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

[dominik]

The Hub-Motors can be serviced.
You can replace the halls, add a temperature sensor and you can change the phase wires.
To repair the winding itself, is also possible, but much more complicated and it needs a lot of time.

A few pics from my 5kw scooter Motor, where I melted the phase wire insulation.
https://www.elektroroller-forum.de/view ... 48#p316448

[jeff-jordan]

...I think there is a fault in the motor. The impedance is too high (short circuit). ...

Hmm... not sure about this assumption.
Why would bob be able to drive until field weakening kicks in?
As soon as the currents of the phases differ (due to a shortcut), you should notice that with growling sound of the motor and strong juddering.
But this could be caused by a faulty MOSFET bridge too...

By the way: If there is some kind of short circuit at the motor, the impedance would be too low, the conductance would be too high.

Anyway, while waiting for a replacement controller, it would be a good idea to measure the phase wires against each other (with disconnected controller) with a good ohmmeter.
But if there really is a shortcut at the winding of one of the phase wires, the controller should have noticed that and generated a corresponding error beep code.

[bob2.0]

I took it apart again.

(You know you are working on a bike too much when it only takes 5 minutes to take it apart!)

Between each phase I measure 0 ohms. I mean, the coil resistance is lower than the contact resistance of the DVM.

I could put a high current through and measure the voltage difference, but what am I expecting? I need to know to apply an appropriate current, and I think I only have a 30A supply.

I measured on my LCR meter 61uH. I was surprised, I thought it would be a mH or something? I measured 61uH on each pair of phase leads. What would you expect?

I also put the capacitance meter on, just to see what it said, and in each case it said -10uF for each pair (negative, which is what I would expect if it was testing at 10kHz, not sure the test frequency, sounds plausible).

Between each phase and the case of the motor was around 20MOhm.

Between the case of the motor and the chassis was 100k~1Mohm. I guess just an 'incidental' contact somewhere.

Between each phase pair out of the controller; 81kOhm.

Between each phase terminal and motor case; 20MOhm.

Between each phase terminal and chassis; 5ish MOhm


.... a mystery?

How can I better test for short circuits between phases, or does the inductance measurement sound about right?

I still get the feeling that a FET is folding up in the controller under high phase current. Just my feeling about it. Again, how could I tell?

Are there any further electrical measurements worth making?

[bob2.0]

......Anyway, while waiting for a replacement controller.......

Sounds like you think I should go ahead?

I have not yet decided.

I'll keep looking for prices. I think the ebay outlet was without VAT, so the aliexpress options would be cheaper.

[jeff-jordan]

These measurements sound ok for me.

It's difficult to measure low impedances, especially with an ordinary multimeter.
If you've got a lab power supply, where you can adjust the current, you can feed a 1A current through the phases and measure the voltage in mV.... which would be equivalent to milli-Ohm then.

By the way: if you order at aliexpress, you might have to pay customs duty and VAT too, before the consignment passes the customs office (as I did here in Germany).

Further measurements (at the controller):

Disconnect every terminal.
Measure the resistance between the red terminal and all the phase terminals, for the upper part of the MOSFET bridges... and between the black terminal and all the phase terminals, for the lower Bridges.
It should always be between 22kOhm and 50kOhm.

In addition, a measurement at the motor:
If you've got 3 multimeters available: put one between each phase of the motor in AC mode. Then: turn the motor by hand (while the E-Odin is jacked up at it's main stand).
You should measure the same voltage between all three phases.
If one of the phases has a shortcut, one or two*) voltages would be significant lower.

*) depending on how the coils are connected, wye or triangle.

[bob2.0]

Measure the resistance between the red terminal and all the phase terminals, for the upper part of the MOSFET bridges... and between the black terminal and all the phase terminals, for the lower Bridges.
It should always be between 22kOhm and 50kOhm.
....
turn the motor by hand (while the E-Odin is jacked up at it's main stand).
You should measure the same voltage between all three phases.
If one of the phases has a shortcut, two voltages would be significant lower.

Thank you for your experience and support.

Excellent ideas. Yes, spin motor as a generator, of course. I should have enough lying around, or at least one pair at a time.

Yes, I have had customs problem even from Germany. But AliExpress and the UK have now an arrangement and all vendors have to charge VAT (it may not be there for Germany). Funny that ebay does not (not over £150). I prefer to be charged by vendor, it makes it more known. But I have had other things come through OK, just have to pay the charges once it is here.

Let me please ask also; I can see ND72530 for sale, but cheaper than this I can find ND72680. I might as well go with 680? Or is there a programming/pin-out difference to our bikes.

Even cheaper is ND72450? Maybe 450 is enough? Very cheap by comparison (half price of 530)

I am away over new year, if I order now then perhaps a good time, it might arrive once I am back. Which one? I am spoilt for choices.

[dominik]

The 72450 and the 72680 have another housing then the 72360 / 72530.
I used a 72450 in the Zündapp Bella build.
https://www.elektroroller-forum.de/view ... 19#p361419

They will fit inside as someone has already done this in the E-Odin and they should share the same small plugs , probably the big plug on the controller side is different or has a different pinout. You also have to change the way it is screwed/fixed in the Odin.

The cheap 72450 has a few functions less than the 72680, but all the basics are there. I only miss the variable regen controlled by the throttle. My 72680 has had this function, the 72450 and the 721800 I have installed are missing this function.
Regen with brake works with all of them.

[jeff-jordan]

As far as I can see, the ND72680 has a different pinout at the 30pin connector.... this won't be a problem if you build a bike from scratch, but you can't use it as a 1:1 replacement with the existing wiring harness.

But you can replace the genuine controller with the ND72530 without messing around with pinout and/or the mount.

[bob2.0]

Thanks. Great information. I really want the minimum of effort!

[bob2.0]

In addition, a measurement at the motor:

I measured each phase in turn, with just the one meter, spinning the wheel as fast as I could by hand to maintain a speed, and all pairs floated between 2.5 to 3.5V AC as the speed varied. I did not see anything 'obviously' different.

Disconnect every terminal.

I left the power lines connected, after checking they were ~5MOhms each to frame and ~10MOhm to each other.

Measure the resistance between the red terminal and all the phase terminals, for the upper part of the MOSFET bridges... and between the black terminal and all the phase terminals, for the lower Bridges.
It should always be between 22kOhm and 50kOhm.

With +ve test voltage to the phase terminal and -ve test voltage to +ve power terminal, I read 41 kOhm for each terminal.

With +ve test voltage to the phase terminal and -ve test voltage to -ve power terminal, I read a few MOhm for each terminal.
(should these be 41 kOhm too?)

With -ve test voltage to the phase terminal and +ve test voltage to +ve power terminal, I read 41 kOhm on each terminal.

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.

I thought I'd think about that a bit first, figure out if I was trying to measure across body diodes in the bridge.

Your thoughts would be welcome.