DRAKE | 4000 series II | Intercom Technology in 2005/2006

Intercom Technology in 2005/2006
Intercom Technology in 2005/2006
In July 2004, Vitec Group plc, the owners of the Drake 4000, Clear-Com
Matrix Plus 3 and Eclipse intercom brands, merged under one Vitec
Group Communications banner. This is just the latest chapter in the
developments of the former Drake and former Clear-Com matrix
products. In this article we explore how the 4000 and Eclipse systems
came to being the leading intercom brands and reveal the next
developments in intercom for 2005/6.
Intercom is needed where many people work together in some common
project, such as making a television programme or directing a live event. For
this to happen smoothly, so that responses to questions, commands, and
instructions pass speedily between teams, the use of telephones soon
becomes impractical. Consider a situation in which the Director is already in a
call and blocks another incoming call from the studio Floor Manager to alert
the production control room that a guest is not in place, and the Presenter
should go to another item. This case of blocking highlights the difference
between a telephone system and an intercom system. Unlike a telephone
system, the intercom system allows multiple callers to get to a destination,
indicates who they are, and provides selective answering.
Two dominant types of intercom systems exist, the party-line and matrix
systems. In a party-line system, a common 2-wire cable ring connects
together low-cost single-key user stations that conference with each other.
The ring connects everyone together efficiently but disallows simultaneous
single point-to-point private calls on the same cable. The 2-wire party-line
system is prone to noise, and cable interfaces have to be tuned to give the
best quality of service.
In a matrix system all users have multiple keys to talk to selected destinations
independently. The cable interconnections are based on balanced duplex 4wires and this makes the cables less prone to interference without any
requirement to tune interfaces. Each user has a panel, star cabled to a central
matrix, which means that each user can send and receive personalised audio
commands. The matrix of inputs to outputs provides a manageable means of
providing quite complex communication options that would not be available in
a party-line system.
In 1993, the world’s first digital matrix intercom, the Drake 3000, was built on
the heritage of the analogue systems before it. Drake were early pioneers in
taking the “wire-per-crosspoint” hard-wired custom matrices and applying
manufacturing standardisation to give broadcast customers the functionality of
a custom wired intercom but in a repeatable modular system. Early hard-wired
systems in the 1970s and 80s used multi-core wired connections from the
user’s panel keys to the control switches in a central equipment rack. Each
key from every user panel sent an earth-free 0-volt switching contact to close
an input audio gate to a destination audio driver. Equipment racks of multichannel circuit cards provided input buffering, level control, switching,
crosspoint logic, output audio and voltage distribution. The analogue Drake
600 series successfully modularised these functions so that customers could
tailor their complex intercom requirements into a set of standard cards and
frames. These wired systems were characterised by a lack of a single fixed
audio distribution bus and relied on crosspoint cards mixing to multiple virtual
earth connections on the output audio cards.
Changes to these hard-wired systems were very cumbersome and required
delicate wiring changes at the back of inter-linked card frames. The whole
game changed with the introduction of the PC to the broadcast intercom
industry. In 1987 Drake were the first to provide a PC GUI (Central
Configuration Facility) to allow users to quickly define and edit a map, or
crosspoint database, without resorting to point-to-point wiring. The newly
developed Drake 6000 series used a Motorola 68000 Central Processor unit
which now defined the routing between synchronised inputs and output audio
bus time slots through a 16-channel multiplexed routing switch card. The
panel key control system now used RS422 serial data format from the user
panels, and dual CPUs provided a level of redundancy not seen in the earlier
hard-wired systems. In both the Drake 6000, and later in the Clear-Com
Matrix Plus, for the first time intelligent user panels had CPU and memory,
giving updateable displays allowing the re-labelling of keys from the central
configuration program.
From the 6000 digitally controlled analogue routing system, Drake Electronics
Ltd developed the 3000 system which was the first in the world in 1992 to
offer a truly digital Intercom. In 3000, the 16-bit digitised intercom audio was
routed digitally from input source to output destination busses through a fourway fast digital audio router, the FRM. The card had a MDAC to provide
digitised audio gain and the use of four routers allowed very short through
times needed to accomplish all system routing in less than 35 ms. Every
“crosspoint” was now an instruction to route input sample(s) from a TDM
source bus to a TDM slot in the output bus with a gain setting coefficients.
The digitisation of audio into addressable digital audio samples made a
considerable difference to the established analogue routing found, for
example, in the Drake 6000, which it replaced. For the first time a 3000 series
digital 128 x128 port matrix was available in just 9 rack units, while previously
such a matrix had required at least one equipment bay. Also the separation of
the routing function from the input and output gave for the first time the
possibility of backing up the routing to give crosspoint redundancy. This was
in addition to making the control system redundant, an industry practice even
The Drake 3000 series extended its claim to being “truly digital” in first
providing user panel connections to the central matrix over co-axial cable.
Diagram: Clear-Com Eclipse matrix – the newest and most advanced digital matrix on the
market today
The multi-core cables in CPU-driven analogue systems needed at least four
pairs to get the serial data and analogue audio to the main frame. In the 3000
series matrix, digitised audio between the matrix and panels used a bidirectional stereo AES/EBU format with key and display control data
embedded in the user bits. This provided a quick connection using industry
standard 75-ohm video cable and gave the additional benefit of two audio
mixes back to the user panel in one cable. A panel user could have a
separate mix of intercom from cue programme but use just one co-ax cable.
During these advances in intercom digitisation there were also changes in
production techniques throughout the broadcasting world. From single studios
handling all in-house and incoming remote communications came larger
productions with multiple outside studios intelligently networked to provide for
large sports or election-type programmes. At the same time developments in
telecom technology allowed audio and data connections to be managed
through the public switched networks. By the time the 3000 series had
evolved into the early 4000 series 1 (1998) provision had been made to
provide connections to ISDN and soon after, E1.
In a networked intercom system, a user can make calls out of the local matrix
through a distant remote matrix to another user. The routing not only of audio,
but also of control data, provides users with both call signalisation and key
labels needed to identify incoming calls. In intelligently linked systems, like the
Drake 3000 and 4000 series and the Eclipse, this routing is managed through
a central PC application setting inter-matrix connections by audio trunks and
linking the PC and matrices over a common data LAN. The trunk lines
transport individual “conversations” between sites. A local Director could send
production direction commands over a single audio trunk line to many users in
a remote system. The use of trunk lines to take pre-mixed audio between one
to many users provides a useful cost efficiency that allows telecom lines to be
Until the advent of ISDN in the mid ‘90s all remote trunking was done through
2-wire conversion of the audio over PSTN and either separate serial data by
modem or by multiplexing audio over data through single high-speed
modems. The latter operation worked well if bandwidth was available. The
advent of dual 64-Kbps ISDN systems offered reliable dial-up communications
with both audio and data in single 64-Kbps streams. The Drake VeNiX
system, which is offered with the 3000 and 4000 systems, provides two
intelligent trunk lines per ISDN-2 service, one per bearer channel. The VeNiX
ISDN CODEC uses fast G.722 7-KHz audio compression and serial data
multiplexed into a single 64-KB ISDN bearer. This gives a dial-up bidirectional trunk line.
G.711= 3.1 KHz BANDWIDTH
Diagram: VeNiX data architecture or network
The central matrix processor switches the ISDN line according to the remote
trunk requirement, auto-dialling the destination VeNiX CODEC as required.
ISDN dialling is very fast at less than 2 seconds to establish a line. The 4000
system code can hold the line open for a configurable amount of time after the
last trunk request, giving near instantaneous remote communications on
demand. The dial-up ISDN telecom trunk lines save costs on any nailed up
digital system providing a very reliable, efficient, and high quality of service
link. Configuration through the Windows© GUI in CMAPSi is simply a table of
potential links with telephone numbers that the system dials on remote calls.
Conferences, or internal matrix party lines, have been used since the earlier
days to reduce the need to make multiple manual connections when all users
need is an “all call” key to talk and listen to a non-predefined group. The
conference is likened to a “software” bus that users can key to talk and listen
to with the proviso that they hear everyone else but themselves. With the
move to extended TV productions, networking the conferences between
systems is essential. The ISDN dial-up conference operation is managed by a
central matrix dialling up the others for a system-wide conference on demand.
Although ISDN provides a cost effective trunking system for low-to-medium
traffic between remote systems it becomes less cost effective for high traffic
connectivity. The 4000 series II provides a 30-channel E1 connection when
the traffic demands many connections between systems, either over a
telecomm link or directly. The Drake 4000 Hi-Que and Eclipse E-Que cards
provide dual 30 x G.722 audio trunk lines per card. The E1 signal is a 2-Mbps
32 x 64KB data stream. Again this uses G.722 fast audio compression, and in
its simplest form provides 30 trunks over a single CAT5 cable. The E1 is a
widely available telecomm digital standard that is often used between PABX
switches. The wide availability, low latency, and plentiful third-party device
support makes the Hi-Que E1 card the connection choice in many site-to-site
Redundancy is important considering that all 30 channels can be lost when a
single E1 is removed. With the 4000 and Eclipse E1 cards, the second E1 port
can be made to mirror the primary E1 port. In the event that the primary E1
link is broken, the secondary E1 link takes all the routes between the same
destinations. The diagnostic systems will then report such failure. The E1
system carries a synchronising signal that can be used for failure recognition.
One such third-party operation with E1 is to provide TDM audio over IP. IP
communication offers the potential for low-cost wide communications. In
practice the widely available Voice over IP provides packetised digital audio
and data over an enterprise WAN or Internet.
The Clear-Com VoICE VoIP interface provides up to four channels of panelto-matrix and/or matrix-trunk to matrix-trunk communication over IP. In wellmanaged and controlled bandwidth, VoIP can offer very acceptable results.
Although in most cases VoIP suffers from IP packet loss, jitter and latency,
with careful design these can be much reduced by modern techniques.
Packet losses through queuing in routers can be kept to a minimum but losses
cannot usually be regained without adding delay. It is not unusual to have
packet losses of 30 ms per 5 seconds. The VoICE interface uses type-ofservice control bits to help lower packet losses and jitter by providing audio
priority in the network. If latency, the delay in transmission, increases beyond
50 ms one way, intelligibility can be prevented, especially when many IP audio
paths are mixed into conferences. VoICE uses real-time protocol with type-ofservice bits in each channel to minimise the latency in real networks. VoICE
also uses digital echo cancellation technology to dynamically configure the
high bandwidth duplex routes to offer real improvements over conventional
communication over IP. VoICE can be used comfortably for all non-critical
intercom. This can include studio centre to remote bureau, inter-area
intercoms, remote operational centres and off-line production centres.
Where more critical communication is required, and IP links exist, a
predictably more stable and lower latency operation than VoIP is TDM over
IP. In this multi-channel trunk operation the 30 E1 audio/data channels from
the Clear-Com matrix are parallel processed with a single IP overhead into
TDMoIP streams. The jitter between adjacent audio channels is much lower,
the bandwidth at G.722 coding is twice VoIP toll quality, and the delays are
lower and predicable. It has been demonstrated that TDMoIP can offer much
better audio quality for remote conference working and is fast becoming the
choice for low-cost multi-channel trunk operations. Telecom E1 standard IP
muxes offer up to 4 E1 ports per IP stream, and as VoICE interfaces also
bridge the Ethernet LAN control data over the same system. This solution
provides the customer with a low-cost centre-to-centre communication system
where many users can communicate over IP trunks with low and predictable
delay that enables conferencing to be achieved.
The use of Telecom standards offers intercom manufacturers such as Vitec
Group Communications easier connection at lower costs through common
standards and wide IC support than by making proprietary solutions.
However, in the field of highly available interconnections where the matrix
frames are not required to operate over telecom lines and many matrix-tomatrix links are required, single high-bandwidth fibre links prevail.
The fibre connection of matrix audio busses provides the possibility to link
multiple matrix frame audio back plane busses into one configurable fibre bus.
For the Vitec Group Communications Eclipse and 4000 matrices, the FibreNet system offers more than 1300 fibre audio time slots throughout the matrix
network which may encompass up to 127 matrix frames. Simple single or
multi-mode fibre links can be duplicated into a countra-rotating dual-concentric
fibre connection offering full redundancy and self-repairing highly available
One of the advantages to the trunk systems when compared to highly
connective fibre systems is that they are efficient in processor overheads,
easily managing the relatively few interconnections. Fibre connectivity
demands a great deal more from the central processing and for this reason
high-specification leading-edge processing is now required to provide reliable
fibre connectivity without having the systems real-time resources collapse.
The Fibre cards for the Eclipse and 4000 use highly capable processing to
provide both control and audio facilities, such as pre-mixing with level control,
into the fibre network. Using the strengths of the 4000 trunking in combination
with the Fibre bussing allows further savings in processor overhead by
premixing multiple audio inputs into a smaller number of timeslots. This can
be used when a remote panel user wants to listen to a collection of audio
ports at one listening station. The user will be able to make level adjustments
to the mix and these will be sent as gain coefficients to the FPGAs on the
Fibre bus cards.
Fibre connectivity requires similar redundancy to that provided with the HiQue E1 trunks. In the 4000 and Eclipse Fibre-Net systems, the Fibre cards
have dual-fibre transceivers which can be wired to fibre pairs in contra-rotating
busses. In this way the fibre rings can be segmented in failure modes and still
perform. Here are some redundancy examples:
If a single fibre connection is lost on the ring, the matrix nodes adjacent to
the failure will loop-back their connections to the failed cable, healing the
If a matrix node is lost on the ring due to a fibre-optic linking card failure,
the nodes adjacent to the failed node will loop-back their connections to
the failed node, healing the ring.
If two adjacent fibre connections are lost on the ring, this will be handled
as for the loss of a single node, where the nodes adjacent to the failed
node will loop-back their connections to the failed node, healing the ring.
If two non-adjacent fibre connections are lost on the ring, the nodes
adjacent to the failures will loop-back their connections to the failed cables,
healing the ring into two separate smaller rings.
If two adjacent nodes are lost on the ring, this will be handled as for the
loss of a single node where the nodes adjacent to the failed node will loopback their connections to the failed nodes, healing the ring.
All Fibre-optic cables
are Duplex
All Fibre-optic cables
are Duplex
Matrix Frame
Diagram: 4000 and Eclipse Fibre system architecture showing dual concentric rings and
“spurring” to another ring.
The possibility to “spur” a fibre ring from any one of the matrix nodes in order
to add a new matrix into the network—for example, to add an outside
broadcast vehicle into the studio system without breaking the network, is an
important feature of the Fibre-Net system. In this case, one matrix operates as
a gateway by hanging on both rings at the same time. The Eclipse matrix, for
example, will then seamlessly cross-connect the audio from one ring to the
other as required.
In these large communication systems, perhaps providing an inter-area
intercom throughout a studio complex or over multiple sites, there is a
requirement that managing the connections effectively is relatively easy. The
use of on-line supervisory diagnostics and monitoring activities helps to
provide for central support. The 4000 system provides a Supervisor panel
operation that allows a target users LCD key panel to copy its control surface
to the Supervisor panel together with the matrix which copies all audio to the
target to the Supervisor panel also. In this way the supervisor can make
adjustment through the LCD key panel’s audio local assignment and mix
controls for the target user, at the same time as listening to the targets audio
Diagram: Refresh LCD key panel – supervisor or target panel
Of course it would be very useful if all of these “panel facilities”, such as
labelled keys, local access to all ports, and call signalisation, were available in
a wireless system. With FreeSpeak™, the Clear-Com
and Drake Digital Wireless system, they are. The
FreeSpeak system provides full-duplex digital audio and
data over a licence-free DECT link from active antennas
from the matrix through E1 distribution. The DECT radio
space is divided into 240 timeslots in the band 1881.792
to 1897.344 MHz. Each audio route uses two timeslots,
four for a full send-and-return duplex communication.
This gives 60 channels divided across 10 active
antennas. Each active antenna handles five beltpack
attachments plus an extra make-before-break hand-over
channel, six channels each.
The unique feature of FreeSpeak is that beltpack users can freely roam
across all available antennas so that for the first time a wired panel user can
key directly to a wireless beltpack by name and the system follows that
beltpack no matter where it is. This is a Vitec Group Communications
patented feature called “Dynamic Port Allocation”.
The central configuration of these key elements in a modern intercom system
—panel key assignments, port grouping into fixed or dynamic groups,
conferences/party lines, and networking through trunks and/or fibre—is vital to
gaining acceptance with the end user. The configuration GUI must give the
non-frequent user the confidence to understand, explore, and make changes
whilst the system is in use. Vitec Group Communications have released the
ECS software platform for both Eclipse and soon 4000 users, which provides
user familiarity through a Windows© XP explorer look and feel. The ECS
software provides graphical panel assignment that lends itself to intuitive click
and drop mouse operation and allows the panel’s controls to be driven from
the PC screen for centralised supervisory control.
Diagnostics are very important in potentially large fibre or trunked networks.
ECS provides full status and error reporting through to the logging PC server.
Client PCs, logged on to the central server, can run history event logs or
search for particular events from the central database. To aid remote
diagnosis, a web client sits on top of the server and allows users to gain
access to the event and status logs over the Internet. The central server
connects to the 4000 (with the latest Processor Cards) or Eclipse systems
over a Dual Ethernet LAN for redundancy in the control sub-systems.
The VGC 24x7 customer support operation makes good use of these remote
diagnostics, providing another level of customer confidence and service
beyond the office hours telephone help line.
The ECS software design lends itself to adding modules for specific tasks.
The existing 4000 Master Control Room lines management module, MCR, will
migrate into the ECS system offering fast assignment of 4-wire audio, AES
audio, and ISDN lines to remote panels completely online.
Diagram: ECS Panel key assignment page showing click and drop key and on-line PC
supervisor capability
The ability to modify sub-systems within an intercom system has made the
Drake 4000 very compliant with customers’ requests for targeted solutions.
The evolution of this customisation led to the Gemini 4000 product for Air
Traffic Control, now a successful branched development at VGC.
Today’s requirements for intercom have come a long way in 20 years.
Modern systems such as the Eclipse and 4000 provide users with immediate
response to system changes, full diagnostics, low-cost routing across telecom
systems, convergence with Ethernet LAN systems, digital wireless mobility,
and smaller nodes sitting within a fibre network, but all sharing the total
system’s resources. With all the changes in the media and arts we can expect
only the most flexible systems with these high intra-system and external interconnectivity options to survive.
Simon Browne, Matrix Product Manager
Vitec Group Communications
Cambridge Research Park
Cambridge, United Kingdom
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