Cut out in "3" but power stayed on and still worked in "2" : NOW FIXED

[jeff-jordan]

...
the bit that seems odd is ..

With +ve test voltage to the phase terminal and -ve test voltage to -ve power terminal, sometimes I read 41 kOhm and sometimes MOhms. It did not seem consistent.
...

That's related to some capacitors that you'll charge/discharge when swapping around the polarity of the tester.

So over all, the static tests for the MOSFET bridges seem to be ok.
No low resistance and no odd behaviour of just one phase terminal.

But things may look different when the currents increase and/or the controller tries to invoke field weakening (which happens above ~ 60 km/h or ~ 40 mph).

By the way, were you able to check the controller settings?
.... using a fardriver BT-module & fardriver app or RS232/TTL and motorc.exe?

[bob2.0]

I have a USB lead into that funny 4 pin inline socket.

I have not yet attempted to use it.

What should I look for? Could the file/settings have become corrupted?

[bob2.0]

But things may look different when the currents increase and/or the controller tries to invoke field weakening (which happens above ~ 60 km/h or ~ 40 mph).

That is the sense I get, like something just not happy at higher currents. Motor runs smooth, cuts out, motor still runs smooth. If there had been a short, I'd have expected a 'brrrrr' from electromechanical lock and torque ripple.

That is one last test I could do on the motor, I suppose, to short phases and see if they lock the motor as expected.

Overall, I think the controller is the weak part of the setup. Considering I have ridden many miles flat out on the motorway, it really has been at its limit. The motor has always been cool and seems to be capable of higher currents. This is my 'feeling'.

I would not trust the controller to give me an accurate self-diagnosis, if it is going wrong anyway. I would check the settings so that I can set a new module to exactly the same conditions, though, so I will look into this. I am not so good with the computer-connection things though.

[bob2.0]

Shorting the leads by hand, holding together with finger pressure, I measured 3.5 Nm torque ripple resistance to turning the wheel.

This was consistent for each pair of phases.

I'd have expected much higher, but I'd guess it shows the motor phase leads aren't really a suitable wire gauge for the phase currents of the motor.

[bob2.0]

It was rusted magnets.

I have now repaired this, details in another thread.

This is what happened;-

• Condensation had accumulated and caused corrosion in the motor. (Previous owner had stored it outside. No evidence of water ingress.)
  - A couple of magnets had rusted on the back, and swollen, which put them in the way of the stator.
  - The rubbing at high speed wore down the magnets and made them hot.
  - They must have exceeded their Curie temperature, and demagnetised.
  - The lack of magnet field at that position caused a Hall Effect error at low speeds. (At higher speeds, maybe because they were moving fast I guess they generated just enough field to trigger the sensor. Or maybe the software error tolerance is higher at higher speeds.)

I identified this by;-

• Noticing binding and (after standing a few weeks waiting for a replacement controller) 'zing zing'.
  - Dismantling the motor and noticing triangular wear patterns on some of the rusted magnets (showed they were lifting up at one end and touching the stator).

I fixed this by;-

• Chipping off the damaged magnets (quite easy)
  - Finding others of a suitable size (not so easy)
  - Checking they were going to be OK (need a Gaussmeter, fortunately I have one)
  - Using a structural adhesive to attach them

It was technically an easy job, but time consuming. The magnets are thin and not that powerful, you can handle them quite easily and there isn't any risk of chopping fingers off with them, they are too thin so not strong. The most difficult bits were actually reseating the bearings and rerouting the motor cables!

If anyone has to attempt this repair, feel free to ask for tips.

[jeff-jordan]

Congratulations for fixing your issue with the bike !!!
It's always a pleasure to see people using their brain, developing skills and finding a solution.

And thank you for keeping us up to date with your replace-the-defective-magnets job .

[bob2.0]

Congratulations for fixing your issue with the bike !!!
It's always a pleasure to see people using their brain, developing skills and finding a solution.

And thank you for keeping us up to date with your replace-the-defective-magnets job .

I now have a spare 72530 controller, too!!

[jeff-jordan]

....
I now have a spare 72530 controller, too!!

Install it, configure it, and enjoy the improvement

[bob2.0]

Install it, configure it, and enjoy the improvement

I need to convalesce first. Get some miles in the saddle and 'heal'. Overcome the trauma of open-motor surgery!