Toyota MGR Rear Traction Transaxle

[james@N52E01]

I was unclear. If 54750 gives you reverse direction when selecting forward, subtract 32768 and you end up with 21982. This should then give you forward rotation when selecting forward. Nothing strange going on.
To my understanding 0-65535 corresponds to 0-360 degrees so adding or subtracting 32768 is equal to changing syncofs by 180 degrees i.e. change of rotation direction of the motor.

That part makes sense, when I try with 21982 though it acts completely different to 54750 (very jerky, albeit in the correct direction now). I’m not sure what could be causing this but as this worked for you both, there must be something else causing this issue. Will have another look at the resolver and try and post some footage next week

[ZooKeeper]

One (of the many) error(s) I made was to connect the throttle BEFORE getting the syncoff setting dialed in. By messing about with the throttle, I had managed to convince myself I had an idea of what would work or not.

Damien's video dispelled any feelings of "I figured this out myself", while I was close in terms of apparent function, I was also incorrect.

[james@N52E01]

Maybe that’s it. Will go back to square one and go through with manualID again from scratch. Might not try increments of 100 though for sanity’s sake

[mdrobnak]

360 * 100 / 65535 = 0.5493247882810712 degrees.
Maybe 5 degrees at a time? so 1000?

-Matt

[ZooKeeper]

Maybe that’s it. Will go back to square one and go through with manualID again from scratch. Might not try increments of 100 though for sanity’s sake

Yea, I would not recommend that to others

One thing I know for sure, both from the great video and my own testing... when the spin changes directions, you crossed OVER the valley to the "other side" and need to work back towards it.

[519 Kustomz]

Hello again all! Finally have some time to revisit this. I have the motor spinning reasonably well but there is some jumping on startup only in forward direction. I decided to try to make my own power stage, so I have made a gate driver and voltage sense board the plugs directly into the version 3 main board. Makes things a lot cleaner! https://youtu.be/0mNpBZH6M0o

[519 Kustomz]

My question is regarding id and iq. When I plot them I was expecting to see iq fluctuate with throttle input and id stay relatively lower. Seems to be backwards? Maybe I’m just backwards?? Any insight is appreciated!

[johu]

Is it just spinning freely? Then not much torque needs to be generated and you spin right up into the field weakening range. id=field weakening.

EDIT: thats a nice gate drive board in the video, do you have some details?

[519 Kustomz]

Yeah absolutely! Here is a couple pics of the gate driver board. It has the sense connections on it as well and is designed to stack your main board. Also included a 5v / ground tap I use for the throttle pedal.

[AdvancedEnergySys]

Got some bad news on the rewiring front, the MGR consists of 48 poles meaning 16poles per phase and 4s4p configuration. I've designated the half dozen transaxles I currently have as engineering samples so if anyone has any recommendations for rewiring I would be happy to hack at the windings without concern for function afterward. My hope is that I'll find one of the new style MG1 stators from a prius that will fit inside the MGR housing as the diameter is the same as at least one generation.

[180jacob]

if anyone has any recommendations for rewiring I would be happy to hack at the windings without concern for function afterward.

Bear in mind I haven't opened mine up yet, this is just based on another photo earlier in the thread, but I would suggest trying to find the three crimps I have arrowed in the picture below. These appear to form a star point so you should be able to cut the crimps and reconnect the resulting 24 ends into a delta for running at higher speeds with lower voltages.

Personally I am holding off on opening up mine for a couple of reasons, I haven't found motors in general to be too fussy about voltage (within reason, maybe this will be an exception?), I haven't got a car to put it in or built a suitable test rig yet, and Damien appears to have got his subframe mounted back in the blue yacht so I am happy to wait and see what happens there. viewtopic.php?p=23617#p23617

[ZooKeeper]

Personally I am holding off on opening up mine for a couple of reasons, I haven't found motors in general to be too fussy about voltage (within reason, maybe this will be an exception?), I haven't got a car to put it in or built a suitable test rig yet, and Damien appears to have got his subframe mounted back in the blue yacht so I am happy to wait and see what happens there.

Similar thoughts here.
1) If the 450H can run 70mph on <500v, why not the MGR? Clearly I am not an EE and do not play one on the net, but....
2) I do not need high speed, 50 is fine.
3) It seems like the Gen3 Toyota stuff can be CAN controlled, if so, maybe the Gen2 as well. If the inverter simply responds to Tq request, BINGO!
4) I am fairly certain the MGR resolver will not be plug-n-play with a Prius inverter, but am willing to try.

[tom3141]

Bit of an update on the MGR that I have in my Suzuki Cappuccino project for those who haven’t seen my project thread.

I’m now permanently boosting the voltage to just over 500V. I’ve sorted a few little issues that were preventing me from testing properly so it’s now fully driveable on the road.

Unfortunately, I still have the same performance issues that Bexander has also had in his MGR powered Clio. Acceleration is good up to around 30mph then drops off dramatically. Top speed on the flat is around 55mph. On hills I can maintain a maximum of 30mph at full throttle. I’ve never been able to put more than about 20kW into the motor. Max DC current I’ve seen at full throttle uphill is around 70A (at ~300V battery voltage). The car weighs around 850kg.

[johu]

Interesting, 300*70=21kW, the stated limit of the boost converter.
You said a variable boost rate did not improve performance?

Speaking of rewiring, star to delta would yield 1.7 times less voltage requirement, 4s4p to 2s8p 2x less and both combined... well.

[tom3141]

I suppose it is possible that originally the limit was the voltage but now it is the boost converter power and the symptoms for both are exactly the same.

Variable boost rate didn’t seem to make a difference although I could do some more thorough testing now that I have the other issues sorted. The performance between 0 and 30mph seemed pretty similar whether I have no boost or run at 500V. If anything it is better at 500v. Above 30mph I’d say performance is marginally better at 500v.

What I can’t really get my head around with these type of motors and FOC is what actually could be limiting the power? Going from 300V to 500V hasn’t made the difference I would have imagined if the problem was back EMF. I guess I just haven’t fully got my head around what’s going on with these compared to DC or AC induction motors.

[tom3141]

If it turns out that this is definitely the limit of the MGR then I’d be up for getting another MGR to rewind and test...

[johu]

All motors are the same in that they need double the voltage at double the speed. With the booster off you can evaluate peak torque at low speed, i.e. when unboosted battery voltage alone can sustain high motor currents. Just keep increasing throtcur but keep in mind you can burn out the motor if you push it too far.

[tom3141]

Did some more thorough testing this afternoon.

High load, full throttle runs from standstill to 30mph up an incline.

With booster at 530V (see plot below):
- throtcur at 3, AC motor current shows 300A
- max DC current of 160A according to my JLD404 (approx 50kw)
- strange behaviour of idc on the plot. Not sure why it does this?
- noticeable drop in torque as 30mph is approached
- DC current drops quickly back below 100A after peaking
- enough torque to spin the wheels when pulling away hard

Without booster:
- feels noticeable slower
- max dc current of 100A according to my JLD404
- appears to be no benefit to this

Full throttle run from standstill to approx 50mph up an incline:

I tried increasing and decreasing the throtcur to 4 and 2. Noticeable difference below 30mph but not above.

[bexander]

Great testing tom3141!

Yes, I experience a very similar behavior of the significant decrease in acceleration when going into field weakening region. Iq drops to 30% of max with little change in speed.
This is not what I expect but I only have a very little understanding of motor drives and FOC.
I expect roughly constant torque and increasing power up to FW (field weakening) and then close to constant power as speed increases and torque decreases. Maybe this is to much of a simplification?

Anyway the MGR is proably designed to provide high torque at standstill to very low speeds and then only beeing able withstand high speed operation as always connected to wheels but not providing much torque at all at this high speed.

The package of the MGR is very convinient but it seem to not be suitable to drive a vehicle up to road speeds.

[Ev8]

So with enough load you are hitting 50kw before passing peak torque. And you can spin the wheels, I think it’s all down to gearing, get some bigger wheels on there, it may be that a rewind is the next thing to try,

[ZooKeeper]

Speaking of rewiring, star to delta would yield 1.7 times less voltage requirement, 4s4p to 2s8p 2x less and both combined... well.

Guessing that would also mean ~2x the current capacity (trading volts for amps)??? I'd be ok do go down that path if performance w/o the booster is miserable *and* it is feasible to rewire to delta.

[MattsAwesomeStuff]

We knew it was little, just judging by the size and weight.

So, it might be suitable for an RWD hybrid to a FWD donor car if you wanted to keep the engine.

Or, just a lighter vehicle.

50kW ain't bad, if it can sustain that with some added cooling still.

Consensus being, it's just hard to get the kind of power out of it that is useful without a really high voltage pack? (Making it very poor for DIY hybriding, as, most DIYers have voltage limits due to the large size of smallest form factor of OEM battery packs).

[bexander]

Did some measurements and I don't think the poor performance during field weakening from the MGR is due to the motor itself but rather an inverter or tuning issue, more info here: viewtopic.php?f=7&t=1683
Will report any findings on this matter.

[ZooKeeper]

Matt, 50kw is about 67HP, an increase of more than 50% from what the majority of VW aircooled engines had from the factory and more than double for a few.

How much power one wants is all about expectations. If you want a vehicle that is performance oriented, even if it is light, 100kw is about the bottom of the range. If, OTOH, one simply wants comparable to factory performance, then <=100kw will get the job done on almost any light vehicle.

Of course the above presumes that you are not going to increase vehicle weight during the conversion, which is a tall order for sure.

[ZooKeeper]

Did some measurements and I don't think the poor performance during field weakening from the MGR is due to the motor itself but rather an inverter or tuning issue, ....

Thanks for all the work! We know from Toyota marketing materials and that the MGR is not intended to be a prime-mover, so I am not shocked that there is a performance decrease with speed. The American preference is for acceleration from a start, which is where the MGR provides good results.