Dodgy PP film capacitors?

[Artlav]

I was making a scope CRT into a vector display, and ran into a peculiar problem.

There were occasional artefacts on the screen - curles and spirals visible for fractions of a second, almost like what you'll see in a cloud chamber.
I couldn't make heads or tails of it, and went all the way from firmware through wiring, to DAC and deflection amps - everything was fine.
Turns out the problem was in a completely different place.

The thing needs 1000V anode supply, of a few mA.
A simple push-pull converter from 12V.
Naturally, it have an output smoothing capacitor.

I used a PP film capacitor, rated for 2000V.
These ones:


And guess what?
They were the problem.
I noticed it purely by accident - having placed some electrolytics in series in it's place to rule out any power supply noise, seeing the problem vanish and then scaling down trying to find the right capacitance.

Electrolytics - no problems.
Two 630V WIMA capacitors - no problems.
Either of the two capacitors above - problems.
Either of the two of them in parallel with anything else that worked - problems!

I guess it arcs on the inside, and causes the voltage to dip momentarily, making the beam wander on the screen in a peculiar fashion.

Is that normal for this type of capacitors?
Or are they perhaps rated in Chinese volts (3 for the price of 1)?

[SeanB]

Probably a 600V film being way overrun. Thin leads as well. Only way is to crack the case off and unwind it to see the massive areas where the foil coat has been arced away. Try using 3 in series, that will likely work well enough, or buy real brand name 2kv caps which are likely to be much bigger.

[Richard Head]

Are you sure those caps are polypropylene and not polyester?