larry@uunet.uu.net (Larry Lippman) (09/17/89)
In article <telecom-v09i0372m04@vector.dallas.tx.us> bmk@mvuxi.att.com (Bernard Mckeever) writes: > One of the early forms of "data" communications that still has a few > applications left is the telegraph. Several of the older alarm > reporting systems used by fire departments cling to this technology > because it still works, and the tariffs for telegraph grade facilities > are cheap and hard to change for political reasons. Your favorite > alarm company may still be using McCulloh systems for home and > business protection. The McCullough Loop system is still around, but it is rapidly disappearing. The McCullough Loop works by running a huge series circuit, which begins and ends in the alarm company central station, through each and every alarm customer location, While there is only a single pair which enters a customer premises, the wires should be visualized more as "in" and "out" rather than "tip" and "ring". The combined resistance of a closed (i.e., no customer sending a signal) McCullough Loop might be several thousand ohms. The alarm company central station excited the loop with 130 volts DC, and used a sensitive relay to detect current flow. Each customer location had a NcCullough "transmitter" which consists of a motor-driven code wheel, which when operating sequentially opens the loop and shorts it to ground. Each customer for each alarm (a customer could have more than one alarm) had a transmitter with a uniquely-coded code wheel, which resulted in a signal not unlike that of Morse Code. The alarm company central station was equipped with a "pen register" which drew lines on moving paper tape to display the code. The design of this pen register can be traced back to before 1850 - the era of Samuel F. B. Morse! When an alarm customer sent a signal, the initial open of the McCullough Loop operated an audible alarm at the alarm company central station, where an attendant would go over to the pen register and attempt to decode the signal. While this was rather crude, it worked well except for one situation - where TWO or more customers were simultaneously sending signals! Since 50 or more customers could be on a given McCullough Loop, this was not uncommon - especially at common times in the morning and evening when customers with intrusion alarms would automatically report "openings and closings". A particular problem arose when the McCullough Loop would go open due to a cable problem. In some cases certain customers would still have alarm protection since the transmitter also shorted the loop to ground, and the alarm company apparatus was also equipped to detect this condition. However, other customers might be totally without protection. It was generally not possible for the alarm company to know where a loop open had occurred, so they immediately called the telephone company, who would in turn dispatch a repairperson to a suspected CO to check each pair from the CO on trial-and-error basis. It was also easy for a burglar with technical knowledge to defeat the target alarm system by first shorting the McCullough loop at a utility pole, and then opening the pair to the subscriber premises. The apparatus at the alarm company central station could never detect this condition, and more than one burglary was effected in this manner without anyone being the wiser - until it was too late. This weakness was never countered until the late 1960's when electronic Loop Integrity Devices were available. The Loop Integrity Device was installed at customer locations willing to pay for this additional security. In operation, added apparatus in the alarm company central station would send a low-speed coded interrogation signal with a unique address for each Loop Integrity Device, superimposed on the normal McCullough Loop current. Since this was DC signaling, the speed was 50 baud or less. When the Loop Integrity Device recognized its address, it would momentarily open its loop as an acknowledgement to the alarm company system. Polling would be interrupted upon receipt of an alarm signal. The Loop Integrity Device was only of limited benefit, since an open loop due to cable failure (accidental or "otherwise") still put some customers out of service. Most McCullough Loop systems have now been replaced with apparatus that uses bridged data lines (so that a short on one subscriber loop will not affect other subscribers) with frequency-shift keying of two-way digital data. The newer systems also allow one alarm unit at a customer location to send a number of individual signals to indicate different alarm conditions. > Today several vendors have T-CXR channel units > for this service, but it is still a basic series [just like old > Christmas tree lights] circuit that causes the Telco nightmares when > the circuit goes open. It was also an annoyance for customers. When my organization still had McCullough Loops for fire and intrusion protection, I was called in the middle of the night more than once to open up the building and meet the telephone company who was trying to localize a cable problem! > Western Union at one time had a huge telegraph network that spanned > the globe. Often telegraph circuits were transported over analog > carrier systems. I can remember the 43A carrier system that combined > up to 17 separate telegraph circuits on one N1-CXR channel [4 khz > bandwidth]. We had two systems of N1 carrier that each had 10 43A > systems on them. From time to time one or both of the systems would > fail when the A1 cable under the river started to go belly up. NY 7 > Telegraph [the control office] would be on the tie line in seconds > arranging a reroute. There is still a lot of 43-type and compatible apparatus in service today. The 43A1 voice frequency carrier data system system was developed in the early 1950's to replace the 40C2 system, which was developed around 1930. While the 43A1 also added 8 channels in the above-voice frequency range of ~3,500 to ~5,000 Hz, the basic 17 channels from 425 to 3,145 Hz were compatible with the earlier 40C2 system. The 40C2 system, which used rather large vacuum tubes (with grid caps, no less!) was a monster where ONE channel occupied about 16 inches of rack space. I saw 40C2 in service during the 1960's, but I am certain that it is all gone by now. The 43A1 system used miniature vacuum tubes and plug-in modules, and then allowed three channels to occupy about 14 inches of rack space. I feel certain that there is still 43A1 in service, but it must be getting scarce by now. The 43A1 system was replaced by the 43B1 system in the late 1960's. The 43B1 finally used transistors, and all 17 channels could fit in less than 36 inches of rack space. The 43B1 was compatible with all channels in the 43A1 system. I am not aware of any WECO apparatus newer than the 43B1, and it is entirely possible that nothing has replaced it since the 43-type system is, by its very nature, doomed to a now rapid obsolescence. The 43-type system has primarily been used for two purposes: low speed (150 baud and less) telegraph channels for baudot telex service and private line teletypewriter service, and for telemetering and control circuits as used by utility and pipeline companies. Both of these applications are changing, however. International telex is going to higher speed ASCII terminals which work at 1,200 baud or more - at least in the U.S. where the international ASCII <--> baudot conversion is performed at the U.S. switching center of the telex carrier. Utility and pipeline companies are migrating toward process computer-to-computer communication which can handle numerous data and control points with a single high speed data channel. However, there is still a large number of 43-type compatible systems in service for the utility and pipeline industry, especially on private microwave radio links. Since actually little WECO apparatus has been sold to these industries, most of the apparatus is made by such vendors as RFL Industries, GTE/Lenkurt, etc. and is functionally compatible with the 43-type system, but is physically different. > Another common data application was the dumb terminal to host > connection, used mainly with time share systems. Connections were > typically 300 baud, and the modems were much bigger than a bread box. > Connections were available in two flavors, dial-up and private line. Data sets for 1,800 bps and below which employed the switched network or private lines all operated on frequency-shift keying, and all dealt with bit-serial data. Well, there was an interesting type of data set - now long since obsolete - which operated as a _parallel_ data set; i.e., it sent data as _characters_ formed by sending simultaneous tones. These data sets fell into two categories, the 401-series which could transmit data at a maximum rate of 20 chars/sec, and the 402-series which could transmit data at a maximum rate of 75 chars/sec. Up to three simultaneous tones could be transmitted. The tones were divided into three groups. The "A" group had four tones which were exactly the same frequencies as the 4x4 matrix DTMF Low Group tones. The "B" group had four tones which were exactly the same frequencies as the 4x4 matrix DTMF High Group tones. The "C" group had three frequencies which were higher than the "B" group. Some versions of these data sets only used the "A" and "B" group tones so that the data sets were effectively compatible with conventional DTMF signaling with a 4x4 matrix. There were also three additional tones, which were optional, and were assigned as "idle" tones for the "A", "B" and "C" group. A common application of these data sets during the 1960's and early 1970's was a direct interface to unit record equipment for sending data encoded in Hollerith character format. 'Don't see that no more. :-) <> Larry Lippman @ Recognition Research Corp. - Uniquex Corp. - Viatran Corp. <> UUCP {allegra|boulder|decvax|rutgers|watmath}!sunybcs!kitty!larry <> TEL 716/688-1231 | 716/773-1700 {hplabs|utzoo|uunet}!/ \uniquex!larry <> FAX 716/741-9635 | 716/773-2488 "Have you hugged your cat today?"