Measuring pole count in BLDC tool [probably solved]

General topics and discussions about the VESC and its development.
hexakopter
Posts: 86
Joined: 24 Dec 2015, 15:24
Location: Germany

Measuring pole count in BLDC tool [probably solved]

Postby hexakopter » 07 May 2017, 11:07

Hey guys,

is there a way to measure the pole count of a motor with the VESC and BLDC-Tool? I have a quite big motor, but don't know the exact pole count and don't want to open it just to get that value. I just know that it should have a lot of poles. :D
I read in that thread (viewtopic.php?f=16&t=568&start=30) that rew said he can read the number of poles from a screenshot of the current plot. I dont know if that was just a joke, because I don't really get how it should be possible to read the pole count from that given information. You can see the peak to peak time offset of both currents and the max amplitude, but I don's see a formula to bring that together to the number of poles. But I will attach screenshots anyways.

To make the screenshots I flipped my eBike and use a const throttle and hit the "Now" button afterwards.
Attachments
2.png
2.png (512.38 KiB) Viewed 1086 times
1.png
1.png (549.27 KiB) Viewed 1086 times
Last edited by hexakopter on 08 May 2017, 17:06, edited 1 time in total.

rew
Posts: 904
Joined: 25 Mar 2016, 12:29
Location: Delft, Netherlands.

Re: Measuring pole count in BLDC tool

Postby rew » 07 May 2017, 19:47

I'd guess 8 or 9. But your motor is of higher quality than the motors that I have so it is much more difficult to see. But because
you only have something like 12 periods in your plot, it is very difficult to verify that guess. I'll make a guess again if you make a new plot.

The question is: Do you see a pattern that repeats every N electrical cycles.

P.S. I have e-bike motors that have 22 poles, so that would not be visible on this plot because you have too few cycles in the plot. Increase the decimation or run the motor faster.

hexakopter
Posts: 86
Joined: 24 Dec 2015, 15:24
Location: Germany

Re: Measuring pole count in BLDC tool

Postby hexakopter » 07 May 2017, 22:37

Hey rew,

thanks for your answer. Why do you think my motor is of a higher quality than your ones? Just because it looks like an almost nice sine and every period looks nearly the same? I did the plot a lot of times to get that nice sine. I need to use a high throttle command from the accelerator throttle to get that result. (so a high current) In the screenshot "Throttle Low Speed" you can see how it looks using a little throttle command at low speeds.

After that I used the current command from the GUI to set 3A and 10A and did the following screenshots. All tests are done with the unloaded motor just fighting against the motor + wheel inertia. I think I would get better results with the plots really riding on the bike, so having a more constant rpm value and higher current. Right now the rpm value varies a lot during the sampling. Would you confirm that?

I can tell you that the pole count should be in the range of 44-48 poles. (so 22-24 pole pairs) I have a favorite number but it would be nice to confirm that with the measured value. (there are different informations on the internet) So I think I have to look for a repeating in the range of 22-24 periods in the graphs. Is that consideration right? That are a lot of periods in one graph and definitely not easy to see a pattern.
Attachments
Throttle Low Speed.png
Throttle Low Speed.png (539.07 KiB) Viewed 1063 times
10A Current.png
10A Current.png (630.28 KiB) Viewed 1063 times
3A Current.png
3A Current.png (650.9 KiB) Viewed 1063 times

rew
Posts: 904
Joined: 25 Mar 2016, 12:29
Location: Delft, Netherlands.

Re: Measuring pole count in BLDC tool

Postby rew » 08 May 2017, 07:34

Yes, your motor is better because all the sines look the same. On my motors I often see that one period every 7th electric cycle is recognizably different. This is difficult to spot on your motor.

The fact that yours has a sine shape does not really matter.

Can you run at say 40% duty cycle?

e-biker
Posts: 18
Joined: 14 Aug 2016, 13:20
Location: Lithuania

Re: Measuring pole count in BLDC tool

Postby e-biker » 08 May 2017, 14:01

If it is a hub motor it should have 48 poles, moust of I have used was 48 poles (MXUS, Cromotor, CRYSTALYTE HS series), exept crystalite 5303 which had 12 poles.

Have you tried spin a motor with your hands and then watch the tacho value in BLDC Tool? It should be pretie simple.

hexakopter
Posts: 86
Joined: 24 Dec 2015, 15:24
Location: Germany

Re: Measuring pole count in BLDC tool

Postby hexakopter » 08 May 2017, 17:04

@rew Have done three new screenshots at 40% duty cycle with a decimation of 1 (to see that it looks like a sine), 4 and 12 (to see if I find any pattern). But from a quick look I don't see a pattern.

@e-biker It is a MXUS hub motor, but I found on the web some people saying it has 48 poles and others saying it has 46 poles
Thanks for your tipp with the tacho. I think that will solve the question. I did a few tests with multiple revolutions.
For 10 revolutions I got a tacho count of 1373 => tacho count of 137.3 per resolution
For 50 revolutions I got a tacho count of 6855 => tacho count of 137.1 per resolution
For 100 revolutions I got a tacho count of 13717 => tacho count of 137.17 per resolution

Interesting that the rounded value is closer to 137 than to 138. My interpretation is, that I need to divide that value by 3 (because of the 3 phases) what results to 46 poles. So now I believe my motor has 23 pole pairs, so an pole count of 46. ;)

Thank you both for your help.
Attachments
Decimation12.png
Decimation12.png (671.2 KiB) Viewed 1037 times
Decimation4.png
Decimation4.png (542.08 KiB) Viewed 1037 times
Decimation1.png
Decimation1.png (493.06 KiB) Viewed 1037 times

e-biker
Posts: 18
Joined: 14 Aug 2016, 13:20
Location: Lithuania

Re: Measuring pole count in BLDC tool [probably solved]

Postby e-biker » 09 May 2017, 06:07

Just remembered that i have one disassembled mxus. I've counted it and your calculations is 100% right - it has 46 poles. Added few photos, maybe yours mxus is the same :)

Image

Image

hexakopter
Posts: 86
Joined: 24 Dec 2015, 15:24
Location: Germany

Re: Measuring pole count in BLDC tool [probably solved]

Postby hexakopter » 10 May 2017, 18:43

Yes, my motor looks similar to yours. (maybe yours is v2 in my is v3, so slight differences) Thanks for sharing the picture and confirm my calculations. I also searched a little bit for pictures of an opened MXUS, but couldn't find one where you see all magnets/poles to count them. :roll: Now we have one.

benjamin
Site Admin
Posts: 280
Joined: 15 Dec 2015, 08:38
Location: Sweden
Contact:

Re: Measuring pole count in BLDC tool [probably solved]

Postby benjamin » 10 May 2017, 19:36

tacho counts per revolution is probably the best way to measure the number of pole pairs. For every electrical revolution the tacho makes 6 counts, so the number of poles should be the tacho counts per revolution divided by 3.

Note that the 2.18 firmware has a small bug in the counting algorithm in FOC mode, so it will counts slightly too few counts. This is fixed in the new firmware.

I'm thinking of implementing a "pole pair counting" terminal command as well, which simply feeds the motor with a constant current vector. Then you can turn the motor by hand and see at how many distinct positions it gets stuck per revolution.

hexakopter
Posts: 86
Joined: 24 Dec 2015, 15:24
Location: Germany

Re: Measuring pole count in BLDC tool [probably solved]

Postby hexakopter » 10 May 2017, 19:47

Ahh, that would explain why I was counting just 137 instead of 138 tacho counts per revolution. I was using FW2.18 in FOC mode.

Great that it is fixed in the new firmware for VESC 6.0.


Return to “General”

Who is online

Users browsing this forum: No registered users and 1 guest