../ Front ./ Phreaking Seattle History / Morris ESS Trial /
UW special numbers the 3ESS list of Bell System Practices ITT electronic transmitter

the number 3 ess

This document is badly organized, and some of the information is wrong. All will be fixed eventually.

The 3ESS at the Museum of
Communications in Seattle, WA

Someday this will be a clickable image map. Until then, you can just pretend. Full-size picture. Other pictures: The Teletype Controllers (TTYCs), a close-up view of the connectors for the Serial Periphal Interface.

At the Museum of Communications I often work to maintain and restore the No. 3 Electronic Switching System. This document will be rife with three letter acronyms. The Bell System had tens of thousands of TLAs, with numerous collisions (there are only 17,576 possible three-letter acronyms). If they could understand it, so can you.

The 3ESS is one of the least written-about of the Western Electric switches. Of course, this is partially due to the fact that there were only a few hundred of them in service for a few years each. Other switches, such as the No. 5 Crossbar, which served more large metropolitan areas, have received correspondingly more coverage. I intend to remedy this. Because I interact with a 3ESS on a semi-regular basis, I will (eventually) include samples of interaction with the system, audio recordings, et ceteraa.

History

The 3ESS was created in the early 1970s. It came after the 1ESS and the 2ESS. The 3ESS has a switching capacity of between 1 and 4500 telephone lines, and was intended for the low-volume rural and semiurban markets. Switches that fill this niche are called community dial offices (CDOs).

In the early 1980s, before the Bell System was broken up, the Federal Communications Commission required Bell to allow telephone subscribers equal access to any long-distance carrier they wished to use. This required extensive changes to the switch generics. The cost of modifying the 3ESS's generic (base program) to support equal access turned out to be greater than the cost of replacing all the 3ESS offices then in use. Thus the effective service life of the 3ESS was less than ten years in most cases. (In the Bell System, the service life of a telephone switch is usually between twenty and sixty years.)

Physical Attributes

The 3ESS is quite a small switch. Even including the distributing frame, it can easily operate inside a semi-truck trailer.

The 3ESS's processor is the 3A CC, which stands for 3A Central Control.

Most parts of the switch are duplicated for reliability. This includes the CC (one is running, the other idle; they switch when there is a problem, or just if someone feels like it) and associated Main Store, the standby CC is made active and the active CC is made standby. When there is a problem (or when the idle CC is being maintained) it is taken Out Of Service. Practically any part of the switch can be made OOS. Before a piece of equipment can be put back in service (INS), it must pass each part of a full diagnostic routine. Sometimes it is possible to force equipment INS, though it isn't the best of ideas.

Each 3A CC has its own main store (MAS). Each MAS memory word is stored with two parity bits, one each for the high and low words. Addresses are 20 bits, allowing for a full megabyte of address space (not necessarily memory), while integers are 16 bits. I have no clue how this works out in memory. Perhaps the machine is basically nibble-based, with four nibbles for integers and five for pointers. (But I'm pretty sure each parity bit is calculated on exactly half of the word.)

Logical Attributes

The software is a custom event-loop, written in EPL, which stands for ESS Programming Language. It's basically a complex macro package for assembly programming. The system documentation includes commented listings of the compiled software. It operates at a normal rate of XXX:SOME_RATE full cycles per second. If the average cycle time drops below XXX:SOME_RATE, something happens. Therefore, all tasks that will take more than a short amount of time (tens of microseconds?) have to be split up and executed piecewise.

At one time, Dennis Ritchie considered porting Unix to the 3ESS, but then decided it wouldn't work out. If that had been done, it is likely (in my opinion) that it could have been more easily upgraded to operate in the modern telephone network environment, and thus the 3ESS would likely still be in service. (The 5ESS has always used Unix, and it is one of the more popular switches in service today.)