I recently acquired another Keithley 196 off eBay that was listed as not powering on, which, could be a lot of things, but hopefully was related to the power supply/rail infrastructure (thinking, most likely the digital logic 5V regulator, primary filter cap, etc).

The unit arrived, and indeed, did not power on at all. No indication that the digital side of things was getting any power at all.

So, opened the unit, and pulled the analog board/enclosure off, and didn’t see anything. So, I pulled the digital board, and wow, the PCB was all brown/burned around the transformer. Sadly, I didn’t take a picture. =( Seemingly, and this is a guess, there was a cold solder joint (either from the factory, or from vibration) on the pin in question, and it heated up over time to the point of burning the pad it was soldered to, which killed that power rail. Tracing out the adjacent pins revealed that the pin in question was related to the 5V digital rail. Viola, we have our issue. Obviously, I’m hoping that the power supply going out didn’t damage any of the digital board.

The biggest issue with the repair was getting the burned pin to take solder. It required a fair amount of sanding with some 400grit paper, then cleaning that up a bit with some 1500grit (didn’t have anything between those two). =P Then, it was just a matter of bodging a wire (in this case, a 20awg stranded wire) between the transformer pin, and the bridge rectifier for the 5V digital rail. Then just tack the wire down with some hot glue (I don’t like wires just flapping about.

I also re-flowed the solder joints for the rest of the transformer pins, as I bet the issue was just poor soldering at the factory… the PCB also flexes a bit in the enclosure, and the transformer weighs a bit (which is why they bolted it to the PCB in the first place). The picture to the left is obviously my repair. I put the unit back together again enough to test (put the digital board back in the enclosure, plugged in the front panel (display, etc)), and powered it up. Bam! Works! So, put the unit back together fully, and tested it out. It’s about 20mV off from my other 196, and my 199. So either they’re both off (which wouldn’t surprise me since I nuked the configuration on the 199 at one point), or this new one is off. One of these days I’ll actually get one of my meters calibrated, but at this point, they’re very accurate, and pretty darn close. Remember, in the 30V range, the 196’s least significant digit is measuring 10uV’s, so 20mV’s is both a lot, and not very much. =)

Job accomplished.