Heya folks, the Old Crow here. I thought I’d relate the first of a few Radio Shack/Tandy stories as told by my synthesizer engineer friend and sometimes partner in crime, Paul Schreiber. Paul worked for the Tandy Corporation from 1977 to 1987 as an electrical engineer. His first story concerns their first direct-connect modem, the TRS-80 MODEM I. In Paul’s own words:
“The demise of Radio Shack/Tandy brings back many mixed emotions. I started there in 1977 as a co-op (intern), making $3.85/hr, etching pc boards for TV games (a la ‘Pong’) in the “R&D building,” a former tire store where the etchant tank was in the women’s bathroom (there were no females in the building, so why not?) The next session, in the spring of 1978, I was in the TRS-80 factory in the repair section (we had a quota of 50 boards/day). After graduation, I went there full-time in May of 1980 (after 1 glorious year at Data General, still the best job I ever had, but that’s another story).”
“My first Tandy design assignment was the MODEM I, the first stand-alone, non-Bell direct-connect modem (there was a Hayes S-100 card that was direct-connect, for you old-timers). I worked on it by myself (most of Tandy’s engineers worked ‘alone’) for about 5 months. The filter design software I wrote on a Model I in BASIC, that would iterate the filter’s group delay response versus picking standard 1% resistor values. The program was very useful, and it ran pretty fast on the 4MHz Z-80 (an average run, based on first picking the caps then running the resistors, took about 45min).”
“I even designed the case , the graphic overlay and wrote the manuals. I applied for 2 patents and much to my surprise both went through without hardly a blink of an eye from the USPO. One patent (for the filter design) would turn out to be instrumental in keeping AT&T/Olivetti out of the ‘laptop’ space 7 years later, when they tried to clone the Model 100, which had my modem design in it.” Note by Crow, here is the patent.
“I still have my “first article” production sample. Good times…sigh…”
“That photo in the 1981 catalog caused me to get yelled at by John Roach, the CEO. I only got yelled at 2 times in 9 years, not bad. The issue was the toggle switch on the right side. The photo was taken in April of 1981, this was not the production unit but a painted, wooden model (the ‘mockup’). That switch selected 300 or 600 baud data speeds over the phone line.”
“Well, during FCC testing for Part 68 (attaching to a phone line), I found out that 600 baud was very flakey. I started to panic a bit. I read the Motorola data sheet for the modem IC like 50 times, looking for clues. Then, one day, in a footnote buried in the datasheet, was this statement: 600 baud operation not guaranteed over PSTN.”
“PSTN?? WTF???? Well, PSTN = Public Switched Telephone Network. It turns out that indeed, due to the filtering the phone company puts in the line to make ‘voice sound good,’ at 600 baud FSK modulation has sidebands attenuated too much by the voice filters.”
“So after the CEO ass-chewing, I had to throw 50,000 manuals and 50,000 panel overlays in a dumpster. Electrically, we just removed the switch from the BOM. Thankfully, production was in June! I learned to read ALL of a datasheet.”
Crow here again. It just so happens the very first PC board I ever etched in 1982 was a 300-baud direct-connect modem using the same modem IC Paul used, a Motorola MC14412. I initially used it by hand-dialing the number, then flipping a switch when the carrier tone sounded. This is how I connected to the university telecomputing network my freshman college year.
In the summer of 1983 I hand-taped my first-ever PC board, a single-board computer based on the “Arduino of its era,” the RCA 1802. The 1802 mailing list guys call it the “Olduino.” The chip was a tad unorthodox, but it was a CMOS device and could run on 1mA of current from a 9v battery, which made for a great portable microcontroller. This board has my modem mounted to it and it ran a hand-assembled-on-paper program to (badly) emulate enough of a Hayes J-Cat modem to tone dial a dialup line, detect the carrier and connect the modem. Parts are missing now, but perhaps I will one day see if it still runs. Cheers!