Fluke 8840 behaviour

[Pat Pending]

The fluke 8840 has a very high impedance on its most sensitive DC volt ranges before the voltage divider networks get switched in.

My question...is it normal on the highest sensitivity DC volts range, with nothing attached (not even leads),
for the readout to appear to charge. I thought this meter autozero'd.

A physicist friend talked about rain clouds with charge having a detectable volts/meter at ground level !

How do other 4W bench DMMs like the 8840s or the HP3478a perform?

[ejeffrey]

I would expect that to be fairly normal.  The bias current of the input amplifier (which is really a leakage current) has nowhere to go, and slowly charges up the input capacitance.  It is a real voltage, not an amplifier offset, so auto-zeroing wouldn't help.  If you connected the probes to a perfect voltmeter, you would see the voltage on the probes slowly charging up.

I have a keithley 2100 that has a configuration option for the millivolt range to be > 100 GOhm or 10 MOhm.  The default is 10 MOhm, but when switched to high impedance mode it charges up exactly like you describe.

[robrenz]

8846A in high impedance mode
Meter is not warmed up but I don't think it has anything to do with the results.
The jump in the graph is the range change at 1.2V

First image first try 13Min's duration max and min volts are shown. (that n is for nano volts)
Second and third image second time (shorted inputs and restarted trendplot) max, time, and min are shown.

[Pat Pending]

Great! Thank you all,

It's good to know that charging appears to be the normal behavior with extremely high input impedance meters.

The 8840 can't measure nV but I do see a 30mV build up over the period of a few minutes.

Heck, I was getting ready with an isopropanol /deionized water scrub, thinking that it was flux residue or finger oils on the main PCB.

[Zad]

This was captured from my Tektronix 4020 / Fluke 8808a with LabView, the probes were just sat on the bench next to each other. The two spikes are where I got up and sat down again. Curious, but probably the sign of a well designed really high impedance front end.

[alm]

On many meters the voltage tends to go up until the meters switches to a range with a 10 Mohm input impedance. Then they get discharged, the meter switches to its lowest range and the process starts again. This can actually wear out relays if left like that 24/7. I tend to switch them to a higher range when not in use.

Auto-zero disconnects the inputs, shorts the ADC input and measures the offset, so this only compensates for offsets within the ADC and related circuitry, not the input voltage. It would get quite messy if you were measuring a voltage and the meter decided to zero out this voltage .