[pc1010]

I am far from being knowledgeable regarding modern engines, but I can't see that half a degree of timing is going to cause the idle to fluctuate!

I don't think that is the point the OP is making. The point is that to stabilise idle you may need to adjust the idle by less than 1 degree in order to get the revs to drop or increase by a much smaller amount than can be achieved using 1 degree of timing. I think others have mentioned that in theory, if you constantly move the idle timing up and down by 1 degree then in effect you get, maybe, half a degree on average. That may still not be enough resolution to control it.

I have experimented with ignition based idle control, using a simple rpm versus timing delta table, with 1 degree of resolution. It made no difference to the stability of the idle, so something more complex is obviously needed for my engine at least.

Yes you have my point, in my car stepper has low resolution and it's hard to maintain desired effects using only air control. The thing you said about moving idle up and down by 1 degree looks like oversampling and theoretically it could work because result of the integral would be a fraction of single degree but I don't know it would be possible to do on Speeduino

[PSIG]

I have experimented with ignition based idle control, using a simple rpm versus timing delta table, with 1 degree of resolution. It made no difference to the stability of the idle, so something more complex is obviously needed for my engine at least.

I think the control algorithms are the question, as if 1° resolution were too coarse, then the result would have been "too good". OEM systems I've seen use many degrees of correction at high rates on small rpm deviations. Each control method would have its own traits of course.

David

[lwoodward96]

+1 to David on this for OEM level ignition based idle stabilisation (which is now entirely interlinked with the torque reserve functions)

I'm also very interested in ignition based idle stabilisation, with +-8 degrees at idle on my Triumph V8 plenty to cause a +-200 rpm change.

So I would have thought that more than 1 degree resolution might be required (welcome to be shown otherwise though!)