"The hall switch generates a square wave. When the magnet approches it arms, goes high. It triggers when the magnet leaves, goes low. The duration of the square wave is the dwell time.."

agreed, but the magnetic field shape is not a square wave, so the distance from the magnet to the sensor will, at some point, begin to affect the width of the pulse....

"If I understood Terry correctly in one of his other posts, the system only uses dwell timing for starting (up to 800(??)RPM), and then reverts to measuring between hall pulses (much more accurate). "

clever.... this makes my concerns about your dwell-approach much less important...

[:-]