The precision one-milliohm short circuit

[Vell]

This happened back in '89 or '90 when I was an electronic midwife, being the first to apply power to newly-assembled switchmode industrial amplifiers. Separate output circuits handled positive vs. negative-polarity outputs, as I recall.
One day, a new one had wrongly scaled output, but only for one polarity.  Each polarity had a one-milliohm, four-terminal current sense resistor, but the scale factor was off by just about two, exactly -- too high.

After probing and measuring, I concluded that the sense resistor was half a milliohm; got and installed a replacement. Same story. Decided to do a four-terminal measurement (as I recall), and both original and replacement were one milliohm.

Now, our PC vendor was not one of the best, and we had occasional shorts. Turned out that there was a one-milliohm short across the sense resistor, likely between the current pads. This short was within 3%, or less.

[MagicSmoker]

Welcome to the forum, but do try to post in the appropriate area. This - and your other posts, actually - really belong in general chat.

[Zero999]

Interesting. I wonder how stable the short was.

How much current was the shunt passing?

[Vell]

How much current? Sorry, have forgotten, but maybe 5 A.

[Zero999]

How much current? Sorry, have forgotten, but maybe 5 A.

Presumably it was a current measuring resistor?

5A would have been a voltage drop of only 5mV, across 1mR. I imagine the amplifier would have needed be low offset, probably a chopper stabilised op-amp.

[Vell]

As I recall, the op amp was not particularly special. Board layout, though, was probably skilled.