Help! Motor fault - diagnosis and repair?

[Slightlypedantic]

A while ago I purchased a Yose Power kit with what appears (from the markings on it) to be a rebranded Aikema 95RX 250W rear hub motor. The problem was that the motor would surge once it had reached the speed limit for each power setting, with the rpm rising and falling rhythmically. This was most noticeable with the bike in the work-stand (no load), as well as on the road. Not unrideable, but a bit annoying.

Yose Power promptly replaced the motor, which cured the fault. I can't fault their customer service!

So I now have a spare motor that I'd like to fit to my mtb. I'm wondering if someone with bit of common-sense and a small soldering iron (me) could fix it. But it needs someone with a lot more knowledge than me to make a stab at what the fault might be. Hall/speed sensor(s) perhaps? Is there anyone out there that could point me in the right direction?

Thanks!

 

[saneagle]

There's nothing in the motor that can cause that fault other than a damaged cable. I bet if you put it back on your bike, it works perfectly.

 

[Slightlypedantic]

There's nothing in the motor that can cause that fault other than a damaged cable. I bet if you put it back on your bike, it works perfectly.

Thanks saneagle. That would be the best possible outcome.

Is there any other way I can confirm that the motor is sound and/or identify and fix a motor cable fault? I have a reasonable digital multimeter.

NB Unfortunately I have very limited working space (too many bikes) so it would be tedious although perfectly do-able to change the motor core back again. If I use the spare motor on another bike I would first have to build it into a rim, so I'd like to make sure beforehand if possible.

 

[saneagle]

You need a bit of motor cable with the female pins in the connector to connect to the motor and flying leads on the other end.

Test for a short to the case:
With your meter set to continuity, put one probe on the motor case, then probe each wire to make sure there is no connection.

Motor phase wires:
Take each pair at a time (3 tests).Set your meter to Volts A/C
, connect your probes to any pair, then spin the wheel backwards to see the generation. Each pair should generate the same voltage at the same speed.

Hall sensors:
You need a 5v source. Use three AA batteries or any 5v charger you can find. Connect the plus to the red wire and the minus to the black. Set your meter to 20v D/C scale, put the black probe on the black wire and measure the voltage between each of the thin yellow, green and blue wires while you turn the wheel slowly backwards.Tou should see the 5v switching on and off several times with each rotation.

If it passes those three tests, there is nothing wrong with it unless there is an intermittent connection, but you should still see that in the tests.

 

[Slightlypedantic]

You need a bit of motor cable with the female pins in the connector to connect to the motor and flying leads on the other end.

Test for a short to the case:
With your meter set to continuity, put one probe on the motor case, then probe each wire to make sure there is no connection.

Motor phase wires:
Take each pair at a time (3 tests).Set your meter to Volts A/C
, connect your probes to any pair, then spin the wheel backwards to see the generation. Each pair should generate the same voltage at the same speed.

Hall sensors:
You need a 5v source. Use three AA batteries or any 5v charger you can find. Connect the plus to the red wire and the minus to the black. Set your meter to 20v D/C scale, put the black probe on the black wire and measure the voltage between each of the thin yellow, green and blue wires while you turn the wheel slowly backwards.Tou should see the 5v switching on and off several times with each rotation.

If it passes those three tests, there is nothing wrong with it unless there is an intermittent connection, but you should still see that in the tests.

Thanks again. I like the sound of that although it'll be a while before I get to it, so watch this space.

 

[Slightlypedantic]

You need a bit of motor cable with the female pins in the connector to connect to the motor and flying leads on the other end.

Test for a short to the case:
With your meter set to continuity, put one probe on the motor case, then probe each wire to make sure there is no connection.

Motor phase wires:
Take each pair at a time (3 tests).Set your meter to Volts A/C
, connect your probes to any pair, then spin the wheel backwards to see the generation. Each pair should generate the same voltage at the same speed.

Hall sensors:
You need a 5v source. Use three AA batteries or any 5v charger you can find. Connect the plus to the red wire and the minus to the black. Set your meter to 20v D/C scale, put the black probe on the black wire and measure the voltage between each of the thin yellow, green and blue wires while you turn the wheel slowly backwards.Tou should see the 5v switching on and off several times with each rotation.

If it passes those three tests, there is nothing wrong with it unless there is an intermittent connection, but you should still see that in the tests.

Hi Saneagle,
Apologies for the delay and thanks for taking an interest. I have now attempted the tests you recommended, repeated to ensure consistency.

Short to case
There was no continuity between any of the nine motor lead terminals and either the motor case or the axle.

Motor phase wires
Back emf generation - increasing with speed, consistent between all three phases.

Hall sensors
The motor is not built into a wheel and so the axle was held in a vice. Due to the internal gearing it was very difficult to spin the motor slowly and precisely. It wanted to sit in specific "detent" positions, presumably magnetic and representing the individual motor poles. The individual "detent" positions are close together, so no scope for precision with the rotor position.

As best I could determine:

White: Off (0V) for a small part of the hub revolution (2 poles) and 5V for the rest of the hub revolution. Presumably this is the wheel speed sensor. This sensor behaves differently to the phase sensors.

Yellow, green, blue: The phase sensors appeared to switch frequently as the hub rotated, however voltage readings could only be obtained at the specific "detent" (magnetic?) positions. I removed the motor from the hub casing so I could rotate the rotor more precisely without the gearing interfering. I detected off voltages of 0V on all phases. On voltages were about 0.7 - 0.8V on one phase, 0.5 - 0.6V on another and around 0.3 - 0.4V on the third. These were the same as measured before removing the hub casing.

Surely all three (four including the white wheel speed one) sensors should behave the same way?

I tried wiggling the motor lead where it enters the axle, and also the 9pin connector. No difference.


Just to recap: In the OP I mentioned that this new motor had been surging and after sending a video Yose Power supplied a replacement, which cured the problem. Sorry, Pedelecs won't let me attach .mov or .mp4 files, not even 8MB. I now want to recover the "faulty" motor, if reasonably possible.

In the bike stand the motor runs smoothly up to maximum speed in each assist level but this maximum speed then keeps varying rhythmically, frequency once or twice a second, as though it is hunting for the right speed. It's fairly noisy when doing this but quiet otherwise. It sounds as though the power is being varied significantly. This surging happens in all assist levels, more pronounced and noisier in higher ones.

Btw, the controller appears to be JYT brand according to Yose Power web site but is unbranded inside and out. It's marked model YCSH-D. The web site claims it is a sine wave controller.


My guess, FWIW but probably a LoB!:
Perhaps the Halls are sending duff/inconsistent/badly timed messages to the controller, which then cannot apply the correct amount of current to each phase at the right timing. So, when running at full speed in the bike stand, perhaps the motor ends up going from near full power to nothing according to the speed sensor, over and over again? I.e. the ramping off of current as the speed limit approaches has become pretty much on/off, on/off... with no fine control of current in between, and not balanced between phases?

If the controller is sending different and erratic amounts of current to each phase, might this account for the noise?

Any thoughts or ideas?

Thanks again Saneagle, your help is invaluable and really appreciated.

 

[saneagle]

The surging you describe sounds a bit like the normal speed cut-off. In each level set on the LCD is a different speed limit. When the motor reaches that speed, it cuts off, then when it slows down a bit, the power comes back in, feeling like a surge; however, if it didn't do that with the replacement motor, it can't be that.

The hall sensor readings look a bit low, but there might be a reason for it. They often need a pull-up or pull-down resistor to fully switch. That resistor is in the controller, so to test them you need to add your own. If you have any resistors, you need something in the range 5k to 10k or thereabouts. It needs to be connected between the 5v wire and the signal wire to pull up or the signal wire and ground to pull down. One of those two ways should get the hall sensor switching all the way to 5v or whatever your supply voltage is. You only need to test one. If you can get it fully switching, the others will be OK too.

Another cause of surging is when you don't push the motor connector in ALL THE WAY to the MARKED LINE. If I had to bet, I'd say that you simply didn't push the motor connector in far enough and you owe Yose the money for the motor.

At this stage, from the tests you've done, I'm going to say that there's probably nothing wrong with your motor. When the controller loses a hall signal, the motor makes an immediate rasping sound and goes very rough, or it completely cuts out. It's a very simple device, intermittent faults are pretty unusual. I've seen/heard of the odd one where a cable wasn't tied down properly so the insulation rubbed away allowing intermittent contact with the case. Main cable damage where it comes out of the axle is very common, and rarely a bad solder joint on the hall PCB. Any other fault would be in the controller and othe parts, but they all worked when you fitted the replacement motor, so the fault isn't likely to be there .

Here's a replacement AKM hall sensor board. You can see the resistor (R1) for the speed sensor, which is why that one switches OK, but there isn't one for the three motor ones. It's pulling up, so try that first with the other sensors.

Page Not Found - Aliexpress.com

www.aliexpress.com

 

[Slightlypedantic]

Thanks, some interesting comments.

Speed cut-off - yes it does feel like that, but very much more pronounced and noisy than the replacement motor, which is only just perceptible and remains smooth and quiet.

I'm fairly confident that the 9 pin plug was fully engaged as I was aware of the importance of that before fitting the kit. Although it's a while ago now and my memory has faded, I'm also fairly confident that I would have checked motor and display connections before considering whether the fault lay elsewhere. You are of course quite right to flag it up as a possibility.

I just tried jumping one of the Hall signal wires to the +5V red with a 15k resistor: result - clean switching, either 5V or 0V, just as you said. Presumably this would give the controller a much better chance of registering the Hall switching reliably rather than perhaps detecting some switching events and not others.

Since the motor itself checks out OK and seemed fine when riding apart from the surging, and a replacement motor cured the fault, to me it seems likely that weak Hall signals are the culprit.

So, one option would be to remove the board and check for poor solder joints that are in common to all three Halls but not the speed sensor. I'd touch them with a 15W iron until the solder melts, then reassemble and test - a bit tedious, and may or may not cure the fault. Another option would be to solder three 15k resistors into the controller, where there is likely to be sufficient space and less dismantling required - this seems likely to work. A third option would be to replace the board with the one you identified on Ali Express - that would be the ideal fix.

Perhaps I'll strip the motor a bit further and see what's involved. Doesn't look difficult.

As I am thinking of changing to a KT controller and display, a different approach might be to get one that can run sensor-less and ignore the Hall sensors? I suppose it might not be as simple as that...

Does that all sound good? Which option would you choose, and why?

Hope I'm not barking up the wrong tree somewhere lol.

Thanks again for your help. It's great that you, and others, with knowledge and experience are so kind and generous with advice on this forum.

 

[saneagle]

The pull-up resistor doesn't have to be in the motor. It's only necessary for testing. Your controller has the resistors in it between the 5v and the hall input wires, so your halls will switch properly when connected to it.

Try putting the motor back on your bike to see if it's OK now. If it still does the surging, then investigate further, but don't waste time and resources otherwise. maybe there was just a bit of dirt in the connector that's shifted now.

 

[Slightlypedantic]

Better and better - no work to do - I like it! So perhaps there was a poor connection after all.

If I connect my spare controller and switch the battery on (no display, throttle or PAS sensor connected), I could rotate the rotor by hand as before and should see the full switching voltage? Or would that be dangerous, i.e. could the motor start to run by itself?

 

[saneagle]

Better and better - no work to do - I like it! So perhaps there was a poor connection after all.

If I connect my spare controller and switch the battery on (no display, throttle or PAS sensor connected), I could rotate the rotor by hand as before and should see the full switching voltage? Or would that be dangerous, i.e. could the motor start to run by itself?

That's a good way to test and is how I always do it. It only works if you can access the hall wires or their connections, which is why I always choose controllers with block connectors, not the waterproof type. With some controllers you can test where the hall wires are soldered, but that means opening the controller, and it won't work if the pcb is potted.

 

[Slightlypedantic]

UPDATE

Hi Saneagle,

Unfortunately my connectors are indeed the waterproof type and the controller is potted. Worse, the controller is integrated with the battery base and the battery socket is on the pcb so, whilst I can remove the controller from the base and see the pcb, it's all covered by the battery once it is plugged in. I could cobble up a temporary cable battery connection but that invites the risk of sparks and damage. So it comes down to reinstalling the motor and seeing what happens. However...

I remembered recently that, when I switched the defective motor internals for the replacement supplied by Yose, I was slightly surprised when the reduction gear assembly slipped freely off the new motor and stayed in the casing, just by gravity, whereas the gear assembly on the defective motor stayed put on the axle when I took it out, even when inverted.

I thought nothing more of this at the time but since there doesn't appear to be any electrical fault I am now wondering whether the surging and rhythmic noise might actually be a mechanical fault due to a defective gear assembly, perhaps imperfect/varying gear meshing caused by misaligned gears, bent axle, or a badly machined hole in the middle that might also explain why it's jammed on the axle.

The frequency of the noise could perhaps have been once per wheel revolution. I have a short 8MB video of it but can't find out how to upload it. The "Attach files" facility doesn't seem to like .mov or .mpeg - it just says "unsupported file type" (= PO). Can anyone help with this please?

I decided to switch the gear assemblies over but not the motors and see what happened. If that didn't reveal a fault, Plan B was to switch the motors but not the gear assemblies. One of those should be conclusive. If both indicate no fault then a previous bad connection as suggested by saneagle, and now resolved, would be indicated.

I then found that the reduction gear assembly on the defective motor is jammed on the axle and will need a puller to get it off. So it appears there is definitely something wrong, but I don't yet know exactly what. Rummaging for a puller... used to have one... have since moved house...

Reduction gear assemblies can be obtained relatively cheaply through an Aikema spares shop on AliExpress, but I want to check first that the axle is straight and the keyway good. Hopefully it will be sufficient to put the motor in the drop-outs and see if it wobbles when rotated. I should be able to see the keyway once the gear assembly is off.

If it's just a case of replacing the reduction gear assembly, that would be pretty good news!

Comments anyone?

 

[saneagle]

The gears staying in the case is nornal for many motors. Sometimes you get a circlip to retain it, other times not. Either way, it can't cause surging. If anything were wrong, the only faults you can get is the gears not driving at all or nasty rattling sounds.

With all these type of motors, it's not uncommon to see small design changes each time you open one up. Bafang CST was pretty bad for that. It used to drive me nuts when I wanted to fix them. It often meant that I couldn't swap good parts from broken ones.

 

[Slightlypedantic]

Ah well, I think it's down to funding a puller. If I can get the gear assembly off the duff motor and substitute the good one (that's the one that stayed in the case) and see if it runs well, I'll know where I am... hopefully. Beyond that it's looking like time to give up! I hate giving up... hey ho.
Thanks again.

 

[Slightlypedantic]

UPDATE
Sorry for the delay - things are a bit spasmodic on the e-bike front!

I managed to remove the gear assembly with the axle clamped in a vice and two large flat screwdrivers to prise the assembly off. Wasn't too hard with more or less even effort on either side, once I had got it moving.

I then swapped the gear assemblies over, so I had changed them but not the motor. All ran fine. Then I swapped the other motor/gear assembly in and that ran fine too. Admittedly not on the road and under load, but the funny noise was quite apparent before with the bike in the stand and no load. So I now have two good motors!

It looks as though Saneagle was absolutely right, and there was indeed a bad connection. Rather embarrassing to think that I may have overlooked the depth of engagement of the motor cable plug. Slightly surprising that Yose didn't suggest checking that before sending a replacement motor. Not that I'm complaining - two working motors is a very good outcome.

My skill at removing motor wheels and swapping internals has increased a lot with all the practice. What used to be somewhat intimidating and tedious is now the work of minutes. That's good too.

With the arrival today of a KT 17A SW controller and LCD3, I should soon be able to try current control, which everyone says is vastly superior to speed control, which I'm sure it will be.

Thanks again, and especially to saneagle.

 

[Woosh]

It looks as though Saneagle was absolutely right, and there was indeed a bad connection.

saneagle is ace at fault finding.

 

[saneagle]

saneagle is ace at fault finding.

Especially when it's the motor connector. I knew that if I shouted it enough times, I'd eventually get one right.

 

[Slightlypedantic]

Always good to look for the simple things first. Next project, install better lighting in workshop lol.

 

[Slightlypedantic]

After taking a break, The KT controller is installed and running nicely.

Thanks everyone for your help.