# Two VESCs, two three-phase windings, one motor stator

14 posts / 0 new
kubark42
Offline
Last seen: 1 day 14 hours ago
Joined: 2020-07-17 18:46
Posts: 52
Two VESCs, two three-phase windings, one motor stator

I've got a motor which I could reconfigure to have two wholly separate three-phase windings, but on the same stator lamination stack. Instead of one three-phase winding with 6 teeth pairs per phase, each pair spaced 60 degrees from its neighbors, it would become two three-phase windings with 3 pairs each, the pairs spaced 120 degrees apart.

The application is for a torque-driven propeller, so control would be purely current.

Magnetically and mechanically I believe everything is sound. Electrically, is there any reason to imagine the VESC would have issues? If so, what kind of pathologies should I be on the lookout out for?

If it helps, I have hall sensors mounted on the motor.

TechAUmNu
Offline
Last seen: 3 days 13 hours ago
Joined: 2017-09-22 01:27
Posts: 564

That's a very interesting motor! Don't see why that wouldn't work if you are running in current control mode. Although probably some weird issues with startup. I would try with some cheapo vescs just in case something really weird happens.

skyline1970
Offline
Last seen: 2 days 23 hours ago
Joined: 2017-09-07 22:28
Posts: 170

just connect one stator side, and observe signal sync via a dual channel osilascope if any matching phase. matching phases could be paralled, otherwise it is a 6 phases motor.

kubark42
Offline
Last seen: 1 day 14 hours ago
Joined: 2020-07-17 18:46
Posts: 52

Here's the motor. I have highlighted the windings.

It's easy to see that it's done similarly to the below, with the only difference that the windings are connected in parallel instead of in series:

benjamin
Offline
Last seen: 56 min 21 sec ago
Joined: 2016-12-26 15:20
Posts: 476

If the windings sets are electrically separated from each other it should be fine to run them on different VESCs on the same stator. In the picture it looks like the 6 wires that come out are connected together with the three copper rings, which would be a problem.

Elwin
Offline
Last seen: 4 hours 53 min ago
Joined: 2021-09-30 16:41
Posts: 70

The motor is wired in delta and all coils go to the three rings. I think you can split in into two motors, by adding 3 extra rings. What I am unsure about is whether the mutual inductance will cause problems.

skyline1970
Offline
Last seen: 2 days 23 hours ago
Joined: 2017-09-07 22:28
Posts: 170

it could be drived by a unity controller as seperate phase windings (parallel fw recomended), or just one controller with paralled phase windings on the motor. just ensure phase sync .

kubark42
Offline
Last seen: 1 day 14 hours ago
Joined: 2020-07-17 18:46
Posts: 52

In the picture it looks like the 6 wires that come out are connected together with the three copper rings, which would be a problem.

The motor would undergo some light surgery, adding a second set of phase rings.

it could be drived by a unity controller as seperate phase windings (parallel fw recomended)

Are you saying that Unity has specific experience with this condition? Elwin and I are both concerned about the unknown impact of mutual inductance. It would be very helpful to know that Unity has found this is either not a problem, or is a problem they know how to solve.

kubark42
Offline
Last seen: 1 day 14 hours ago
Joined: 2020-07-17 18:46
Posts: 52

Hmm, no joy on the experiment. The windings were removed and the o-scope looks good, but it doesn't run. It just chatters agitatedly.

The VESC FOC settings page fails to detect the new R and L, but I can detect them with the command line measure_res and measure_ind. The measurements are double, which is to be expected for removing half the parallel windings.

However, both the UI and the CLI fail to detect the flux linkage. And yet, when connected to the o-scope we see that the new motor setup yields identical results to before, which we expect since, again, the teeth are in parallel.

So I'm a bit of a loss about what is going on here.

Thoughts?

skyline1970
Offline
Last seen: 2 days 23 hours ago
Joined: 2017-09-07 22:28
Posts: 170

ok, if u,v,w (A,B,C in the picture)  is short circuit or not. seem to be it is not Y connected. if so just connect end windings to make a Y connection, rpm will be half in that case.

you could make a virtual gnd to make a ref. point to measure phase sequences u,v,w. another one on the other stator. virtual gnd.s to be connected on  o-scope . then simply observe signals, while turning the motor with a motorized screw driver.

skyline1970
Offline
Last seen: 2 days 23 hours ago
Joined: 2017-09-07 22:28
Posts: 170

no connection on A,B,C ?

short circuit on A,A' ?

short circuit on B,B' ?

short circuit on C,C' ?

kubark42
Offline
Last seen: 1 day 14 hours ago
Joined: 2020-07-17 18:46
Posts: 52

Works on one phase! No idea what changed, I disassembled the system, took it home, upgraded to latest 5.3 beta, and as soon as I clicked the button to detect params, it all went well.

More tests to follow. Just need to get another VESC and see if 2x3 works, too!

kubark42
Offline
Last seen: 1 day 14 hours ago
Joined: 2020-07-17 18:46
Posts: 52

The motor starts up fine. But I'm not sure what to think when both the HobbyWing (duty-cycle control, trapezoidal waveform) and the VESC 100/250 (current control, FOC)  run simultaneously. Here's the video: https://vimeo.com/647759550

Description:

• First I run the ESCs individually
• Then I start the HobbyWing (non-FOC) and bring the VESC (FOC) online. Notice that the system barely speeds up, but the sound quality changes.
• Then I shut the HobbyWing off. Notice that the motor almost comes to a complete halt before the VESC kicks in and accelerates to a higher speed than before
• Finally, I run the VESC and then bring the HobbyWing online. The result is a brutal deceleration. It's hard to see in the video, but the sound is unmistakable!

TechAUMNu points out that this behavior might be perfectly normal, since the duty cycle control of the HobbyWing imposes a speed, and the current control from the VESC provides more torque but in the no-load situation the additional torque doesn't lead to much more additional speed.

skyline1970
Offline
Last seen: 2 days 23 hours ago
Joined: 2017-09-07 22:28
Posts: 170

could you find the matching phases? u, u',  v,v'  w,w'