top of page

The #5 Crossbar Telephone Exchange

Preamble

 

The Western Electric (AT&T) Crossbar exchange comes in many colors. First, there was the No. 1 Crossbar system installed in 1938, Troy Avenue, Brooklyn NY. In 1941 Bell developed the XBT (XBar Tandem) toll switching system, based on the #1 Crossbar. A Tandem office was a kind of regional switching center, routing calls between end offices.

 

The #4 Crossbar toll center was developed to meet the need for more long distance calling in 1943. Next, the #5 Crossbar exchange was developed and installed in Media, Pennsylvania in 1948.

Crossbar switches in a telephone  exchange

Fig 1, Crossbar switches in an exchange [Davis]

Wait, don’t let this crossbar word salad scare you off. The compilation reveals the large variety of systems all with the moniker “Crossbar system or exchange” that Bell engineers developed from 1933 to 1948 and beyond. This is a short list with many derivatives not listed. 

 

The goal of these pages is to focus on just one system, the Bell System’s #5 Crossbar. Understanding how a call progresses from start to finish on one system provides important insights into all the other crossbar designs.

 

Background

 

By 1933 the panel exchange had been providing service in North America metro areas for ~12 years with good success. However, panel had several disadvantages. First, maintaining the motor driven line finders, selector switches, and sequence switches was a constant challenge. Second, the difficulty in manufacturing many unique parts (replacement parts too) was becoming very costly. These, and other reasons, moved engineers to look for alternative switching methods. 

 

So, at that time, exploratory work on the application of crossbar switching for Private Branch Exchanges (PBX), local, and toll switching systems was started by Bell engineers.  The crossbar exchange was built on the foundation of the crossbar switch.

 

The crossbar switch is a remarkable invention with an interesting development story. See Crossbar Switch for a summary of the switch’s history and how it operates, with explainer videos. It became an international success with the L.M. Ericsson company being a major innovator and provider of switches and crossbar exchanges worldwide.  See Endnote 2. 

 

The No. 1 Crossbar was the forerunner of the No. 5 Crossbar office. No. 5 differs from the No. 1 with an improved switching plan, versatility, and maintenance approach. They each had their own sweet spot subscriber size.

 

Of all the large switches used in telephone exchanges, the crossbar device has a minimum of mechanical movement to close a connection. Compare this to panel, rotary or SxS switch movements. Small movements yield big maintenance cost savings!

 

Common Control

 

One of the primary design features was the extensive use of common control, more so than in the 7A rotary and panel systems.  "Subscriber dialing did not immediately and directly control the switches. Rather a relay-based “marker”, the brains of the outfit, controlled all the crossbar switches based on the dialed digits and other factors.

 

There are several different types of markers too (dial tone marker, completing marker, …). For this coverage, marker is used broadly to encompass them all. Also, there may be ten markers in an office to share the load.

 

A marker could have 1,500 relays and was the most advanced “logical thinker” of its day (see Endnote 1 -relay computers). The sequence switch (used in panel, rotary) was replaced by relay logic circuits spread among the common control circuits.  

 

During call establishment, the marker, active for less than one second, makes decisions and sets or “marks” the switch paths.  A marker may be invoked at least twice to help establish the talking path. Standing before a live marker during heavy calling (a chorus of chunk, chunk, chunk…) is very impressive and noisy! [Adam]

Marker relays in 4 racks in a #5 Crossbar office

Fig 2, Marker relays in a #5 Crossbar office

The marker’s tentacles reached most system components. Without explaining any of the detail in Fig 3, just notice the controlling pathways from the marker to other components.  Markers rule!

Number 5 crossbar diagram showing the Marker  control paths

Fig 3, #5 Xbar diagram showing marker control focus [Adam]

The switch fabric

 

A single crossbar switch may be duplicated and arranged into a much larger switch fabric and thereby create an uber switch. Fig 4 shows a basic idea of a fabric switch.

 Three stagecrossbar  switching fabric

Fig 4, Three stage crossbar switching fabric [Fundamentals]

A very basic call connection is shown. There are nine, 3x3 crossbar switches in this fabric.  Notice that the call’s talking path may route through a switch twice. Panel, rotary and SxS certainly don’t use this approach.

 

This switch example is not optimized, and some calls could be blocked under heavy calling. In aggregate, this is an equivalent 9x9 switch function but not identical to a 9x9 monolithic switch.  See fabric for more on the blocking problem. In a crossbar exchange the marker is responsible for closing/opening individual fabric switches to make the talking paths. 

 

In Fig 3, the Line Link and Trunk Link Frames are crossbar fabrics and may each have hundreds of 200-point crossbars.

 

Making a call on a #5 Crossbar system

 

Physically observing a call’s progress through a Step-by-Step exchange is very interesting. This author has traced call progress – switch by switch – on a 4-digit SxS PBX during a low traffic quiet time. Some readers may have done the same on a panel or 7A rotary system.

 

In-person call progress tracking is a fun exercise. It brings the switching theory to life when you see the switches work as you expect them to. However, doing this with a Crossbar system is boring! There is very little physical movement of the crossbar switches. Plus, a marker’s decisions for switch control are not easily predictive compared to the more deterministic switch actions of, say, a SxS or 7A rotary call.

 

So, to follow a caller’s progress on a #5 Crossbar lets take an animated journey through a virtual exchange, with sound effects. The animation is under construction…

Conclusion

 

 Exchange designers prized the crossbar because it added so much flexibility. To leverage this markers are needed to intelligently route calling traffic. The combo provides excellent switch utilization. 

 

The switch remained prevalent for ~70 years. It was the last generation (and so was the #5 end office) of the great electromechanical telephone switches.  Britain's last crossbar, at Droitwich, was decommissioned in 1994. See Endnote 2.

 

It competed for installations with SxS in many cities with Step winning out for most small installs due to the cost overhead of the marker and other common equipment.

For a narrated walkthrough of a Number 5 Crossbar office see [CMoS]. 

 

 

Endnote 1-- The Relay Computer

In 1939 Konrad Zuse (Berlin Germany) completed building the world's first programmable relay-based digital computer, the Z2. It had 800 relays and 16 bit words. In 1941 Zuse completed the Z3 with 22 bit words supporting floating point math. It had relay memory and a clock cycle of 5.3 cycles per second.  The Z3 had ~2,300 relays.

 

The original Z3 was destroyed in the second world war. The Deutsches Museum in Munich has the only working replica, built by Zuse in the 1960s. The image is the Z3 at the Museum. 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

In 1940 George Stibitz, a researcher in the mathematics department at Bell Labs, along with Bell engineer Samuel Williams, designed a relay-based Complex Number Computer.  It had 450 telephone-type relays and stored intermediate results in ten crossbar switches. This was several years after the #1 Crossbar marker had been invented.
 

Endnote 2

TXK (Telephone eXchange Crossbar) was a range of Crossbar exchanges managed by the British Post Office, then BT, between 1964 and 1994. The acronym TKC, more appropriate, was not used.

 

The Plessey Company became Britain's largest manufacturer of telecommunications equipment, including most of the country's crossbar switches and exchanges. Wikipedia.org/wiki/TXK

 

References
 

Adam, A.O., No. 5 Crossbar Marker, Bell Laboratories Record, Nov 1950.
 

Busch, A.J., “The Crossbar Line-Link Frame”, "Bell Laboratories Record, May 1939. (#1 Crossbar)
 

CMoS: A walk through a No. 5 Crossbar office, Connections Museum of Seattle https://www.youtube.com/watch?v=XKyXCZV_faY
Narrated by Ed Mattson. At 7:11 Ed describes a working marker, “Sounds likes music to me.”
 

Davis, R. C. "The Crossbar System." Bell Laboratories Record, February 1939. (#1 Crossbar)
 

Fundamentals of Telephone Communication Systems, Western Electric, 1969.
 

Graupner, W.B., “Equipment Arrangements for No. 5 Crossbar Markers”, Bell Laboratories Record, Sept 1950.

Shipley, F. F. "Crossbar Toll Switching System." Bell Laboratories Record 22, April 1944, p. 355.

Z series of electromechanical computers

 https://technikmuseum.berlin/en/exhibitions/permanent-exhibition/computers/ 

Z3 all relay, floating point, computer
bottom of page