Mike, I believe the spring is bent due to a function of two factors.Steve: Put on your thinking cap. Why is the spring bent?
A) The safety plunger is too small in diameter for the tunnel in which it resides. This gives it considerable "wiggle room" side to side/up and down. There is a lot of "play" to the plunger when it is in the tunnel in the slide, and this is likely allowing the entire apparatus to tilt when the safety drum is rotated. Over time, the spring will become bent behind its connection with the plunger as it's rocked back and forth.
B) The safety plunger is manufactured too short considering this play. The laws of angles and vectors takes effect here... a short pin in a loose tube will have a much greater amount of displacement vs. a longer pin because the longer one will have a greater distance between the contact points within the walls of the tube. At the end of the deflection from the portion extending from the end of the tube, the shorter pin will have a much greater vector angle vs. the longer one. I'm not sure how else to describe this constraint.
I feel like a rude student disagreeing with the wise professor when I disagree with your hypothesis the detents are machined out-of-alignment with the tunnel cut into the slide of the PPK in question. I replaced the firing pin/safety cylinder in the slide sans plunger/spring/extractor piece to reposition it and took the best pictures I could "down the tube". Observe:
A bit closer, best resolution I could muster:
It's difficult to see, but the detents in the cylinder align very well with the spring channel. Mind you, there is still a bit of side-to-side "slop" afforded to the cylinder due to the tolerances of the safety pin through the cylinder, but I believe this is taken up when the safety plunger assembly is replaced in the slide. I believe it's much like reaming a pre-pilot drilled hole; the reamer will follow the hole no matter if the reamer is offset from the hole. In this case, I think the plunger will align the safety cylinder and take up the side-to-side play.
Of course, I may be completely wrong in all regards but don't think so. I hate to be arrogant and disagree with a true master, but my eyes, fingers and tiny mind tell me otherwise.
Perhaps the best bet is to hope for the ability to secure new parts. Should this be the case I will definitely swap them out and try a new approach. And if it works, this thread will turn to an experimental one. My current safety cylinder will be remachined with slightly deeper "fire" detent, my plunger will be remade out of hardened carbon steel to a diameter much more closely matching the tunnel. The "slop" in the tunnel is most definitely a factor in this problem which seems to be creeping up more and more as we discuss it.