I was suspecting the LCM all this time to be honest. I thought that the GT1 was giving fits too. See a few posts above - January 29th - the "issues" diagnosed with GT1. The software "knew" something was wrong, but it was pointing in the wrong direction... Maybe the "hiccup" in the ballast created some weird signal enabling the crash recognition program every time the low beam came on. I don't know?