[rmccomiskie]

I've switched my fuel flow input back to the ECU signal. Everything is working just fine except the high GPH when coasting. I repeat, the ECU signal works just fine. I've dialed in the VSS and the uS/gallon. I just don't understand why coasting behaves the way it does.

I've tracked the high GPH reading from instantgph() back to the external interrupt routine processInjClosed(). In that routine, the value returned by elapsedMicroseconds() jumps from a low value to a high value when I'm coasting.

The elapsed time is ~500-700uS at 30mph steady and increases with higher speed and acceleration. Seems normal. When coasting, however, it decreases to about 250-300uS where it suddenly jumps up to more than 11000uS. As the car slows further and the engine rpm decreases, the value starts rising from 11000 to 15-18000uS. It seems to follow engine rpm inversely. The lower the RPM, the higher the elapsed time. It acts like it switches and starts measuring the wrong part of the pulse. The point at which it switches coincides with the ECU signal going flatline.

I'm stumped! When the signal goes flatline, I expected that the external interrupt routines would not fire. However, I checked and found that the routines are firing normally when coasting and the signal is flat. How can that be?

Question: What signal voltage level is needed to trigger the external interrupt pins? The normal ECU signal is negative going. Looking back at my scope images of the flatline, I see that there are some verrry small positive blips. Could it be that they are triggering the external interrupts opposite of what is expected?

Grasping at straws here.