Only time I felt good after successful debugging was when I found an obscure bug in the Cyrix x86 clone (non-aligned DWORD push across page boundary with lower page not present). Most of the time you find it's your own bug then sit around wondering what else you f***** up.
I love the product cycle. You work out all the known bugs, start feeling cocky about your code, chant "ship it, ship it". Your product manager says ok, you sober up from the RTM party, then nervousness increases as the time for retail availability gets closer. On the day of retail you wait for the phone to ring, check for disaster headlines on PCWeek/Infoworld/etc. Even if your part doesn't cause a s***storm, the old saying "I'd rather be lucky than good" keeps coming to mind.