New (to me) HP 3421A DA/C Unit, some questions

[Garth Wilson]

(05-11-2017 06:52 PM)4ster Wrote:  Finally, I would be interested to hear what have you done with your 3421. I had a project that involved simultaneous truthing of a batch of six oxygen cells for rebreather diving that the 3421 would have been perfect for. For that application I built a test rig using an HP 3468 meter and a CMT-200 DAC running a home made six channel switch box, so that project is done. I suppose if I could find a 4465A (Option 40) board I could use the DAC to run lawn sprinklers at my house. ;-)

My first contact with it was in 1983 when I was working in applications engineering at a company that made VHF and UHF power transistors, mostly for military communications and radars. I estimated there was about $100,000 of RF equipment per engineer in the lab, and much of it could be interfaced to, if not also controlled by, a computer, but usually wasn't. We had a lot of repetitive measurements to do on parts though. For example, Litton or other client company would be eyeing one of our parts for a particular job they had for it, and they would set up an engineering contract to evaluate the part's performance under the conditions they dictated. Part of my job was to build an amplifier that used it, and test a small lot or two or three and report the results. It got pretty repetitive and boring to do by hand, setting levels, reading meters or screens, switching things, etc..

Half the engineers had their own HP-41's. I was still using TI which was not interfaceable. I had a close friend there with a 41cv though (and lots of accessories! ), and the HP sales engineer came by and dropped off a 3421A for us to try. This friend took his lunch time with me and set up the 3421A at a workbench that was otherwise already set up for a job, took only 20 minutes to write a program, and had it working quickly! We put in a requisition for the 3421A and an HP-41cv with HP-IL and maybe a module or two. The company approved the 3421A, but said it was too easy for the 41cv to walk away in someone's pocket, so they ordered an IBM PC for several times the money. After three weeks of fiddling with it, the two engineers who were trying to use it were still trying to get the PC to talk to the IEEE-488 card!

A couple of years later, I was the only engineer at a tiny company that made aircraft communications equipment, and the boss decided to get into the headset market. He had the first batch of 500 coming, and I wondered how we were going to test them. Things were still unsteady enough that the testing had to be pretty thorough, not just put it on your head and talk to yourself and say, "Yeah, it works." I built a small anechoic chamber (which we called the "coffin" because that's about what it looked like) and I set up testing of the earphones and mic, partly using the 3421A, along with a few other pieces of equipment. Results were printed out, matched to the serial numbers of the headset being tested. By that time I had a 41cx (and HPIL-to-IEEE488 interface converter and plenty of other accessories), and that served as the initial controller, running a 20-page program that accessed six files. The test setup morphed over the next few years and I got my 41 back. Surprisingly, when we went to a 68000-based HP computer for the controller, the speed did not increase much, because much of the time was still spent waiting for filters to settle and readings to come back from the equipment.

I do have a pair of 3421A's here at home that I got inexpensively on eBay but have not used so far.