If you can find somebody to rewind the motor for you, it should be possible to run the motor/diff at lower voltages, where the Aussie controller will certainly be able to handle it. There was a bit of discussion in another thread (here http://www.diyelectriccar.com/forum...ar.com/forums/showthread.php/random-motor-questions-feel-free-join-42675p8.html) and one thing that did surface was an unwillingness of rewinders to do these jobs. If you can find somebody willing to rewind the motor to a lower voltage, there are some guys here who would like to know, so please do not forget about them

Regards
Dawid

 

They are:
-50kw(68hp) from 4,610-5,120 rpm

And they cost sub $1k from Car-part.com all day long.

Do you have any more info? Maybe a part number to search for? Picture?

I feel like this might be a great find. Especially with a new HV AC motor controller available. Thanks for sharing

($1k puts it within my "buy it and try it" range, if a guinea pig is still needed)

 

OK-Go to http://Car-part.com
Enter the following in the appropriate slots:
-Year: 2007
-Vehicle: Toyota Highlander
-Part: engine
-Leave it on: "All Areas/Select an Area"
-Sort it by whatever method you prefer (distance, price, etc...)
-Enter your postal code
-Press enter.

This brings you to the engine selection page. Click on the one that says:
"Hybrid, elec, Rr (4WD) "

This should give you an ocean of results. I just did this search and found them as low as $450.

 

It would be sooooo easy to mount that.

 

Here is where I got most of my information. The rear diff motor is called the "Motor Generator Rear" (MGR)


http://pressroom.toyota.com/pr/tms/toyota/document/2010_Highlander_Hybrid_Product_Info.pdf

Some other random googling found me the voltage range the motor typically saw.

 

At 5k RPM and the given diff ratio it would spin the wheel at ~730 RPM.
This is OK, if you have huuuge wheel like the Highlander has, but for normal car conversions it's a bit too slow. (I'm living in Germany with non speedlimit highways )
I also expect the motor rev's much less than 5k at only 350V...

On Car parts are also some Lexus diffs to buy. I wonder if their diff ratio is different?!
But definitely a nice find
-Olaf

 

I also expect the motor rev's much less than 5k at only 350V...

-Olaf

The tech specs from the 2010 highlander PDF linked to above states a nominal voltage of 288v of NiMh

Edit...

You're absolutely right, the nominal voltage is 288v but there is a "inverter/converter" unit that bumps it up to 650v for the motors.

 

Hmmmm, some as low as $450 with only 21k miles...

Seems powerful enough for a small direct drive vehicle, where the high drive ratio would be an advantage. Might make for a good "starter EV" if you can get the right controller or make a pack with higher voltage.

Controllers listed under "Power Inverter." I think.

 

Hmmm, That is a pretty cheap 100kW power section for a 3 phase motor. Perhaps it can be hijacked and given its own control board. Then try stuffing the power from the larger controller into the smaller rear drive assembly. You'll know its turned up to much when something breaks

 

Yeah, now that you mention it, it is one helluva bargain.

But, it's also a $1000 brick waiting for me to apply power to the wrong pin. At least as an OEM module it's guaranteed to be fairly tough. Full EMF shielding and all that jazz.

One could potentially be used to power both a front, and rear mounted "rear module" in an AWD car.

Anyone have any idea about what signals it is likely to require/expect? Likely to be all CAN-based?

 

What is the maximum rpm of the rear motor? I'm pretty sure it quite a bit higher than 5120 rpm. How fast does the power decline above 5120 rpm? I know that with series wound motors power falls off slow enough to have useable power well above peak power.

 

The hook with synchron motors is: they have a fixed kV. You'll need the 650V to make it spin 5120rpm and more volt to spin faster. Permanent mag motors usually have lower max RPM ratings than asynchron motors due to centrifugal forces on the magnets. I expect a physical max of 6k rpm for this one. ( wild guessing tho )

You also must be very careful to overrate PM motors, you can easily overheat the magnets
-Olaf

 

http://www.greencarcongress.com/2005/08/rx_400h_heading.html

That article claims 10,752 rpm max for the rear motor.

 

Specs seem to be

Electric rear motor Power (kW@rpm) 50@4610-5120
Torque (Nm@rpm) 130@0-610
Maximum (rpm) 10,752
Maximum voltage (v) 650


Two sources mention (2) Rx400h and a 2010 Highlander hybrid



from:
http://www.reviewcentre.com/cars430.html

http://pressroom.toyota.com/pr/tms/toyota/document/2010_Highlander_Hybrid_Product_Info.pdf
http://www.greencarcongress.com/2005/08/rx_400h_heading.html


Question is, can anything be done with it?

 

Tritium is Australia has a new AC controller that should handle this. Downside is it only go to 450V, so I guess you lose about a third of your power if you cannot get it rewound for a lower voltage. If you know a friendly rewinder, it will be worth your while.

Dawid

 

A 6000 $AUD, 150kw capable controller for use with a 50kw 1000$USD motor??

 

Price and value are not the same thing

 

Motor is 450 for less than 60 miles away.

However, I cannot justify that controller. Also, a boost converter might cost a bit too, however, if I did consider the AC option, id probably just build a 450v pack

 

At least it's almost impossible to kill the inverter with such a 'small' load. At peak load it wouldn't draw more than 80A.
Someone should buy one and look inside. It's not easy to disassemble a PM-motor, but maybe he would only need to rewire the coils from serial to parallel or from wye to delta. That would raise kV a lot ( same RPM with less V )
If it's not possible, resell it again
-Olaf

 

I'm new to this so someone should check my logic, but here's how I see it:

For BLDC motors (that's what this is, right?), if the controller can provide enough power, they will make constant torque up to max rpm, so that's where max power will occur.

If one were to use the tritium controller and the motor revs to 10752 at 650 V, then at 450 V, it would rev to 450/650*10752 = 7543 rpm. With 6.86:1 diff and typical tire size, that's 75-80 mph, so more than enough.

Ignoring controller/motor loss, 50kW @ 4610 rpm (3279V) requires 179A because 50,000W/(650V*4610rpm/10752rpm) = 179A. My guess is this is peak power because the batter pack is only 288V so above this, full current can't be supplied... or something. Also, the pack is only rated for 45kW, so it makes sense that it can't keep supplying current beyond about this voltage.

For a sanity check, 50,000W/(4610rpm*2pi/60) = 104N*m (76lb-ft), so pretty close to constant torque all the way from max of 96lb-ft at 0 rpm.

With 180A up to 450V, the motor's power should increase to about 80kW without too much more stress than its OEM application, right? Also, the tritium controller only supplies 300A max so it's not a complete waste, but not a perfect match either. Plus I imagine the motor could be pushed beyond 180A for short periods.


Even if this is correct, this motor/controller/diff combo would only work for a very small car. Max wheel torque would be 96lb-ft*6.86 = 659lb-ft. A Subaru Justy 2wd in 1st gear with the smallest engine available puts up to 804lb-ft to the wheels ignoring losses: 59lb-ft*3.071*4.437 = 804lb-ft. A typical compact car, say a mid 90's impreza, sends to the wheels 110lb-ft*3.785*4.11 = 1711lb-ft.

I don't know what I've decided from all this, maybe just use the diff with a better motor since the size and ratio seem well suited to EMs.

I seem to be on a different page than a few others here, so feedback would be appreciated.

 

Bringing this back to the top... Based on Corbin's thread, I'm optimistic that quite a bit less than max torque will be enough to get up a 5-6% grade as long as the car isn't too heavy. I would also worry about steeper hills in the city since there would be no way to gear down (I live in seattle), but I figure it can probably take a bit more current for short periods than toyota rates it for.

So the point is, I'd be ready to buy this motor/diff and the tritium controller if I had some reasurrance that these figure were in the ballpark. Also, the tritium controller requires 3 hall effect position sensors, could I expect these to be built into the motor? If not can I add them myself? Is there a chance of a 100k NTC thermistor being in the motor somewhere or is there some place I could add one that would give a reasonable estimate of when the motor is about to fry?

Finally, if these questions make it obvious that I shouldn't dive in, somebody please tell me that too. I'm hoping that the tritium line is plug and play enough for me to get going then learn the details and customize things as I go along.

 

Googling around on the Highlander MGR, I found a picture of one opened. Look near the bottom of this page:
http://www.vibratesoftware.com/html_help/html/Diagnosis/Reference/CVT_Transmissions.htm

 

Cool You think they could have put more copper in there? And do you know if it is liquid cooled?

Thanks,

major

 

So I actually bought one. It's not liquid cooled, but there is a drain plug that leads to the compartment in those pictures, so I suspect there's oil in there which would greatly improve heat transfer to the casing. The is also suggested by the open bearing. I guess I could pull the plug and look, but I don't see any reason just yet.

The main sticking point so far, aside from unknown torque capability and Kv, is that there are no hall effect position sensors, instead a resolver. James at Tritium has been very easy to work with and has suggested that if I can get him the operating frequency and primary:secondary turn ratio for the resolver he would buy one and build this capability into the next revision of their controller. Sounds like work on this would start in a month or two. I'm planning to use an audio frequency generator for my computer and an oscilloscope tomorrow or this weekend to sort out these parameters. Any suggestions for how to do this would be greatly appreciated. I have no experience with electronics, but an EE here at work said this strategy would more or less work.

The backup plan is taking the thing apart, removing the resolver, and shipping it to Australia.

Does anyone know an engineer that works for Toyota?

 

So I actually bought one.

Double cool

Without the liquid cooling, I wonder how useful it would be for an EV application. But it looks like a really solid machine

The main sticking point so far, aside from unknown torque capability and Kv, is that there are no hall effect position sensors, instead a resolver.

I know they make a reslover to encoder converter. Don't know if that would be of any use. I thought this would be a PMSM. In which case, I am not sure the converter gives you the Z marker. I'll see if I can get the info on that.

BTW, I have run BLDC machines in V/f mode open loop. Not saying this would be worth a crap for a traction motor. But if you put an inverter to it at a set frequency, it rotates smoothly, you can pull a wave form off the resolver and, I think, pretty much determine its characteristics.

Yeah, a long shot, but WTH.

major

 

I'm just hoping I get to find out about the cooling issue. If it becomes a problem, I think I can find solutions, maybe welding up more fins or coolant pipes between the fins.

Isn't this a PMSM? It has permanent magnets on the rotor that are supposed to be in synch with the field... I guess I don't totally know the difference between that and BLDC.

Anyway, I don't think I even have to rotate the motor much to characterize the resolver. My understanding is that a resolver is basically a transformer with 2 secondary coils at 90deg and a rotating primary coil. Thus, 10khz should give plenty of resolution even at 10k rpm which would be 60 small peaks per broad peak (broad peak = 1 rotation), I just need to make sure that this isn't too low a frequency, which would require too much current to induce any voltage in the secondary, thus frying the thing, or that it isn't too high which would create too much impedance to create the signal. I can then turn the motor by hand and look for the peak voltage in the secondary to figure out the turn ratio. I am planning to watch the input signal and one of the output signals with the oscilloscope, would there be a need to watch the output signals simultaneously?

Not sure how I will measure current, but basically I'm figuring if I can get a clear signal at 10kHz with the headphone jack, then that's what I'll call the operating frequency.

 

I can then turn the motor by hand and look for the peak voltage in the secondary to figure out the turn ratio. I am planning to watch the input signal and one of the output signals with the oscilloscope, would there be a need to watch the output signals simultaneously?

O.K. Where do you get position?

I'll tell you. I got no idea what Toyota uses for a control. But if there are PMs on the rotor, I think they need position. But maybe for a rear end boost, maybe they don't.

Don't get me wrong, I am just trying to learn about resolvers. I've been like dancing around them for years. And using encoders. I figured one of these days, I would have to deal with the buggers.

major

 

Sorry, that was just to get a few characteristics of the resolver to send to Tritium to help them work the capability into their next controller.

The resolver does indeed give position as a function of the ratio of the two output voltages, which are in quadrature due to their mount position. I assume that the controller will, as part of its setup run, send a signal to the resolver while sending a slow drive signal to the motor. By watching the output signal of the resolver, it would determine whatever it needs to about the motor. I think it does the same thing with hall sensors in its current configuration, just doesn't have the capability to drive a resolver.

I hope that it doesn't matter that the resolver doesn't sense magnet position directly and that as long as they know the angle of the rotor they can figure all they need to with their setup routine.

 

I hope that it doesn't matter that the resolver doesn't sense magnet position directly and that as long as they know the angle of the rotor they can figure all they need to with their setup routine.

Yeah,

I hope so too

major

 

If the motor portion of the case actually carries fluid, it wouldn't be hard to add a small cooling loop. A couple of fittings in the housing, some hoses, a small radiator, and a pump and you would be set.

It looks like there is a little ATF colored fluid sitting in the center of the the housing in the far right pic.

 

Thank you for the suggestion, that would be much more effective, I'll look into it when the time comes. The controller is also water cooled and it doesn't seem atf would hurt it, so maybe I can put them in the same loop.

There's also a thermistor inbedded in the windings, so I'll figure out the type and constants at some point so I should be able to monitor temperature pretty well and figure out if I need to do anything at all.

FYI if you save those pictures then open them they have pretty high resolution... which is maybe obvious but just thought I'd point it out. Looks like 8 poles, maybe there's other info there. I have the pinout too if anyone wants it, although it's not very useful without thermistor and resolver parameters.

 

I finally got an Rx400H MGR down under here Ryan, not sure I've got the necessary smarts or pocket depth to do much with it but I got it past AQUIS which is an achievement in itself!

 

Nice! I haven't written much about my project yet since I don't have much substance to report at this point, but controller, batteries, and charger should be here in a few weeks, so hopefully progress will be a little quicker after that.

I'll definitely post when I have an idea of the torque, speed, and heat dissipation capability of the motor.

Do you have a car in mind? My saturn sw came in a bit under 1900 lbs/ 860kg with the IC components out, so I think with the electric stuff and 6.5kwh of headways it will be around 2200 lbs. I think the motor will handle that fine, but I have some concerns about how many more batteries it will lug around.

-Ryan

 

I've been sniffing around 1998 model toyota Celicas Ryan, ready availability in the used car market & under 1000kg as an EV. There are some smaller alternatives but sharing the road with 2 tonne SUVs in a modified suzuki cappuccino scares the daylights out of me.
Sounds like I'm 6 to 12 months behind you progress wise, but it's good to have a pathfinder.
I can't help dreaming about what it would be like to have 2 of these puppies in a vehicle for a nice AWD EV but that's getting seriously ahead of myself.
Hope your conversion goes well. It's a big project, hope I've got the where with all to pull it off.