High voltage (AC mains) differential measurements with 1:100 scope probes

[burino]

I have a Rigol DS1054Z scope and I wonder if is it possible/safe to troubleshoot the primary side of a 230V mains-connected Switch Mode Power Supply (for instance the Power switching transistor) using two high voltage 1:100 oscilloscope probes such as these Hantek T3100, using the scope A-B differential signal:
http://www.amazon.com/Hantek-Oscilloscope-Voltage-Passive-2000VDC/dp/B0030L0X58/

The Rigol RP2200 10x probes coming with the scope are not safe to use for mains voltage measurements, as they are only rated for 300V AC. 

The alligator ground clips cannot be removed from the Hantek probes, so where would I connect them to minimise noise injection? The primary side of a SMPS has a hot ground which is not at earth (0V) potential, so I certainly cannot connect them there. Is it safe to connect the 2 alligator clips together and ensure that the connection point doesn't come in contact with anything else while the Power Supply under test is live? Would the measurement I get be accurate?

Would a much more expensive differential probe be better/safer than 2 separated probes? I know it would in terms of common mode noise rejection, but does it matter at the SMPS operating frequencies (10KHz - 1Mhz)?

Thank you very much for your help.

[uncle_bob]

Hi

The "ground" is the problem. The switcher is likely operating at a high enough frequency that a two probe A minus probe B differential setup will be a challenge. The probes you show are *safer* for a high voltage circuit, but they are not *safe* in an absolute sense. Anything that comes along the mains feed will hit those probes. It is not unheard of to see crazy spikes.

Other ways to do this:

1) Get a nice big isolation transformer with all the proper leakage and shielding specs on it. Invite a large friend over to carry it to your bench for you This will let you feed one side of the supply with a grounded feed. Still not ideal, but not quite as crazy.

2) Get a proper differential probe with high voltage certification. These probes have an external (analog) difference amplifier and proper protection. You can buy quite a few 1054Z scopes for the price of one of these beasts.

3) Consider what you are trying to look at. Is all the "stuff" you want to see above 10 KHz? If so, a couple of high voltage (> 2KV) capacitors may be what you want to get first. That gets rid of the nasty mains voltage (high pass filter) and gets things down to (maybe) op amp territory.

Lots of things to think about.

Bob

[burino]

The "ground" is the problem. The switcher is likely operating at a high enough frequency that a two probe A minus probe B differential setup will be a challenge. The probes you show are *safer* for a high voltage circuit, but they are not *safe* in an absolute sense. Anything that comes along the mains feed will hit those probes. It is not unheard of to see crazy spikes.

The ground is one of the problems indeed. That's why I asked if connecting the two probe ground clips together is a good idea or it would be better/equivalent to connect them to the earth conductor of the device under test or leave them disconnected (I guess not a good idea as they would probably pick up noise).

Thanks.

[uncle_bob]

Hi

You will only really know how the noise "works" when you get hooked up to the power supply. By then it's a bit to late to buy a different probe If you do hook to any part of the supply side, check what is feeding it. I've seen an amazing number of wiring errors over the years. Hooking to a ground that didn't get wired is not much protection at all ....

Bob