It's not the LEDs directly draining the battery. The problem is that the LED don't completely shut off which prevents the BCM from sleeping. It's actually the BCM draining the battery because of the low current flow to the LEDs.
With a load resistor on the backup lights the BCM can properly control the LEDs and turn them off when they need to be off.
I saw a schematic a few years ago - danged if I've been able to find it again - and, IIRC, it showed a hookup wherein the low side of the filament was being switched (to ground, to turn on the light). When the light was off, the negative side of the filament would be at +12V.
Another signal came off the low side of the filament, and went elsewhere, presumably back to the computer (where a high would indicate light off).
With an LED in place of the incandescent filament, the voltage drop across the LED, when off, could be different which might cause the computer to misinterpret the status of the light, which in turn might cause the computer to never go into standby, its lowest current draw level.
If I ever find the drawing again, I'll tag it for later reference. (And study it to better understand why it was connected that way.)
Anyway, the point is, with this extra parasitic load designed into the circuit path, I can sort of see the LEDs causing the computer to stay awake thereby running down the battery.
Doug
.
