Sunday, October 17, 2010


After hacking away at this thing FAR MORE than I had intended to, the diode test fixture works, again. Having displaced the guts enough to tackle some metalwork, I bored out the power supply connector hole to accept a DB-9 connector that will support all power feeds, though I've retained the banana connectors for the variable supply for convenience.

After plumbing that in, I applied power only to learn that I'd lost a channel. Following a bit of hair pulling (from the front, naturally), I isolated a poor solder joint on one of the three breadboards. Closed it up and discovered that while I'd fixed the channel I was working on another channel took a shit. Traced that to the same board, repaired it.. gently tucked everything away and closed the lid. You'll just have to live without gutshot of the most current incarnation, sorry. Imagine a lot of wires that have been cut and spliced numerous times and three ugly breadboards.

Here's a couple face shots:

These numbers are meaningless in terms of absolute voltage. I have traded accuracy for precision. Here is a shot that demonstrates the sort of swing one gets from a change in temperature, my finger throwing off the readings of channel D (it took about 20-30 seconds to drop that far and it hadn't finished dropping).

That is why all matching of semiconductors takes so long, one must allow the part to thermally stabilise or any readings are essentially useless.

No comments: