Fluke 8600A steady rising reading in DCV range even with no leads attached

[nightfire]

Could use some input here for first thoughts: My trusty (at least, of so far) ebay bargain https://www.eevblog.com/forum/testgear/fluke-8600a-used-buy-what-things-to-check-for/  has just begun to act up.

After measuring some resistances I went back to DCV, and wanted to measure some voltage drop in USB Power supplies. This was when I noticed that the value was not plausible, and steadily rising (20V range).
Pulling leads- no change. Numbers still rising.
Changing ranges: Numbers (in absolute value) not changing, but still rising, only decimal dot moves ;-)

As I want to learn and understand about the nature of this, and also about the way a multimeter does its measurements (dual slope according to fluke docs I read in this case) some discussion about the parts in question would be nice and educational.

Would this behaviour be a sign of some other part injecting some leakage current into the cap that is used for the integration of the measured voltage?

[Kleinstein]

The 20 V range input voltage should drop to zero, as this range is already 10 M input resistance with the Fluke8600. For the 2 V and 200 mV it is normal that the input voltage with open leads is undefined and tends to drift. This is brecause of the very high impedance ( e.g. 100 GOhms range) and still some biasing current and some filter capacitance at the input.

A possible defect would be an open contact with one of the relays / switches. The error description looks like an open input for the ADC (inluding the input buffer) and thus not effect of the divider.
I doubt this failure has something to do with the ADC itself, more like the switches for range / function switching.

[nightfire]

Good point with the switches.

I had left the DMM in DCV setting, switched to 20 MOhms and it also happened. Switched to Ohms range, back to DCV, and in the 20MOhms range and below nothing happened, only in the 2V and 200mV range settings the readings steadily raised.

Played back and forth with the DCV and Ohms mode switches to actuate them and hopefully clean a bit of eventual oxidation, and it got sporadically a bit better, but not on the last two lower ranges.

And this would not be like the normal voltage you measure with open leads, these would oscillate around some low value usually in the mV range- here I would be seeing numbers of >100Volts (theoretically) in the uppermost range, and after switching a range down, the exact same numbers would stay. (Only the decimal point would be moved at this then)

[Kleinstein]

The 200 mV and 2 V range are high impedance (could have to option for lower impedance) and in this case it is normal to set the drift with an open input. The problem is if you see such drift also with a shorted input or a low impedance voltage source.

[nightfire]

Hmm, I am confused a bit- just turned that d***ned thing on again, and behaved perfectly.
On DCV and open inputs, it showed the expected readings as of a few mV, and with shorted inputs/test leads is went exactly 0.
Switched to Ohms, tested some stuff, values were ok and plausible.
Got back to DCV, and no sign of the former erratic behaviour, with values steadily rising no mattter what.

Could I chalk this off due to old switches that really should make some acquaintance with deoxit etc., or is there maybe some more serious underlying matter?

[Kleinstein]

 A drity or hanging switch (maybe relay contact) is a very plausible explaination. Sometimes it is enough to operate the contact a few times.