Inside Look of 2018 On-Board Charger (OBC2) OBC Gen 2

[stoatwblr]

I think it's as simple as a poor quality capacitor failing spontaneously well within its supposed limits, or perhaps being under-speced for it's role in the circuit or a bit of both.

There's always the old chestnut of moisture getting into the HV container. It might even be a contaminant in the potting mix causing breakdown over long periods

It could be something as simple as tin whiskers. Lead free designs were causing major problems with space-rated hardware where I work well into the 2010s and we found that conformal coating thickness/uniformity was utterly critical in keeping this under control

 

[kriiise]

It's not clear to me exactly where this tiny spark was, but it's near a gate (left) lead. Perhaps it was a gate component arcing internally.

Yeah, i drew the arrow on my phone without zooming properly. The sparks came from the tiny bits and bobs to the left (in the image) of the gate lead(?) of the transistor. It may well have been from inproper cleaning after the last blow. All i did was spray a bunch of electric cleaner in the areas where i could see marks from the last blow.

replace the damaged transistors again, then test and as necessary repair every gate component before re-applying power.

Is there any way you can explain how to test the gate components for someone with a very basic knowledge of PCBs? Someone who calls the smaller parts for bits and bobs If you'd have to explain 6 months of college level engineering, then i get it if you can't be bothered :lol:

 

[coulomb]

Is there any way you can explain how to test the gate components for someone with a very basic knowledge of PCBs?

It's a matter of testing the components that drive the gate. Usually, that's resistors, diodes, driver transistors, and chips. The resistors can often be measured directly with a multimeter, but it helps if you understand how circuits work so that you won't be surprised when other circuit elements cause a resistor to read lower than its nominal value. Diodes can also be tested with a multimeter; they should read about 0.6 V in one direction (0.4 V for the larger ones) in the forward direction, and open circuit (on the diode range at least) in the reverse direction.

Chips and transistors usually have diodes in them, and for this sort of testing you can treat a bipolar transistor as two diodes in anti-series. Most chips and MOSFETs have diodes as well. Gates should read open circuit. The diode checking works well, because fault currents usually cause diodes to fail, usually short circuit, but occasionally open circuit or high leakage (medium resistance).

After these checks, I usually like to perform a powered test of the gate driver at DC. So with the circuit powered up and a multimeter measuring the gate to source or gate to emitter voltage, I'll short something to cause the gate voltage to change. Usually this requires the big transistors not to be soldered in place yet. In other words, you take out the old transistors, do the test, then solder in the replacements.

If this sounds complex, it gets much easier with time and familiarity. Not something you can really convey in a post, sorry.