High Voltage Scope Probing

[Randy222]

My need is to hunt down noise and harmonics that come from VFD based motor drives, using scope FFT.
I looked at some differential HV probes for the task, some are very expensive.
Beyond price, it's not clear to me how an active diff probe may mask what I am looking for, or perhaps even introduce artifacts.

So, what about a simple common mode divider?
R1, R4 1Meg
R2, R3 10k
M1 is a VFD & motor
Mains is 240Vrms 60Hz w/ center tap

And then I wonder about common mode on scopes that now use a wall wart to power them, like the little Rigol 800/900 series? Does a channel "gnd" run all the way back through wall wart to mains gnd (aka "EGC"). NOPE. But they do have a gnd "lug" in the rear that is zero ohms to channel BNC outer gnd connection.

Thoughts on this design?

[EPAIII]

Thoughts? Well unless that is a step-up transformer and the secondary Voltage is over 300 VAC and assuming those are 10X or 100X probes, I would:

Remove R2 and R3.

R1 = R4 = 0 Ohms.

Fuses are OK, I guess, but plain wire would work as well.

Set Ch1 gain = Ch 2 gain.

Set scope inputs to display Ch1 - Ch2.

Works exactly like a differential probe. But is a lot less expensive.

[EPAIII]

Oh, and I would connect the probes to a scope. I doubt that you will see anything without that.

[Randy222]

The xfrmer is a step-down to 240Vac(rms) 60Hz with a grounded center tap. Std mains in US.
Yeah, was not expecting to get readings from unconnected probes.
The probes would be 1x. I would be probing each side of the motor to see which side is creating the noise I am looking for,
The divider is to knock down the voltage going to probes. As-is, each half is 0.118mA, so the probes should be seeing around 1.18Vac(rms)

[DL2XY]

Stay with the voltage dividers, 240V is too much peak for most 10:1 probes.

To get usefull measurements you will need a Line Impedance Stabilization Network between transformer and DUT.

[oz2cpu]

this page explain a few good things about scope voltage probe DIY for high voltage
at higher speeds too

https://webx.dk/oz2cpu/hi-volt-probe/hi-volt-probe.htm

[trobbins]

Nowadays you could also 'go battery' and avoid mains connection and parasitics, as the example Rigol is 12Vdc powered.  Even before that an option could be to use a laptop with an external soundcard and 100:1 scope probe with spectrum analyser software for assessment up to 96kHz.

[wasedadoc]

I would be probing each side of the motor to see which side is creating the noise I am looking for,

in which case a differential measurement is NOT what you should do.  If you see the noise you cannot tell which side, or both, it is present on.  Just probe each side w.r.t. ground independently.