Metrahit 26S mV range displays "over limit" when no leads are connected.

[peertux]

Hi,

If the leads on my Metrahit 26S are not in contact with each other in the mV or lowest V range the number on the display ramps up from 0 to OL within a couple of seconds. If I short the leads it displays 0v, measurement are also correct.
The higher V ranges work as it is supposed to do and display 0v when the leads are not touching each other.



I suspect that the stacked capacitors that are connected to the resistor array in the photo are the ones that cause this, when I measure them incircuit they bounce between 0.043uF and 0uF.
The soldering on the capacitors next to them also seem to be a bit off (flux residue)compared to other parts on the board.



Could the capacitors be the culprit or do I need to look somewhere else?

[EEVblog]

I would presume it's a high impedance mV input, in which case that is normal, it's just charge build up on the input. It goes away if you short the leads.
Many high end meters have this, it's actually a deliberate "feature"!

[KhronX]

Spec sheet says mV range input impedance is ">20M ", as opposed to 10-11M on the other ranges so that may well be the case 

The 100mV, 1V and 10V ranges on my Keithley 2015, where the specified input impedance is ">10G " behaves the same way - direct over-range on the first two, and the 10V range, after maybe 1-2mins now shows -5.7v and slowly creeping further.

[mmagin]

I would presume it's a high impedance mV input, in which case that is normal, it's just charge build up on the input. It goes away if you short the leads.
Many high end meters have this, it's actually a deliberate "feature"!

The "feature" as I understand it is that they precharge the input after shorting it for auto-zeroing, before connecting it back to the input, so as to not steal charge from the input to charge the amplifier input capacitance.  Interestingly, they all seem to precharge slightly extra, so the disconnected input drifts up instead of down.  (At least in my HP 3478A and 3457A meters anyway).

I suppose it would be difficult to design and calibrate it to reduce the drift by more exactly precharging the input but does anyone know why all the implementations seem to have this slight upward bias instead of a slight downward bias?

[Kleinstein]

The AZ precharge is there to reduce the input bias current. It can still have both directions, depending on the individual unit and adjustment.

A hand held meter might even omit such features as precharge in the higher end (e.g. 6 digits) meters. So input bias can be higher. An interesting point might be measuring the input bias, e.g. by having a 1 M or 10 M resistor at the input or watch how fast a 1nF (or so ) low leakage cap is charged.

For a high Z input this type of drift is normal. I would expect the bias to be 100 pA range, maybe even lower.

[Alex Nikitin]

Out of curiosity I've measured the input current and input impedance on 300mV DCV range of my MetraHit 28S (the spec just says the same as for the 26S - >20M).

The input current at 0V measured with the Keithley 617 - about 1.8pA. I've also measured +/-100mV and 0V voltage source directly (the meter was spot on, with less than 5uV error on these three points) and through 1Gohm resistor.

The measured values with 1Gohm in series with the input were about +101.7mV, -98.3mV and +1.7mV with the noise variations on the readings about 0.5mV p-p. Essentially the "input impedance" is over 50Gohm with the difference created by the input current only. It takes some minutes for the voltage to settle completely with 1Gohm resistance in series, indicating a rather largish input filter capacitor on the mV range - probably around 100nF.

I am not sure if the input stage on the 26S is the same, but it might well be the case.

Cheers

Alex

[Vtile]

If it is all about the same as my 100Meg input AMM, it also shows all electic stray fields nearby (mainly static charges). In example it do have direct correlation if I sweep my feets on the floor (just some regular home flooring) or walk past my poor and cheap work chair (with plastic fabric) or put my foot on top of power cord of any device.

Power Electric engineers and technicians sometimes call these phantom voltages. With ie. passive analog meters (or cheaply build DMMs or some meters that are for electric work might have intetional low impedance mode that have some added [we get parallel connection etc. I don't know how they do it .. ask electronics dudes] resistance between measuring leads ) you typically not see any of these.