Now that I think about it, I wonder if the governor might be part of the problem. If you were normally flying around burning RC holes in the sky, you tend, I think to always be driving the prop with a positive forward load. If the aerodynamic load decreases, then the motor simply spins faster (assuming the throttle stays constant. So the gear teeth would tend to always be contacting on one surface (until the throttle is lowered, but we are talking about relatively slow variations of the gears slopping back and forth on each other.
With choppers, the blades represent a pretty big flywheel and so I am also thinking the gear load would tend also to be positive.
Now we come to our application, where our flywheel is pretty minimal (especially for all those who like those light props!
) and I know a lot of you have really boosted the governor gain up there. So I bet the motor torque is really jumping back and forth as the ESC tries to keep the rpm constant. Which means, I think, that the gear teeth are constantly rattling back and forth (at high frequency for sure so you probably can't easily hear it). However the brass pinion probably is just being eaten alive by the constant whacks from the steel teeth.
Seems reasonable to me since that gear was eaten up pretty darn fast! I can't imagine a metallic gear wearing away so fast without something going on at high frequency. Even my old Delrin/aluminum setup lasted a fair length of time. I didn't have an issue with the steel pinion /delrin gear afterward.
My solution would perhaps back off a bit on the governor gain (as long as the rpm plots look reasonably flat).