Electrical – Oscilloscope ground clip : short circuit

Equipment with a 3 prong plug generally has a capacitor called a Y-cap connected between the metal frame (which should be tied to earth ground) and each side of the line.

This is in order to bypass high frequency EMI. It also allows a small leakage current to flow into the case.

If the case is actually grounded, no problem. If the 3rd prong of the plug is floating for some reason you will get a small spark when connecting the chassis to ground. So either your scope or the chassis of the power supply is not connected to earth ground.

Regulations concerning Y-caps ensure that there's not enough leakage for a severe electrical shock, but you can feel a tingle if you come between the two. In this situation a shorted Y-cap can be a big safety hazard, but fortunately these caps are designed not to fail short.

John

I'm not sure what your 13 kHz ripple is, but it seems more like power supply ripple than a ground loop. If your HV supply is a positive supply, then the PMT 'ground' is going to be ripple-y. The floating ground lead on your scope presumably has a lot of parasitic capacitance, since it goes to the chassis and everywhere inside. When the ground lead is connected to PMT ground, there's not much capacitance on the anode, but when the ground lead is on the anode, there's a big chunk of capacitance to charge and discharge with the ripple. That's my speculation.

In general, what you want when dealing with pulses is a "pulse discriminator", which lets you put conditions on the height, length, etc. of the pulses you're interested in. Since you're pulse counting, you probably don't need anything fancy, a simple comparator will do, but it will need to have sufficient negative input range. You can build a simple one using an op-amp, but you can get ones off the shelf that are faster and with logic level output. Alternatively you could use an inverting amplifier topology to flip the signal, but you will still need negative input range and sufficiently high bandwidth to pass pulse waveforms through.

Edit: One more thing, isn't the max voltage on the R1166 1250V?