Two-signal common mode spike suppression for scope
John Duffy wrote 06/28/2026 at 18:42 • 0 pointsI'm working on a test setup where I've got a problem with a common mode spike. It's a high voltage (8kv!) capacitor bank discharging in about a microsecond, so just the parasitic capacitance to ground (I assume), causes enough dV/dT to disrupt measurements. I'm measuring voltage and current with a 1000:1 divider and a shunt, respectively, with coax running to an oscilloscope. Those signals are ~ a couple V once on the coax. The power supply for the bank is floating, so theoretically there should be no ground current on the coax at all, but even with a shorted input to the coax, I'm seeing a couple volt spike on the scope. I can't seem to post a schematic of the setup here, I'll see if I can in a comment. The coax/scope ground is connected to one terminal of the capacitor bank, and the divider/shunt are referenced to the same ground. A separate, 16awg ground lead made no difference, I'm guessing any wire length has too much inductance given the ~10kV/uS but I don't see a practical way to get a low-z ground between the scope and test jig (few feet, for safety given the voltages).
I've tried wrapping the coax around a ferrite core to make a CM choke and it maybe gave me a few dB, but I need like 20dB reduction to get acceptable snr.
Since there's 2 signals, would two separate CM chokes work, or would the ground current not distribute in a way to cancel? Or could I put three windings on one core (one ground going one way, two signals going the other)? Or is there a better way to do this? I don't want to build a bunch of active circuitry to do fully differential measurements or buy a couple hundred dollar probe.
ask
Discussions
Become a Hackaday.io Member
Create an account to leave a comment. Already have an account? Log In.
what is the purpose, context would help
Not sure what you purpose is, but if you are talking about a start point spike, that is correct, it should. incandescent light bulbs come on instantly for a reason, they don't start glowing slowly as the carbon heats, it draws extra current, it spikes a basic meter at the start. A choke on a capacitor is basically a capacitor on a capacitor or a choke on a choke. try this. get any line tester roll up a 20 metre lead and plug anything in, the lead in a coil is a choke, it will create a magnetic field, but the power will slowly rise as the copper warms reducing impedance, just because there is no iron using carbon to create an electromagnet, just like the lead, it still produces a magnetic field. the chokes cause spikes on an oscilloscope just like the light bulb.
or are you building a deliberate spike device?
Are you sure? yes | no
very much a deliberate spike :)
8Kv/uS, it's approximating the spike they put into multimeters to test input protection. The whole event is over in about 50uS, not remotely long enough for thermal effects.
I've got about 10' cables now, with a choke on it, but that doesn't seem to be enough inductance, or not the right move
Are you sure? yes | no
of course its long enough for thermal effects, the cable is a sink. and the amount you are talking as an input is massive, so the fact the sink is so low means you are as close as you will probably get. you will never get a zero fluctuation. the closest will be carbon cables. its just physics thermodynamics energy to heat path of least resistance. on high voltages we lose 20-30 of electricity to the power lines, even in a closed loop you will lose energy to magnetic field attenuation. like an mri is a very pin point magnetic field, yet attenuation means you still cannot be in the room with a pacemaker or metal on you, not just the person in the mri machine.even most test equipment is calibrated for these things which on a visual appears perfect, but its not.this is just attenuation
Are you sure? yes | no