Jasen Betts wrote:
On 2018-08-19, Cursitor Doom wrote:
Gentlemen,

One of the drawbacks of attempting to fix vintage stuff is the expected
voltage readings given in the service manuals of the day. The manuals
usually state that the readings given were measured with analogue VMs of
a certain ohms-per-volt rating - most commonly IME 20k. Consequently if
you measure with a modern DVM with stupendously high Zin you're screwed
and will get unrealistically high values. That's never worried me as I
keep a vintage AVO for just such circs. All the British service manuals
seem to reference 20k OpV AVOs. However, I'm currently TS on a mid 70s Tek
scope the manual for which states the readings given are valid for a
meter with a Zin of between 100k and 200k (specifically a Triplett 630NS
see link).

Anyone come up with a solution to the problem of making voltage readings
on high impedance parts of a circuit with a meter of a different Zin to
that used by the people who wrote the service manual?

if too high add parallell resistance.
if too low select a higer voltage range.

In the old analog meters, there basically was only a current meter,
often 50uA full scale, and when measuring voltage a suitable series
resistor is switched in to make it draw 50uA at the full scale reading.

So a 10V range would have a total resistance of 10V/50uA = 200k
(which would be the resistance of the meter itself plus the series R).
At 10V measured voltage there is 50uA through the 200k resistance.

When looking at this, any range will have a resistance of 1V/50uA per
volt of range, hence "20K per volt". The 100V range will be 2M.

This is no longer true for a modern DVM. They usually have a 10M
series resistor on the input with selectable resistors to ground to
make a voltage divider that outputs the desired voltage for the ADC.

So, depending on the range you select, the input resistance will be
10M plus a small value that will get smaller when you select a higher
range.

Therefore there is no fixed "K per volt" input resistance anymore, and
selecting a higher range will not result in a higher resistance.


However, as already can be seen, the "20K per volt" is not really
telling the input resistance to be used in the measurement.
It depends on the selected range, and available ranges vary between
meters. One may have ranges of 10-30-100 and another maybe 10-50-200.
When you need to measure a 24V testpoint, on one meter it may be on
the 30V range (and thus 600k resistance) and on another meter it would
be the 50V range (and thus 1M resistance).
It is expected that the person doing the measurement understands how
this could affect the result, if it does at all.
(when measuring a supply voltage, there should not be a noticable
difference. when measuring in a high-impedance signal circuit, there
could be)