Why does a calibration procedure specify the use of an analog voltmeter

[TheUnnamedNewbie]

I have acquired some old Philips PM2527 multimeters that I am restoring as a project. I started calibrating one of them yesterday (using a calibrated keithley 2000 to compare to).

However, in one of the steps, they specify the use of an analog voltmeter. I have one, so that is not a problem, but I was wondering why they might do so. In my experience, the resolution on an analog voltmeter is far inferior to any digital meter with more than 2 digits, let alone something like a 5 1/2 or more digit bench meter.



So what reason do the authors have to specify this? And is that still applicable in modern day?

[oldway]

Checking if a voltage is less than 1mV  is not really a calibration operation even if it is part of a calibration procedure.
No need of high precision.....an analog voltmeter has often a low DC voltage range.....and is easier to use if this voltage is varying.
A digital voltmeter can have an offset and an analog voltmeter has not. (0 of the scale can be ajusted)
But, of course, nothing wrong if you use a digital voltmeter.

[Kleinstein]

That part of the circuit might be noise sensitive. Some DMMs do inject some RF noise back, somewhat similar to AZ OPs and this way might interfere with the circuit. In the old days, DMMs might have more capacitance to ground.

[Richard Crowley]

Analog meters are also more user-friendly when "tuning" for minimum or maximum.  Where it isn't the exact VALUE of the measurement that is of primary importance, but the "SLOPE" of the function.  That is why higher-end DMMs include a "bar-graph" as a crude substitute.

[CopperCone]

Imagine it as a mechanical low pass filter that will nicely bounce between the peaks, so the information is much easier to identify then trying to see numbers bouncing around. Bar graphs suck in comparison.

It would be interesting to see the transfer function of a mechanical display.

And yes, modern equipment is sparse on HF filtering. If you want an example, plug a multimeter (like a bench one) into a spectrum analyzer. There is infact a market for devices that don't use MCUs, or have a disable function, particularly for cryogenic work, where the noise might heat the sensitive equipment (for real!).

[David Hess]

Sometimes the specified readings depend on the relatively low input resistance of an analog meter but I doubt that is the case here.

[TheUnnamedNewbie]

Interesting! I'll have a look at what I do then. I might as well use my analog meter, since I pretty much never use it anyways... good excuse to get rid of the dust.