Mulitmeter non-zero reading, nothing plugged in, EEVblog #99

[twistar]

Disclaimer: This is my first post here. Apologies if I've put it in the wrong place or broken any etiquette. I'm sure there is information on this topic available online but with a bit of searching I wasn't able to find exactly what I needed through google searching or searching these forums.

In  EEVblog #99: $100 multimeter shootout 4/6 of the multimeter under review read close to 0 mV (within 1 mV) throughout the test. However 2/6 of the multimeters read ~350 mV. See 5:00 into the video. Why is this the case?

I've also seen another strange behavior with multimeters. Say, for example, cable leads are plugged into the multimeter (DC voltage measurement mode). If have seen that if I attach one of the leads to equipment ground (conductance to third prong of US outlet) and then move around the other lead I can also see spurious voltages. This was taking place in a physics optics lab environment which had all sorts of low and high frequency and moderate power electronics all over the place.

Are these two effects related? My guess is that the second effect can be explained in terms of stray capacitance between the lead which is "flapping in the wind" and various voltage sources within the room.

The first effect is a bit more puzzling to me. If there was stray "static" charge inside of the multimeters I would think it would rapidly discharge across the measurement resistance in the multimeter. It's also especially striking that some multimeters show the effect right next to other multimeters which do not show the effect.

[EEVblog]

Disclaimer: This is my first post here. Apologies if I've put it in the wrong place or broken any etiquette. I'm sure there is information on this topic available online but with a bit of searching I wasn't able to find exactly what I needed through google searching or searching these forums.

In  EEVblog #99: $100 multimeter shootout 4/6 of the multimeter under review read close to 0 mV (within 1 mV) throughout the test. However 2/6 of the multimeters read ~350 mV. See 5:00 into the video. Why is this the case?

It's high impedance input mode (many Gohms instead of the normal 10Mohms).
Charge just builds up on the input cap, and this is normal and actually a feature.
All high end bench meter have a mode like this, and some handheld ones.

[twistar]

Thank you very much for your response! I've been looking up about high impedance input modes on meters and found a little bit of information but I am still not 100% certain about the behavior seen in the video.

I think I have an overly simplistic model in my head for the multimeter voltage mode operation. If I imagine an input impedance of 10 Gohms and say an input capacitance of 100 pF then I would still expect only a 1 s RC constant.

Thus if charge built up on the input capacitor I would expect it to discharge across the resistance in a second. Even if I'm off by an order of magnitude or two it would still discharge noticeably over the course of a minute.

So it seems to me like some noise in the room or something must be someone continuously building up the charge on the cap.. This is why I brought up the idea of stray capacitance to power sources in the room. Could be an AC signal rectifying on something in the meter.. I am worried I am way off base in my goose chase for what might be creating this signal in the disconnected meter.

I guess I'm trying to build a circuit model in my head to explain this behavior. Maybe it is too difficult for me to try to do this because at such high impedance very small currents can create large signals so everything just gets finicky..

Any additional direction is very much appreciated!

[MosherIV]

I think a fundamental mistake in your thinking is that you are trying to model the behaviour as a closed circuit.
The random voltage you are seeing on the mV range is due to stray picked up voltage and yes capacitance is holding it there. However, there is no closed circuit. Not in real life.
Yes, you can model it as a closed circuit but the circuit path resistance will be very very very high.

[EEVblog]

With a 1 Gohm impedance you are going to be picking up everything, stray capacitive coupling and electric fields. Trust me, it's normal.
If you have a seriously high impedance source you are measuring then there is an art in shielding things to avoid these issues.

Try:

[beanflying]

We were discussing induced voltages a while back in the TEA thread. For interest I put the 34401A on GOhm input and hooked up a set of un terminated 1M leads casually draped around some test gear. I was able to trigger the 10V relay with a but of fiddling with lead placement near the Counter and Generator.

I had an issue a while back too with a small SWplugpack running a 3W LED causing a DC offset at well over a metre while running some tests on a DC Calibrator with the same meter. No idea of if it was the Fluke or Agilent or the leads that was p[icking it up but the SMPS is in the bin. Switchmode supplies generally SUCK when it comes to sensitive measurements.

Stress less but know it is an issue 

[magic]

Thus if charge built up on the input capacitor I would expect it to discharge across the resistance in a second. Even if I'm off by an order of magnitude or two it would still discharge noticeably over the course of a minute.

So it seems to me like some noise in the room or something must be someone continuously building up the charge on the cap.. This is why I brought up the idea of stray capacitance to power sources in the room. Could be an AC signal rectifying on something in the meter.. I am worried I am way off base in my goose chase for what might be creating this signal in the disconnected meter.

It's leakage from the ADC or input amplifier.
This input signal goes to some electronics, those electronics contain various voltages inside, no insulation is perfect, some tiny current "leaks" and charges the input capacitance. Sometimes it drifts towards some specific value, sometimes it just keeps rising until it gets out of range.
There are no gigaohm resistors in those devices either. It's only a declaration that input resistance is no less than 10G or whatever.

[EEVblog]

There are no gigaohm resistors in those devices either. It's only a declaration that input resistance is no less than 10G or whatever.

Yes, it's just the FET input plus whatever leakage from any protection devices.

[IanB]

I think I have an overly simplistic model in my head for the multimeter voltage mode operation. If I imagine an input impedance of 10 Gohms and say an input capacitance of 100 pF then I would still expect only a 1 s RC constant.

Thus if charge built up on the input capacitor I would expect it to discharge across the resistance in a second. Even if I'm off by an order of magnitude or two it would still discharge noticeably over the course of a minute.

If it is going to discharge there has to be a closed circuit. Where is the return path for the current?

[Kleinstein]

The input amplifier and protection circuit tends to have some leakage current of usually some 10s of pA.  This input current can slightly depend on the input voltage, thus looking a little like a resistor in the > 10 GOhms range. However this can be nonlinear and temperature dependent and differ from unit to unit.
So the 10 GOhms are only a rather crude description of the input impedance. It is more like that the input current does not change by more than 100 pA if the input voltage is changed by 1 V.

The 100 pF number is at least in the right order of magnitude and this can also vary between meters (up to some 5 nF in some older meters - faster ones tend to have lower capacitance).

[magic]

Thus if charge built up on the input capacitor I would expect it to discharge across the resistance in a second. Even if I'm off by an order of magnitude or two it would still discharge noticeably over the course of a minute.

If it is going to discharge there has to be a closed circuit. Where is the return path for the current?

Wait, what? Are you suggesting that a capacitor can't be discharged by a parallel resistor?
Is it a revival of that "does current really flow through capacitors" thread?