.. to replace an old Farnell 30V, 1A supply that killed three prototypes. Transient over voltage, to >45V for a few tens of nanoseconds, when switching the output on and off. How they managed to generate that short a pulse using a circuit built round 741's and 2N3055's is a mystery to me.
Inductance + dI => ringing? What was the PSU's max output?
But I'm surprised that short a transient wasn't removed by whatever is across the UUT's rails.
The circuits, when undamaged, we're drawing about 20mA at 30V. The absolute maximum for the chip was 33V or so, can't remember exactly. The spikes only arose when the DC switch contacts bounced, and the feedback loop was momentarily broken as the sense was O/C while the output was not. I think it was a really bad move switching the sense terminal as well as the output, not a good design.
I suppose the decoupling on the prototype reduced the amplitude, but increased the length, of the over voltage transient. The effect on the chip wasn't a hard fail, the circuit just behaved weirdly (wrong voltages, low gains) but you expect that to some extent with a prototype. The transient didn't happen every time, and it was only when the third prototype worked correctly at first, then didn't after being power cycled, that I started to suspect the supply.