Hmmm. That makes me wonder if the two pins for the sender got reversed. If so, that would explain both the long glow time, and the dead temp gauge (assuming it remains down at 40C all the time).