When May worked at Intel, the IC's would encounter unexplained state changes from time to time.

He looked into it - and eventually realized the problem was the ceramic/clay semiconductor packaging was slightly radioactive, causing single event upsets. Intel switched to less alpha-heavy packaging, plastic, and the problem went away. That's impressive debugging!

I knew him more from his cryptography and privacy work (and his curmudgeonly Usenet posts on how downtown Santa Cruz was going down the drain, in his opinion), but this feat always impressed me.