Originally Posted by bugblatterbeast
I'm guessing the old relay had a reverse connected diode in parallel with the coil. This is fine if the low side driver in the ECU is BJT based (with a properly designed snubber) but Subaru probably used a n-channel fet in the ECU or put a zener clamp on that line to ground. You end up with the parasitic body diode of the fet (and/or the zener clamp) and the clamp diode in the relay forming a pretty low impedance path for negative going transients between V+ and ground.
They probably fixed the issue by tossing some resistance in series with the relay's clamp diode.
I'll have to remember that - I'm building a DIY RPM triggered circuit that drives a relay, and only have it set up with a diode across the coil. I have a 16 volt zener set up to protect the circuit, but should probably add a resistor as well.
For $15 for the updated relay, perhaps I can convince my dealer to just give me one. Who knows, they may just do it...