Brushless Controller Diagnosis & Repair

[danielrlee]

I thought I'd try my hand at diagnosing and repairing an old 6-FET sensorless controller that sort of crapped out on me a couple of years ago when it started juddering under load. I've tested the FETs as instructed by the Grin controller diagnosis guide and none of them are obviously faulty from basic continuity testing with a multimeter.

I then tried hooking the controller up to a controller tester and I noticed that the phases aren't syncing correctly:

[youtube]-rZ9E7W-qgc[/youtube]

The video was shot without the hall sensor leads connected, but yields exactly the same result when they are, plus a 5v hall feed confirmation.

So, does anyone have any ideas about what could be wrong with this controller?

Thanks

 

[danielrlee]

Bump....

The controller isn't anything special, just an ebay jobbie. In my limited experience, anything not obviously 'dead' can be difficult to troubleshoot. Any experts here suspect whether it's even worth the time or effort?

 

[pff7]

I got the same tester , as of now you got it working better than I did...not so good instructions.

My FETS tested out OK also--tested them with simple meter right on the board.....test red,green,black throttle wires going to the throttle, should get around 5volts, this is where my controller didn't seem to produce voltage(also have a yescom controller but 48v0lt one,) Need to try out my BK oscilloscope with this but haven't used it for a while.

When I tested the throttle with controller on bike though....throttle tested OK....weird but when controller off bike it didn't seem to put out the 4-5volts(between red and black wires)?????

 

[d8veh]

Only a guess, but the controller has a pattern of pulsing to initialise the starting of the motor. It then needs feedback from either the motor halls or back emf to determine the pulse timing, so it's just stuck in start-up mode. If you connected it to an actual motor, it would be different.

 

[pff7]

I connected throttle to controller(not motor) and now read 4.8 or so volts after connecting power to controller, but going to wait for new throttle. Original throttle is suspected to be worn out, been in the rain to much.

dv8 or other---do you know the symptoms of a defective controller as opposed to defective hall sensors?

 

[d8veh]

If your mosfets were completely blown, they wouldn't cycle like that. The timing comes from the software. It can't change to a different pattern from what it's programmed for. Look at all the capacitors on the pcb. Do any have slightly domed heads?

Juddering under load is most likely to be a faulty connection somewhere, like the phase wire bullets. It could also be a dodgy mosfet or a duff capacitor, especially if you overvolted it at any time.

 

[danielrlee]

Only a guess, but the controller has a pattern of pulsing to initialise the starting of the motor. It then needs feedback from either the motor halls or back emf to determine the pulse timing, so it's just stuck in start-up mode. If you connected it to an actual motor, it would be different.

I'm not sure if you've used one of these motor/controller testers before, but every other (working) controller I've tested with it shows a nice snappy phase sequence with perfect synchronicity. I'm fairly sure that the strange pattern displayed isn't to do with sensorless timing feedback.

If your mosfets were completely blown, they wouldn't cycle like that. The timing comes from the software. It can't change to a different pattern from what it's programmed for. Look at all the capacitors on the pcb. Do any have slightly domed heads?

Juddering under load is most likely to be a faulty connection somewhere, like the phase wire bullets. It could also be a dodgy mosfet or a duff capacitor, especially if you overvolted it at any time.

Thanks d8. I haven't had a chance to open the controller up yet (a 10 week old baby soaks up time like a sponge), but you've given me a couple of things to start with there. Is it possible for a FET to be partially damaged where the gate doesn't open fully? I think it's safe to rule out a faulty connection though since the judder occurs just as the motor starts to turn. With an unloaded motor however, it spins even while I give the bike a vigorous shake.

EDIT: And yes, it is a 24v controller which I ran on 12s lipo, so your theory's entirely possible.

 

[danielrlee]

Found a chance to open 'er up and straight away noticed two capacitors that have significantly puffed. One had even cracked open at the bottom and started spewing its contents onto the board. I don't have any suitable replacement caps rated for 63v in my parts bin, so have ordered a few.

 

[d8veh]

Just before I read that, I was going to suggest that you try a higher valur capacitor in the controller or an extra one between the battery wires. Hopefully, new capacitors will sort you out.

 

[danielrlee]

It's been ages since I posted anything on here, but for the first time in months I found some time to do a bit of tinkerin'.

I finally swapped out the puffed caps on the for new ones that I'd ordered ages ago, but the controller showed no change with the weird hall sequence. I'm not going to spend any more time on it since they don't cost much new, but I just wanted to thank you d8veh for your help. To be honest, it's been nice to get busy with a soldering iron again and this was just the thing to start me off.

For any new parents reading this now, or in the future...... life does get easier!