Cap Leakage Tester

[Aurgelme]

Hello,
I need to test 650V volt electrolytic capacitors in some welding machines. So now I have started building a tester according to drawings from M Caldeira (youtube Electronics Old and New by M Caldeira)

This tester goes up to around 340V, but I want to test at 500V.
Have changed C1 to 10uF/650V and Q2 to 900V/15A 200W N-fet I had lying around.

Since I don't know much about calculating components, is there anything else that should be changed?

[johansen]

are you sure those are 650v capacitors?

most everything i've ever found uses 2, 400 or 450v capacitors in series to get the dc bus from a 380 to 480v ac supply.

anyhow you can certainly charge them up slowly using a milliamp meter to measure the current, and then disconnect the capacitor and measure the voltage 10 minutes or 2 hours later with a 10M ohm volt meter, the difference in voltage then translates directly to leakage current.

[Aurgelme]

Yes, there are only parallel capacitors from Chemi con, marked 650V 4700uF and cost around $300 each.
The data sheet also shows how to check insulation resistance, and I have an insulation tester that goes up to 5kv.

When measured between the terminals that are connected to each other and to the mounting clamp on the insulating sleeve covering
the case by using an insulation resistance meter of 500Vdc, the insulation resistance shall not be less than 100MΩ

But this was mostly for the fun of being able to build this instrument. Later I will restore an old jukebox with a tube amplifier, thinking of keeping most of the original parts including electrolytes.

[Victorman222]

Most of the calculations here are Ohms law. All resistors and switches need to have their maximum working voltage rating above 500V(or whatever is your supply voltage). VR-B1 will have 410V on it if the supply voltage is 500V and one of the resistors is shorted by the switch. R1 needs to be able to handle the pulse currents from charging the filter capacitor for reliable operation (wirewound or special film resistor types). R11 could be bumped up to 5W or a capable pulse power rated type. If the current limiter fails R7 could dissipate 21W at the beginning of capacitor charging, so probably choose something like a wirewound that can handle the pulse and wouldn't fail burning. Take care!

[Aurgelme]

Thanks for the hints,
I have to learn to start calculating on this, since I'm completely new to building electronics without finished drawings and pcb kits.

I have mostly worked on repairing (old) 380V systems and industrial machines, so I have tested quite a few volts through the body, but everything has gone well, even we as boys had fun peeing on the electric cow fence and the grandchildren don't seem to to have been injured.