Can you parallel 3 phase motor controllers together?

[zombiess]

Just thought about this. What happens if you parallel the phase wires of two identical controllers together with identical settings, same throttle (easy as hell to split a throttle to multi controllers). Would it work, why not? If they each had their own separate battery pack it would fix and shared ground issues would it not? Could they be run from a single pack?

Two or three 150V 18 FET controllers in parallel would be a nice packaging arrangement for a delta mode modded motor. 36 FET's. Not sure if these controllers use an external clock source, but if so it's not that difficult to build a single clock source to link the two into sync (done it with other micro controllers) if having PWM out of sync is much of an issue, if not, bonus. Two 150V 18 FET controllers each with their own 125V 10AH packs sure would pack a wallop and make for a fun ride and be pretty cost effective, 125V@120 battery amps/360 phase amps if it works. There will also be illegal speed runs on the highway if this is possible.

What kind of controllers are out there for 150-300VDC? I was thinking something that could burst around 200A. They seem to get pretty pricey at this range

yes I'm totally serious. I did just take a large amount of a prescription substance so I'm planning my own demise and kicking around some ideas. Setting a record is a happy thought right now. Should to sign off, getting double vision but feel awesome.

Will those on here with a fuller understanding of these controllers please step up and tell me what could go wrong and the challenges involved . This setup would probably go to Bonneville and use active cooling on the motor. I'd like to burst around 80+ KW into the motor I'm getting.

I know most people on this forum don't know me well, but this is the crazy shit I come up with when I start to think outside to dodecahedron. I'm too high on Carisoprodal (I'm such a lightweight due to never taking drugs before) at the moment to think of the failure mode if any (body diode cond. I've given up, frock money frock everything, I have some new goals and they involve ludicrous speed. Might as well use my high IQ for something beside looking at porn and pulling my pud.

 

[Philistine]

Might as well use my high IQ for something beside looking at porn and pulling my pud.

Don't think of it as a zero sum game in that way, pull the pud, surf porn, and think about other things. You really can have it all.

Can't help you on your question, but I really wish I had some of what you have rolling, sounds sweet......

 

[amberwolf]

I am not certain, but I think that differences in controller timing for PWM/etc will turn one or more of them into plasma as they create shoot-thru from upper FETs in one controller to lower FETs in another, as they drift in and out of sync.

Even if they're all the same model, sequential "serial numbers" from the same batch, timing from each is likely to vary some, as will rise/fall times of gate drivers, etc. The latter probably won't matter as much, possibly at all, but the former would matter a whole lot.

The only way to guaranteed reliably do it is to cut out the MCU section from all but one, (not off the board, but electrically isolate it), and run all of them from one controller's MCU section.

 

[bobc]

quick answer no
to "hard parallel" inverters requires synchronisation to a few 10s of ns. Softer paralleling with inductors sharing the load would still benefit from synchronisation & the inductors would have to be massive.
A possibility is to ditch the star/delta connection and have an inverter on each end of each motor winding (6 wires coming out of the motor). This uses the motor's own inductance to cope with being unsynchronised, but because of the magic of 3 phase you don't get a doubling of volts - so no-one has ever bothered....

 

[zombiess]

quick answer no
to "hard parallel" inverters requires synchronisation to a few 10s of ns. Softer paralleling with inductors sharing the load would still benefit from synchronisation & the inductors would have to be massive.
A possibility is to ditch the star/delta connection and have an inverter on each end of each motor winding (6 wires coming out of the motor). This uses the motor's own inductance to cope with being unsynchronised, but because of the magic of 3 phase you don't get a doubling of volts - so no-one has ever bothered....

Figured the pulse sync would be the issue. Don't wanna mess around with external inductors, just messy. I'm not star delta switching, just the standard delta mode, it was just an idea to give it extra current without getting a crazy expensive controller.

 

[John in CR]

There's a bunch of us who wished that would work. I'd parallel 4 of the $30 controllers tomorrow for a quick and easy 12kw controller for $120.

EDIT-
Something that I've speculated might work is to separate the windings at both ends of each phase to separate the actual magnet wire strands and make parallel groups for each phase with each group electrically separate. Then run a separate controller for each group 3 phase group. They would all run off the same set of halls, so the commutation would be virtually identical, but the sticking point is the high likelihood that the resulting parallel coils wrapped around each tooth would interfere with each other too much and cause problems since the PWM pulses are highly unlikely to be exactly in sync.

If your stator slot count divided by 3 isn't a prime number, then you could rewind the motor for 6 or more phases. That would not only split the controller load, but it would also give you smoother operation. In the motor section there's a guy how built a motor and ran it with something like 7 or more RC controllers. He said it ran much better that way than wired for just 3 phases and powered by a big Kelly. He was running a motorcycle with that motor, and was selling instructions for exactly how to build the motor. Another advantage is that he was able to run very cheap brushless controllers, but still have good startup torque. My problem is that my slot count is 51 for 17 for each phase, a prime number, which means I would have to use 17 controllers to pull it off. Actively ventilated and 17 tiny RC controllers inside the motor could be really cool, but that kind of project is a bit too involved and tedious for my taste. Maybe you're game for being the first to do it with a hubbie.

John