Debugging the Dead Power Supply

I don't think I mentioned it (and am too lazy to scroll down to check) but a friend of mine stopped by the other day and I was showing her all the stuff I've built so far. I loaded the switching power supply with 3 ohms to demonstrate the noise I get, but later I realized the -5 volt output was no longer working … odd. Tonight (Friday, by the way, and yes, I have had a few drinks — a chilled shot of Ouzo, a Saranac Pale Ale, and a Genesee can at home) I checked the power supply and inexplicably, the LM324 died — the op-amp that controlled the pulse-width modulation stayed at the positive rail for no reason. I pulled it and threw it away, even though it might have just as well been a loose connection. Replacing it worked just dandy. I tried connecting every LED — both strings, all three colors — and the input drew 0.53 amps for 0.76 amps out (in power, that's 6.36 watts in for 3.8 watts out for an efficiency of 59.7% — cool: almost 4 watts of output through LED's!) Anyway, I decided that the switching noise wasn't going to be much of a problem. Even with all three colors on one string lit, the noise is noticeable but ignorable … it adds an element of irritation. If I go to just one color lit (which, technically, is the load under bit modulation) the noise becomes nearly inaudible except very close. I have my fingers crossed.