Today I decided to try fitting a couple of "spare" controllers, neither of which I've actually used yet. Turns out both have some issues, though I expected that with the second one.
No problems each time I went back to the repaired 12FET generic "ecrazyman", so the motor, connectors, and wires are all ok.
The first is a Methods' 100V 100A 18FET "infineon", one of the (I think) XC846-chip types, from before they were programmable, IIRC. (it has the header but software available to us doesn't talk to this one, AFAIK. Haven't actually tried yet). It has a switch on it for low or high voltage use (it is where I borrowed the idea for the one on my Lyen 6FET that is on DGA).
I'd repaired this quite some time ago, to use for the 2010 Undead Race, but I had problems after playing around with it on the Fusin and DayGlo Avenger, trying to find the Fusin's phase/hall color combo for it. I had blown another FET or more (cant' remember now), and then fixed it in time for the UR. Didn't end up using it as instead I think I used the 12FET generic one I'd replaced the melted shunt and blown caps on from Ianmcnally, but I had worked out the hall/phase combo for the 9C for it...which did not work when I hooked it up; all I got was grindiness, even off-ground.
So, I played with variations and tried the simpler method of finding the right combo, but that method didn't work. I drew up a full 36-spot grid for all possible combos,
and went thru them, and found several rough-running forwards and reverses, and a couple of almost-normal ones for each direction. But even those were MUCH noisier than normal, even once at speed. Off-ground they would at least spin, and drew only about 1.6A once at speed, even at full speed (~40MPH off-ground). But it wouldn't even start the wheel moving when on-ground even without me on the bike.
So something is still wrong with the controller. Dunno what yet; will have to start scoping it out to see what the waveforms look like, and see if they tell me which FETs are still bad or whatever.
Also, I found some oddities:
--Regen braking does not work in forward motion. I shorted BK to GND with a jumper, and verified it's connection with a meter when it didn't appear to work. Brake signal input does stop the controller, but it does not actively brake the wheel (whcih does work on the 12FET EC). There is no throttle regen, either.
--Regen DOES work in reverse. I added a switch to let me enable reverse (so I can back the bike up, finally!) from X3 to GND (note there is no DX3, just X3). During testing of braking, it actively braked the wheel when X3 was grounded.
--neither X1 nor X2 do anything noticeable in off-ground testing. Most likely the controller is not programmed for anything to be different in throttle control with them.
I did go ahead and put the 3-speed switch on the bike...on the little nub of tube I have in front of the steering pivot. Right now it is wired with X3 reverse as top position, and middle as "default" speed, and bottom as X2. The actual switch is the one that came with the first Fusin kit from Dogman; I broke the throttle assembly off of it a long time ago, but the switch is still good.
If I can't use it with this controller, I can hook it up to the 12FET EC or something instead.
Unfortunately unless I can enable forward regen, I will not be able to use this controller on this bike, because I MUST have regen braking to help me slow down--the front rim brakes are completley insufficient to do it (and even with regen it is still insufficient sometimes--it takes far too long to stop).
But at least I have learned by accident that X3 is the same as DX3, and should enable reverse on the 12FET EC, too (it also doesn't have a DX3 pad).
Next I tried a Crystalyte "48V" unit 12FET or 15FET (is as big as an 18, but PCB doesn't run full length)
from Karma, which already had a known issue that it required starting the wheel moving forward before it would begin running. I have not finished troubleshooting, but so far only tested one combo--the very first combo I tried worked audibly better than any of the combos for the 18FET, but not as well as the 12FET EC.
On-ground testing started to move but ground to a stop instantly. This might be a false positive or it might be that I will need to take it out on the road to test it, throttling up and pedalling hard at the same time. If I have to pedal hard with no assist to startup, I can't use this controller on this bike--my knees won't take that for long.
Regen either does not work or is not enabled on this controller, or perhaps it does not have that function avaialble. Haven't explored it yet. Brakes just turn off the motor power.
The end cap has two switches on it, neither of which were wired to anything (there are nubs cut off on one switch, and various nubs on pads inside, probably from a CA-DP connector that isn't on there anymore). The second switch with no nubs might have once been soldered to the red wire that is spliced together in a loop now, which appears to be this controller's equivalent of the "ignition" wire.
I'll play with this more later; I had to stop to feed the dogs and myself.
I did find this LED spotlamp that Texaspyro sent to be very useful while doing the wiring combos. I just put it into a clip-on desk lamp, and clipped that to the bike frame to point it at the controller area: