Thanks Chris. Yes, waisted spark is what I want to do. But I think that there is another way of tuning the timing: ignition pulse ionisation current sensing. Very easy to do by hand, but not to build yourself into a computer. This is because you've got to decide what part of the curent waveform it is you are after, and artificial intelegence is not yet available to the hobbiest. But if my ignition computer is based on a formula rather than a 3D map of advance vs. RPM vs. load, then it would be easy to tune the timing curve while I'm driving the car. I know, not the best way of doing it, but my BW35 is not either...

So now my question is just what order of equation would be fine for defining the advance curve. Keep in mind this is for normal driving, not performance. Would it be enough to define both centrifugal and vacuum advance as y = mx + c and just add the 2? Isn't that (with max and min limits) what happens in a distributor? That is how it is described in my service manual. A bonus would be if I add an equation for temperature compensation as well. I'm an electronic engineer and do not have the background in engine mechanics, but if someone can point me to a web site with technical information then that will be great as well.

Thanks, and have fun...

PS: Here in New Zealand the cost of labour is too expensive, so I will probably not go for extended dyno testing.