Brushlless motor stalling

[songflytta]

Hi

I'm fixing up and Electric bike for a friend and I'm having a problem with the motor stalling. I'd be grateful for any help anyone is able to supply.

The controller is one I picked up from Ananda drive. It worked great the first few times I tried it, however I stalled the motor for the first time this morning. (on a hill, the motor tried to retain power and then gave up and allowed me to roll back) When I got back on flatter ground I found the throttle didn't work at all. So I lifted the wheel and still nothing occured when the throttle was pressed.

If i roll the motored wheel back a little then I can get a little torque when I apply the throttle, If I then lift the wheel I can get it to spin, and it will continue spinning at full power. At this stage if I do a little jump start and throw myself onto the pick and put the motored wheel down then I can get it to take off. If I release the throttle at this stage and then reapply it then the motor power will resume. But if I allow the wheel to slow or stop, reapplying the throttle has no effect.

If I get on the bike and pedal to get up to speed and apply the throttle there is also no effect. It appears the motor needs to roll back a little before it will start to grab.

Is this likely to be a controller problem or a motor problem? I had a look inside the controller as I'd read things about shunts causing this problem. Unforturnately all the circuitry in the controller is covered with a gel, so I won't have any luck there. could I have perhaps broken a winding or hall sensor in the motor? Or if there is a controller problem is there a way of "reseting" the way it trips out the motor?

The other thing that I noticed was when I did get the motor running I could slow it down (provided I didn't release the throttle entirely and attempt to reapply) and when it was running quite slowly it seemed that power wasn't being applied consistently to the motor, there was instead one spot in the rotation where it appeared to struggle (based on the sound it was making)

As I mentioned I'd be more than grateful for any help, I have a few tools that I can do tests with such as an oscilloscope and DC power supply, but i'm not quite sure what to test. for extra info, the motor is 200W, planetary geared at about 5:1, Hall effect sensors, three phase power.

Thanks
Lawrence

 

[wrobinson0413]

.

 

[dnmun]

i guess i don't follow the part about how you need to roll backwards in order for the motor to run. rolling backwards under throttle i thought was gauranteed to blow the high side FETs.

i have one of those little ananda controllers too, actually 2, they are potted in goop that is impossible to remove. lemme know if you want one since i only need one. are you rich? hehe.

 

[songflytta]

Hi, and Thanks for your replies

I would guess that you blew one of the highside fets. Had similar symptoms when I accidently modified one of my highside mosfet control signal to be off permanently. The wheel as you rotate it will have torque except at certain spots in the rotation. If you get the wheel rotating, the inertia will allow you to coast over the dead spot and continue to rotate.

Ok, that sounds not so good. If i've blown a highside FET how would I test for that (since torque is rather difficult to measure) I have a little ac power hall effect transducer used to measure power in my house hold appliances. This morning I put this around each of the phase cables in turn as the motor was running, I found that each of the cables seemed to be supplying approximately the same power to the wheel. (ie i didn't seems as though one phase wasn't doing anything) but that said my sensor is particularly accurate, so I might use an oscilloscope instead when i have a spare hour.

i guess i don't follow the part about how you need to roll backwards in order for the motor to run. rolling backwards under throttle i thought was gauranteed to blow the high side FETs.

yes I thought the bit about rolling back was a bit odd too. I'm only talking about a fraction of a turn here, but it seems thats what it wants.

i have one of those little ananda controllers too, actually 2, they are potted in goop that is impossible to remove. lemme know if you want one since i only need one. are you rich? hehe.

I'd love another controller if thats whats broken, Alas i'm not rich, otherwise I'd own an electric bike myself instead of fixing up one for my friend. I'm sure we can work something out though.

Let me know your thoughts
Thanks
Lawrence

 

[wrobinson0413]

.

 

[HighTekBikes]

When troubleshooting brushless motors, you should always check the hall sensors. There is a detailed procedure in this forum somewhere, do a search. But basically, you apply power to the sensors by plugging the sensor connector into the controller. Disconnect the power wires to the motor. Power up the controller and using a meter or scope, probe the three phase output wires. They will go from 0 to about 5V. Use the ground as a reference, either the black wire on the connector or battery ground. Be careful not to short anything. Rotate the wheel very slowly, while monitoring. For a geared motor, you have to turn the motor backwards, so the internal gears will engage and rotate the magnet assembly. If any of the sensors are stuck high or low, it's bad. If you see nothing on any of them, check for 5V on the red wire. Even if you find one, check the controller also.

 

[wrobinson0413]

.

 

[dnmun]

now why didn't i think of that? it was on the tip of my tongue, just did not think to say it. hehe

good advice from 2 guys who know controllers, post up a picture of your little ananda controller and especially all the plugs so when can show you where to put the voltmeter probes.

was it running ok before you stalled it on the hill and rolled back?

 

[songflytta]

Hi Guys

Thanks for all your help, This has been fantastic. I've been away for a couple of weeks and I'm only just getting a chance to look at this again now. I'll run through all the tests you've mentioned and let you know the result. One thing that may interest you is a couple of tests I did just before i went away. I put my CRO on all the phase leads and took a photo of the output while the motor was running -- included below.

Thanks - i'll let you know how it goes
Lawrence

--------

View attachment 2

 

[wrobinson0413]

.

 

[songflytta]

Hi Guys,

I've done all the test you suggested. Here are the results

Results to test suggested by wrobinson0413 » Thu Sep 03, 2009 2:02 pm
I've checked all the integral body diodes as per your suggestion. Here are the results:

Pos Lead - On Black of power input
Neg lead – Yellow: 0.537V
Neg lead – Blue: 0.536V
Neg lead – Green 0.531V

Neg Lead - On Red of power input
Pos lead – Yellow: 0.5327V
Pos lead – Blue: 0.5357V
Pos lead – Green 0.53V

All Fets seem to be good

Again all the values measured should be very close. If you have a damaged mosfet, this test usually catches it pretty quickely. Either it will be a dead short or it will be blown open for most failures.

Results to test suggested by HTB_Terry » Thu Sep 03, 2009 3:03 pm

Hall sensor Supply: 5v

I refer to motor positions in this experiment. I had 12 motor positions in total ie, each were approx 30 degrees apart. A "Click" could be felt as I rotated the motor slowly, there where perhaps 30 to 50 Clicks per motor position.

Testing Blue sensor
Motor Position 1: 5.027
2 ‘Clicks after position 1’ 0.027V

Motor Position 2: 5.027
2 ‘Clicks after position 2’ 0.027V
4 or 5 ‘Clicks after position 2’ 5.027
7 to 10 ‘Clicks after position 2: 0.0257
A few more “clicks’ : 5.027
A few more “clicks’ : .0257
A few more “clicks’ : 5.027
A few more “clicks’ : .0257

IE Four brief dips in voltage between position 2 and 3
Motor Position 3:
Four dips between position 3 and position 4

Testing Yellow sensor
Motor Position 4 to 5:
Down (0.027V) consitently with 3 UP spikes (5V) evenly dispersed between 4 and 5

Testing Green sensor
Motor Position 5 to 6:
Started down, 4 up points evenly dispersed between 5 and 6, Due to my slightly uneven numbering spacing though this is consistent with the other phases

Answer to Qn by dnmun » Fri Sep 04, 2009 2:07 am

was it running ok before you stalled it on the hill and rolled back?

Yes, it was all good before that


Comment to Dnmum:
Thanks for the offer of the controller, I'm running off 36V at the moment so I'm not sure how suitable this substitute may be. Also it seems the results of the tests above indicate that the controller is not the issue. Thanks anyway though

Results to test suggested by wrobinson0413 » Fri Sep 25, 2009 12:56 pm

If you measure the A-B, B-C, and C-A phase to phase dc resistance by passing a known current at at known voltage, you should be able to see if one of the phases is shorted.

I did this with a multimeter, although I'm not quite sure of the validity of this as the windings are, of course , inductive. The results of the multimeter test are below ( all values in Ohms)

Phase to Phase resistance testing with DMM
Red probe Yellow, Black Probe Green: 0.18 drops to 0.16 holds at 0.16
Red probe Green, Black Probe Blue: 0.20 drops to 0.14 holds at 0.15
Red probe Blue, Black Probe Yellow 0.17 drops to 0.14 holds at 0.15

Time to stabilize, approx 3 to 5 seconds

Black Probe Yellow, Red probe Green: 0.18 drops to 0.15 holds at 0.15
Black Probe Green, Red probe Blue: 0.18 drops to 0.14 holds at 0.14
Black Probe Blue, Red probe Yellow: 0.16 drops to 0.14 holds at 0.14

Similar time to stabilize

I do have a Power supply, rated at 13.6V with a max current of 6A. Rated to a 50% duty cycle. 6A seemed like a lot so I wanted to ask you about this before I did it. Since its about a 200W motor, thats 66W/phase. Since the Phases are running at 36V and you saw on the picture that each phase is on for maybe a 20-30% duty cycle. 66W/(36V*30%) = 6.1A so I'm thinking that the windings should be happy to take this. Does that sound feasible to you?

Thanks
Lawrence

 

[wrobinson0413]

.

 

[HighTekBikes]

If all the motor hall sensors are changing state, then you probably blew a mosfet in the controller. When you rolled back, the controller got the wrong info from the sensors and turned on the wrong mosfets.

Just buy a cheap replacement controller someplace, it's good to have a spare anyway. I have some misc. controllers including from Ananda if you are interested you can contact me.

-Terry

 

[wrobinson0413]

.

 

[songflytta]

It is too bad that he doesn't have another motor to test the controller on or another controller to test the motor on so that he can isolate which side is the source of the problem.

I thought about a test that may work for this. Could you confirm if this is valid

I will connect the yellow hall effect wire and phase cable outgoing from the controller into the Blue phase and hall effect wire into the motor. Blue from controller into green on motor, Green from controller into yellow on motor. Then i can look at the wave forms again and see if the squared off wave remains with the yellow phase of the controller, or is linked to the yellow phase of the motor.

Is this likely to damage anything however?

Cheers

 

[songflytta]

If the body diodes are still intact as what I saw from his measurements, it is less likely that a mosfet is destroyed though not improbably. He certainly has something funky going on, but I am not convinced that it is a power mosfet. It is too bad that he doesn't have another motor to test the controller on or another controller to test the motor on so that he can isolate which side is the source of the problem.

I just did the test where I connected up the controller phases to a different colour phase of the motor (hall effect sensors too)

Controller Motor
Yellow Blue
Blue Green
Green Yellow

The flat spot in the curve on the output stayed with the yellow phase of the controller (which was connected to the blue phase of the motor). This makes me think that it must be a controller problem.

 

[wrobinson0413]

.