Methods for testing electro-mechanical relay-based telephone exchange systems

Maintenance Test Frames

An electromechanical telephone exchange is a complex beast. A large amount of apparatus must be maintained in proper working condition, so first-rate maintenance processes and tools are essential.

Troubleshooting involves responding to alarms, circuit testing, progress monitoring, measuring, adding/removing subsystems, and more. Maintenance systems started evolving with the first automatic exchanges in the 1890’s.

In the 1950’s exchange maintenance personnel managed a central test desk (or frame) embedded with alarm and status lamps, testing circuits means, a way to remove/add subsystems (markers, senders, decoders,…; explained elsewhere) and a trouble card recorder.

Also, it was common practice to proactively test for trunk/device/system operational health both manually and automatically. Personnel were “looking for trouble.” This effort was made to seek out and correct faulty equipment before the defects create trouble under service conditions.

Each generation of exchange had its own style of “test desk”. Below are snapshots of various testing apparatus for several exchange types.

Panel test desks

Fig 1 shows a “tea wagon” as it was called, a portable test set. In this case, the craftsperson is testing a panel incoming selector. This may be a proactive test or one to fix a known issue. You can see some cords connected to test points on the frame for control and measurement.

Fig 1

Portable testing of a panel switch using a tea wagon in 1928

Each panel exchange had a test desk with access to all incoming and outgoing trunks (2 wires each, talking paths). Fig 2 is a partial shot of such a test desk. The rows of jacks access the trunks. Any trunk could be tested for fidelity by a craftsperson connecting the trunk to measuring equipment. Notice the large Weston meter for making measurements.

Fig 2

Panel trunk testing area

Fig 3, test desk, shows a dial and typical punch-key board. The keys are used to enter codes for testing. More on this below.

Fig 3

Portion of a panel test desk

Crossbar test desks

Testing tools and processes improved with each new generation of exchange. Fig 4 shows a portion of the test desk for the #1 Crossbar system at the Connections Museum of Seattle. This system was designed in 1938. Notice the large panel of indicator lamps on the far left side.

Fig 4

Test center for #1 Crossbar Office

The #5 Crossbar exchange and test center was the most advanced of the electromechanical systems and was designed in 1948. It included a new “trouble recorder” for indicating alarms and status of the exchange using punched cards. In a way, the durable cards replaced many ephemeral indicator lamps that were used in the Panel and #1 Crossbar systems.

Fig 5

Maintenance desk with trouble recorder (left side)

The recorder punches pre-labeled cards with holes to mark conditions and states of interest. The trouble cards have two sections that provide space for a total of 1,080 punch indications -- 18 rows of 60 columns each. The actual number of holes appearing on a particular card, however, varies widely with the type of trouble being reported. Sure, they are eye charts but very handy for debugging.

Fig 6

Example of an un-punched trouble ticket card (left half shown)

These cards are analyzed by maintenance craftspeople who use the punched locations and a set of charts and circuit diagrams to locate and repair the troubled condition. Reading cards is an art, and it may take years of experience to quickly deduct a cause of trouble. The terms on the cards refer to times, relays' states, subsystem IDs/states, call progress and other pieces of relevant data to service personnel.

Fig 7 shows a craftsperson entering a test code into the punch keyboard based on the contents of the trouble card. The trouble card recorder is partially hidden, on the bottom left side.

Maintenance desk with trouble recorder crossbar office

number 5 crossbar explainer, Example of an un-punched trouble ticket card (left half shown - trouble ticket

Fig 7

Master test frame at #5 Crossbar office in Englewood, NJ

Fig 8 is a closeup of a card recorder located at [CMoS]. On the upper portion can be seen the hole punching solenoids.

Fig 8

Trouble recorder with punched card waiting for removal

For a deeper look on how a card was used to resolve a problem visit the Appendix below.

A test keyboard

All large exchanges had some configuration push buttons for entering test codes and other numbers. The keypad was an essential tool for maintenance. Fig 9 shows the diagram of the button layout that is used in Fig 7. Keys on the test panel may be operated to simulate any calling line location, class of service, and called number.

Fig 9

Layout diagram of master test frame keyboard in Fig 7

For those familiar with crossbar exchange terminology, certain terms specific to a #5 office will be recognizable.

Of course, there is so much more to testing and repairing these complex systems.

This section emphasizes the importance of maintainability in telephone exchanges and their associated equipment.

The Appendix to follow runs through a debugging adventure for a failure in a marker subsystem using a trouble ticket card. This is a good place to exit if this is not your cup of tea.

number 5 crossbar: Layout diagram of master test frame keyboard

Appendix

Sherlocking a Marker failure

You are working the late shift as a craftsperson in a #5 Crossbar office. Things are relatively quiet with no warnings or alarms needing attention. Sipping some coffee, you hear the sounds of the trouble recorder punching out a card and dropping it into the pickup well. Oh no, trouble, time to get cracking. The card was automatically generated due to a failure of some kind.

Here is the right half of the full card that drops.

Punched trouble card (right side only)

With years of experience every code on the card is like a second language to you. The first thing you notice is, in the top row, a marker's TK relay has not operated. Yikes, not a marker!

Here is a closeup of the top row (the image scan is a bit blurry). A black circle (card hole) means a relay has operated and no hole means not operated. All the top row indicators are relay states.

explainer, trouble ticket punched 5 crossbar office

Sherocking a marker failure

Closeup of trouble card under study

There are about 1,500 relays in a marker, so the fault could take some time to resolve. You put this marker into "out of service" state and start troubleshooting.

The marker's TK relay (Test checK) operates when the marker reaches a defined call progress milestone. If operated, the marker has performed many of its assigned tasks up to this point. If not, something has caused a fault and terminated. "Getting to TK", says a lot about the marker's health.

Next, you take out the sequence diagram for the TK relay to see what conditions causes it to operate. Well, looks like ten other relays need to operate before TK operates. An “x” means operated in the context of this diagram. The JCK and TK are test points.

TK relay sequence diagram

Each of these ten relays (JCK1, JCK0, CHE, LCK…) are related to the progress of the marker’s operation. So, if TK has engaged then many process steps have had success.

You are wondering what relay is not operating in the TK engaging chain. Looking at the card again you notice that JCK, LCK and others are punched (this is good) but TCHK is not punched (bad).

Digging into a relay sequence diagram, you find the causes for TCHK to operate. After some detective work you decide to clean a contact on the TCH relay that engages TCHK [Endnote A].

Making your way to the marker in question, and with a burnishing tool, you mount a ladder to clean the suspect TCH relay’s contacts. Did this fix the problem?

Back at the test frame, you run a test putting the marker through its paces. A status card is dropped, three cheers, the TCHK and TK relays did engage! Compare the figure below to the similar one above. Problem resolved. Time for another cup of coffee.

Explainer, crossbar trouble ticket closeup of TK LCK JCK and TCHK punch holes

Explainer diagram for TK relay sequence diagram #5 crossbar office

Marker test success – relays TCHK and TK have operated

Marker health: Relay operations for "Getting to TK"

The video to follow shows a Completing Marker at the Connections Museum of Seattle. A small portion of the 1,000+ relays is selected and this contains the LCK, TCHK, FAK, RK3 and TK relays and others. Additional relays relevant to operating relay TK are off screen.

References and Acknowledgments

Bell System: BSP 300-160_I3 No.5 CROSSBAR circuit description, SCD-10-01 (TK relay)

G. H. Duhnkrack, Analysis of No. 5 Crossbar Trouble Recorder Cards, Bell Laboratories Record, June 1955.

L. M. Allen, Routine Tests in a Panel Office, Bell Laboratories Record, May 1929.

C.V. Taplin, Maintenance Center for the Crossbar Toll System, Bell Laboratories Record, Oct 1944.

C.W. Haas, No. 5 Crossbar- Marker and Transverter Testing, Bell Laboratories Record, May 1954.

L. N. Hampton and J. B. Newsom, The Card Translator for Nationwide Dialing, Bell System Technical Journal, Sept. 1953

CMoS: Figures 2, 3, 4, 7 and video from Connections Museum of Seattle. Many thanks to Sarah Autumn for selected media.

Endnote A

A crossbar telephone office contains nearly 2,000,000 precious-metal contacts for establishing the connections for 10,000 subscribers' lines. About 1,200 relays, with an average of seven contacts per relay, are involved in establishing a call from one subscriber to another.

When the operations per call are multiplied by the 50,000 average daily calls that may be handled by such an office, and by the number of days in a year, the total number of relay contact operations per year is over a hundred billion. The operations for the bulk of the individual contacts vary from about 50,000 to 15,000,000 annually.

Contacts that open/close often may become eroded. Burnishing a contact can help restore its health until the relay needs to be replaced. Some contacts may have a 'snubber' circuit attached to improve reliability, see Endnote B.

Derived from: Swenson, Contacts, Bell Laboratories Record, Feb 1949.

Endnote B -- Protecting the contacts

When a the current through a relay's coil is changed, the coil's voltage follows this equation.

Inductors (coils) resist a change in current, generating a rise in voltage. The faster the rate of change of current (di/dt) the more the coil's voltage increases. So a coil's voltage may reach (+-) 1,000 volts, or more, if di/dt is sufficiently rapid. This can't be good for reliability.

Typically, one relay's contacts control the coil current (on/off) in another relay. So, the voltage across the contacts will also increase as the contacts open since the current is trying to go to zero rapidly.

The Western Electric 186 series of networks comes to the rescue. They were installed by the millions to protect relay contacts The combination of a series resistor and capacitor "snubbs" the rising voltage, reducing arcing, and thus extended the useful life of the contacts.

In the 1950's diodes became available and were often used in place of the 186-type snubber. Bottom line, whatever the means used, some relay contacts needed protection from arcing else contact integrity would suffer.