frequency in their transmissions within a designated frequency
slot. The receiver determines which channel is to be recovered by
tuning to the desired frequency slot. It is important that each
transmitter’s frequency limits be strictly obser ved; any
transgressions would create interference in the neighboring
channels. (Figure 2)
USER
USER
A
USER
B
C
ADJACENT FREQUENCY BANDS ARE
TYPICALLY LEFT UNUSED TO PROVIDE
EXTRA GUARDBAND
FREQUENCY
88MHz
108MHz
Figure 2. Illustration of frequency multiplexing, showing unused frequency bands to provide separation.
Using the conversational analogy, this might be like providing a
set of booths, one for each speaker; if they speak quietly enough,
all “transmitters” can broadcast simultaneously, and a listener
may “tune in” by listening at the desired booth.
Almost all wireless applications are subject to frequency band
constraints; national and international regulatory bodies, e.g., the
FCC in the United States, license the transmitter to specific
frequencies or restrict its class to specific bands. Wired applications
like cable TV also use frequency separation to allow simultaneous
transmission of hundreds of channels (both analog and digital).
Keeping within the specified frequency constraints has numerous
ramifications for component selection. For example, some
component in the system will be used as a precise frequency
reference. It could be an absolute frequency reference, like a crystal,
7
or it might contain a circuit that receives and “locks on” to an
external reference frequency. Components in the transmission path
must have carefully limited spectral content; this can be done
through filtering—but it is also necessary to control component
linearity, so as not to generate incidental “out of band” harmonics
and other spurious frequency components.
CDMA—Carrier Division Multiple Access—Continuing the
conversation analogies, suppose that 10 people are trying to carry
on 5 simultaneous one-on-one conversations in a small room. Suppose
further that one pair agrees to converse in English, another in French,
the others in Chinese, Finnish, and Arabic—and all are
monolingual. If you were a member of the English speaking pair,
you would hear a din of background “babble”, but the only intelligible
information would be in English. So it’s easy to see that all 5
conversations could take place simultaneously in the same room
(though in practice, everyone would probably get a headache).
This is essentially a description of the underlying idea of carrier
division multiple access. All users transmit and receive over the
same frequency band, but each pair is assigned a unique code
sequence. The digital bit stream you wish to send is modulated
with this unique code sequence and transmitted. A receiver will
receive the combined modulated bit streams of all the transmitters.
If the receiver demodulates this composite signal with the same
unique code, it essentially performs a cross-correlation operation:
the bit stream that was modulated with the same code sequence
will be recovered; all the other transmitted signals that were
modulated with different codes will be rejected as “noise”.
Modulation with the special code tends to spread the spectrum of
the initial digital bit stream over a much wider bandwidth, which
helps improve its immunity from interference. Despite this spectral
spreading, spectral efficiency can be maintained, because multiple
users can share the same bandwidth. Adding more users simply
leads to the appearance of increased noise in the channel
Examples of CDMA systems include the IS95 Digital Cellular
standard in the US and numerous military “spread spectrum”
communications applications (an additional advantage of
modulating the transmitted signal with a unique signal is that it is
essentially encrypted; a receiver cannot recover the transmitted
message without the unique modulation sequence). Though
CDMA systems involve greater digital complexity, the performance
requirements for their analog components are reduced. However,
because multiple transmitters will be broadcasting in the same
channel at the same time, it is usually desirable to minimize the
contributions to background and spurious noise by transmitter
components.
SDMA—Space Division Multiple Access: Returning to the
conversation analogy, another way to carry on simultaneous oneon-one conversations in the same room is to move to opposite
corners of the room and speak in relatively hushed tones. This
captures the spirit of SDMA. In wireless applications, signal
strength falls off rapidly with increasing distance from the
transmitting antenna. At a great enough distance, the signal can
be considered to have faded completely, from which point a new
transmitter could reuse the same frequency or time slot for a
different signal (Figure 3). In broadcast radio, the same
frequency can be reused in different cities, provided that they are
far enough apart.*
*The attenuation of signal with distance is a strong function of frequency: The
higher the transmitter frequency, the faster the rolloff.
8
8
5
6
4
1
3
7
CELL 8 CAN RE-USE THE
FREQUENCY USED BY CELL 1
ANTENNAS ARE CENTRALLY
LOCATED IN EACH CELL
2
Figure 3. Illustration of geographical multiplexing, showing
honeycomb of cells with base-station antennas at centers.
The concept of channel re-use with distance underlies the term
“cellular telephony.” Cell size is determined by the area of effective
coverage by a given transmitter, and the same frequencies can be
reused in other cells. In practice, however, patterns are designed
so that adjacent cells will not re-use the same frequencies.
Conventional antennas radiate in all directions, producing a
circular coverage area and the “honeycomb” cellular pattern in
Figure 3. Modern technology has added new dimensions to the
concept of SDMA with the development of focused, or beam
steering antennas. Phased-array technology can create a focused,
directional signal transmission pattern aimed at either an individual
target receiver or a particular target area (e.g., a specific highway
at rush hour). This can allow more rapid re-use of frequency
spectrum, thereby effectively increasing total capacity for wireless
applications.
Advanced digital communications systems use combinations of
these multiplexing schemes to effectively pack as much capacity
as possible into the available transmission channel. For example,
GSM cellular phones use TDMA, FDMA and SDMA to allocate
traffic. Even many wired applications make use of TDMA and
FDMA protocols. Although these multiplexing arrangements
typically add to the system complexity, the effective increase in
channel capacity more than offsets increases in component cost.
THE NEAR/FAR PROBLEM
In previous installments, we have discussed the impact of error
rate and modulation scheme on the required dynamic range in a
digital communications system. However, in many applications,
the multiplexing arrangements create the ultimate demands on
dynamic range in the communications receiver.
In any application, the strength of the received signal is a function
of the strength of the transmitted signal, the distance from the
transmitter, and numerous environmental factors relating to
the transmission medium (be it wireless or wired). Most
communications systems are designed to work over a variety of
distances, and so have to be designed to accommodate a large
variation in power of the received signal.
Consider, for example, a cellular telephony application. The
receiver circuitry must be designed to recover the weak signal
resulting from a transmission while at the very edge of the “cell”.
This capability to recover weak signals is often referred to as a
receiver’s sensitivity. To recover such weak signals, it seems
appropriate to include gain stages in the receive circuitry.
Consistent with good, low-noise design practice, one might expect
to put the gain as early in the signal path as possible to quickly
boost the signal above the noise floor of subsequent stages.
Unfortunately, this same receiver must also be capable of receiving
the signal transmitted by a user standing directly under the base
station’s antenna. In the case of GSM, for example, this signal can
be up to 90 dB stronger than the weakest signal. If the receiver
Analog Dialogue 31-1 (1997)
has too much gain in the signal path, the strong signal can saturate
the gain stages. For modulation schemes that include amplitude
information (including AM and QAM), this will essentially destroy
the signal. Phase and frequency modulation approaches may be
more tolerant of this clipping, depending on the circumstances.
(The clipping will still create distortion products which are
sufficient to cause problems, even in phase-modulation schemes.)
A basic approach to addressing the near/far dynamic range problem
is to use variable/programmable gain stages in the receive signal
path. Automatic gain control (AGC) allows the gain to be adjusted
in response to the strength of the received signal. An important
design consideration, though, is how rapidly the gain needs to be
adjusted. For example, in ADSL (asymmetric digital subscriber
line—see sidebar) modems, the received signal strength changes
as outdoor temperature changes affect the line impedance, so time
constants of minutes would be tolerable. On the other hand cellular
phone receivers must be designed to track the signals from fast
moving vehicles that may be moving into or emerging from the
shadows of buildings or other signal obstacles, so very rapid gain
changes are required. TDMA systems put an additional demand
on gain-ranging circuitry, because the near/far signals could be
located in adjacent TDMA time slots; in this case, the circuitry
would have to change gains and settle in the transition period
between time slots.
FDMA systems offer a different kind of near/far challenge. Here,
the worst case to consider is recovery of a weak signal in a frequency
slot next to strong signal (Figure 4). Since both signals are present
simultaneously as a composite at the input of a gain stage, the
gain is set according to the level of the stronger signal; the weak
signal could be lost in the noise floor (in this case, the noise floor
could be thermal noise or quantization noise of an A/D converter.)
FAR
In CDMA systems the near/far problem is a little more difficult to
describe. Since all signals are simultaneously transmitted in the
same frequency space, filtering cannot be used to discriminate
against unwanted signals (though it is still used to eliminate signals
in adjacent bands). CDMA employs demodulation using a carrier
unique to the desired signal to extract the desired from the
unwanted signals; signals modulated with a different carrier appear
as background noise. The ability to successfully recover the signal
is set by the total noise energy—including that of the other
carriers—in the band. Since filtering can’t be used to discriminate,
the best situation to strive for is to have all signals arrive at the
base-station antenna at equal power. To achieve this, many CDMA
systems communicate the received power levels back to the
transmitters so that power of the individual signal components
may be adjusted to equalize power levels at the base-station receiver.
To help reduce their near/far problem, TDMA systems could also
use this kind of power control, though it tends to require a moresophisticated (i.e., costly) handset.
b
ASYMMETRIC DIGITAL SUBSCRIBER LINE
USER B
POWER
ADSL is one of the many technologies competing to bring
broadband digital services into the home. The concept
underlying ADSL is to take advantage of the twisted-pair wires
that already provide almost universal telephone service to homes
in the United States. Other services providing a two-way flow
of information, such as ISDN (integrated services digital
network), require an additional, dedicated wire to provide
service.
USER
A
RECEIVER
ANTENNA
NEAR
USER A
Most radio designs feature a cascaded series of filters and gain
stages (some of which may be variable) to remove/attenuate strong
interferers, then amplify the desired signal to a level that can be
readily demodulated. Wideband radios, however, attempt to
simultaneously recover all the signals in one receiver; they cannot
use analog discrimination filters; accordingly, wideband receivers
typically have the most stringent requirements on dynamic range
in their analog circuitry and converters. Interestingly enough, even
applications where you think you have the communications channel
to yourself can suffer from simultaneous near/far signals. For
example, in ADSL modems, the system must be designed for the
scenario where the near-end echo (leakage from the local
transmitter) appears as an interfering signal that is actually up to
60 dB stronger than the desired receive signal.
USER
B
Figure 4. Near-far effect calls for the ability to handle wide
dynamic range between adjacent channels.
ADSL uses frequency-division multiplexing (FDM) to convey
modulated digital information in the frequency space between
20 kHz and 1.2 MHz, above the frequency space occupied by
conventional voice traffic. This frequency separation allows an
ADSL modem to operate without disturbing a phone call
occurring at the same time—an extremely important feature.
Even if subsequent stages have a low enough noise floor to provide
dynamic range to recover the weak signal, there must also be a
very stringent constraint on the dynamic linearity of the gain stage;
harmonics or other spurious responses of the strong signal that
wind up in the wrong frequency bin could easily obliterate the
weaker desired signal. To reduce this interference problem, most
FDMA systems attempt to filter out unwanted signals early in the
receive circuitry. The ability to discriminate against unwanted
signals in adjacent frequency bands is usually referred to as a
receiver’s selectivity.
The ANSI standard for ADSL provides for simultaneous
upstream (outgoing from the home) and downstream (incoming
to the home) transmission using either FDM (separating the
upstream and downstream signals in frequency) or echo
cancelling. Echo cancelling uses sophisticated signal processing
(analog, digital, or both) to separate the strong transmitted
signal from the weaker received signal, passing only the received
signal to the demodulator. Using the conversational model, this
is analogous to a person who can effectively talk and listen at
the same time.
FREQUENCY
(OR TIME)
Analog Dialogue 31-1 (1997)
9
ControllerBoard
SystemAllowsFor
EasyEvaluationOf
General-Purpose
Converters
by Mike Byrne & Ken Kavanagh
A critical task when designing a system using an analog-to-digital
converter is the actual choice of the converter to be used.
Constraints such as power, price, and size—along with system
requirements of resolution and performance parameters will
narrow the choice. However, the designer may still be left with a
number of converters which at first glance will perform in the
application. One option is to simply to choose a converter based
on data-sheet information; but the prudent engineer will actually
want to see how the part performs under the conditions required
by the application. At issue, though, is how to spend as little time
as possible on the evaluation, since much of it will be spent
evaluating converters that will be eliminated from consideration.
Evaluating converters is not an easy task. Hardware design work
and careful PCB layout are required to allow the part to achieve
its specified analog performance. Even if the hardware problem is
solved by use of a standard evaluation board, software routines
will be required to control the part and read back the data. In
today’s diverse market, the wide variety of available converters,
especially those from different manufacturers, are housed in
proprietary pinouts and generally require different software
routines for control and access.
Given this background, any tools that enable the designer to
evaluate the converter more easily are welcome. Such tools speed
up the process of selecting the converter by saving the designer
the time and effort needed to develop a system just to evaluate the
part. Besides having the converter chosen more quickly and easily,
the design process should be further speeded up with a quick and
easy means of testing the part in the application. To be useful, the
test system should effectively be self-contained, run on a platform
readily available to the designer, and fulfill two distinct functions:
provide a quick means of demonstrating the performance of the
converter, and provide a development platform for evaluation of
the converter in the user’s application.
Those were the design goals for an evaluation system for mediumand high-speed general purpose analog-to-digital converters, using
a standard PC as the platform for evaluating the devices. The PC
has two ports for external communications—a serial RS-232 port
and a parallel printer port. Because neither port can operate at the
speeds required to run an ADC at throughput rates in excess of
10,000s of samples per second, direct connection between the PC
and the ADC is not possible if the ADC is to be operated at
specified throughput rates.
Thus, to facilitate controlling the ADC and running it at specified
speeds, additional processing power is required external to the
PC. A memory buffer area is also required to store the data,
gathered at the higher throughput rates, before it is uploaded to
the PC at the lower rates compatible with its porting speed. One
10
approach is to build evaluation boards with processing power and
memory. But this has the disadvantage of requiring a processor on
every evaluation board; and the evaluation board becomes less
flexible for use as a development board.
The approach adopted by ADI’s General-Purpose-Converter
Applications Group in Limerick was to simplify the evaluation
boards and move the memory, processing power and PC
communications to a separate controller board. This controller
board, labeled EVAL-CONTROL BOARD, provides the link
between the RS-232 port of the PC and the evaluation board for
the device of interest. It provides a means of controlling a
converter’s (daughter) board and operating a converter at its
maximum throughput rate, while demonstrating the results on a
PC. The Eval-Control Board is capable of being used with a suite
of daughter evaluation boards, each geared to the specific needs
of a particular converter or family. The user buys one Eval-Control
Board and can connect daughter evaluation boards to this
controller board via a standard 96-lead connector. The system
configuration is shown in Figure 1.*
HIGH SPEED RS-232 LINK
PC COMPATIBLE
OR
PORTABLE PC
EVAL-CONTROL BOARD
12V AC TRANSFORMER
EVAL-AD7xxxCB
Figure 1. Bench-test setup.
All daughter evaluation boards that are compatible with the EvalControl Board have designations that end with the letters CB. To
date, there are 10 daughter ADC evaluation boards available for
use with the controller board; their performance range is from the
12-bit, 600-ksps AD7892 to the 16-bit 200-ksps AD976A.*
In keeping with the requirement that the setup be self-contained,
a standard 12-V ac transformer with 1-A current capacity is
available to power the system. From it, the Eval-Control Board
then generates all the power required for itself and the daughter
board (Figure 2). The board’s functions are based on an Analog
Devices ADSP-2111 digital signal processor (DSP). The
ADSP-2111 features two high-speed serial links, a 2K × 24 internal
program-memory RAM and a 1K × 16 internal data-memory
RAM. The processor controls the sampling rate and the data
acquisition from the evaluation board. For serial interfaces, it is
capable of taking data at a bit rate of 8 MHz; for parallel interfaces,
it has a word rate of 3.2 MHz.
Communication between the Eval-Control Board and the PC is
controlled by an 8051 microcontroller. The UART port on the
8051 connects to the COM-1 serial port of the PC, and information
is transferred via a standard RS-232 cable. The 8051 is connected
to the host interface port (HIP) of the ADSP-2111. The software
for the appropriate daughter evaluation board is downloaded to
the ADSP-2111 via the 8051. Similarly, when the ADSP-2111
has acquired and processed the data samples, they are transferred
back to the PC for analysis via the 8051.The Eval-Control Board’s
8K × 16 bit RAM allows as many as 4000 samples to be stored
before they are transferred up to the PC for processing. It also
includes a 14-bit single-chip DAC and ADC (AD7869). Thus it
can acquire analog data, and generate analog output signals if the
user so wishes.
The evaluation software, which runs under DOS 4.0 or higher,
allows the user to control the operation of the daughter board,
*To request technical information on this system, Circle 4
Analog Dialogue 31-1 (1997)
DIGITAL VDD
DIGITAL GND
12V AC
INPUT
ANALOG VDD
ANALOG GND
POWER
SUPPLY
SECTION
DATA BUS
ADSP-2111
RS232 TO TTL
LEVEL SHIFT
14-BIT
DAC
14-BIT
ADC
ANALOG VSS
ANALOG +12V
ANALOG –12V
ANALOG +17V (UNREGULATED)
ANALOG –17V (UNREGULATED)
+5V OR 3.3V/
GND SELECT
RS232
CONNECTOR
POWER
SUPPLIES
AD7869
DAC/ADC
ENABLE
8051
MICROCONTROLLER
5V/3V
LEVEL SHIFT
SPORTS
5V/3V
LEVEL SHIFT
DATA BUS
96 WAY
CONNECTOR
8K X 16 RAM
ADDRESS BUS
HIP
ENABLE
5V/3V
LEVEL SHIFT
ADDRESS BUS
DECODER
FLAGS INTERRUPT
ENABLE 5V/3V SHIFTERS
5V/3V
LEVEL SHIFT
Figure 2. Block diagram of the Evaluation-Control Board.
which contains the device under test. The samples are gathered at
rates up to the maximum throughput rate of the converter, and
then uploaded to the PC for display and analysis. The front-end
PC software has a number of screen displays; examples of these
are shown in Figure 3.
Eval-Control Board can operate as a true development platform.
To adapt and develop the software for this purpose, it is first
necessary to understand how the software for the Eval-Control
Board works.
SOFTWARE
The software comes in two main parts. The first, written in C
code, runs on the PC; the second is the DSP code—it runs on the
evaluation control board and is based on the ADSP-21xx code.
We consider here as an example a 12-bit ADC with a parallel data
bus. Before examining the 21xx code, one should understand what
happens when the user selects commands from the PC software.
a. Sine wave and its digitized spectrum.
b. Histogram analysis
Figure 3. Examples of data presentation.
The software runs with a series of function keys, hot keys and onscreen buttons. With it, the user can select sample rate, number of
samples to be taken, analog input range, mode of operation and
control of on-chip registers. The software can also perform a
window function on the captured data, carry out a fast Fourier
transform (FFT), compute signal-to-noise ratio, and display the
results. It can also display the spread of codes for a dc input and
compute the mean and standard deviation of the distribution. The
software allows the captured data to be stored to a file, from which
it can be exported for processing by other software packages. Data
can be taken from a file (rather then captured from the evaluation
board) and processed off line.
The Eval-Control Board thus fulfills its requirement of
demonstrating nominal performance using the complete set of
software routines for the above functions. Once the designer has
moved to the evaluation phase, these routines can be adapted to
evaluate the device under application-specific conditions. Here the
Analog Dialogue 31-1 (1997)
Each daughter evaluation board requires its own software routines;
they are stored on the PC and downloaded to the Eval-Control
Board whenever that specific evaluation board is to be used. The
board-specific software files are called host-interface-port (HIP)
files. When the user selects the program to be loaded, the HIP file
for that program is loaded to the Eval-Control Board over the
serial port of the PC. After a program has been loaded, the DSP
program causes it to wait until it gets a command before doing
any further operations. When the user selects an operation from
the PC, a four-byte code is sent to the DSP via the serial port. The
code is pre- and post-formatted with an “S” and “E” character
respectively. These allow the software to interpret what function is
required to be carried out (e.g., take samples, write to a register
etc.) When a command is sent to the Eval-Control Board the ‘S’
and “E” bytes are stripped by the microcontroller and the 4 bytes
are loaded into data memory (DM) addresses 0x3FE1, 0x3FE2,
0x3FE3 and 0x3FE4, as diagrammed here.
S
0x3FE1
0x3FE2
0x3FE3
0x3FE4
E
HIP Register Addresses
The microcontroller also writes a value of 0x03 to adö3ess 0x3FE0
to indicate that data has been received. These addresses are called
the HIP registers. There is also a HIP status register at address
0x3FE6. This register is checked to see if any new data has been
written to the HIP registers, by either the DSP or the PC. The
four-byte code sent to the DSP can be seen if the command line
option -e+ is added to the program name running on the PC. For
instance, if the program being used is AD7892.EXE, the command
line is AD7892.EXE -e+.
The first step in writing code for an application is to write all the
initial setup routines for the DSP—the interrupt vector addresses,
the serial port configurations and initialization of any flags that
may be needed. (That part of the code, which depends on the
specific application, will not be shown in this article.) Once the
DSP has been set up, the program should wait for commands to
11
be sent to it from the user. This is done using the Wait_Confirm
subroutine as shown below and ALU input registers AX0 and AY1.
AR is the ALU output register.
wait_confirm:
AX0=DM(0x3fe6);
{load HIP status register into AX0}
AY1=0x0001;
{load AY1 to check HIP against register 1}
AR=AX0 AND AY1;
{Has anything been written?}
IF EQ JUMP wait_confirm; {If not then repeat the loop}
RTS;
{return when the commands are received}
This routine simply monitors the HSR6, HIP Status Register
(0x3FE6) and doesn’t return until the LSB of the register is a 1,
indicating that something has been written to the HDR1 register.
This diagram identifies the contents of the four data bytes:
0x3FE1
S
NUMBER OF
SAMPLES
0x3FE2
0x3FE3
0x3FE4
TIMER OR
REGISTER
HIGH BYTE
TIMER OR
REGISTER
LOW BYTE
TYPE OF
COMMAND
E
HIP Register Functions
Once the data is received, the timer register can be loaded. The
manipulation of the data to rebuild the 2 bytes into one 16 bit
word (by shifting the high byte 8 bits left and combining), and
store it in data memory, is shown here:
AX0=DM(0X3fe2);
AX0=DM(0X3fe3);
AX0=AR; AY1=SR0;
AY0=AR;AR=AY0-1;
DM(Tcount_Reg)=AR;
SI=AX0; SE=8; SR=LSHIFT SI(LO);
AY1=0X00ff; AR=AX0 AND AY1;
AR=AX0 OR AY1;
DM(Tperiod_Reg)=AR;
Finally, the address of the parallel port (in this case the edge
connector), and the start address of the memory location where
the data is to be put, need to be set up. This is done using the
ADSP-2111’s pair of data-address generators (DAGs). DAGs
consist of three registers, identified as I (Base/index address of the
memory block), M (incremental step to next location), and the L
(the length of the memory block; The L block is required for return
to start of a loop involving a circular buffer; in this example it will
be set to zero). The code looks like this.
I0=0; M0=1; L0=0;
{address of memory locations for data}
I1=0x2000; M1=0; L1=0; {address of parallel port }
A loop can now be set up which will read in a data sample every
time an interrupt occurs. Since reading samples takes but a fraction
of the time of the whole operation (i.e., reading and uploading the
data) it is possible to fill the memory with data and only select a
portion to upload.
AY1=1; AR=8191;
{set sample increment and number of
samples req’d}
MSTAT=0x20; IMASK=0x01; {start the timer and unmask the interrupt}
luup: IDLE;
luup1: AR=AR – AY1;
IF NE JUMP luup;
{wait for the interrupt to occur}
{decrease the count}
{repeat until finished}
In this example the code is simplified to use only the timer
interrupt. This can be done by waiting for the timer, starting from
a preset time, to expire—thus causing an interrupt. Data can then
be read from the data bus and, after a short delay (if required),
start the next conversion. The code below shows how the timer
interrupt vector address should look when it appears in the vector
table. The interrupt service routine is also shown.
timer_int: jump ISR;RTI;RTI;RTI;
ISR: AX0=DM(I1,M1);
DM(I0,M0)=AX0;
RESET FL0; SET FL0;
RTI;
12
{interrupt vector}
{read data}
{Write to memory}
{Convst}
When using this technique the first data value to be returned must
be ignored, because it will have been read in before the conversions
were started; hence it is not valid. Once all the data has been read
in, it is time to upload the data to the PC. After turning off the
interrupts and resetting the DAGs controlling the data memory,
one can check to determine how many samples are required to be
uploaded, using this code:
IMASK=0x00; MSTAT=0x10;
{mask all interrupts and stop the timer}
I0=1; M0=1; L0=0;
{reset DAGs to start of Data Memory }
AR=DM(0x3fe1); MY0=256; MR=AR*MY0(UU);
{Read number of samples required}
AY0=MR0;
{store count}
The data from the 12-bit ADC should be sign-extended to fit the
16-bit-wide data registers. This can be done by examining the data
and, if the MSB is a 1, OR the data with 0xF000 to set the highest
4 bits to 1 as shown here.
MORE_TO_ECHO: AX0=DM(I0,M0);
AY1=0x800; AR=AX0 AND AY1;
IF EQ JUMP OVER;
AY1=0xF000; AR=AX0 OR AY1;
AX0=AR;
{read the data value}
{check the MSB }
{if it’s a 1 set the upper bits to 1}
The data is now properly formatted and can be transferred to the
PC via the host interface port. This is done by calling the OUT_HIP
routine as part of a loop:
OVER:
CALL OUT_HIP;
AR=AY0–1;
AY0=AR;
{upload the data value }
{decrease the count }
{and copy it back to the
{register}
IF GE JUMP MORE_TO_ECHO;
{if we’re not finished then
{repeat}
GG:
AX0=0x00FF;
{all the data has been
{so indicate this to the PC}
DM(0X3fe3)=AX0;AX0=0;
DM(0x3fe0)=AX0;
JUMP cloop;
{return to start and wait for
{next command}
OUT_HIP
SI=AX0; {input the data to the barrel shifter}
SE=–8; {set the number of bits to shift by}
SR=LSHIFT SI(LO);
{and shift the data}
DM(0X3FE4)=SR0;
{ HI byte }
DM(0X3FE5)=SI;
{ LO byte }
AX0=1; DM(0X3FE3)=AX0; { indicate that data
is available }
CALL WAIT_UPLOAD; {and upload it to the PC}
RTS;
{return when finished }
WAIT_UPLOAD: AX0=DM(0x3FE6);
{check contents of HIP
{register 6}
AY1=0x0001;
AR=AX0 AND AY1;
IF EQ JUMP WAIT_UPLOAD; {repeat until ready}
AX0=DM(0x3FE0);
AY1=0x0003;
AR=AX0 - AY1;
IF NE JUMP WAIT_UPLOAD;
RTS;
Once the required number of samples has been sent to the PC,
the program should return to the start of the main loop and wait
for further instructions.
The last few routines are the standard set; they will rarely need to
be changed when modifying the Eval-Control Board’s DSP code
to meet application-specific conditions. All the routines shown
above have been used in the HIP files that control the various
evaluation (daughter) boards that interface to the PC via the EvalControl Board. They are shown here to give an indication of how
existing programs are modified for additional functionality.
b
Analog Dialogue 31-1 (1997)
By Albert O’Grady & Jim Ryan.
An analog transmitter is a field-mounted device that senses a
physical parameter such as pressure or temperature and generates
a current proportional to the measured variable in the standard
range, 4 to 20 mA. Providing the output as a current in a twistedpair loop has many advantages: the measurement signal is
insensitive to noise and is not affected by changes in loop resistance;
transmitters meeting the standard are interchangeable; and the
power required to energize the transmitter circuits can be derived
from a remotely supplied loop voltage. Figure 1 shows a
conventional transmitter circuit, consisting of a power supply, a
current-manipulating transmitter, and a receiving controller.
MEMORY
SENSORS
4 TO 20mA
D/A
CONVERTER
MEASUREMENT
CIRCUIT
COMMUNICATION
SYSTEM
Figure 3. Intelligent Transmitter.
Component selection for smart analog transmitter designs:
Figure 4 shows a circuit that implements the smart transmitter of
Figure 2. The following sections discuss smart transmitter design
factors and tabulate alternatives. Besides low cost, the all-important
constraint is that the entire circuit consume less than 3.5 mA (the
“low alarm” setting, 0.5 mA below the 4-mA signal floor), to permit
the transmitter to be loop-powered.
50Ω
2.2µF
DVDD
SENSORS
RTD
mV Ω TC
0.01µF
VCC
AVDD
ANALOG
TO
DIGITAL
CONVERTER
100kΩ
4.7µF
4.7µF
DRIVE
CLOCK
LATCH
DATA
MICROCONTROLLER
MCLK IN
AD421
GND
MCLK OUT
CS
DATA OUT
SCLK
DATA IN
DGND AGND
TRANSMITTER
POWER SUPPLY
0.01µF
COMP
VCC
AD7714/
AD7715
LV
BOOST
REF OUT1
REF OUT2
REF IN
1.25V
REF IN
AMBIENT
TEMP
SENSOR
DN25D
3.3V
0.1µF
CONTROLLER
SENSOR
MICROPROCESSOR
A/D
CONVERTER
LOOP POWER
Build a Smart Analog
Process-Instrument
Transmitter with LowPowerConvertersand
aMicrocontroller
COM
C1
1kΩ
1000pF
LOOP
RTN
C2
C3
Figure 1. Analog Transmitter.
Transmitter design has responded to the requirements of users
for improved performance and versatility, plus reduced cost and
maintenance. A second-generation “smart analog transmitter” has
a microprocessor (and data conversion), to provide remote memory
and computing power (Figure 2). It can condition the signal
remotely before converting it to current and transmitting it back
to the controller. For example, it can normalize gain and offsets,
linearize sensors having known nonlinearities (such as RTDs and
thermocouples) by converting to digital, processing with arithmetic
algorithms in the µP, converting back to analog and transmitting
on the loop as a standard current. This reduces the control room’s
signal processing burden, a big advantage if a large number of
signals must be dealt with.
MEMORY
MICROPROCESSOR
D/A
CONVERTER
4 TO 20mA
The AD7713, AD7714, and AD7715 meet these criteria and are
suitable for use in the front end of any smart transmitter.
The AD7714 is a complete multi-channel (3 differential, 5 singleended) analog front end targeted for low-frequency applications.
It can accept low level signals directly from the transducer (Figure
5), has an on-chip programmable-gain amplifier (PGA),
configurable for gains from 1 to 128. Front-end instrumentation
amplifiers are unnecessary in most applications using the AD7714.
MEASUREMENT
CIRCUIT
Figure 2. Smart Transmitter.
AVDD
IN+
The third generation, “smart-and-intelligent” transmitters, add
digital communications, which share the same twisted-pair line
with the traditional 4-20-mA “dc” signal (Figure 3). The
communication channel allows both analog and digital versions
of the measured variable to be transmitted over the twisted pair,
as well as control signals and diagnostic data relevant to the
transmitter, such as calibration coefficients, device ID, and data
relevant to fault diagnosis. Transmitter faults can be diagnosed
remotely—very useful for transmitters in hazardous locations.
The Hart protocol is the de facto communication standard used
by smart transmitters. It employs frequency-shift keying (FSK)
modulation, based on the Bell 202 Standard. Data is transmittedat
1200 bits/s, switching between 2.2 kHz (“0”) and 1.2 kHz (“1”).
(See sidebar, page 15.)
Analog Dialogue 31-1 (1997)
+5V
EXCITATION VOLTAGE = +5V
AVDD
DVDD
CHARGE
BALANCING
A/D CONVERTER
1µA
OUT–
BP01
OUT+
AIN3
IN–
AD7714
AIN1
AIN2
AIN4
AIN5
SWITCHING
MATRIX
SENSORS
A/D
CONVERTER
Figure 4. Details of a Smart Transmitter.
A/D converter: Other major criteria for ADC selection are:
• a high level of integration to reduce front-end component count
• high resolution to meet required system resolution & accuracy
• Single-supply operation on either 3-V or 5-V supply
• Calibration features to allow the removal of component or
system errors due to drift with time and temperature.
AUTO-ZEROED
PGA
BUFFER
MODULATOR
DIGITAL FILTER
MCLK IN
A = 1–128
AIN6
24kΩ
1µA
CLOCK
GENERATION
SERIAL
INTERFACE
MCLK
OUT
REGISTER BANK
AGND
REF IN (+)
15kΩ
REF IN (–)
POL
AGND
DGND
DOUT
BUFFER
DIN
SCLK
OPTOISOLATORS
MICROCONTROLLER
Figure 5. Interfacing a transducer interface to AD7714 ADC.
13
The reference can be derived from the sensor excitation voltage
for ratiometric measurement applications.
Its sigma-delta architecture is capable of up to 24 bits of nomissing-codes
performance. Operation is from a single 3- or 5-V
.
supply, consuming 650 µA (< 5 µA in power-down mode). The
AD7714 has a differential reference input. Its input signal range is
from 0 to +20 mV up to 0 to +2.5 V, unipolar, depending on the
PGA gain setting, and ±20 mV to ± 2.5 V bipolar. Its serial interface
can be configured for easy-to-isolate 3-wire operation by a
microcontroller in a smart transmitter circuit.
The µC can do calibrations periodically, removing gain and offset
errors—and drifts with time and temperature—in the device itself
or the whole system. The calibration features include self-,
background-, & system calibration. On-chip calibration registers
allow OEMs to do calibrations at the factory, store coefficients in
memory, and rewrite them to the device in the field.
The AD7715 is essentially a 16-bit, 1-channel version of the
AD7714, with all of its features, including the PGA and input
signal ranges, differential reference, calibration features, 3-V or
5-V single-supply operation, and 3-wire serial interface.
The AD7713 is a 24-bit complete analog front end for lowfrequency measurements. Its two low-level differential analog input
channels can take signals directly from a transducer. It also accepts
a high-level (up to four times the reference) single-ended input. It
has a differential reference input, and two on-chip current sources
are available; they can be used for excitation in 3-wire and 4-wire
resistance temperature detector (RTD) circuits. Gain settings,
signal polarity, and RTD current control can be configured in
software using the bi-directional serial port. The AD7713 is also
capable of self calibration, system calibration and background
calibration to remove zero- and full-scale errors.
ADC Selection Table†
ADC
Resolution Input channels
Supply
voltage
AD7714 16/24
3 fully differential
or 5 single-ended
3 V/5 V
AD7715 16
1 fully differential
3 V/5 V
AD7713 16/24
2 fully differential
and 1 single-ended
5V
Current
Package
consumption* options
500 µA @ 3 V 24-pin DIP,
670 µA @ 5 V 24-lead SOIC,
28-lead SSOP
450 µA @ 3 V 16-pin DIP,
650 µA @ 5 V 16-lead SOIC
1.1 mA
24-pin DIP,
24-lead SOIC
*Current consumption values are based on using a 1-MHz master clock.
†For technical data on these devices, consult our Web site, www.analog.com, use Faxback (see
p. 24), or circle 5
Microcontroller: The microcontroller (µC) is the engine of the
smart transmitter; it controls the entire transfer of information
from sensors to loop current. The µC’s memory and processing
capability makes possible periodic calibrations, signal conditioning,
error correction, temperature compensation, and linearization, all
at a remote location powered by loop voltage. Besides low cost
and low power consumption, a µC for a smart transmitter should
have these features:
• memory. It should contain enough ROM and RAM for
implementing all software functions (boot program plus data
processing) without external memory; this reduces component
count, board space and power consumption in the system.
• serial communications port to provide an internal interface with
the input ADC and the output DAC for control and data
transfer. Galvanic isolation is easily implemented with a few
optoisolators.
• low clock speed to minimize power consumption, which is
generally directly proportional to clock speed in CMOS devices.
14
The table suggests µCs with sufficient on-chip memory and lowenough power consumption to be powered from the loop and to
implement the “smart” functions required by a typical transmitter.
Microcontroller Options
Microcontroller*
80L51
On-chip
ROM
4K × 8
MC68HC05
6160 bytes
PIC LC 54
512 bytes
EEPROM
MC68L11
µPD780xx
16K
8K to 32K
On-chip
Power consumption
RAM
128 bytes × 8 1.7 mA (Vcc=3 V, Fclk=3.58 MHz)
50 µA (Vcc=3 V, Fclk=32 kHz)
10 µA typ in power-down mode
224 bytes
0.8 mA (Vcc=5 V, Fclk=100 kHz)
0.7 mA (Vcc=3 V, Fclk=100 kHz)
32 µA @ 5 V, 20 µA @ 3 V in powerdown
25 bytes
1.8 mA(Vcc=5 V, Fclk=4 MHz)
14 µA (Vcc=3 V, Fclk=32 kHz
5 µA in power-down mode
512 bytes
2 mA (Vcc=3 V, Fclk=500 kHz)
256 bytes
120 µA(Vcc=5 V, Fclk=32 kHz)
to 1024 bytes 64 µA (Vcc=3 V, Fclk=32 kHz)
1.8 mA (Vcc=3 V, Fclk=5 MHz)
*These microcontrollers are not Analog Devices products.
D/A converter: In the smart transmitter, the DAC is the means
of driving and controlling the loop current. The examples we will
consider here involve 2-wire, remotely powered transmitters.*
The D/A converter and current loop control circuitry should have
comparable resolution to the front-end conditioning circuitry and
the ADC; and the DAC should be monotonic, since the process
variable (PV) may be part of a control loop. These, together with
supply current low enough to keep the total remote drain from
the loop supply below 3.5 mA, are the minimum requirements.
The specified current drain of the DAC circuit must of course
include that of a precision reference and the quiescent current of
an output amplifier. Operating temperature range must be
sufficient to maintain required overall specifications without
excessive drift in an industrial environment. Other requirements:
• A highly integrated chip for overall component reduction
• High resolution to meet system resolution and accuracy needs
• Single-supply operation on either 3-V or 5-V supplies.
The AD421 and AD422 monolithic ICs are designed specifically
for industrial current-loop control applications. Both devices offer
the functions essential to remote transmitter operation.
The AD421* is a loop powered current-control sigma-delta DAC,
implemented in BiCMOS, for high resolution and accuracy. It is
complete with two precision references and a voltage regulator.
The DAC has 16 bit resolution for operation in the 4-to-20-mA
(16-mA span) program range, and an extra bit for optionally
programming alarm currents in the range 0-32 mA. The DAC
output provides the setpoint to a loop-current control circuit within
the AD421.This circuit servoes the loop current to track its setpoint
by measuring the return current and manipulating the current
drawn at its Boost terminal. The 1.25-V and 2.5-V precision
references, laser trimmed and temperature compensated for low
drift, eliminate any need for stand-alone references. They can be
used as reference inputs to the DAC (2.5 V) and an ADC.
The AD421 incorporates an adjustable voltage regulator, which
powers the whole remote transmitter circuit, including the AD421
itself. The regulator has selectable settings for 3-V, 3.3-V and 5-V
*In most industrial applications power and signals (or process variables) share
the same wiring to remote locations. The 4-20-mA interface can be used
either to transmit measured variables or to send a command signal to a valve
or actuator. Actuators may use more power than can be effectively derived
from a 4-20 mA loop signal, so they are normally connected as four-wire devices, two for command signal and a separate pair for powering the device.
Analog Dialogue 31-1 (1997)
Besides the advantages of integration, another benefit is that the
device’s current-loop control performance is specified, avoiding
the need to calculate error budgets involving multiple devices.The
AD421 features an SPI type three-wire interface that allows easy
and efficient interface to most microcontrollers and requires a
minimal number of opto-isolators if galvanic isolation of the loop
from the transducer is needed for intrinsic safety.
The AD421 was designed for use in both smart and intelligent
transmitter applications. Smart transmitters (Figure 4), though
digital in operation, simply produce analog loop current
proportional to process variable over a 16-mA (4-20-mA) span.
Intelligent transmitters add another dimension of functionality (see
sidebar). The transmitter can both send a PV signal on the current
loop in analog form, and it can transmit and receive digital
information by modulating the current in the loop. The digital
signals are sent and received using a modem circuit, which
translates and transmits digital levels as modulated currents and
translates received modulated current into digital ones and zeros.
The AD421 can be interfaced with a stand-alone modem; its
current control loop section provides a suitable input node for the
modulation signal to be coupled from an external HART modem,
such as the AD424 or the 20C15 from Symbios Logic. The separate
modem is digitally interfaced to the microcontroller or a UART,
as shown in Figure 6.
50Ω
3.3V
0.1µF
DN25D
2.2µF
0.01µF
AMBIENT
TEMP
SENSOR
DVDD
AVDD
1.25V
REF IN
ANALOG
TO
DIGITAL
CONVERTER
AD7714/
AD7715
MCLK IN
100kΩ
4.7µF
BOOST
REF OUT1
REF OUT2
REF IN
4.7µF
VCC
MICROCONTROLLER
LV
AD421
GND
MCLK OUT
CS
DATA OUT
SCLK
DATA IN
DGND AGND
0.01µF
COMP
DRIVE
CLOCK
LATCH
DATA
COM
C1
LOOP POWER
SENSORS
RTD
mV Ω TC
VCC
1kΩ
1000pF
LOOP
RTN
C2
C3
CC
HART
MODEM
BELL 202
WAVEFORM
SHAPER
HT20C12/20C15
BANDPASS
FILTER
Figure 6. Complementing the smart transducer of Figure 4
with a discrete HART modem.
A new device, the AD422,* reduces the physical complexity of
intelligent transmitters using the HART protocol. It combines the
AD421’s functional blocks (voltage regulator, DAC, current control
loop and references) with a HART modem and several supervisory
circuits (watchdog timer, alarm input and reset generator)—all
on a single chip! It is a highly integrated solution specifically
targeted to design of intelligent transmitters (Figure 7), with a
greatly reduced component count.
b
*For technical data on the AD421 and AD422, consult our Web site,
www.analog.com, use Faxback (see p. 24), or circle 6
Analog Dialogue 30-4 (1996)
DN25D
50Ω
3.3V
2.2nF
1µF
DVDD AVDD
BOOST VCC LV HART
IN
REF IN
SENSORS
RTD
mV Ω TC
AMBIENT
TEMP
SENSOR
1µF
10µF
ANALOG
TO
DIGITAL
CONVERTER
AD7714/
AD7715
MICROCONTROLLER
MCLK OUT
CS
DATA OUT
REF OUT1
REF OUT2
REF IN
10nF
COMP
DRIVE
VCC
MCLK IN
LOOP POWER
operation, and can be programmed to any voltage from 3 to 5 V
with a suitable external resistor. An external depletion-mode pass
transistor is required to implement this regulator function; it must
be capable of supplying the total current required by the
transmitter. The AD421 is available in a small-outline surfacemount package, which fits nicely in crowded explosion-proof
housings.
GND
RXD
AD422
TXD
RTS
CD
SCLK
LOOP
RTN
SDATA
DLATCH
CLATCH
WDO
WDI
SCLK
DATA IN
MCLK IN
1kΩ
1nF
ALARM
RST
HART
OUT
1.2nF
DGND
MCLK OUT
COM C2
AGND
C1
0.15µF
0.15µF
Figure 7. Complete intelligent transmitter with AD422
HART PROTOCOL
The de facto standard for communication over 4-to-20-mA
current loops in industrial applications is the Highway
Addressable Remote Transducer (HART) protocol, originally
developed by Rosemount, Inc., but now supported for the
general community by the HART Foundation. The protocol
is adapted from the Bell 202 frequency-shift keying (FSK)
telephony standard: the loop current is transmitted at 1200
bits/second as one of two phase-continuous frequencies,
1.2 kHz—“mark” (1) or 2.2 kHz—“space” (0). Suitably
filtered, the ac signals do not affect the measurement’s dc value.
HART implements a master/slave protocol; a remote “slave”
device responds only when addressed by the master.
As smart transmitters are remote instruments with no means
of communication other than the analog PV signal, they could
not be interrogated for status information (but an extra alarm
bit is available to output currents less than 4 mA or greater
than 20 mA). Intelligent transmitters, however, can
communicate with the control room interactively so that
detailed status information can be polled at any time. HART
circuitry can replace an existing smart or analog transmitter
installation without having to run new cabling—a major
advantage because a large proportion of existing transmitter
installations can be upgraded simply by replacing the existing
transmitters with HART-compatible transmitters. The
intelligent transmitters also allow the performance of process
control loops to be improved. For example, the control room
can remotely “trim” the transducer output. Transmitters are
often capable of measuring two process variables (primary and
secondary process variables) instead of the single PV that can
be transmitted via a simple 4-to-20-mA interface. Intelligent
transmitters can send information on two PVs, plus other
relevant information. HART configurations can also feature
digital communication alone; the analog current is used solely
as a vehicle for the digital information.
The protocol has many layers closely related to the seven-layer
OSI model. The AD422 from Analog Devices is a solution at
the physical layer; the others are implemented in software.
15
New Product Briefs
(For information use reply card or see back cover)
Amplifiers, Buffered Switches & Muxes
4-␮A Op Amps
Ultrafast Op Amp
Dual OP281 & Quad OP481
Rail-rail outputs, 1-mV VOS
AD8009 has 1-GHz BW
256-point, 200-MHz AD8116:
5500 V/␮s SR, 175 mA Output Crosspoint has output buffers
The OP281 and OP481 respectively
comprise one and two pairs of very low
power op amps having fast recovery times
(useful in comparator applications) sink/
source outputs without increased supply
current when at the rails. Maximum supply
current is 4 µA per amplifier (5 µA from
–40 to +85°C). Operation is guaranteed with
single supply from 2.7 to 12 V, and the
devices are specified at +3, +5, and ± 5 V.
The AD8009 is an unusually fast, lowdistortion monolithic current-feedback
operational amplifier, characterized by
1-GHz typical small-signal bandwidth (G=+1),
440-MHz large-signal bandwidth (G=+2),
5500-V/µs slew rate, and 10-ns settling time to
0.1%. It has a SFDR of 44 dBc at 150 MHz
and a 3rd-order intercept of 18 dBm at
150 MHz.
The low power consumption enables
operation for years on batteries, for remote
sensors, and for applications such as security
systems and medical instrumentation. The
OP281 is available in 8-pin DIP, SOIC, and
TSSOP packages; the OP481 is available in
14-lead DIP, TSSOP and narrow-body SO.
Prices (1000s) start at $2.44 for the duals,
$3.19 for the quads.
Faxcode 2075* or Circle 7
b
6ⴛ16 Video Switch
It can furnish 175 mA of load current, will drive
4 back-terminated video loads, and has
excellent video specifications, viz., differential
gain and phase errors of 0.01% and 0.01°, with
150-Ω load. The AD8009 operates at ± 5 V,
drawing about 14 mA.Typical applications are
in pulse handling, video, RF/IF, and high speed
test and instrumentation. Housed in an 8-pin
SOIC, it operates over the –40 to +85°C range.
Price in 1000s is $2.99
Faxcode 2079* or Circle 8
b
The AD8116, the world’s first 16 × 16 video
crosspoint switch IC, and the first of a family,
routes video signals from any of its 16 inputs
to any of its 16 buffered outputs (which are 3statable and can drive 150 Ω), as programmed
by an 80-bit serial word. It has –3-dB
bandwidth of 200 MHz and switching time of
60 ns for 0.1% settling. It can route many types
of signals, including compressed or
uncompressed composite or component video
signals. Its uniquely high speed and compact
size are helpful for building large wideband
switch arrays. Applications include broadcast
video, surveillance and video conferencing,
video on-demand, and inflight entertainment.
It is packaged in a 128-lead TQFP (14 mm
× 14 mm) and is specified for 0 to +70°C. It
requires a ± 5-V supply, and dissipates only
900 mW. Prices start at $105 in 100s.
Faxcode 2070* or Circle 9
b
xDSL Driver/Receiver VGA: Linear Gain (dB) Fast Buffered Muxes
Differential 500-mA drivers,
Low-noise preamplifier pair
AD604/605 duals have low
noise, 40-MHz bandwidth
700-MHz AD8170/74 (2:1/4:1)
750-MHz ’80/’82 (2:1, 2ⴛ2:1)
The AD816 comprises two 500-mA current
drivers and two low-noise (4 nV/√Hz)
voltage-feedback amplifiers (with 70-mA
drive capability), combining drive and
preamplification on a single monolithic chip.
It functions as the line interface to and from
the telephone line in telecommunications
applications such as ADSL, providing up to
26 dBm of differential line drive. It can also
be used in CRT monitors, and its voltageFB amplifiers are useful in active filters.
The AD604 and AD605 are highperformance wideband (40-MHz) 2-channel
variable-gain amplifiers (VGA). Gain is linear
in dB as a function of control voltage. They
can be used to control signal gain &
attenuation in AGC applications or as a timegain amplifier in medical ultrasound. The
AD604, optimized for low noise (0.8 nV/√Hz
RTI), offers gain ranges from 0 through
+48 dB and +6 through +54 dB. In addition,
control-voltage sensitivity can be scaled from
20 dB/V through 40 dB/V. The single-supply
AD605 has gain ranges of –14 to +34 dB and
0 dB to +48 dB.
The AD8170, -74, AD8180, and -82 muxes
select between wide-bandwidth input
signals, and switch them in <10 ns. Typical
applications are in large video switcher and
router arrays, scanners, picture-in-picture
video, and MPEG video for PCs. All four
devices have buffered inputs & outputs. The
AD8174 switches 4 signals; the AD8170
switches 2 signals, with a 250-MHz BW
(G=+2). In both, a current-feedback
amplifier buffers the output; its gain can be
set with external resistors. The AD8180
switches two signals with a bandwidth of
750 MHz; the AD8182 is a dual 2:1.
They are available in plastic DIP and
SOIC packages for operation from –40 to
+85°C. The AD605, available in A and B
grades, has 16-pins; the AD604, operating
from ± 5-V supplies, has 24 pins. AD604/
605 price in 1000s starts at $11.90/$8.90.
Circle 11
b
Power consumed is low (± 5 V,<10 mA), but
the ’7x deliver >50 mA and the ’8x deliver
>30 mA. They operate from –40 to +85°C,
are housed in 8- and 14-lead PDIPs and
SOICs. Prices (1000s) are $1.95/$2.95
(’80/’82) and $2.25/$3.95 (’70/’74).
The thermally overload-protected driver can
apply 40 V p-p differential output to a 50-Ω
load at 1 MHz or furnish 1 A of peak current
to a 15-Ω load with a 1% duty cycle. The
AD816 is available in low thermal resistance
power packages, operates from –40 to
+85°C, and is specified with ± 15-V supplies,
drawing 59 mA max of quiescent current. It
is priced at $7.25 in 1000s.
Faxcode* 1978 or Circle 10
b
b
Circle 12
All brand or product names mentioned are trademarks or registered trademarks of their respective holders.
*For immediate data, visit our WorldWide Web site: http://www.analog.com. In North America, call ADI’s 24-hour AnalogFax™ line, 1 (800) 446-6212 and use Faxcode.
16
Analog Dialogue 31-1 (1997)
New-Product Briefs
(For information use reply card or see back cover)
A/D and D/A Converters, Volume Controls
12-Bit ␮power D/A
Two 16-Bit, +5-V ADs 8-Bit, 200 MSPS ADC
2.7 to 5.5-V, 100-µA rail-rail
AD7392/93: pin-compatible
AD7722: ⌺-⌬, serial/parallel
AD976/A: 2ⴛ faster 2nd source
AD9054 uses only 500 mW,
has 380-MHz full-power BW
With a maximum IDD of 100 µA, the 12-bit
AD7392 and 10-bit AD7393 are the newest
members of a family of DACs having ultralow power consumption. Operating on a 2.7
to 5.5-volt single supply, they offer a pincompatible choice of parallel-loading 12 or
10-bits, allowing a cost/resolution tradeoff
without board layout changes. Typical
dissipation in 3-V operation is 165 µW; in
shutdown, supply current drops to 0.1 µA.
The AD976, pin-compatible with the
ADS8705, is a 16-bit low-power +5-V
single-supply successive-approximation
ADC. It samples at up to 100 ksps, and the
AD976A version samples at 200 ksps. The
device has an on-chip clock and 2.5-V
reference, accepts inputs up to ± 10 V, and
dissipates only 100 mW. Operation is from
–40 to +85°C. Packaging is in 28-pin DIP,
SOIC, and SSOP, and prices for AD976/A
in a plastic DIP (1000s) are $20/$26.
The AD9054 is an 8-bit monolithic A/D
converter optimized for high speed, low
power, small size, and ease of use, using
an innovative bit-per-stage folding
architecture with many advantages over
traditional “flash” converters. Its output is
interleaved between two TTL-compatible
output ports at one-half the clock rate,
reducing interf ace speed and buffer
memory cost. 80-MHz SINAD (200 MSPS)
is 43 dB. Applications include RGB video
processing, digital data storage read
channels, medical imaging, direct-IF &
undersampling in digital communications.
The output range is rail to rail (0 to VDD),
and the (multiplying) reference input can be
tied to VDD. Both devices are available for
–40 to +85°C. Packaging is in 20-pin PDIPs
and SOICs; AD7393 versions are also
available in a 1.1-mm high tiny TSSOP-20,
for PCM card applications, and in a SOIC
for –40 to +125°C for automotive
applications. Prices (1000s) start at $2.95
for the AD7393 and $5.06 for the AD7392.
Faxcode* 2068 or Circle 13
b
The AD7722 is a complete sigma-delta ADC
with signal bandwidth up to 100 kHz and
output word rate up to 220 ksps (14-MHz
clock). It has an on-chip reference and can
interface in either parallel or serial mode
(a pair of converters can share the serial bus).
Input range is 0 to +2.5 V or ± 1.25 V. It
draws 375 mW, is available in a 44-pin PQFP
for –40 to +85°C. Price (1000s) is $20. For
data on both ICs, Circle 14
b
It operates on a single +5-V supply and has
an internal 2.5-V reference.The analog input
range is 1 V p-p. Two speed grades are
available in a 44-pin TQFP—200 MSPS and
135-MSPS sample rates. Respective prices
in 100s are $48 and $28.
Faxcode* 2093 or Circle 15
b
12-Bit A/D Choices
Triple 8-Bit Video D/A Vol/Bal-Control ICs
Simultaneous dual AD7862
Single serial 3.8-␮s AD7895
360-MHz, 196-bit ADV7129
Complete with PLL and V ref
6, 4-channel SSM2160, 2161
Clickless operation, serial in
The monolithic AD7862 comprises two
250-ksps 12-bit simultaneous sampling
ADCs, each with multiplexed dual inputs.
It includes an on-chip reference and parallel
digital interface and requires a +5-V supply
(60 mW dissipation, 50-µW in power-save
mode). Available grades are A & B (–40 to
+85°C) and S (–55 to +125°C). There is a
choice of ± 10, ± 2.5, and 0 to +2.5-V input
ranges. Housing is in PDIP, SO, SSOP, &
Cerdip. Prices (1000s) start at $11.
The ADV7129 is a complete analog-output
video DAC on a single CMOS chip, designed
to drive ultra-high-resolution (2 K × 2 K)
color monitors. It includes three 8-bit video
DACs (RGB), a programmable pixel port,
and an internal voltage reference. An onboard phase-locked loop clock generator
provides high-speed operation without
requiring high-speed external crystal or clock
circuitry. The 192-bit input port enables full
24-bit true color pixel data to be displayed
at pixel speeds of up to 360 MHz.
The SSM2160 and SSM2161 are complete
6- and 4-channel clickless digital volume and
balance control ICs. They accept 6 (or 4)
audio signals and either amplify (up to
31 dB) or attenuate (over 100 dB) each one.
A Master Control sets the attenuation, while
a Channel Control enables balance to be
adjusted. They are principally applied as the
main volume/balance controller for a DVD
player, multi-channel audio amplifier, hi-fi
receiver, professional sound mixer, or communications voice processor.
Its analog outputs are RS-343A/RS-170
compatible, and its digital inputs are TTLcompatible. It is fully controlled through its
8-bit MPU port by on-board command
registers. It uses a single +5-V supply, has
an internal voltage reference, and is available
in a 304-pin PQFP package for 0 to +70°C.
Its price in 100s is $765.
Audio characteristics include THD+N of
0.01% and 0.001% IMD. Both devices can
operate from single or dual supplies; e.g., the
SSM2160 operates from +10-V to +20-V
single supplies or ± 5 to ± 10-V dual supplies.
They are available in PDIP and SO packages
for 0 to +70°C operation. Prices (large
quantity) are as low as 60¢/channel.
The AD7895 is a fast 12-bit single-supply
ADC (+5-V, 20-mW) with automatic powerdown. It converts in 3.8 µs and has a highspeed, easy-to-use serial interface. It is
available in 8-pin mini-DIP and SOIC for –
40 to +85°C in A/B grades, and a choice of
± 10, ± 2.5, & 0 to +2.5-V high-impedanceinput ranges. Price (1000s) is $4.95/$6.80.
For data on both devices and 14-bit upgrades
(AD7863 & AD7894), circle 16
b
Faxcode* 1854 or Circle 17
b
Faxcode* 1848 or Circle 18
b
All brand or product names mentioned are trademarks or registered trademarks of their respective holders.
*For immediate data, visit our WorldWide Web site: http://www.analog.com. In North America, call ADI’s 24-hour AnalogFax™ line, 1 (800) 446-6212 and use Faxcode.
Analog Dialogue 31-1 (1997)
17
New Product Briefs
(For information use reply card or see back cover)
Power Management Supervisory Circuits
Step Up/Step Down Batt-Chge Controllers 1-30 Volt Supplies
ADP3000 switching regulator Secondary side off-line
Step-up, step-down, fixed or
Micropower, high-frequency
ADP3810/11: LiIon/NiCd, NiMH adjustable µpower sw. regs.
The ADP3000 is a versatile step-up/stepdown switching regulator that operates from
an input supply of 2 to 12 V in boost mode
and up to 30 V in buck mode. An adjustable
and 3 standard fixed-output versions are
available for 3.3, 5, and 12 V. The 3.3-V
version can deliver an output current of
100 mA at 3 V from a 5-V input in stepdown, and 180 mA at 3.3 V from a 2-V input
in step-up configuration.
Few external components are required; and
operation at a high switching frequency
(400 kHz) allows the use of small inductors
and capacitors. Operating in the pulsefrequency mode (PFM), it consumes only
500 µA. It has an on-chip low-battery detector,
and the output current limit is user-settable.
The ADP3000 is available in 8-pin DIP and
SO-8 packages for 0 to +70°C. Prices (1000s):
$2.14/$2.28 (PDIP/SOIC).
Faxcode* 2028 or Circle 19
b
The ADP3810 and ADP3811 precision
battery-charger controllers combine userprogrammable current limiting and precise
voltage control to charge lithium-ion, nickelcadmium, and nickel metal-hydride batteries
(popular batteries for portable computing
and communications gear). They combine
a precision 2-V reference, control input
buffer, under-voltage-lockout comparator,
output buffer, and over-voltage comparator
in an 8-pin SOIC for –40 to +85°C.
Charging currents are programmable from
typically 100 mA to 1.1 A. The ADP3810
features on-board precision resistors for
± 1% overall accuracy when charging Li-ion
batteries; it is optionally available for 4.2,
8.4, 12.6, and 16.8 V. The ADP3811 is
programmable via external resistors, for a
wide range of user-selected voltages. Prices
(1000s) are $2.57/$2.43, respectively.
Faxcode* 2069 or Circle 20
b
The ADP1110 and ADP1111 are members
of a family of step-up/step-down switching
regulators with 3.3, 5, and 10-V fixed
outputs, as well as adjustable-output
versions. The 5-V, 10-V, and adjustable
versions are pin-for pin replacements for
existing 1110 and 1111 devices; the 3.3-V
versions are currently available only from
Analog Devices. The ADP1110 has an input
supply range from 1.0 to 30-V, permitting
use on single cell batteries; the ADP1111’s
input range is from 2-12 V in step-up, and
2-30 V in step-down.
Typical applications for the ADP1110 are
in cellular telephones, pagers, cameras, laserdiode drivers, and handheld inventory
computers. Both devices are available in 8pin plastic DIPs and SOICs for operation
from 0 to +70°C. Prices in 1000s for PDIP/
SOIC are $2.42 /$2.57 for ADP1110, $1.93/
$2.07 for ADP1111. Circle 21
b
Switching Regulators ␮P Supervisory ICs
␮P Supervisory ICs
High-efficiency synchronous
ADP1148 step-down family
Low-cost ADM708ARM
Available in ␮SOIC packages
ADM690AA, ADM691AA
Now in SO, ␮SOIC & TSSOP
The ADP1148 is a high-efficiency (>95%
at 200 mA) synchronous step-down
switching regulator with operating supply
range from +3.5 V to +18 V. It is available
for 3.3-V, 5-V and adjustable outputs; the
5-V and adjustable versions are pin-for-pin
replacements for existing 1148 devices; the
3.3-V version, for use in low-voltage systems,
is currently available only from Analog
Devices. The ADP1148 switches a pair of
external complementary MOSFETs at
frequencies up to 250 kHz.
The ADM708ARM is a low-cost µP
supervisory circuit for monitoring +5-V
supplies/batteries and microprocessor
activity. It is the first of the ADM70x family
to be available in a tiny 8-pin µSOIC package
(one-half the area of an SO-8). It provides a
pre-trimmed 4.4-V power-fail reset, manual
reset de-bounce, and reset pulse-generator
for a microcontroller. It is complete, requires
no external components, and draws only
200 µA of quiescent current.
The ADM690A and ADM691A are the first
small-package versions in the ADM69x
family of microprocessor supervisory circuits.
(More are coming!) The ADM690 is
now available in 8-pin DIP (-AAN), SOIC
(-AARN), and µSOIC (-AARM); the
ADM691 is available in 16-pin DIP
(-AAN), wide SOIC (-AARW), narrow
SOIC (-AARN), and TSSOP (-AARU).
The ADP1148 has an automatic powersaving sleep-mode (2 mW at 10 V in) when
load current drops below the level required
for continuous operation. In logic-controlled
shutdown mode, only 20 µA are required.The ADP1148 is available in 14-pin
PDIPs and SOICs for –40 to +85°C.
Respective prices (1000s) are $3.18/$3.31.
Faxcode* 2023 or Circle 22
b
It can be used in all kinds of microcontroller
applications, including printers, modems,
PC settop boxes, drives, telecom cards,
routers, etc., and is well-suited for mobile
applications. Its electrical specs are the same
as those of the ADM708AR, and it is form,
fit, and function-compatible with 708CUA.
The ADM708ARM is specified for –40 to
+85°C. Price in 1000s is $0.84 (i.e., 84¢!)
Faxcode* 1866 or Circle 23
b
The functions of both devices include
Watchdog timer, V CC Reset, Automatic
backup battery power switching, and Power
fail monitoring. In addition, the AD691A’s
design includes fast onboard gating of Chip
Enable signals and low line status indication.
All A version devices allow the Vbatt terminal
to float when not connected. All versions
operate from –40 to +85°C. Prices are low,
starting at (1000s) $1.44 (ADM690AARM)
and $1.60 (ADM691AARM).
b
Circle 24
All brand or product names mentioned are trademarks or registered trademarks of their respective holders.
*For immediate data, visit our WorldWide Web site: http://www.analog.com. In North America, call ADI’s 24-hour AnalogFax™ line, 1 (800) 446-6212 and use Faxcode.
18
Analog Dialogue 31-1 (1997)
New-Product Briefs
(For information use reply card or see back cover)
Mixed Bag: Communications, Video, DSP
Low-Cost SHARC
CCD Signal Processor GSM Receiver IF
ADSP-21061: 120 MFLOPS
1 Mbit dual-ported SRAM
AD980x: 12/10-bit, 6 MSPS
AD6459 is a 3-V subsystem:
Complete on a chip, PQFP-64 Mixer, Linear amp, I/Q Demod
The ADSP-21061 is the first in a family of
low-cost SHARC™ (Super Harvard
ARchitecture Computer) DSPs. It features
a high-performance 120-MFLOPS/120
fixed-point MOPS core. Its 1 Mbit of dualported on-chip SRAM stores up to 32 K 32bit words. To eliminate bus bottlenecks, six
DMA channels support background I/O
transfers. An integrated host interface
supports 16-bit or 32-bit µP. Two high-speed
serial ports (40 Mbits/s) interface to
industry-standard peripherals. The ADSP21061 is pin-compatible with and runs the
same source code as other SHARC DSPs.
The AD9807/AD9805 is a complete singlechip charge-coupled-device (CCD) imaging
decoder and signal processor.It integrates all
of the signal conditioning and correction
needed for most mid-range scanner
applications. Its three color-component
inputs can be directly ac-coupled from the
CCD, and its output is a processed 12/10bit parallel word. Included is circuitry to
perform 3-channel correlated double
sampling (CDS) and adjustment of the
programmable gain. After conversion to
digital, on-board DSP circuitry corrects for
pixel rate offset and gain, and odd/even CCD
register imbalance errors. A parallel control
bus acts as a simple interface to µCs.
It is a high-performance solution for pricesensitive applications in audio, graphics,
communications, video, conferencing, and
imaging. Development support packages
and samples are available now. ADSP-21061
prices are $64 in 1000s, $49 in 10,000s.
Circle 25
b
Operation is specified from 0 to +70°C on
+5-volt analog and digital supplies, with
dissipation of 530 mW max and packaging
in a 64-pin PQFP. Its price in 1000s is $25.
Faxcode* 2021 or Circle 26
b
The AD6459 is a 3-V, low-power receiver
IF subsystem. Optimized for operation in
GSM, DCS1800, and PCS1900 receivers,
it is used in digital communication systems
(cellular, satellite, and private mobile radio)
to convert a signal from a high IF (up to
500 MHz) to a lower IF (5 MHz to
50 MHz), amplify the signal, and provide
demodulated I and Q outputs. It consists of
a mixer, an IF amplifier, I and Q
demodulators, a phase-locked quadrature
oscillator, a precise AGC subsystem, and a
biasing system with external power-down.
It interfaces to AD7013, AD7015, and
AD6421 baseband converters, will operate
on supplies from 2.7 to 5.5 V, and draws
8 mA at mid-gain (and only 2 µA in Sleep
mode). It is packaged in a 20-lead SSOP for
operation from –25 to +85°C (–40°C for
supply ≥3.3 V). Its price (1000s) is $5.95.
Faxcode* 2079 or Circle 27
b
Video Encoder
Pulsewidth Modulator 155-Mbps Transceiver
AD724: RGB to NTSC/PAL
PLL helps reduce system cost
AD9561: for high-resolution AD6816: A complete ATM chip
copier- & printer applications for UTP#5 copper or fiber
The AD724 is a low-cost RGB to NTSC/
PAL encoder; it converts red, green, and blue
color component signals into their
corresponding luminance (baseband phase
and amplitude) signals in accordance with
either NTSC or PAL standards. It also
makes available composite video output. All
3 outputs can simultaneously drive 75-Ω,
reverse-terminated cables. It is an improved
version of the AD722 and is pin-compatible
when its RGB inputs are properly bypassed.
The AD9561 is a second-generation highspeed, digitally programmable pulsewidth
modulator (PWM). Its output pulsewidth
is proportional to an 8-bit digital word (from
5 ns to 100% FS). Two additional control
inputs determine if the pulse is located at
the beginning, middle, or end of the clock
period. Pulsewidth and placement can be
changed every clock cycle up to 60 MHz.
The AD724 lowers system cost, by virtue of
its on-chip PLL, which enables the use of a
low-cost FSC crystal for generating all the
timing. However, it will also accept FSC or
4FSC clocks, and it can accept CSYNC or
HSYNC/VSYNC. It uses a single +5-V
supply, is packaged in a 16-pin SOIC for
operation from 0 to +70°C. Price (1000s) is
$4.55, and less than $3 in 10,000s.
Faxcode* 2056 or Circle 28
b
It is principally useful in copiers and printers
that feature very high resolution and/or
image quality (“photorealistic” is the
appropriate buzzword). Pulse modulation is
the only cost-effective method to obtain both
high spatial resolution and wide gray-scale
dynamic range (see Analog Dialogue 27-2,
pp. 10-11). The AD9561 operates from a
+5-V supply and is housed in a 28-lead
SOIC. It operates over the 0 to +70°C
temperature range. Price in 1000s is $7.40.
Faxcode* 2027 or Circle 29
b
The AD6816 provides a single-chip physical
layer solution for interfacing an
asynchronous transfer-mode (ATM) usernetwork interface IC to either a Category
#5 unshielded twisted pair (UTP) system
or a fiber optic system.The AD6816 provides
line equalization and baseline restoration,
line driver, clock recovery and data retiming,
local reference clock oscillator, and
frequency synthesis functions. Typical
applications are in network interface cards
and switching-hub multiport boards.
Complying with ATM UNI 3.1 and
SONET/SDH OC-3/STM1, it offers a
lowest power, complete solution for analog
functions, and allows partitioning between
analog and digital functions for a lowest-cost
system solution. It is housed in a TQFP-44
for 0 to +70°C and +5-V supply (95 mA
max). Price (10,000s) is $23.45.
Faxcode* 2080 or Circle 30
b
All brand or product names mentioned are trademarks or registered trademarks of their respective holders.
*For immediate data, visit our WorldWide Web site: http://www.analog.com. In North America, call ADI’s 24-hour AnalogFax™ line, 1 (800) 446-6212 and use Faxcode.
Analog Dialogue 31-1 (1997)
19
Ask The Applications Engineer—24
by Steve Guinta
RESISTANCE
Table 2. Color code for carbon-type resistors
Q. I’d like to understand the differences between available resistor types
and how to select the right one for a particular application.
A. Sure, let’s talk first about the familiar “discrete” or axial-lead
type resistors we’re used to working with in the lab; then we’ll
compare cost and performance tradeoffs of the discretes and
thin- or thick-film networks.
Axial Lead Types: The three most common types of axiallead resistors we’ll talk about are carbon composition, or carbon
film, metal film and wirewound:
• carbon composition or carbon film-type resistors are used in
general-purpose circuits where initial accuracy and stability
with variations of temperature aren’t deemed critical. Typical
applications include their use as a collector or emitter load,
in transistor/FET biasing networks, as a discharge path for
charged capacitors, and as pull-up and/or pull-down elements
in digital logic circuits.
Carbon-type resistors are assigned a series of standard values
(Table 1) in a quasi-logarithmic sequence, from 1 ohm to 22
megohms, with tolerances from 2% (carbon film) to 5% up
to 20% (carbon composition). Power dissipation ratings range
from 1/8 watt up to 2 watts. The 1/4-watt and 1/2-watt, 5%
and 10% types tend to be the most popular.
Carbon-type resistors have a poor temperature coefficient
(typically 5,000 ppm/°C); so they are not well suited for
precision applications requiring little resistance change over
temperature, but they are inexpensive—as little as 3 cents
[USD 0.03] each in 1,000 quantities.
Table 1 lists a decade (10:1 range) of standard resistance
values for 2% and 5% tolerances, spaced 10% apart. The
smaller subset in lightface denote the only values available
with 10% or 20% tolerances; they are spaced 20% apart.
Table 1. Standard resistor values: 2%, 5% and 10%
10
11
12
13
15
16
18
20
22
24
27
30
33
36
39
43
47
51
56
62
68
75
82
91
100
Carbon-type resistors use color-coded bands to identify the
resistor’s ohmic value and tolerance:
digit
–
–
0
1
2
3
4
5
6
7
8
9
–
color
silver
gold
black
brown
red
orange
yellow
green
blue
violet
gray
white
none
multiple
0.01
0.10
1
10
100
1k
10 k
100 k
1M
10 M
–
–
–
# of zeros
–2
–1
0
1
2
3
4
5
6
7
–
–
–
tolerance
10%
5%
–
–
2%
–
–
–
–
–
–
–
20%
• Metal film resistors are chosen for precision applications where
initial accuracy, low temperature coefficient, and lower noise
are required. Metal film resistors are generally composed of
Nichrome, tin oxide or tantalum nitride, and are available in
either a hermetically sealed or molded phenolic body. Typical
applications include bridge circuits, RC oscillators and active
filters. Initial accuracies range from 0.1 to 1.0 %, with
temperature coefficients ranging between 10 and 100
ppm/°C. Standard values range from 10.0 Ω to 301 kΩ in
discrete increments of 2% (for 0.5% and 1% rated
tolerances).
Table 3. Standard values for film-type resistors
1.00
1.02
1.04
1.06
1.08
1.10
1.13
1.15
1.17
1.20
1.22
1.24
1.27
1.29
1.32
1.35
1.37
1.40
1.43
1.46
1.49
1.52
1.55
1.58
1.61
1.64
1.68
1.71
1.74
1.78
1.82
1.85
1.89
1.93
1.96
2.00
2.04
2.09
2.13
2.17
2.22
2.26
2.31
2.35
2.40
2.45
2.50
2.55
2.60
2.65
2.70
2.76
2.81
2.87
2.92
2.98
3.04
3.10
3.17
3.23
3.29
3.36
3.43
3.49
3.56
3.64
3.71
3.78
3.86
3.94
4.01
4.09
4.18
4.26
4.34
4.43
4.52
4.61
4.70
4.80
4.89
4.99
5.09
5.19
5.30
5.40
5.51
5.62
5.73
5.85
5.96
6.08
6.21
6.33
6.46
6.59
6.72
6.85
6.99
7.13
7.27
7.42
7.56
7.72
7.87
8.03
8.19
8.35
8.52
8.69
8.86
9.04
9.22
9.41
9.59
9.79
9.98
Metal film resistors use a 4 digit numbering sequence to
identify the resistor value instead of the color band scheme
used for carbon types:
4991
1ST DIGIT
2ND DIGIT
MULTIPLIER
TOLERANCE
20
1ST DIGIT
EXAMPLE:
RED-YELLOW-ORANGE-GOLD
WOULD HAVE A RESISTANCE
OF 24,000Ω with 5% TOLERANCE
2ND DIGIT
3RD DIGIT
THE INDICATED RESISTANCE HERE
IS 4990Ω. FOR VALUES LESS THAN 100,
“R” INDICATES A DECIMAL POINT,
e.g., 49R9 49.9Ω
NUMBER OF ZEROS
Analog Dialogue 31-1 (1997)
• Wirewound precision resistors are extremely accurate and
stable (0.05%, <10 ppm/°C); they are used in demanding
applications, such as tuning networks and precision
attenuator circuits. Typical resistance values run from 0.1 Ω
to 1.2 MΩ.
Here are a couple of examples that make the point.
1. Assume both resistors have an actual TC of 100 ppm/°C
(i.e., 0.01%/°C). The resistance following a temperature
change, ∆T, is
R = R0(1+ TC ∆T)
High Frequency Effects: Unlike its “ideal” counterpart, a
“real” resistor, like a real capacitor (Analog Dialogue 30-2),
suffers from parasitics. (Actually, any two-terminal element may
look like a resistor, capacitor, inductor, or damped resonant
circuit, depending on the frequency it’s tested at.)
THERMOCOUPLE
EFFECTS WITH
EXTERNAL LEADS
OF DISSIMILAR
METAL
SERIES L RESISTANCE
SHUNT C
For a 10°C temperature rise, both Rf and Ri increase by
0.01%/°C × 10°C = 0.1%. Op amp gains are [to a very good
approximation] 1 + RF/RI. Since both resistance values,
though quite different (99:1), have increased by the same
percentage, their ratio—hence the gain—is unchanged. Note
that the gain accuracy depends just on the resistance ratio,
independently of the absolute values.
2. Assume that RI has a TC of 100 ppm/°C, but RF’s TC is
only 75 ppm/°C. For a 10°C change, RI increases by 0.1%
to 1.001 times its initial value, and RF increases by 0.075%
to 1.00075 times its initial value. The new value of gain is
(1.00075 RF)/(1.001 RI) = 0.99975 RF/RI
Factors such as resistor base material and the ratio of length
to cross-sectional area determine the extent to which the
parasitic L and C affect the constancy of a resistor’s effective
dc resistance at high frequencies. Film type resistors generally
have excellent high-frequency response; the best maintain their
accuracy to about 100 MHz. Carbon types are useful to about
1 MHz. Wirewound resistors have the highest inductance, and
hence the poorest frequency response. Even if they are noninductively wound, they tend to have high capacitance and are
likely to be unsuitable for use above 50 kHz.
Q. What about temperature effects? Should I always use resistors with
the lowest temperature coefficients (TCRs)?
A. Not necessarily. A lot depends on the application. For the single
resistor shown here, measuring current in a loop, the current
produces a voltage across the resistor equal to I × R. In this
application, the absolute accuracy of resistance at any
temperature would be critical to the accuracy of the current
measurement, so a resistor with a very low TC would be used.
For an ambient temperature change of 10°C, the amplifier
circuit’s gain has decreased by 0.025% (equivalent to 1 LSB
in a 12-bit system).
Another parameter that’s not often understood is the selfheating effect in a resistor.
Q. What’s that?
A. Self-heating causes a change in resistance because of the
increase in temperature when the dissipated power increases.
Most manufacturers’ data sheets will include a specification
called “thermal resistance” or “thermal derating”, expressed
in degrees C per watt (°C/W). For a 1/4-watt resistor of typical
size, the thermal resistance is about 125°C/W. Let’s apply this
to the example of the above op amp circuit for full-scale input:
Power dissipated by RI is
E2/R = (100 mV)2/100 Ω = 100 µW, leading to a temperature
change of 100 µW × 125°C/W = 0.0125°C, and a negligible
1-ppm resistance change (0.00012%).
Power dissipated by RF is
I
A different example is the behavior of gain-setting resistors in
a gain-of-100 op amp circuit, shown below. In this type of
application, where gain accuracy depends on the ratio of
resistances (a ratiometric configuration), resistance matching,
and the tracking of the resistance temperature coefficients
(TCRs), is more critical than absolute accuracy.
RI = 100Ω
INPUT
0-100mV
RF = 9.9kΩ
0-10V
TO A/D CONVERTER
G = 1 + RF /RI = 100
Analog Dialogue 31-1 (1997)
/R = (9.9 V)2/9900 Ω = 9.9 mW, leading to a temperature
change of 0.0099 W × 125°C/W = 1.24°C, and a resistance
change of 0.0124%, which translates directly into a 0.012%
gain change.
2
R
Thermocouple Effects: Wirewound precision resistors have
another problem. The junction of the resistance wire and the
resistor lead forms a thermocouple which has a thermoelectric
EMF of 42 µV/°C for the standard “Alloy 180”/Nichrome
junction of an ordinary wirewound resistor. If a resistor is
chosen with the [more expensive] copper/nichrome junction,
the value is 2.5 µV/°C. (“Alloy 180” is the standard component
lead alloy of 77% copper and 23% nickel.)
Such thermocouple effects are unimportant in ac applications,
and they cancel out when both ends of the resistor are at the
same temperature; however if one end is warmer than the other,
either because of the power being dissipated in the resistor, or
21
its location with respect to heat sources, the net thermoelectric
EMF will introduce an erroneous dc voltage into the circuit.
With an ordinary wirewound resistor, a temperature differential
of only 4°C will introduce a dc error of 168 µV—which is
greater than 1 LSB in a 10-V/16-bit system!
This problem can be fixed by mounting wirewound resistors
so as to insure that temperature differentials are minimized.
This may be done by keeping both leads of equal length, to
equalize thermal conduction through them, by insuring that
any airflow (whether forced or natural convection) is normal
to the resistor body, and by taking care that both ends of the
resistor are at the same thermal distance (i.e., receive equal
heat flow) from any heat source on the PC board.
Thin film on Good matching (<0.01%)
Often large geometry
ceramic
Good TC (<100 ppm/°C)
Good tracking TC (2 ppm/°C)
Moderate cost
Laser-trimmable
Low capacitance
Suitable for hybrid IC substrate
Thin film
Good matching (<0.01%)
on silicon
Good TC (<100 ppm/°C)
Good tracking TC (2 ppm/°C)
Moderate cost
Laser-trimmable
Low capacitance
Suitable for hybrid IC substrate
Table 5. Substrate Materials
Substrate Advantages
Glass
Low capacitance
AIRFLOW
AIRFLOW
RIGHT
Ceramic
WRONG!
Q. What are the differences between “thin-film” and “thick-film”
networks, and what are the advantages/disadvantages of using a
resistor network over discrete parts?
A. Besides the obvious advantage of taking up considerably less
real estate, resistor networks—whether as a separate entity, or
part of a monolithic IC—offer the advantages of high accuracy
via laser trimming, tight TC matching, and good temperature
tracking. Typical applications for discrete networks are in
precision attenuators and gain setting stages. Thin film networks
are also used in the design of monolithic (IC) and hybrid
instrumentation amplifiers, and in CMOS D/A and A/D
converters that employ an R-2R Ladder network topology.
Silicon
Sapphire
IC INSTRUMENTATION
AMPLIFIER
VIN –
Table 4. Resistor Networks
Advantages
Low cost
High power
Laser-trimmable
Readily available
Thin film on Good matching (<0.01%)
glass
Good TC (<100 ppm/°C)
Good tracking TC (2 ppm/°C)
Moderate cost
Laser-trimmable
Low capacitance
22
Disadvantages
Fair matching (0.1%)
Poor TC (>100 ppm/°C)
Poor tracking TC
(10 ppm/°C)
Delicate
Often large geometry
Low power
R2
R3
A1
R1
RG
Thin film networks are moderately priced and offer good
matching (0.01%), plus good TC (<100 ppm/°C) and tracking
(<10 ppm/°C). All are laser trimmable. Thin film networks are
manufactured using vapor deposition.
Tables 4 compares the advantages/disadvantages of a thick film
and several types of thin-film resistor networks. Table 5
compares substrate materials.
Low capacitance
Suitable for hybrid IC
substrate
Suitable for monolithic Low power
construction
Capacitance to substrate
Low capacitance
Low power
Higher cost
In the example of the IC instrumentation amplifier shown
below, tight matching between resistors R1-R1', R2-R2', R3R3' insures high common-mode rejection (as much as 120 dB,
dc to 60 Hz). While it is possible to achieve higher commonmode rejection using discrete op amps and resistors, the
arduous task of matching the resistor elements is undesirable
in a production environment.
Thick film resistors are the lowest-cost type—they have fair
matching (<0.1%), but poor TC performance (>100 ppm/°C)
and tracking (>10 ppm/°C).They are produced by screening
or electroplating the resistive element onto a substrate material,
such as glass or ceramic.
Type
Thick film
Disadvantages
Delicate
Low power
Large geometry
Large geometry
A3
R1'
R2'
R3'
A2
VIN +
Matching, rather than absolute accuracy, is also important in
R-2R ladder networks (including the feedback resistor) of the
type used in CMOS D/A converters. To achieve n-bit
performance, the resistors have to be matched to within 1/2n,
which is easily achieved through laser trimming. Absolute
accuracy error, however, can be as much as ± 20%. Shown here
is a typical R-2R ladder network used in a CMOS digitalanalog converter.
b
R
R
2R
R
2R
R
2R
R
2R
2R
RFB
Analog Dialogue 31-1 (1997)
Worth Reading
SERIALS
COMMUNICATIONS DIRECT—Systems and IC solutions
for demanding markets: Volume 2, No. 2 (December, 1996,
8 pages). Features articles on Broadband, low-cost ADSL chipset,
cable modems. And more . . . FREE, Circle 31
APPLICATION NOTES
Amplitude modulation of the AD9850 Direct Digital
Synthesizer, by Richard Cushing [2 pp., AN-423]. Circle 32
High-speed data acquisition using the RTI®-2100 Series,
by Jim Maxwell [8 pp., AN-421]. Circle 33
NEW DATA SHEETS
AD8400/AD8402/AD8403 1-/2-/4-channel digital potentiometers.
Circle 34
AD73311 Low-cost, low-power CMOS general-purpose analog
front end. [Codec for consumer and telephony applications
includes 16-bit A/D and 16-bit D/A channels.] Circle 35
ADDC02808PB 28-V, 200-W pulsed DC-DC converter with
integral EMI filter. Circle 36
ADP1173 Micropower DC-DC converter. Circle 37
ADP3302 High-precision anyCAP™ dual low-dropout linear
regulator. Circle 38
OP184/OP284/OP484 Precision rail-to-rail input & output single/
dual/quad operational amplifiers. Circle 39
RTI-2100 Series Real-time data-acquisition system and RTI2100 Series Software Software support—DOS and Windows.
Circle 40
5B08/5B08-MUX 5B Series: 8-channel backplanes. Circle 41
IN THE LAST ISSUE
Volume 30,Number 4,1996,24 Pages
For a copy, circle 42.
Editor’s Notes, Authors
Single- and dual-axis micromachined accelerometers (ADXL150, ADXL250)
EMC, CE Mark, IEC801 . . . What’s it all about?
Integrated digital video encoders—studio quality video at consumer video prices
Selecting mixed-signal components for digital communication systems—II
Voltage regulators for power management
New-Product Briefs:
Three new op amp families
A/D and D/A converters
DSPs and Mixed-signal processors
Mixed bag: Circuit protectors, Temperature to current, Switched-cap regulator
Ask The Applications Engineer—23: Current-feedback amplifiers—II
Worth Reading, More authors
Potpourri
NEW PATENTS [not available from Analog Devices]
5,568,145 to David Reynolds for MOS current source layout
technique to minimize deviation • • • 5,572,166 to Barrie
Gilbert for Linear-in-decibel variable-gain amplifier • • •
5,574,392 to Edward Jordan for Asymmetrical ramp generator
system • • • 5,574,454 to James Wilson and Ronald Cellini for
Digital phase-locked loop utilizing a high-order sigma-delta
modulator • • • 5,578,224 to Theresa Core for Method of
Analog Dialogue 31-1 (1997)
making micromachined device with ground plane under
sensor • • • 5,583,290 to Stephen Lewis for Micromechanical
apparatus with limited actuation bandwidth • • • 5,583,713
to Peter Real, Mairtin Walsh, Kenneth Deevy, Patrick Griffin, and
Philip Quinlan for Pipelined demodulation and ADC
conversion scheme for disk drive servo system • • • 5,585,757
to Douglas Frey for Explicit log-domain root-mean-square
detector • • • 5,587,689 to Derek Bowers for Voltage controlled
amplifier with a negative resistance circuit for reducing
nonlinearity distortion • • • 5,589,785 to Patrick J. Garavan for
Low-voltage CMOS comparator • • • 5,589,791 to Barrie
Gilbert for Variable gain mixer having improved linearity and
lower switching noise • • • 5,589,792 to A. Paul Brokaw for
Resistor programmable temperature switch • • • 5,591,996
to Geoffrey Haigh and Scott Munroe for Recirculating charge
transfer magnetic field sensor • • • 5,592,120 to Wyn Palmer
and Fernando Viana for Charge pump system • • • 5,594,266 to
David Beigel, William Krieger, and Susan Feindt for Integrated
circuit (IC) with a two-terminal diode device to protect
metal-oxide-metal capacitors from ESD damage.
b
MORE AUTHORS (continued from page 2)
Jim Ryan (page 13) is a Senior
Engineer with the Applications
Group in Limerick, Ireland, where
he provides customer support and
is involved with new-product
definition for CMOS converter
products. He holds a BSc. in
Electronic Systems and an MEng.
in Computer Systems, both from
the University of Limerick, and has
worked previously as a designer of laboratory instrumentation. In
his leisure time he enjoys rugby and hill-walking.
Albert O’Grady (page 13) is a
member of the Applications Group
at Analog Devices, in Limerick
Ireland, in support of both analogto-digital and digital-to-analog
converter products. He holds a
BEng. from the University of
Limerick. In his spare time, Albert
enjoys reading and plays
badminton and tennis.
Dave Robertson (page 7) is a design engineer in the Analog
Devices High-Speed Converter group in Wilmington, MA. His
photo and a brief biography appeared in Analog Dialogue 30-3.
Steve Guinta (page 20) is the Technical Training Manager for
ADI’s Central Applications Group in Wilmington, MA. His photo
and a brief biography appeared in Analog Dialogue 30-2.
b
Notice: In the last issue (Analog Dialogue 30-4), the article “Voltage Regulation
for Power Management” (pp. 13–15) showed a circuit to maintain two regulated 3-V outputs, using a Lilon battery (Figure 2). The circuit was designed
and built by Hendrik Santos. The ADP3000, used in that circuit, was designed by James Ashe; and the ADP3302 was designed by A. Paul Brokaw
and Giovanni Pietrobon.—Ed.
23
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