DIY Electric Car Forums banner

1 - 8 of 8 Posts

·
Registered
Joined
·
42 Posts
Discussion Starter #1 (Edited)
How to change Motor perceived direction on Sevcon

Hi All,
Hopefully someone can chime in.
I Converted from a Kelly to a Sevcon (ME0913 Motor) and was having all sorts of issues but have finally figured out that my unexplainable runaway acceleration and over current fault issues are caused by my encoder direction and motor direction are inverted. That is, every time I tried to accelerate to 100rpm the 3 hall sensors would report the motor was moving at -100rpm and then continue to accelerate to -5000rpm in order to try and reach 100rpm. Obviously it would never get there.

I figured I need to swap M1 and M2 on the motor to reverse the motor direction but that only caused the motor to shake as it must now be misaligned with the encoder. So.

My question:
Anyone think that if I switch hall B and Hall C I will reverse the reported RPM direction and not change the Motor direction?
Is there a bit in Sevcon I can flip with DVT that can just invert the direction reported?

Any help would be huge.
Thanks all.
 

·
Registered
Joined
·
42 Posts
Discussion Starter #4
OK, I used the following approach to solving the issue.
I am still not 100% sure exactly how I fixed it but this is what I think I did.
I swapped encoder lines and ran a trace from the help menu.
I did this until I found that When I rotated the wheel the forward direction the RPM readout was positive.
I then found the motor would not spin any more so I then swapped two motor phases and it allowed the motor to rotate.
I am not sure how but also somewhere in there I changed the encoder from UVW to UVW inverted only.

I honestly don't know what that did. (invert the square wave?)

Either way. At this point the RPMs are now positive when moving forward and neutral braking works.

Foot braking still doesn't work along with my foot brake analog sensor.
But atleast now it is not all cattywompus.
 

·
Registered
Joined
·
1,469 Posts
The motor has 3 phases and there are 3 hall sensors attached internally with the windings to measure magnet position as the shaft rotates and used for motor commutation. Unless there was an error at the factory it should not be necessary to mess with the hall sensors to get the motor to spin--that timing is set and fixed with epoxy.

An encoder could be added to either the motor or the drivetrain shaft to measure shaft angular position, to sense rotational direction, and to calculate shaft speed.

It's not clear where your encoder is mounted, and i'm not familiar with your controller, but it sounds like you solved the problem with a little bit of trial and error, and didn't let out the magic smoke...
 

·
Registered
Joined
·
42 Posts
Discussion Starter #6
Yup, no magic smoke. Took it out for an extended spin today and everything seemed great. I agree about the hall sensors. I have no encoder. They do make encoder versions of my motor but I have the hall sensor version. My guess is that when I converted from Kelly I looked and the kelly wiring guide says to swap Hall lines, so UVW are matched with Hall sensors ACB respectively.
I think sevcon wanted UVW to ABC. So when I rewired it was unhappy.

Happy Now :)
 

·
Registered
Joined
·
1,469 Posts
How to reverse direction with Hall Sensor Commutation

For motors with 120 electrical degree Hall sensor spacing, if the physical wiring between a 3-phase motor and the drive controller with Hall commutation feedback needs to change, in order to achieve a specific motor direction for a given input command, then you only need to

(1) Flip the two outside motor phase connections (phase: A and C or U and W) at the controller.

(2) Switch the top two associated Hall connections (Ha and Hb, Hu and Hv, or S1 and S2) at the controller.


For a motor with 60 degree Hall spacing, then it can be converted to 120 degree spacing by inverting the polarity of the output voltage of the S3 or W sensor (U-V-W inverted).


This convention uses a positive hall voltage to correspond to a positive back emf voltage in the phase winding, for the desired direction of rotation.

Another NEMA motor convention defines motor windings as "U-V-W" for positive direction in clockwise direction viewed from the shaft end. Motors using "A-B-C" windings have positive for CW viewed from the hall sensor wire exit (opposite) end.

Looking at the ME0913 drawing, it is labelled as U-V-W but using CCW as the positive direction--so it is not following the NEMA convention. Your controller was likely made for an U-V-W motor and the Motenergy motor is wound as an "A-B-C", and this was the root of the problem.


Just a little fyi, the use of the word "encoder" (even in the title) was really confusing and a distraction of the actual issue at hand, i.e. the phasing of the hall sensors.
 
1 - 8 of 8 Posts
Top