scooter hub motor - hall sensor placement

[tef]

hi,

I hope that somebody can help me with troubleshooting my 3kw brushless scooter hub motor.
The third controller is still not working.

the story:

This summer I bought a used scooter with 60V lead acid batteries.
The original controller let the hub spin but the speed was less than 10km/h up the hill.
I did not test further because my A123 cells were on the way and I had to remove the slas and upgrade the original infineon controller for higher voltage (fets, caps,...).
The modified controller shorted after a few metres.
I thought that something went wrong with the controller upgrade (fets not properly isolated,...) and replaced the defective fets and rechecked everything.
But the controller shorted immediately after just "touching" the throttle.

The Hall sensors were switching between 0 and 5V and I could not detect shorts between the phases of the motor (resistance is equal (200mOhm), no suspicious cogging).

After spending days with modifying and repairing, I decided to give up with this controller and ordered a new 24fet 4110 from lyen.
I found the correct hall/phase combination and tested with very conservative current settings.
Result of test under load: shorted mosfets

I thought that the low resistance or inducance of the motor is a problem for those infineon controllers and ordered a Kelly KBL09401B.
I went through all 36 hall/phase combinations a couple of times with 60 and 120 degrees hall spacing.
There is no combination which lets the 13'' motor spin faster than a few rpm.
If I get the motor to spin, it sounds rough and the controller starts to report a hall error after some turns.
At that time the halls were still ok.

Afterwards I managed Somehow to damage the halls during trial and error hall/phase matching, so I opened the motor and replaced them successfully with honeywell ss41.
I even rewired everything and upgraded the phase wires, but no change. The motor just turns slowly and sounds rough as before.
BTW, the middle of the three halls is reversed!! I read somewhere that this is an indicator for a 60 deg hall placement?
There is a position where all three sensors put out 5V!

The last thing I did is to attach a scope and measured the back EMF voltages together with the hall signals.
I labeled the phases x,y and z and attached one scope channel to the halls and one to the phases.
The wheel was turned backwards.



I just know the basics of EE and don't know how to go on.

Can somebody please tell me if there is something suspicious (placement of hall sensors)?
I attached a screenshot of the measurements and pictures of the stator with the halls.

tef

 

[tef]

I connected the scope again and captured screenshots of all hall signals together.

View attachment hallSignals1.BMP




Could the yellow hall signal measured on chanel 3 (purple) be the problem?
The image I attached in the previous post shows that the peaks of the curves (yellow hall) are shifted!!

 

[amberwolf]

If the purple trace is on the same vertical scale as the others, then it's not outputting even half of what they are, or rather, it is not going as far to ground as they are. Either the hall itself could be bad, or if it uses a pull-down resistor in the controller, that could be what's bad (poor solder connection, etc).

 

[tef]

thanks for the reply.

Purple is the digital channel, it has just a different vertical scale.

 

[tef]

To reach the 120 degree spacing I removed the middle hall sensor and epoxied it the same direction as the other two, so all sensors are facing the same direction.
I'm confused because I read that the middle sensor has to be inverted for 120 degree setting?
Anyway, the motor spins much better now but during higher rpm (loaded and unloaded) the kelly controller reports "frequent reset" and "internal reset" errors and shuts down.

The oscilloscope shows that the middle hall sensor (green) does not switch immediately.
What could be the reason? Is it likely that the sensor is defective or just misplaced (interferences?)?

View attachment IMAG071.BMP

 

[Gordo]

Edited to agree with John's post below;
You are not alone with this puzzle. I too have a 60V ZAP scooter (X-treme) with a blown controller. I also purchased a Lyen Controller, which will not run the motor. John in CR also has one of these 3KW motors and a blown controller. He has a 36FET controller on the way from China and will report what happens with it. I am now leaning towards getting the original controller from Mt. Chen, who exported these scooters to NA. Several people believe that the very low inductance of these motors is the problem. Lyen's only comment, from many emails, is "this controller runs all motors." So now we have 2 people with this 3KW motor, with Lyen controllers which can not run them. I would like Lyen to answer how 2 people can be wrong?
Lyen follows a very clever script. I told him which motor I have and asked him to tell me which controller to use. He gave me 2 choices. Now I have a controller which will not run this motor, he points out that I made the choice, so the cost of replacement is for my account. I will not be buying another Lyen controller.

 

[John in CR]

I haven't tried any other controllers yet on mine. I just didn't want to run my 24fet on it, because that was the last high power controllers I had at the time and wasn't willing to risk it. I'm in the final stages of frame mods and will launch hubmonster in the next week now that I have a variety of power controllers. First though, it sounds like I need to try a throw away controller on it first and definitely use chokes on the phase wires, at least until I'm positive about wiring and function. If programmable, start with the lowest current settings too.

The controller I blew on the same looking motor was the factory controller, but it blew because I modded the shunt to double the current, so it blew from abuse. I haven't connected another controller yet. My motor looks the same, with the only difference being dual sets of halls and dual hall wire plugs. I tried both and one seems to have slightly advanced timing compared to the other, but the controller blew before I had a chance to find out much.

I've blown controllers on that first throttle pulse exactly like that with my other motors, and they run fine with regular controllers. Arlo1 has blown lots of controllers with Colossus the same way. He measure Gordo's motor at about 30uH on each phase, which compares to a couple hundred on an X5304, so chokes are warranted at least till we get the motors spinning up well. After this thread I'm definitely winding some toroidal coils with double 8ga before even thinking of running the monster ventilated 36fet I got from SteveO. I think I'm going to remove half of the 12 shunts he put on it too, so I can program it low enough for testing.

There's some info somewhere in the adding halls to RC motors threads that explains the flipped center hall and physical angle of sensor placement, and how that relates to phase angle for the controller.

Unfortunately I've never used a scope and have no training, so I need explanation to me and can't provide answers.

The only thing I have to say is the hall board wiring looks a bit funky and seems to stick up too much. Could it be shorting out against the cover.

Together we're going to get this solved. Hopefully with no more blown controllers.

John

 

[tef]

hi,

@gordo
Did the mosfets short on your lyen controller or did you have other problems?

@john
I ordered the kelly because it has current limiting on the phases and they sell this motor too, so I hope I won't need additional inductance.
I will repair and try my 24fet again and post the results after I solve the hall problem.
Can you remember if one of your hall sensors is flipped?
Will you mount your motor to a bicycle?

 

[Gordo]

I am assuming the MOSFETs have not shorted as I see no physical evidence. It still tries to turn the motor, but no matter which combination of phase/halls I try, it does not turn. Both sets of HALL's check out on the tester as do the phase's.

John's motor and mine look identical. Both have 2 sets of HALL's. They are spaced differently than yours. Two of each set are in the slots and the center one is in the center of the lamination's. Our original controller is still available for $200 plus shipping.
As we seem to be in an exiting time with controller development (and I have 2 spare rigs), I will wait a few months to see what shakes out.

 

[tef]

View attachment IMAG071.BMP

I switched the hall wires and found out that I get the abnormal signal always on hallinput A of the controller.
The hall sensors are ok.

What could be wrong with the controller? Does somebody have an idea?

 

[Gordo]

It looks like a capacitor on the A leg is charging up. Have you tried switching the A leg to a different input on the scope? Maybe the scope is giving you BS? I also not that as you see a little positive spike on the C channel, you see a corresponding negative spike on the A channel

 

[tef]

Have you tried switching the A leg to a different input on the scope?

Yes, I swapped the inputs.
The result on channel B is the same, so the scope and the probes are ok.

Thanks for the suggestion.

 

[Gordo]

If we look at your first screen shot, we see a normal trace for all 3 HALL's. If I read the sequence correctly, you have inverted the center HALL and now you get a different trace. Seems to me you should put that center HALL back the way it was. You are now seeing it turn on slowly instead of quickly.

 

[tef]

hi,

You are correct, I swapped the middle hall.
The reason why the first screenshots look normal is that I used the hall tester from lyen and rotated the wheel manually.
With the controller attached I get the error on Hall A as you can see in the last image.
I even connected a new hall directly to the controller hall inputs and triggered it manually with a magnet.
The result is the same, input A is delayed.

In the meantime the whole project is getting really frustrating and I'm not sure anymore if I'll ever finish the scooter .

 

[Gordo]

What does the trace look like if you turn the wheel backwards?