Agilent 34401A probe arcing when measuring 500VDC

[Luciano2572]

Hello all;

I was doing some fixes on my tube amp, and when I went to measure a voltage of 475VDC (relative to amp chassis), I noticed a small arc on the tip when I touched the positive probe to the terminal under test. I was able to measure it, but I could hear some interference on the amp.
Measuring another -60VDC point, resulted in a loud pop from the speakers so I didn't even get to measure.

Amp is NOT grounded to mains earth; it is powered thru a bucking transformer which is also not grounded.
Agilent 34401 is also not grounded to mains earth (I do not have earthing at this building at all)
One thing, I have three phases on my lab, so the DMM is on one phase and the amp was on another just due to the outlets I used. But still..

Is there a catch somewhere I'm missing? I know we have to pay lots of attention to grounding and shorts on oscilloscopes but on dmm's?
Attached a picture of the amp under measure just incase.

Many thanks in advance!!

[Kleinstein]

The 34401 has some capacitance (some 700 pF)  in parallel to the inputs. With a high voltage this can be enough to cause a small spike and upset the amplifier to produce a a spike at the output.
There is also some capacitance towards ground / mains.

When using the AC mode that is an additional capacitor, though with resistors in series. So usually no spark but a longer pulse to the circuit.

[Gyro]

The 34401A is a bit of a pricey instrument for probing around in tube amps, especially with a dodgy grounding situation. Regardless of what is going on (check the Lo input to ground rating too).

You'd probably be better risking a handheld instead. Fluke 101?

Remember that the meter has a 10M input resistance on that range, in addition to the input capacitance, which would certainly cause pops on high impedance nodes (Edit: -60V grid bias?)

[floobydust]

The small arc is due to the multimeter's input capacitance, they all have some. The resulting burst of EMI can cause some DMM's to reboot or crash.
Old Fluke DMM's had safety recalls because if it, over 18 years ago Fluke 170's 2002 recall and there's another for the 87 family, where it would crash when connected to HV and go blank or read a silly value.
 
The 34401a DMM-portion is floating but has capacitance to earth-ground. It's not an industrial or "tough" multimeter, Cat. II rated. There was a problem (might have been an HP recall) with C109 0.01uF 1kV being cheap and failing short circuit, I replace them on all older 34401a's.

For tube gear under 1000V, I manual switch to the 1000V range (never let it autorange down to mV) and add a series 10k resistor to the probe to stop arcing. Probing the output tube plates can cause drama, the most dangerous place to poke around.

[TimFox]

In the heyday of tube amps, it was common to use a VTVM (vacuum-tube voltmeter) to probe high-voltage nodes.  These meters included a 1 megohm resistor at the probe tip, before a shielded cable to a 10 megohm voltage divider at the meter circuit input.  This isolated the meter and cable capacitance for measuring nodes with high-frequency content.

[Luciano2572]

Hi folks!

Sorry for the late reply - thanks for all the replies! I'm learning my way into tube amps - I was able to fix this one (was fairly easy, a blown 10Ohm resistor, then verified the voltages but ended up blowing the fuse somehow when measuring with my fluke 179)

I've noted down - fix the range to 1kv and possibly 10k ohms resistor in series! Thank you very much

Probing the output tube plates can cause drama, the most dangerous place to poke around.

Why is that? Should I avoid that completely or is there any tips for that?

I am currently an electronics engineer student, but I have had 0 classes regarding valves - and as I'm into audio and recording, I'm trying to chase that by myself. Currently looking for good materials to study on

Thanks again everyone

[james_s]

People have been killed doing that. You really have to be careful with tube gear, especially audio power amps that have plate voltages in the 400V+ range, if you make a mistake you may not get a second chance. I avoid live testing unless I absolutely have to.