BLDC phase timing

Joe Perez

100 W
Joined
Apr 6, 2011
Messages
187
Location
New York City
I was reading through this thread the other day, in which it is pointed out that a certain RC-based motor with external IR position sensors permits the relationship between the sensors and the motor to be varied, thus permitting on-the-fly mechanical adjustment of phase timing.

I'm relatively new to the world of high-power brushless motors, but something about the aforementioned observation bothers me: Why is this necessary?

No, not "why is it advantageous to adjust phase timing as a function of RPM", but rather, "isn't this already a standard function of the controller?"

Take a look at a simple automotive ECU, such as the Megasquirt. That device is based on an 8 Mhz 6800 (not 68000) derivative CPU, which has got to be one of the slowest microprocessors still in current production. And yet it's perfectly capable of dynamically controlling spark advance on a multi-coil engine to within 1° of accuracy at > 10,000 RPM. The math involved here really is quite trivial, even if you're having to work predicatively (which you'd need to on a BLDC motor, given the requirement for accurately neutral timing at very low speeds. All you need to know is cycle time, and from that you can start a timer when you see "TDC" go by, to wait slightly less than one full cycle before activating on the next go-round.

I ought to just go out to the garage, hang a few scope probes off the bike and answer my own question, but it'll be a week or so before I'm back home and I figured perhaps someone here might be able to offer some insight. Do the BLDC controllers which we typically use in everyday eBike applications incorporate any sort of advance mechanism on phase timing, or do they really just sit there waiting for a sensor transition before turning on each pair of FETs?
 
Back
Top