A letter to Todd (microcontroller developer) outlining how the Honda engine position sensor problem should be solved:
Todd,
I think the Honda ECU cares about the order of the cam signals but not where the sequence starts. So long as we send A, AB, B we can start anywhere in the sequence.
My guess is that the Honda ECU works a lot like the VW ECU. Once it figures out which cylinder is at Top Dead Center (TDC) on the compression stroke using the cam position sensors then it knows through the crank position sensor what happens and when until the engine shuts down. When the engine starts back up it goes through the learning process again--which cylinder is at TDC on compression (comes through cam sensor), set ECU reference point, use crank position sensor to track cylinder position from there on.
The part that seems to really matter is the crank position sensor information. It might not be that hard to convert VW to Honda information. The VW crank has a 54-2-2-2 architecture. As that file I sent you explains, this means the sensor wheel has 54 teeth with 6 teeth missing split 2 teeth at a time at 120 degree intervals. The Honda has a 12 tooth crank position sensor or one-fifth the number of teeth.
So, we could count 5 VW crank position sensor teeth then produce a Honda crank position tooth. Then we could produce Honda cam position signals every 8, 16, and 24 Honda crank position teeth--cam A at 8 teeth, cam A/B at 16 teeth, cam b at 24 teeth. All we're really left with is a counting problem at this point. There's no need to worry about frequency as counting the VW crank position sensor teeth will take care of that by default.
Best wishes,
Jake
Open source code is forthcoming for the Honda crank and cam position sensors.
5/1/09
After contacting Moates they directed us to a sensor signal simulator they developed for Honda ECU testing. As it turned out, Moates' device was doing the same thing we are trying to do with respect to crank/cam signals. Fortunately Moates also had a forum link for their device: http://forum.pgmfi.org/viewtopic.php?f=2&t=15570&st=0&sk=t&sd=a&start=30. This allowed us to learn how to send our signals to the Honda ECU in such a way that they would be recognized. So, after a bunch of head scratching, some toying with microcontroller code, and test and evaluation on the bench we have a solution that should work.
Pictures of the translator as well as reports on test results to follow shortly.
