Hammering "thunk" noise from my 10" hub motor

Joined
Nov 2, 2021
Messages
72
I’ve been looking all over but can’t find another behaving like mine.

When i quickly twist the throttle, the noise happens just as I release the throttle.

When free spinning, no noise. When very slowly release the throttle, no noise.

In the below video, to prevent the scooter from lunging forward, I stepped on the brake. As soon as I release the throttle, you can hear the “thunk” noise.

https://www.youtube.com/watch?v=IoC2iT5M9dA

Any ideas on what’s causing the noise?
 
owhite said:
I've had problems like that from a) controller damage and b) bad phase wire connection.

How did you figure it was controller damage? Any way to test it? or did you just replaced the controller?
 
spinningmagnets said:
I'm probably wrong, but...do you use two stout torque-arms?

Here's my torque arm.

I'm using the plate provided by the QS. The hub bolt sits pretty snug in it. I'm pretty sure there's no play when power is applied.
torque.jpgtorque2.jpg
 
If you take the hardware off and find any markings on it or your frame from movement, you'll know for sure. "Pretty snug" is not enough; it has to be "perfectly snug".

It doesn't take much movement to cause a pretty big noise; you may not even be able to see the amount of movement it takes, while staring at the axle during the noise...but you could also video it close-up with the camera mounted hard to the frame so it can't move relative to the axle, and zoomed in enough to fill the screen with the "scene of the crime" so to speak. ;) It may be visible that way if it's moving. I dont' see this movement in your original video, but I can't see the axle itself behind the frame. I also can't see the other end of the axle on the other side of the bike.


Also note that if the axle key or keyway inside the motor is not a tight fit, the stator itself could be rocking back and forth inside the motor on the axle (not all axles are welded or pressfit; my QS205's are not). You can't see that from the outside; it rquires at least taking the sidecover off (which means you can't spin the motor with the controller, though you can manually rock the whole wheel back and forth with the axle mounted in the dropouts as it is now, just minus the sidecover). I'd check the dropout-axle mounting before doing that, though.

This kind of play is what broke both axle ends off on the first QS205 I got from anohter ES member, while he was out for a ride....
 
All that said: The noise sounds electromechanical, not purely mechanical.

It sounds like the controller is holding a current on the stator, which is applying a force on the coils against the magnets, then when throttle is released it releases that current.

If it's a tiny drive current, then the motor should be either harder to spin or easier to spin when present, than when it is not, so if you turn the bike upside down so the wheel is in the air, you can manually check the spin-resistance of wheel with throttle completely released vs just slightly turned but not enough to spin the wheel. If it's this, then adjusting your throttle vs drive settings in the controller (if it's possible) to ramp down to zero before throttle off should fix this.

If it's an FOC controller, it may be setup (or designed) to keep a waveform going thru the stator all the time for position identification purposes, and it might shut that off when there is no throttle signal, or upon engagement of a separate throttle-switch-off signal.

Does your throttle have a switch in it (like Domino's do), that is wired to something on the controller? If so, how exactly is it wired? And what is the input on the controller setup to do?

If it's the FOC position-sensing waveform/current, then it shouldn't affect how easy or hard it is to manually spin the wheel. But it may be possible to adjust out the thunk, depending on what settings are available in the controller software, so it gradually releases this signal instead of suddenly.

A torque-control throttle (that changes phase current rather than voltage (speed)) could cause such a thunk if the controller settings are such that the lowest throttle input still results in some current to the motor, or the throttle itself still outputs some voltage at it's lowest setting just before off, but the controller's off-threshold is below that.



It is also possible it is a phase/hall combination problem: Some motor/controllers I've used have made noises vaguely similar to this (the noise changes depending on the motor design, the frame it's bolted to, etc) when the combination is incorrect, but appears to be working (timing is advanced or retarded). Shifting the phase wires all over by one usually fixes this (whcih direction requires testing). If the controller no-load full-speed current is low (a couple of amps or so, depending on motor/controller size), it's probably correct combo. If it's high, it's probably wrong.
 
amberwolf said:
If you take the hardware off and find any markings on it or your frame from movement, you'll know for sure. "Pretty snug" is not enough; it has to be "perfectly snug".

It doesn't take much movement to cause a pretty big noise; you may not even be able to see the amount of movement it takes, while staring at the axle during the noise...but you could also video it close-up with the camera mounted hard to the frame so it can't move relative to the axle, and zoomed in enough to fill the screen with the "scene of the crime" so to speak. ;) It may be visible that way if it's moving. I dont' see this movement in your original video, but I can't see the axle itself behind the frame. I also can't see the other end of the axle on the other side of the bike.


Also note that if the axle key or keyway inside the motor is not a tight fit, the stator itself could be rocking back and forth inside the motor on the axle (not all axles are welded or pressfit; my QS205's are not). You can't see that from the outside; it rquires at least taking the sidecover off (which means you can't spin the motor with the controller, though you can manually rock the whole wheel back and forth with the axle mounted in the dropouts as it is now, just minus the sidecover). I'd check the dropout-axle mounting before doing that, though.

This kind of play is what broke both axle ends off on the first QS205 I got from anohter ES member, while he was out for a ride....

Thanks for taking the time to help with this. I'm pretty sure it's not loose in the torque arm. I've taken the hub off the bike and clamped it down on the vice grip (the bolt). I would hold on to the tire while slightly twisting the throttle. The noise is definitely coming from inside the hub.

I opened the cover a few days ago to inspect the mechanical part of the hub. Didn't take the whole thing apart, just the cover. Didn't see anything loose. I might have to open it up again to check the stator. I'm not sure about the keyway inside the motor but i will be looking for that too.

Do you have any videos or pics of the keyway inside the motor? Want to make sure I know what to look for.

Here's what it looks like when i took the cover off.

https://www.youtube.com/watch?v=rGEJeMqlbTY
 
amberwolf said:
All that said: The noise sounds electromechanical, not purely mechanical.

It sounds like the controller is holding a current on the stator, which is applying a force on the coils against the magnets, then when throttle is released it releases that current.

If it's a tiny drive current, then the motor should be either harder to spin or easier to spin when present, than when it is not, so if you turn the bike upside down so the wheel is in the air, you can manually check the spin-resistance of wheel with throttle completely released vs just slightly turned but not enough to spin the wheel. If it's this, then adjusting your throttle vs drive settings in the controller (if it's possible) to ramp down to zero before throttle off should fix this.

If it's an FOC controller, it may be setup (or designed) to keep a waveform going thru the stator all the time for position identification purposes, and it might shut that off when there is no throttle signal, or upon engagement of a separate throttle-switch-off signal.

Does your throttle have a switch in it (like Domino's do), that is wired to something on the controller? If so, how exactly is it wired? And what is the input on the controller setup to do?

If it's the FOC position-sensing waveform/current, then it shouldn't affect how easy or hard it is to manually spin the wheel. But it may be possible to adjust out the thunk, depending on what settings are available in the controller software, so it gradually releases this signal instead of suddenly.

A torque-control throttle (that changes phase current rather than voltage (speed)) could cause such a thunk if the controller settings are such that the lowest throttle input still results in some current to the motor, or the throttle itself still outputs some voltage at it's lowest setting just before off, but the controller's off-threshold is below that.



It is also possible it is a phase/hall combination problem: Some motor/controllers I've used have made noises vaguely similar to this (the noise changes depending on the motor design, the frame it's bolted to, etc) when the combination is incorrect, but appears to be working (timing is advanced or retarded). Shifting the phase wires all over by one usually fixes this (whcih direction requires testing). If the controller no-load full-speed current is low (a couple of amps or so, depending on motor/controller size), it's probably correct combo. If it's high, it's probably wrong.

"Electromechanical" What a cool word!!!
I did changed some throttle settings, will look into it again. There's no switch in my throttle.
This is the throttle i'm using.
Screen Shot 2022-05-03 at 12.13.35 AM.png

I will look into changing some settings before opening the hub again.

Thanks!
 
amberwolf said:
If it's a tiny drive current, then the motor should be either harder to spin or easier to spin when present, than when it is not, so if you turn the bike upside down so the wheel is in the air, you can manually check the spin-resistance of wheel with throttle completely released vs just slightly turned but not enough to spin the wheel. If it's this, then adjusting your throttle vs drive settings in the controller (if it's possible) to ramp down to zero before throttle off should fix this.

I was able to slowly and slightly turn the throttle enough to not spin the wheel but i can hear something spinning very slowly inside the hub.

What is the throttle and drive settings do you recommend trying?

I have the KLS7145N kelly controller.
 
kingkongshrimp said:
I was able to slowly and slightly turn the throttle enough to not spin the wheel but i can hear something spinning very slowly inside the hub.
Unless yours is different than it looks from the video, it's not a geared hubmotor, so there can't be anything actually spinning inside the motor.

You might be hearing the controller rotating the magnetic field around the phases trying to start it moving; if the coils are not mounted in a way that keeps them completely still (like potting), then they may vibrate from the changes in the field as it pushes on the magnets.


If you're hearing an actual object rotating inside the hub, there could be something (screw, etc) floating around loose in there (well, actually stuck to the magnets) and being dragged around by the field (not very likely).
 
kingkongshrimp said:
Do you have any videos or pics of the keyway inside the motor? Want to make sure I know what to look for.
I'll get some pics of my QS205 stators and axles; one still has it's broken axle in it, the other has no axle in it, and two spare axles/hardware with it.
 
kingkongshrimp said:
What is the throttle and drive settings do you recommend trying?

I have the KLS7145N kelly controller.
I don't know enough about the Kelly settings to give useful recommendations.


All I can remember about their software when I tried to use it to verify setup in a controller I have here is that the way they do their settings makes no sense; the numbers and ranges appear to be randomly chosen by lottery or something, as they don't match numbers and ranges anyone would actually ever use for the things they are there for. :roll: They even use wierd sets of numbers to toggle things on and off instead of just *toggling things on and off with a button*, or using a zero and a one, etc. :?

FWIW, there don't seem to be any controller manufacturers capable of making simple-to-use setup software that just works. Grin's Phaserunner/etc might be an exception to that; it's the only one I haven't seen threads discussing program/setup/etc problems with of this type. I don't know what is so hard about starting with the end-user in mind and designing the software around helping them use it to do the job they need to do, instead of whatever most programmers' development method is. :/
 
Back
Top