Brushless Hub Motor Jerking/Seizing, but only if throttle is engaged right as the wheel stops spinning?

[dev_null]

A bit of context, I have a Caroma D3 750W that I've been upgrading. Currently, I'm running the following:

Controller:
JN Electric Bike Controller 24/36/48V Brushless Motor Controller 30A Sine Wave Controller With Three Mode For E-bike Scooter 750W 1000W
Electric LCD Display, S866 Cycle Computer 24V 36V 48V Electric Scooter Control Panel with Plug for Electric LCD Display Display

Throttle:
Mokuo 22mm/0.87" Electric Bike Handlebar with 3-Speed Control Funcation,E-bike Throttle Grip Made of Rubber and Metal Material Fit for Universal Electric Bicycle Scooter,Titanium Color

Second Battery:
Airuxuan 48V Ebike Battery 10Ah Electric Bike Battery with Waterproof Case,54.6V 2A Charger, XT60 Connector and BMS for 100-800W Electric Bicycles Motor
Connected with stock battery via:
Electric Bike Dual Battery Discharge Converter 30A 24-72VDC Dual Battery Pack Switch Balancer Double Battery Parallel Module with XT60 Plug for 500w Motor

It's mostly running fine, except for one weird quirk that I only just barely figured out how to reproduce.

Twisting the throttle from a complete stop of about 3 seconds or more is fine. Throttle while in motion seems fine. However, if I twist the throttle while the wheel is about to come to a stop from free-spinning, the motor seizes up and either jitters and then engages normally, or just keeps jittering rapidly until I release the throttle. It's more common if I twist the throttle slowly, and happens nearly every time if I turn it just enough to engage the motor (right at the cutoff point). While it's doing this, the phase wires get warm. The seizing at some point while riding was strong enough to twist the axle through the parallel flat sides/drop-outs, spooling up the motor wires in the process (some of the wires became exposed when this happened, so I stripped the sleeve all the way off, taped off the exposed areas with electrical tape, then wrapped the whole bundle back up using this tape. It did not appear to me that any significant damage was done to the conductors themselves, but this event was caused by this issue so it couldn't be the culprit).

I should mention the above testing was done with the bike suspended and the wheel free-spinning. When actually riding, this would happen pretty often when accelerating from a short stop or low speed (not just as I'm about to come to a stop). This may indicate the issue could be related to the resistance the motor is encountering.

I've pretty much ruled out issues with the wires and throttle (unless there's some weird situation like wire colors not matching to the same color); as I said, everything works perfectly except in this very specific scenario, which I can reproduce almost every time. I've gone through adjusting every setting on the LCD Display, I even attempted to open up the hub motor but it took enough effort that I didn't want to risk damaging anything and quickly bolted it back up. From the research I've already done, I believe it could be related to issues with the magnets, or the hall sensor failing at determining the proper voltage at that specific rotation speed.

Regardless, I'm at a loss and I'm hoping someone here has a solution, or can point me in the right direction. Maybe I'm dumb and doing something I shouldn't be doing or using incompatible components, I am not experienced in this stuff and I'm kinda just figuring it out as I go.

 

[amberwolf]

If the controller is sensorless or sensorless capable, and the halls are not working (or it isn't reading them even though they're wired up; I had one like that), you could have this sort of issue.

But it would probably also have issues starting from a complete stop if the rotor were in anything other than the perfect position, too. (the same is true of almost anything else that could interfere with hall operation--it would almost certainly affect any startup or operation at or below the speed at which the problem does happen).

It's more likely to be a bug in the controller firmware that doesnt' account for some specific condition that occurs at the problem time.

 

[amberwolf]

The seizing at some point while riding was strong enough to twist the axle through the parallel flat sides/drop-outs

This being able to happen means you need a pair of good, perfectly fitting (not well-fitting, but perfectly fitting) and properly-mounted torque arms, or clamping / pinching dropouts. If you poke around with those terms here on ES you'll find lots and lots of posts and pictures of good ones (and bad ones).

 

[dev_null]

This being able to happen means you need a pair of good, perfectly fitting (not well-fitting, but perfectly fitting) and properly-mounted torque arms, or clamping / pinching dropouts. If you poke around with those terms here on ES you'll find lots and lots of posts and pictures of good ones (and bad ones).

Whether or not they were good drop-outs before, they're definitely not good anymore. That's a problem for another day though.

If the controller is sensorless or sensorless capable, and the halls are not working (or it isn't reading them even though they're wired up; I had one like that), you could have this sort of issue.

But it would probably also have issues starting from a complete stop if the rotor were in anything other than the perfect position, too. (the same is true of almost anything else that could interfere with hall operation--it would almost certainly affect any startup or operation at or below the speed at which the problem does happen).

It's more likely to be a bug in the controller firmware that doesnt' account for some specific condition that occurs at the problem time.

It could very well be a controller issue, that's basically the only thing I'm aware of that I haven't ruled out. It's Amazon so I could just order a replacement and see if that fixes the issue, but I'm really not looking forward to resplicing all the wires again.

 

[dev_null]

Quick update, I was playing around with the motor/hall wires today and I seemingly fixed the issue (although at a cost).

If I disconnect only the hall blue wire, and then connect the learning wires, the motor no longer seizes and stutters. However, it now seems to accelerate slower, and is making a clicking sound which turns into a buzz/whirr at higher RPMs. Obviously neither of these changes are preferable.

I'm working with the seller of the controller as well; it seems they wanted to try to troubleshoot the issue before sending out a new controller. I also installed torque arms which seem pretty solid, thank you for that suggestion!

 

[amberwolf]

Disconnecting a hall means it almost certainly now runs sensorless, which can mean that any of the hall sensors is not working, or that their signals aren't what the motor expects, wrong phase / hall combination, etc.

But I'd expect that kind of thing to cause the same behavior whenever the halls aren't working, so it is probably one of the few(?) controllers I've seen that attempts to derive a third hall signal from the two still working, and that i just doesn't do this well for your motor/system. If htis is the case, then the blue hall is one of the two working ones, since disconnecting it makes it not try to read them at all anymore and just fallback to sensorless.

 

[dev_null]

Interesting. I did notice that it doesn't run at all until I connect the learning wires; my understanding is that "teaches" the controller to use the current configuration? I might try doing the same thing with the other wires just to see what happens. Either way, this basically confirms the controller is the issue, right?

 

[amberwolf]

If you disconnect a hall and it runs "better" then it means there is a "problem" with the hall signals.

That can be a motor fault with bad halls, wiring fault anywhere between motor and controller, or an internal controller fault with connections or wiring or electronics.

 

[dev_null]

The seller is offering a 40% refund on the controller as compensation for repairs, which I accepted (comes out to $15, not a big deal). I think the next logical step is to crack open the motor and take a look at the hall sensors. Looking at videos and pictures it doesn't seem too difficult but I don't want to break anything. Any advice on what to look for, what to avoid doing, etc.? The last time I tried, I managed to get the cover off of one side, but the other side wouldn't come all the way off. It's basically this style of motor.

 

[amberwolf]

Before opening it, you should test the sensors to find out if they are producing the right signals. There's a number of posts around here about testing hall sensors, and replacing them if necesssary, as well as what typical part numbers are used.

If the controller was the problem then there's no reason to open the motor for this problem.

If the motor is the problem there's no reason to replace the controller for this problem.

 

[dev_null]

Gotcha, thanks for the advice! I'll look around and try some tests, and report back if I make any breakthroughs in case someone with similar problems ends up here.

 

[TommyCat]

I'll look around and try some tests

Testing BLDC motor's Phase Wiring - Hall Sensors and Wiring. - Electricbike.com Ebike Forum


Regards,
T.C.

 

[dev_null]

Update: I replaced all 3 hall sensors and that seems to have fixed the issue! Thanks for all the help!