Q: proper measurement methods of voltage in the kV range

[David, K3TUE]

If I were making a sizable plate PS (1-5kV at 0.1-0.5A) I would want to put a voltage meter on it (more knobs + more meters = more better).  But the idea of sending 5kV up to a panel meter seems, well, unwise.  As best I can figure you could measure this voltage by send a much lower voltage to the panel meter after proper reduction.  Here are the 2 methods I can think of:

E method:
You could make a voltage divider with 2 x 1W resistors in series.
With 25M on the (+) side and 25k on the ground side,
you could measure the V drop across the 25k resistor (0-5V).
(NOTE: Eventhough these numbers are 0.1% off (actually 0-4.995V), I figure, since meters seem to run 2-5% accurate, the additional inaccuracy is insignificant at 5-2% of 2-5%)

I method:
You could measure the current through a 2.5W resistor.
With 10M on the (+) side and the meter on the ground side,
you could measure the current through the resistor to ground (0-500uA).
(NOTE: I am sure a similar inaccuracy will result here due to the resistance of the meter, but I suspect the effect of the inaccuracy will be similar to that noted in the E method.)

My question is, which method is more commonly used and presumed to be the better method?  Or is there another method that is even more common and better still that I did not think of?

[WA1GFZ]

I put 6 1 meg resistors in series with the positive lead and mounted the meter behind the panel with a 1/4 inch plexiglass window.

[David, K3TUE]

I put 6 1 meg resistors in series with the positive lead and mounted the meter behind the panel with a 1/4 inch plexiglass window.

And measure the current through them all or the voltage drop across one?

[KB2WIG]

To amplify  'GFZ's method,Ya dont want to put too much juice accross the resistors..   having 6 (equal value) across a 2.5KV supply  works out to 420V per resistor ;~~ %15 below the 500 V Murphy's Law design working voltage...

I use 1kv hookup wire across the 'bottom' resistor of the VD network to the meter terminals. Currently, there is nothing I have over 2.3KV so I don't feel a need to isolate the Plate V meter from the rack.....  Still, its not a bad idea... a few plastic spacers and a sheet of plexie...... The Plate I is pretty standard,  works fer me... ...  klc

[Don, W2DL]

Problem with using 6 - 1 megs in series, low end through meter to ground izs if the ground terminal of the meter comes off ground, or if the meter opens (granted, not a big chance of either) you then have full plate volyage, 2.5 kv or whatever, on the meter. Much better idea, put the 6 - 1 meg resistors in series with two 10K 1 watt (or more) parallel connected, the low end of these connect to ground. Connect meter from junction of 1 megs and 2 - 10K's in parallel to ground. Now, even if one of the 10 K resistors opens up you still have a meter essentially at ground potential. Chances of both 10K's opening up? very, very small - but that's why to use 2 in parallel. Since the Ma meter used as a voltmeter (reading 1 megs resistor current) has very low resistance the 2 - 10K's won't affect accuracy at all. I had a meter open up once, was very glas the HV then didn't appear across the terminale!

73
Don, W2DL

[WA1GFZ]

That is why it is mounted behind the panel and 1 ma should not kill you.
Actually better to use more low value resistors. Resistors have voltage ratings as well as power ratings.
I tend to stay out of the final when it is alive. My old 4-1000A rig stood the hair on my arm once when I was reving the motor with the hood open. Almost got me raw plate power no current limiting.