Teardown Tuesday: Xenon HID Ballast

I was hoping to have more insight into this... but there's too much potting to make it worth the hassle.To make a long story short, my headlight died on my bike a couple weeks ago. I've replaced the headlights with HID lights. Shh... don't tell the authorities. I have to say that they are a crapload better than the original halogen bulbs.Well, there's three things that can go wrong:

1. Power isn't being fed to the assembly
2. The ballast took a crap
3. The bulb burned out

Given that I have two of everything (also swapped the high beam) it makes the diagnosis easy.I just unplugged the high beam's ballast and attached it to the running light and it lit up fine. I also measured the voltage at the source and sure enough I had 13.8V just the way it should. The ballast it was. Until the replacement showed up I swapped things around so I had a working running light and the dead ballast was attached to the high beam.Since then I got a new ballast and here's the old, dead one.That'll knock you on your ass. Watch out.Being an automotive thing, everything will be potted to hell. Same with the back side:The gap in there isn't anything to worry about. There was an additional backing board that was sealed to silicon potting stuff.A bit of screwdriver work freed up a bit. Nothing of real interest down on the ass-end of the board. The only things that are mildly interesting are the long, non-solder-masked traces that got hit with the wave solder to increase the current carrying capacity of the traces.There's an isolation slot cut in there as well to separate the high (23KV) section from the low voltage side on the left.Interesting. There's a PIC16 on the riser board. No doubt this guy is running the show.In many ways it's interesting to see that everything has a microcontroller in it now. I was looking around for an in-circuit programming connector but I can't see one at first glance. I wonder if I could pull the programming off the chip itself. Hmmm... Here's the data sheet for the PIC16F716 in case you're interested.Connected on through to the final GIANT output inductor (the big box with the wires coming out) is a bank of four IRF840A power MOSFETs. I would have expected them to be ganged up in parallel, but from an examination of the board it seems that's not the case. There's another small riser board next to them that control each of the gates individually. Go figure.The flow, as best I can tell, goes something like this:

1. Input
2. Filtering caps
3. First stage boost converter running through the transformer
4. A bit more filtering
5. High power MOSFETs driving the final...
6. Inductor/capacitor module to provide the high voltage output

All orchestrated by the microcontroller.Sorry that I don't have more info. I was hoping it would've been more interesting, but I didn't really know what to expect.