This is a 6502-controlled automated test equipment (ATE) setup I made and programmed at my last place of work in about 1990. It is shown testing one of the products on a board of about 2.25" x 4.5" near the front edge of the workbench. This setup is much more compact than previous ones I had done which used some high-doallar Hewlett-Packard and other IEEE-488 equipment stacked in a 19" rack. This new setup had signal generators, DMM, loads, amplifiers, power supplies, switching matrices, etc.. in a 3U 19" rack cabinet. It's all controlled by a Cubit 6502 SBC running Forth. The program was about 10,000 lines of compact code. It would do scores of tests on dozens of different models of the products, yet the program and all test parameters compiled to about 38KB with Forth headers (or about 28KB without). This project is what really got me going in Forth on the 6502. Using Forth, I was able to spend my time figuring out hardware problems instead of debugging software. (Forth does make hardware problems easier to figure out.) After I released the ATE to the production floor for everyday use, no one ever found any bugs in it. I know they used it for at least 8 years after I made it, and they may still be using it today. They paid me to service it a few times after I left the company, for things like an IC or relay that went south, or a cable that became undependable from being flexed tens of thousands of times.

Same ATE, with the case open, during development. I made two sets. The original plan was for up to seven, but that became unnecessary. This used STD bus (which, BTW, stands for "simple to design," not "standard.") The boards are about 4.5 x 6.5", with 56 contacts at the edge connector.

A close-up of the operator interface. With the earlier setup using the HP-9000-series computer, the test operator had a large screen and keyboard, which proved troublesome. The idea here was to only give the operator access to what (s)he needed to do. As it turned out, it would have been nice to have the LCD a little bigger than 16-character, 1-line-- maybe a 40x2. It fit the bill pretty well though. Failing test results always got printed out so the paper would get rubber-banded to the unit to go back to repair. Optionally, all test results could be printed, or archived for later download to a PC to analyze with something like Lotus 1-2-3.