"Donald E. Kimberlin" <0004133373@mcimail.com> (04/08/90)
Responding to: "Gil Kloepfer Jr." <think!ames!limbic!gil@eddie.mit.edu> Kloepfer responded to my query: > P.S. Any other members of the Databit Alumni Association out there? Saying: >Actually, yes I was! I was a technician for Databit for a year about >eight years ago ... I never knew what I was working on until ... >...reading the Digest and chatting with a telco friend of mine. I'm surprised, Gil, that a telco person would have any familiarity with Databit TDM's, as they were a specialized unit built to handle Telex and TWX switched teletype network functions. There were _far_ more Databit TDMs around than many people ever knew, in networks all over the world. In fact, there are probably still many in use, dating to about 1971. How do I know? I was the Databit sales- man for Africa and the Middle East, and there alone something in excess of 1,000 are probably still being used in the public Telex networks. In the U.S., the prime user was Western Union, to provide not only trunk telegraph lines between exchanges but also what WUTCO called "blockhouses" as unattended concentration points for groups of Telex and TWX subscriber lines. I'd like to hear if that's the sort of environment you found them in; I was working overseas so only had hearsay knowledge of the WUTCo application. (Amendment: Others, like RCA, ITT, WUI and all the "International Record Carriers" had lesser deployments in remote locations as well.) Then, Kloepfer asks: >Could anyone out there tell me what those 208A (I think that was the >number) 300 baud modem boards were for? Well, the number you quote would only be the suffix of the whole part number. The Databit had a variety of low-speed line cards to suit whatever the need. (In fact, we found out that in each different nation, the detail Telex subscriber line signaling varied enough that we had to have a new card for every new country..just part of the job I found out I had to do to sell American technology overseas. Seems the factories let their engineers make little "improvements" in each nation they went to. That, of course, not only added to my challenge but also kept a Telex machine for Kenya from being usable in Botswana, too! The "300 Baud" modem cards were primarily for domestic use when and where telcos couldn't (and wouldn't) provide DC telegraph circuits. By providing the telegraph companies with this interface, they could avoid hassle from the telco, saying, "just give us a two-wire VF pair." The modem was actually a 103-type, run at either the 50 Baud of Telex or the 110 Baud of TWX. Running at so much less than its maximum speed and able to handle losses of 30-35 dB on the VF line, they just sang away. Then, Kloepfer writes: >A few years ago, I could rattle off the number of every board I'd work >on, what problems I'd need to fix most often, and the kinds of stress >tests they'd need. The modems mentioned above, by the way, were composed of a few op-amps, TTL chips, and a bunch of discrete components. There was one gain resistor which needed to be set by hand using a resistor substitution box. And there you describe some of the unique construction of the Databit equipment. In its earliest 1971 product, there was no such thing as a microprocessor, so the designers built a 5-card set of "common equip- ment." It was, pure and simple, a computer processor. But, there were also no CMOS chips, so the equipment had to be built of TTL chips. I do recall being around Databit when CMOS came along, and the usual trade press "hype" stories started that _all_ TTL chip supplies would be gone in a year. The boss almost had a heart attack, but fortunately it turned out to be just an earlier story of the hypes we all take as normal nowadays. As to making the modems from op amps, well, we didn't have any LSI modems, either, so...if you want to sell a product, you build it from what you have. As you can fathom, 50 or 110 Baud was all that was needed, so they could be pretty simple. But, Gil mentions "stress-testing" cards in Databits and such. In fact, the President (who had been the engineer who dreamed it up) was a fanatic for burn-in and QC. My international units went through a whole week of keying a single "Fox message" in and out of every port looped back (a test much telegraph equipment can't pass on static test), BUT in a _rapidly_cycling test chamber from (outside on Long Island) ambient to 130 F. That test chamber had electric heaters to raise the temp fast and fans to cool it right down. From my perspective, the Databit was overdesigned, overbuilt and overtested to the point of perfection. Reason: My units went for a 9500 mile plane ride to some customers. Without fail, we uncrated them, powered them up, and they played flawlessly. We didn't want to have to send a serviceman out there...and we rarely did. The only times I know of was so he could deliver a contractual "training course." About the only thing I ever knew of that hurt a Databit TDM was physically destroying it (oh, you could crunch backplane connector pins, but the cards so rarely required moving, that wasn't much). One day at the plant, we got a TDM from WUTCo somewhere for "warranty re- pair." It was shipped without a carton, and by the time it got to the plant, it had the shape of a football, and WUTCo tag said, "doesn't work." We had a bit of a time with WUTCo over that one! So, how about it? Any more Databit alumni out there? Donald E. Kimberlin, Databit Alumni # 000001, Class of 1971