You are here

VESC 6 Hall Sensor Issues

12 posts / 0 new
Last post
josh
josh's picture
Offline
Last seen: 4 months 1 week ago
VESC Original
Joined: 2017-05-24 12:15
Posts: 92
VESC 6 Hall Sensor Issues

Hello, everyone,

I recently got some new "sealed" outrunner motors for my setup and noticed some strange things happening with the sensored setup. During the sensored RPM range, the motors run very noisy and when riding, there is a noticeable vibration until sensorless rpm is reached. I don't recall having this with the previous motors (Alien Power System) but I don't see anything wrong with the design of these new motors. 

Analyzing the VESC signals, there is a lot more noise in the phase currents during the sensored operation. Also, feeling the torque of the motor in my hand, it is not a constant smooth torque but is instead choppy and segmented. There seem to be some dead zones based on the rotation position. If you guys have any idea what is causing this effect, I would love to know and try to fix it. See images below for captured data:

 

Noise Capture
Notice the transition between sensored and sensorless operation.

Holding Wheel Noise
Captured while holding wheel during uneven torque

Noise Capture Detailed
A higher resolution capture of the transition

Let me know your thoughts.

Cheers!

frank
Offline
Last seen: 1 week 23 hours ago
VESC BronzeVESC FreeVESC GoldVESC OriginalVESC PlatinumVESC Silver
Joined: 2016-12-27 20:19
Posts: 847
Same behavior in FOC or BLDC? Foc should work a lot better with sensors. What motor do you use? Frank
josh
josh's picture
Offline
Last seen: 4 months 1 week ago
VESC Original
Joined: 2017-05-24 12:15
Posts: 92

Thanks for the reply frank. I have not tried BLDC, only FOC. The motors are just like these APS motors but with hall sensors: http://alienpowersystem.com/shop/brushless-motors/alien-6355-hev-outrunn... I purchased some samples from the manufacturer directly. 

The only thing I can think of is if the hall sensors are not getting a good read on the magnets. Maybe the magnets aren't tall enough and don't pass right over the hall sensors. Is there any way you can see the hall sensor data in real time through VESC Tool?

Thanks!

josh
josh's picture
Offline
Last seen: 4 months 1 week ago
VESC Original
Joined: 2017-05-24 12:15
Posts: 92

I tried configuring the motors in BLDC mode and when I ran the Hall Detect, it returned an error:255. Not sure what this means. I can still run the hall config on FOC with good result received. Any thoughts?

Thanks.

josh
josh's picture
Offline
Last seen: 4 months 1 week ago
VESC Original
Joined: 2017-05-24 12:15
Posts: 92

I have found out some things with this issue. To start, the hall sensors on the new motors don't seem to have a pull-up resistor on them based on the waveform in the scope. I assume the VESC has some pull up resistors on the port since the addition of resistors on the motor made no difference while running on the VESC. See images below:


Hall effect signal of problematic motor during hand spin


Hall effect signal from good motor during hand spin


Test Setup

Additionally, the motors were tested with the rotor bell removed and it was found that the good motor had latching hall effect sensors and the bad motor did not. i.e. the latching would only change state if showed an opposite magnetic field, non-latching would change state if a north face was brought near it and again if it was removed. This leads me to believe most BLDC motors use latching hall sensors for a more accurate reading? I will be taking more measurements of the motors on the scope while running on the VESC tonight.

 

josh
josh's picture
Offline
Last seen: 4 months 1 week ago
VESC Original
Joined: 2017-05-24 12:15
Posts: 92

I ran some tests with the motors running on the VESC. I connected the oscilloscope to the ground and the three hall wires and here are the results:

 


Single Hall sensor during motor operation

 The picture above shows the single hall sensor signal. Notice the widths of the high and low pulses. 


All three sensors during motor operation

The picture above shows the signals from all three hall sensors. Since the all have a longer on time than off time, the state changes don't line up properly. Each state change should be equally spaced, but notice how close they are near the cursor and how far they are near the center of the image.


Signal from a non-problematic motor

The image above shows the signal from a motor that runs smoothly during sensored mode. Notice how the high and low widths are the same and the state changes are equally spaced. 

 

Since the problematic motors used non-latching hall effects, they were staying active longer than intended. The misaligned timings made the motors run very inefficiently and with the noise and vibration I mentioned earlier. I have just put an order in for some Honeywell latching halls and should have them in a few days. I will replace the ones in the motors with these and report my findings. I hope this thread will be useful to someone else having a similar issue.

Cheers!

ThierryGTLTS
Offline
Last seen: 6 years 4 weeks ago
Joined: 2017-09-06 14:18
Posts: 116

Hi,

I have glued 3 Honeywell latching halls on my TMotor U8 and they work well, better than the ones on my Maxon EC 90 Flat.

But perhaps they are closer to the magnets.

Have a Nice Day.

Thierry

protontim
Offline
Last seen: 6 years 10 months ago
Joined: 2017-09-25 23:31
Posts: 19

Thierry,

Could you show a picture of how you did that?  I am very curious.

Thanks,

Tim

 

ThierryGTLTS
Offline
Last seen: 6 years 4 weeks ago
Joined: 2017-09-06 14:18
Posts: 116

Send me your e-mail address by PM, please.

I will send you a photo by e-mail.

How do you do such nice graphs (TMotor Flame on other tread)?!

Have a Nice Day.

Thierry

josh
josh's picture
Offline
Last seen: 4 months 1 week ago
VESC Original
Joined: 2017-05-24 12:15
Posts: 92

Hello Thierry,

Sorry I did not see the replies until now. The graphs in the VESC tool was done in the sampled data tab under data analysis. The other ones with the black background were done on an oscilloscope with several probes. In order to get a nice capture, I used a trigger on one of the hall signals.

Cheers!

Nickelberry
Offline
Last seen: 5 years 10 months ago
Joined: 2017-11-24 11:39
Posts: 1

Hey Josh, thanks for posting all this info I have found it very useful, it looks like I have the same motor with the same problems, did changing out the Hall sensors fix the problem? if so can you share the model number of the sensors you used? 

Kent Tabor
Offline
Last seen: 1 month 1 week ago
Joined: 2020-07-02 17:51
Posts: 7

Did this ever work out for you?

Kent