DIYEFI.org Forum

Well, you don't get an IC interrupt for the "missing tooth" do you, so no code runs then, you are just forced to schedule around it. Or do you have a cunning way to fake it somehow? Even if you are faking it, it just boils down to estimation and prediction at the end of the day. No interrupt is no interrupt afterall.

stops... thinks... realises...

You meant on the interrupt BEFORE the missing tooth didn't you. Such that there is twice as much time. Thats a good point for wheels that have missing teeth, but in the case where they don't, its not a lot of use. You could certainly schedule some "once per rev" stuff into that slow in the missing tooth decoder if there is a need for such code to run.

Actually, I was not thinking of how to use it, it was just as a counter example to what you were saying: there is one instance where you have more time if you have a missing tooth wheel. And yes that was for the interrupt before the missing tooth.

As you say you could predict when you have more time to do some computations and do them there. Since this would be in a specific wheel ISR, you should know where you are in the cycle once the engine is running and you have sync and if you don't have sync you have plenty of time because the engine is not running or starting or you lost sync and you need to do some damage control. I'm not sure this is practical or reliable but it might be interesting to keep in mind if CPU cycles become scarce.

Jean

No need for a counter example :-) I was speaking generally, denso has no missing teeth for example... but for a 36-1 it would hardly be worth the effort to consider it really.

Agreed.

The only point where it would be worth messing with is a custom wheel with as many missing teeth as firing locations, but if you're THAT close to the wall you're going to make a custom wheel for each advance curve, you have to rethink what you're doing anyway.