Mini Motor Electrical/Mechanical problem

[luisgouveia]

Hello guys,

Yesterday my bike stopped.

As always, I picked up the bike in the morning and started riding it to work.

15 seconds after starting the ride, the bike stopped. The motor was still cold.

Batteries are ok and I guess the controller is ok.

The motor, however, changed its behavior. Now, when I move the bike backwards, I feel lots of resistance to the movement (even with the batteries disconnected).

I would like to ask you if this could be a mechanical problem in the motor? Have someone in the forum experienced such a problem?

Thank you!
Luís

[Bikemad]

Hi Luís, the minimotors should have more resistance to being turned backwards than if they are turned in the normal forward direction because the wheel has to turn the motor and the gears via the locked up internal freewheel when turning backwards.

However, if this resistance is much more than it used to be, try disconnecting the three thick phase wires going to the motor and see if the resistance to turning the wheel backwards returns to normal. If disconnecting the wires actually reduces the resistance, it would indicate that the speed controller itself has become damaged causing one or more of the FETs to fail in a short circuited state.

If the resistance is not affected by disconnecting the phase wires it could just be the normal resistance that you have not really noticed before. If a hall sensor has been affected by moisture, or has simply failed, the minimotor cannot just continue to operate in sensorless mode because the internal freewheel prevents the motor from being turned, which is necessary in order to initiate sensorless operation.

Did the motor stop while it was under power with the throttle still being operated, or did you fully release the throttle (at a junction etc.) and then notice that the motor did not respond when you reapplied the throttle again?

Do you have the horn wired up, and if so have you heard any beeps?

How many cells (what voltage) were you running on when it died?

Alan
  

[luisgouveia]

Hello Alan,

Thanks for your quick answer. I will answer your questions looking forward to help diagnosing the problem:

1) Yes, there was much more resistance than it used to be. I disconnected the three thick phase wires and it returned to normal just as you said.

2) The motor stopped while it was under power with the throttle still being operated.

3) Unfortunatelly I hadn't my horn wired up at the time, and so I didn't heard any beeps.

4) I was running at exactly 38.2V when it died (2 LiPoly 5S).

I was assuming that the motor was dead, but after reading your email I understood that the problem is in the controller.

I've tested the motor today with another magic controller and the motor is fine. I bought a second controller (the one that burned) because my first one wasn't programmable. I cannot ride the bike with this first controller because I may burn the motor this time. It delivers 30 Amps to the motor. The one that burned was programmed to deliver only 12A. I opened the controller to see if anything was looking bad but for my surprise, and as an electronical engineer, I can say that everything looks fine. I can send you some pics if that will help. Do you recommend me to replace any component? Thanks a lot!
 

[Bikemad]

I opened the controller to see if anything was looking bad but for my surprise, and as an electronical engineer, I can say that everything looks fine. I can send you some pics if that will help. Do you recommend me to replace any component? Thanks a lot! 

Luís,
As the controller has failed, and has not been overvolted or abused, I would like to think that it would be replaced under warranty.
If you purchased the controller direct from GM in China, send an email to tom@goldenmotor.com explaining that your BAC-028xP controller has failed and also tell him when you purchased it or give the order confirmation number etc.
If you purchased it from another dealer, I would email them instead.

If, for some reason, you are not able to get a replacement controller under warranty and are thinking about repairing the failed controller, it's unlikely to be as straightforward as just replacing the failed MOSFETs (FETs):

This testing procedure is for use with a digital multimeter in the diode test-range with a minimum of 3.3 volt over d.u.t. (diode-under-test)


Connect the 'Source' of the MosFet to the meter's negative (-) lead.

1) Hold the MosFet by the case or the tab but don't touch the metal parts of the test probes with any of the other MosFet's terminals until needed.

2) First, touch the meter positive lead onto the MosFet's 'Gate'.

3) Now move the positive probe to the 'Drain'. You should get a 'low' reading. The MosFet's internal capacitance on the gate has now been charged up by the meter and the device is 'turned-on'.

4) With the meter positive still connected to the drain, touch a finger between source and gate (and drain if you like, it does not matter at this stage). The gate will be discharged through your finger and the meter reading should go high, indicating a non-conductive device.

Such a simple test is not 100% -- but is useful and usually adequate.

When MOSFETS fail they often go short-circuit drain-to-gate. This can put the drain voltage back onto the gate where of course it feeds (via the gate resistors) into the drive circuitry, possibly blowing that section. It will also get to any other paralleled MosFet gates, blowing them also.
So, if the MosFets are deceased, check the drivers as well!
 

It will be interesting to know the outcome regarding a warranty replacement, but at least we know this particular failure has not been caused by regen, because the minimotors don't have it!   

Alan
 

[luisgouveia]

Hello Alan!

Thanks a lot for your reply. I really appreciate your concern.

I was just trying to avoid the costs of sending the controller back to Holland, but you're wright, there's maybe nothing I can do about it.

"If you purchased it from another dealer, I would email them instead."

I bought it from your dealer in Europe, Devi-Motion. I really appreciate if you can send that email.

Purchased on: 20 May
Sent to Portugal on: 1 June
Order number:  1082500_00189

Thanks a lot Alan. I keep a good contact with Dennis, but if you'll contact him yourself it will be better, obviously.

Regards,
Luís

[Bikemad]

"If you purchased it from another dealer, I would email them instead."

I bought it from your dealer in Europe, Devi-Motion. I really appreciate if you can send that email.

Luís, there seems to be a slight misunderstanding here, I don't work for GM, I just help out on the Forum.

I wrote that comment the way I would have said it, but it was just a "figure of speech" and was not meant literally.
Perhaps it would have been better if I had written it so it made more sense:

"If you purchased it from another dealer, you should email them instead."

Sorry for any confusion this may have caused.

Alan

[luisgouveia]

Hello Alan!

Lol, it's ok!

Indeed I always thought you worked for GM!

Thank you!
Luís

[MonkeyMagic]

Alan you really have to stop wearing that GM t-shirt mate

 

[Leslie]

It will be interesting to know the outcome regarding a warranty replacement, but at least we know this particular failure has not been caused by regen, because the minimotors don't have it!  

Alan
 

I think mini motors just go from high RPM whirrr to DOINK "knee jerking halt" inside the motor when the regen braking is applied.  Thats got to be not right IMO.

[Leslie]

Alan you really have to stop wearing that GM t-shirt mate

 

It took me a good two years to take it off.  Admittedly I admire loyalty and this is a good thing.

I think it's official for a long time standing the magic controller regen or not regen is not up to the task.  

However I am full aware of what I have purchased and the hubs are great to say the least.  Both melted HBS and still functioning HBS has been such good motors.  The things I did to my HBS motors was incredible and take full responsibility for melting one.

As for warranty I am sick to death of the word.  If I see it again Im going to shave my head, chant Hare Krishna and go insane.

Honestly give up and await for GM to fix the bug what ever it maybe.  Any replacements are going to be the same and just be insult to injury.

I on the other hand when I get my pies Im going to look at the hardware and possibly grab the firmware and analyze the code.

I just order 100 pcs bidirectional 60v 400 watt 8p/m  bi-directional TVS's that can go straight onto the phase wires,  I might try 10 of these per phase replace the relevant caps with 100v parts, remove the internal controller and box it, anchor fets to a big heatsink and mod it to handle 45 amps via shunt then limit the software down to 30 amps max.

If that doesn't work then there is nothing that can work.  If I don't screw the job I may get a great controller.  If I screw it up which I suspect I will there will be no bothering Gary at GM CA as it wont be his doing.

Then I look at a Keywin or Infineon from Renascence.


Another thing maybe to find out what the hell low and high braking pads are on the Magic Controller are, hoping one is just motor cut off with no regen, or just risk not using any brake switches all together..

[MonkeyMagic]

Hi Leslie

Unfortunately you won't be able to touch the firmware mate, unless you can read binary...

I don't think GM will release that. What you can possibly do if you end up modding that controller is listen to the serial comms from the usb cable into the pie for current settings, I'm tipping it would use binary strings. Anyway what I am thinking of doing is listen to the packets sent to the controller, record data for each setting I wish to change to and from.

It would be max around a 1 second delay I figured, as you will need to switch the power connection to the wheel, plus re-direct the 8-pin plug via transistors to the PIC or whatever microcontroller you are using serial input/output; then allow for a small delay before enabling the throttle etc. You can select your options from a character, graphic or even a touch screen if you are really nifty.

I need to set some time aside to monitor strings to and from the controller...

Honestly, I think you will be happy with the internal controller Les- unless something (touch wood) happens then I doubt it's worth going to an external controller or modding your new one voiding any..... ok I won't say that word..... you get it

Man you would be salivating by now waiting for those pies !!

*Leslie sitting at the table, with a bottle of sauce waiting for his pies....

[Leslie]

I found this image on google images.




I get the feeling this sums up how I feel ATM.

[Leslie]

Honestly, I think you will be happy with the internal controller Les- unless something (touch wood) happens then I doubt it's worth going to an external controller or modding your new one voiding any..... ok I won't say that word..... you get it

Man you would be salivating by now waiting for those pies !!

*Leslie sitting at the table, with a bottle of sauce waiting for his pies....

Yes one Pie Gary tested is able to get up to 25 amps via software, and the other cant go higher than 20 amps, the 20 amp Pie concerns me as I have seen what usually cause these bottle necks.

I have the Ping packs and it gets razor close at 62v full charge to 63v. This is why I think the bidirectional 60v zener Transient Voltage Suppressors straight onto the phase wires maybe what will make the difference.

I think I will shelve the 25 amp Pie at first and go to work on the Pie I think will fail me first and have some fun.

Honestly if both Pies test as working IMO that W........ word has been served, Gary is a good guy and I feel it not his fault if any products fail due to the engineering faults..

Plus Yuan has a point with the non GM packs other than the 48v Lifepo4 GM sells.  The voltages are different and I feel one is cutting it too close with 16 cell lifepo4's so to some degree some Modding intervention should be required to make my planned config reliable.


Yes agreed you should use the source code.  I don't know why some are so secretive with such data it could only serve us and GM well if it was more open source.  But hey I can download an entire operating system source code, in essence a huge collage of art in comparison.

[Leslie]

STP140NF75

APPLICATIONS
; HIGH CURRENT, HIGH SWITCHING SPEED
: SOLENOID AND RELAY DRIVERS
: AUTOMOTIVE 42V BATTERY DRIVERS

If you look at the recommended apps for the magic controller fets it is only advised for driving 36v batteries for the hybrid cars I believe.

Ive been reading heaps on unclamped inductive load testing of fets.  And what I am understanding spikes may occur at higher temperatures.  This still has nothing to do with the possibility of 63v caps fusing.

Unclamped Inductive Switching (UIS)
Test and Rating Methodology
Introduction
This application note will review the basic principles
surrounding unclamped inductive switching (UIS). It
will examine what it is, the typical UIS ratings
reflected on data sheets and how designers can
properly use them. The main purpose of this
application note then is to supply designers with
useful tools and information needed to appropriately
deal with UIS related issues in their circuits.
The Need for Power MOSFET Avalanche
Ruggedness
Power MOSFET’s inherently have extremely fast
switching speeds. As a result, designers often use
them in high speed switching circuits which take
advantage of this capability.
Using MOSFETs in high speed switching circuits can
lead to device stress not normally encountered in
slower switching circuits. In fact, switching speeds
may be so fast that at device turn-off, small parasitic
inductance in the circuit can lead to significant over
voltage transients (Figure 1). This is due to the fact
that when current through an inductor is abruptly
turned off, the inductors magnetic field will induce a
counter electromagnetic force (EMF) resisting the
change. If the resulting voltage transient is large
enough, the MOSFET may be forced into drain-tosource
avalanche, V(BR)DSS.
Figure 1. Drain-to-Source Over Voltage Transient
During Turn-Off
The peak over voltage transient during turn off can
be determined by the following equation.
Vspk = L * di/dt + VDD (1)
where
Vspk = peak over voltage transient voltage
L = load inductance
di/dt = rate of change of current at turn-off
VDD = supply voltage
According to equation (1), the faster the switching
speed and or the higher the load current the more
likely a device is to experience an over voltage
transient. Currents and switching speeds may be so
high in some circuits that even low parasitic
inductance may be enough to force devices into
avalanche and possible device destruction.
Due to their inherently fast switching speeds, it is
clear that power MOSFETs need to be designed and
manufactured to insure that they have adequate
avalanche ruggedness for today’s high performance
circuits.
Avalanche Ruggedness Test Method
The avalanche ruggedness of a device can be
measured using a test circuit that performs an
Unclamped Inductive Switching (UIS) function like
the one shown in Figure 2. This type of circuit
mimics the actual application where unclamped
inductive loads are present. A device is considered
rugged if it survives the test at the specified test
conditions.

Slowly Im coming to grips with the engineering math.  It seems to all be dependent on switching speed derating above 25c and inductive spikes.  So not only constant loads are effected by thermals but avalanche ruggedness is effect as well.

Noting you observe more power at the lower voltage settings.  It may be well advised that changing default software settings to the correct values can be paramount to controller reliability too.

Im going to purchase a hand held oscilloscope and test the FM speed at various voltage settings as well.  Curiosity will drive me to do all this eventually but I want to just ride the thing and enjoy it the most.  So to a large degree unclamped inductive testing could prove in a non functioning controller.  Good enough reason to find some means to clamp the voltages at the phase wires and find some means to limit the temperture of the fets as much as possible.