I wish I had remembered this thread existed before embarking on my latest motor project...but now that I've re-found it...
I need to re-read this entire thread again (just finished reading it the first time now), but while I do that, and try to actually absorb some of the information that seems to be bouncing off right now, perhaps some input on my "project":
I've got a huge (20lbs+) brushless powerchair motor that's about the same size stator as something like the X5 series, I think, but it's wound to directly drive a small powerchair wheel (14", I think) at either 4 or 8MPH at 24V (really 28V, two SLA in series).
Built in it has SIN/COS analog sensor output...which is useless (without translation hardware) for ebike controllers (which is what I have avaialble to drive it). I added digital hall sensors at 120 degree physical intervals, gluing them temporarily with just superglue to hand-carved recesses in the material covering the windings between stator teeth.
It's a 45-tooth, 40-magnet, Wye-terminated motor. Presumably it is 15 poles? No idea on number of turns/etc, or inductance (yet, I have a plan to test that later once I find my signal generator).
Presently, the no-load full-throttle current at 40V is about 1.6A, and at 56V is about 1.7A.
There are pics and videos and further data in it's thread over here:viewtopic.php?f=30&t=32838
if needed, or at request I can repost them over here in this thread.
What I would like to find out is what the optimum placement of sensors might be (for timing and high-load situations), if not where I already have them.
At present, there is definite clicking (of changing fields, not mechanical) at each part of a rotation equal to magnet passing stator tooth (or more than one tooth). That might simply be how it works with an ebike controller using trapezoidal commutation, and it might be smoother with a sine controller (but I don't have one). But if it can be made better by simple hall repositioning, I'm up for trying that, if anyone has a fairly certain idea of what might work better. I can't reposition them much, or I'll break the leads off (already did this to one and barely fixed it), and I have no spares except what is in other motors.
What I had planned was to install them on a ring on the outside of the mtor, but there isn't enough flux leakage for them to pick up, so they had to go inside on the stator.
There *is* another magnetic encoder source, which is the ring used for the analog SIN/COS output. I could probably mount them inside there where they can detect the domains on the ring, but I dont' yet know how many there are; if it's less than the number of magnets on the rotor, I'm not sure how things would work with an ebike controller.