I have a setup as drawn below: DC source of 15V connected to high voltage (HV) source of 1500V connected to photomultiplier tube Hamamatsu, R1-R3 are resistors builtin the socket of PMT (there are actualy 10 of them and some capacitors...). R4 is load resistance to convert PMT current to voltage. I am measuring voltage with oscilloscope. Oscilloscope is powered from isolation transformer. Input DC voltage is also powered from power supply with transformer, so the grounds should be isolated.
My problem is shown below on images: if I connect oscilloscope probes in one polarity I get expected result (photon excited pulses), image 1. But if I change polarity I get weird ripples on 13kHz, image 2. I believe it is the ripple from HV power supply. But I don't understand why am I measuring it. If I change probe polarity I would expect just the change of polarity of measured signal (peaks up, not down).
This makes me puzzled most: it is the same when I use batteries as 15V DC instead of regulated power supply.
Why do I get these results? I want to count pulses but it seems that when I try to do so with ADC and Arduino powered from USB I am counting just the ripples or some similar noice, not the peaks. Is it problem with some ground loops? How can I avoid this? Using instrumental amplifier with R4 in negative feedback loop but with similar result.