Desulfator Issues Resolved
1. Added a 2N3904 transistor off the AVR PWM. This allows the pulse off my PWM pin to actually drop to zero. Instead of having a nominal 5V.
2. Swapped out wimpy inductors for high amp versions. I'm using 220uH and 1000uH still, but these can take over 3A each. They are bigger and more expensive, but show no signs of overheating.
3. Reversed firmware from being 95% duty cycle down to 5%. As usual I had reversed the counter idea which was another reason I was blowing up parts. I'm probably going to drive the pulser at a higher duty cycle, but I'll wait for the printed circuit boards to arrive before experimenting with firmware again.