HP-3468A, Display Voltage VDC

[Garth Wilson]

I can imagine the problem being caused by slight leakage between pads that should not be connected. A steady DC reading will not be caused by AC noise in the area as someone suggested. Did someone get in there with a soldering iron and water-soluble flux at some point? Water-soluble fluxes have been a disaster. They are corrosive if you don't clean them off. But just as bad for our company's sensitive analog circuits, water-soluble fluxes penetrate the pores in the PC board where it's not masked, and they produce conductive paths, even if the board appears to be clean. And further, those conductive paths don't act like a quiet resistor, but make a noise like frying bacon in the audio signals we deal with. I always specify in PC-board manufacturing instructions that the holes in my soldermask files are not to be opened up any bigger, and that any pair of adjacent pads absolutely must have soldermask running between them (unless they're connected already anyway). This is more of a flux mask than a solder mask. The first time I encountered the problem 30 years ago, we had just changed assembly houses for a particular board—a rather simple one with no soldermask—and the new assembly house was using water-soluble flux because of the EPA making life difficult for them otherwise. A technician who was testing the product before shipping to customers ran into a problem right away and couldn't figure it out. He asked me for help. It didn't make sense. I had him remove various components from a net until there was nothing at all left on it, and still there was maybe 50K between that trace and another. I told him, "Then there's only one possibility left. Take your solvent and really scrub that thing. When you think you have it super clean, do it again." That fixed it.