DIY Electric Car Forums banner
1 - 20 of 134 Posts

·
Registered
Joined
·
91 Posts
Discussion Starter · #1 ·
Hello Everyone!

I am in the process of building an EV using the Highlander Hybrid rear differential(s). The goal will be to use these units for an AWD system.

There has been a previous thread about that here:
http://www.diyelectriccar.com/forums/showthread.php/any-use-highlander-rear-diff-motori-43920.html

I really admire Ryan800's initive in building his Saturn. I'm hoping to get more performance out of these motors by using them to/beyond their maximum capability. If something breaks, that's how we learn stuff. :p

Generally, I'm a person who likes to see physically what I'm dealing with, so attempting to run this thing as a "black box" doesn't work for me.

Here are the basic specs for the unit:
-50kw(68hp) from 4,610-5,120 rpm
-96 lb ft of torque from 0-610 rpm.
-Gear reduction of 6.86/1
-Weight: 94lb

Other than these basic specs, there isn't much information available for the Highlander Hybrid rear differential. However, there is some good information about the Prius. I think there is enough similarity between the second generation Prius traction motor and the HH motor that this information can be shared.

I plan to take apart one of these units as well as share other info I've found on them. If anyone has any questions or would like more details/measurements of stuff, I will be happy to post that as well. :)
 

·
Registered
Joined
·
91 Posts
Discussion Starter · #2 · (Edited)
To start out with, here are some exterior pics of the rear differential.
Not too much exiting stuff here.
Top right: Right side. That black thing to the right of the output shaft is a piece of cast-iron. It is static and removable; I suppose it's for vibration :confused: There are two connectors on the upper right side. The one on the left is a 6-pin connector for a resolver. The one on the right is a 3 pin connector for a temperature sensor. (I will go into more detail later)

Top left: Front
Second row, left: Left side of motor
Second row, right: Front of motor
3rd row: Top of motor
4th row: Bottom of motor
 

Attachments

·
Registered
Joined
·
91 Posts
Discussion Starter · #3 · (Edited)
In this set of pics, some easily removed parts and the right side cover are removed.

The first pic shows the 3 phase connector/extension. It's purpose appears to be an adaptor to the Toyota 3 phase wiring harness. The internal pins are the same as what's built into the motor.

The second pic shows the counterweight. It is a fairly roughly machined part that simply threads into the cover plate.

The third pic shows breaking open the right side cover plate. An ATF-like oil is draining out.
 

Attachments

·
Registered
Joined
·
91 Posts
Discussion Starter · #4 ·
Ok, here's the inside of the right cover plate. The sensor mounted to it is a resolver, that appears to have some clocking adjustment.

For those unfamiliar w/ resolvers, they are extremely accurate absolute position sensors. Basically they are rotary transformers, and work by being fed a sine wave, that is picked up by two coils positioned 90 degrees apart from each other. These coils feed back the sine and cosine of the current position. If you look closely at the rotor, you will see an oval shaped piece of metal (between the rotor and bearing) This piece varies the air gap, thus changing the signal at the output coils.

Generally resolvers are found in high accuracy servo applications, like CNC machines. They are very robust and can also function well in an electrically noisy environment. :cool:

My feeling is Toyota chose these for a number of reasons, and replacing them merely because I didn't know how they work would be possible, but both expensive and troublesome. :(

Their signals can be translated to digital quadrature signals as one option. There are IC's made by TI that can do this for about $12. Other than that, I'm going to look into directly using their output in a DIY motor controller.

The other pics show how the sensor ouputs are run through the case. Each connector has plugs on both the inside and outside, with an o-ring seal between it and the case. BTW, one of mine was broken by a ham fisted junkyard removal. :mad: Does anyone know where to get parts like this?
 

Attachments

·
Registered
Joined
·
91 Posts
Discussion Starter · #5 ·
One of the difficulties I had with this motor is finding any decent information about it. The real name for the Highlander Hybrid rear motor is either "Q211" or "MGR"

Finally, I stumbled upon some amazing testing done at Oak Ridge National Laboratories. :cool: Here is where they are on the internets...
http://www.osti.gov/bridge/

All the testing was done on various generations of Pruises and Camries.
They did complete tear downs to measure all imaginable details, as well as testing torque, maximum speed, back EMF, gear losses, etc. Also, they did not simply use the Toyota controller - they used another design and double checked that the motors output. All the documents are free, and here are some good documents to check into:

"Evaluation of 2004 Toyota Prius Hybrid Electric Drive System"

"REPORT ON TOYOTA/PRIUS MOTOR TORQUE CAPABILITY, TORQUE PROPERTY, NO-LOAD BACK-EMF, AND MECHANICAL LOSSES – REVISED MAY 2007"

"REPORT ON TOYOTA/PRIUS MOTOR DESIGN AND MANUFACTURING ASSESSMENT"

"EVALUATION OF THE 2007 TOYOTA CAMRY HYBRID SYNERGY DRIVE SYSTEM"

Another really good paper is about the development of the Highlander Hybrid electric motors:

"Development of Traction Drive Motors for the Toyota Hybrid System"
www.e-mobile.ch/pdf/2005/321.pdf

There are many similarities between the motor used in the Q211 and the second generation Prius. Part of my goal in taking apart this motor is to see how closely they match.
 

·
Registered
Joined
·
482 Posts
Ok, here's the inside of the right cover plate. The sensor mounted to it is a resolver, that appears to have some clocking adjustment.

For those unfamiliar w/ resolvers, they are extremely accurate absolute position sensors. Basically they are rotary transformers, and work by being fed a sine wave, that is picked up by two coils positioned 90 degrees apart from each other. These coils feed back the sine and cosine of the current position. If you look closely at the rotor, you will see an oval shaped piece of metal (between the rotor and bearing) This piece varies the air gap, thus changing the signal at the output coils.

Generally resolvers are found in high accuracy servo applications, like CNC machines. They are very robust and can also function well in an electrically noisy environment. :cool:

My feeling is Toyota chose these for a number of reasons, and replacing them merely because I didn't know how they work would be possible, but both expensive and troublesome. :(

Their signals can be translated to digital quadrature signals as one option. There are IC's made by TI that can do this for about $12. Other than that, I'm going to look into directly using their output in a DIY motor controller.

The other pics show how the sensor ouputs are run through the case. Each connector has plugs on both the inside and outside, with an o-ring seal between it and the case. BTW, one of mine was broken by a ham fisted junkyard removal. :mad: Does anyone know where to get parts like this?
e*clipse,

That looks like the same resolver as in the Prius. Here is a rather old video of me spinning the prius motor using 24volts. I also got it to push a car around at golf cart speeds with 48 volts from 4 lead acid batteries on their last leg.

I am using an eval board from Analog Devices to do the resolver to digital conversion.

My project got put on hold for quite some time but now I am back at it. My goal is to see how much power these Toyota motors can really put out.

I have been doing quite a bit with the Prius inverter as a battery charger and as DC to DC converter to do some battery cycling. I recently got another analog devices eval board to do simultaneous analog conversions that I needed for measuring the phase current outputs for FOC of the motor.

I hope to get the time for working on the motor control again very soon.

Good luck
Jeff
 

·
Registered
Joined
·
91 Posts
Discussion Starter · #8 ·
Very nice work Jeff! :cool:

I see a lot of opportunity for cheap Pruis transmissions. There are soooo many Priuses out there. Most people won't have a clue what to do with them, so they should stay pretty cheap. I checked on prices for the Prius and Camry (70kW) transmissions, and they range from about $500>$1000.

Did you try to get the motor running open loop?

What year of Prius is it from? Did you get measurements of the rotor?

Have you figured out a way to bypass most of the gearing stuff for a simple single speed drive?

Good luck with that - I'm extremely interested in how it goes. :)
 

·
Registered
Joined
·
91 Posts
Discussion Starter · #9 ·
Here is some info from the ORNL paper "REPORT ON TOYOTA/PRIUS MOTOR DESIGN AND MANUFACTURING ASSESSMENT"

It shows the differences between the first and second generation Prius motors, along with giving some important dimensions.

First, the output curves. Note the increase in power from 33kW to 50kW. Torque increases from 350Nm to 400Nm. This was achieved without increasing the size of the motor. These numbers are tested in the other ORNL papers.

Second, note the 2004 stator and rotor. I'll post a pic soon which shows the MGR stator and rotor.

3rd & 4th pics are the stator and rotor dimensions and weight. I'll verify the rotor dimensions on my MGR soon...

The last pic show the rotor stampings from the 1st & 2nd generation Prius motors. Changing to a V shaped permanent magnet was instrumental in increasing the output power. Technically, these motors are "Interior Permanent Magnet Synchronous Motors." They rely on reluctance (like a switched-reluctance motor) to provide power past the base speed.
 

Attachments

·
Registered
Joined
·
91 Posts
Discussion Starter · #11 ·
Here is the stator and rotor of the MGR with the right side cover plate removed. Compare this to the 2004 stator in the previous post. Like I said, I'll get rotor dimensions to verify the similarity.
 

Attachments

·
Registered
Joined
·
482 Posts
Very nice work Jeff! :cool:

I see a lot of opportunity for cheap Pruis transmissions. There are soooo many Priuses out there. Most people won't have a clue what to do with them, so they should stay pretty cheap. I checked on prices for the Prius and Camry (70kW) transmissions, and they range from about $500>$1000.

Did you try to get the motor running open loop?

What year of Prius is it from? Did you get measurements of the rotor?

Have you figured out a way to bypass most of the gearing stuff for a simple single speed drive?

Good luck with that - I'm extremely interested in how it goes. :)

I agree that they are cheap and may have some untapped potential.

The video is from a Gen1 prius that I got for free as a reject from a local transmission shop. It may have some defect that I haven't found yet.

To spin the motor I observed the BEMF of the motor in relation to the resolver output on the scope to get a reference point. I then modulated the phase voltages to track the BEMF based on the resolver readings I was getting. No current control at this point.

However the big discovery for me was when I put the motor in a car under a load and had some problems. I was trying to drive the motor like a BLDC motor which worked fine without being under load on the bench. Come to find out the reluctance torque is very significant even at low speeds. By advancing my phase voltage about 30 degrees everything worked much better. This phase shift also matches up with the stall torque graph in one of the ORNL report. Still very crude but it made me research these internal magnet motors a little more.

I am now working with GenII motors and inverters. I am making some progress but it is slow going.

I have some crazy ideas on what I can do with these things. I am far from having a mature controller but the experimenting is a great hobby.

The package I am working towards is a liquid cooled 100kw peak drive train + 20kw battery charger, 12V DC/DC, electric air conditioning compressor for about $1000 by using salvaged Prius components.

Jeff
 

·
Registered
Joined
·
91 Posts
Discussion Starter · #14 ·
To spin the motor I observed the BEMF of the motor in relation to the resolver output on the scope to get a reference point. I then modulated the phase voltages to track the BEMF based on the resolver readings I was getting. No current control at this point.

However the big discovery for me was when I put the motor in a car under a load and had some problems. I was trying to drive the motor like a BLDC motor which worked fine without being under load on the bench. Come to find out the reluctance torque is very significant even at low speeds. By advancing my phase voltage about 30 degrees everything worked much better. This phase shift also matches up with the stall torque graph in one of the ORNL report. Still very crude but it made me research these internal magnet motors a little more.

Jeff
VERY interesting! Thanks for the tip. :) I was going to feed the encoder a sin wave from my signal generator and watch the output on my scope. I had no idea how to correlate that with the actual rotor magnets.

Have you found that phase advance to be constant, or change (with speed, etc) ?
 

·
Registered
Joined
·
91 Posts
Discussion Starter · #15 ·
Ok, here are some very interesting (and somewhat discouraging :() dimensions. This rotor is significantly smaller than the Prius rotor.

Please take these numbers as "comparative" They were measured with my shop calipers, so I wouldn't use these dimensions for anything critical, like bearing fits. I also apologize for the non-SI units.

Rotor laminate OD: 5.825"
Rotor laminate width: 1.8" Width with both endcaps: 2.368"
Bearing OD's: 2.675"
Drive spline OD: 1.375"
Drive spline ID: 1.235"
Drive spline tooth #: 30
Resolver rotor width: 0.156" (made of laminates)

The laminates had 8 flats on the OD, corresponding with 8 notches on the encoder side endcap and 8 flats on the spline side endcap. There are 8 highly magnetic regions in the rotor.
 

Attachments

·
Registered
Joined
·
91 Posts
Discussion Starter · #16 ·
Well, THAT throws a big monkey-wrench into the whole "it's the same as a Prius motor" theory. :mad:

Just to make sure I wasn't too insane, I verified the specs (right off a Toyota Highlander spec sheet )

Motor Generator Rear (MGR)
-Function
Drives rear wheels, regeneration during braking
- Type
Permanent magnet motor
- Max. Voltage
AC 650V
- Max. Output
68 hp (50 kw) @ 4,610 – 5,120 rpm
- Max. Torque
96 lb.-ft. @ 0-610 rpm
I guess it sort of works out if you plug these #'s into the classic Horsepower sanity check equation, IF the rpm is high enough.

HP = Torque (ft-lb) * n (rpm) / 5252
Torque = 68 hp *5252 / 4610 rpm = 77 ft-lb
rpm = 68 hp * 5252 / 96 ft-lb = 3720

However, 96 ft-lb * 610 rpm/ 5252 = 11 hp.

Something's not right here. Could they have included the gear reduction?

Any suggestions from any BLDC motor experts out there?? :confused:
 

·
Registered
Joined
·
482 Posts
VERY interesting! Thanks for the tip. :) I was going to feed the encoder a sin wave from my signal generator and watch the output on my scope. I had no idea how to correlate that with the actual rotor magnets.

Have you found that phase advance to be constant, or change (with speed, etc) ?
I forget the resolver excitation frequency I used. For some reason I think I read 7kHz somewhere but I am not using that. My documentation is sketchy at best and I will need to revisit the resolver at some point since my Resolver to Digital board has the Loss of Signal Error LED constantly on but it still gives me the position so I moved on.

You can spin the motor by hand to produce back emf while apply the resolver excitation. Once I had the bemf to resolver alignment I could do the commutation electronically based on rotor position without regards to phase current just to get it spinning. This would be like running a DC motor with voltage control.

As far as the phase offset needed to take advantage of the reluctance torque being constant I think it is pretty close to fixed but don't take my word for it. I think that is why the ORNL report has the stall torque graph in it to show the interaction of magnetic and reluctance torques.

I forget all the technical jargon but Id=0 does not produce max torque for the minimum amount of stator current for these motors. In some Id Iq graphs showing the Max Torque Min Current for IPM motors there is a slight bend. A straight line (which would represent a fixed offset) looks like a good enough fit for me. The field or flux (not sure which is appropriate) weakening part of the control happens when I run out of voltage headroom and it looks as simple as increasing the Id magnitude with respect to Iq.

My Prius transaxle has two motors in it. Once I get the phase current control working the way I want it to I can use one motor to load the other and perhaps get some power efficiency mapping while adjusting the phase shift angle. You could do the same with two of your motors.

The main thing that I am working on right now is trying to figure out the best way to capture the analog signals coming out of the inverter that represent the phase currents. They are very noisy.

I am going to finish my battery charger development before I get back on the motor control part. However all the things I am having to learn for this piece (like how to measure currents and how to program the micro processor) will apply.

Regards
Jeff
 

·
Registered
Joined
·
3,141 Posts
I'm not an expert on BLDC motors but I found some references that show a peak torque at a lower RPM and then a lower rated torque at the highest RPM where power is greatest.
http://motion.control.com/thread/1276775704
http://www.linengineering.com/line/contents/stepmotors/pdf/BrushlessDC_Basics.pdf
http://en.wikipedia.org/wiki/Talk:Brushless_DC_electric_motor
http://www.appliancemagazine.com/editorial.php?article=551
http://what-when-how.com/electric-m...-load-and-peak-torque-values-electric-motors/

It seems unusual for the peak torque to be at such low RPM, but the rated torque (77 lb-ft) is only a bit less than the peak torque (96 lb-ft). It may be related to the higher power losses at higher RPM (and higher voltage), where heating and high frequency core losses may reduce the safe level of continuous torque.
 

·
Registered
Joined
·
482 Posts
VERY interesting! Thanks for the tip. :) I was going to feed the encoder a sin wave from my signal generator and watch the output on my scope. I had no idea how to correlate that with the actual rotor magnets.

Have you found that phase advance to be constant, or change (with speed, etc) ?
Another tip to align rotor and resolver is to apply a dc voltage to two phases at a time and you can step the motor like a stepper motor while recording the resolver outputs.
 

·
Registered
Joined
·
91 Posts
Discussion Starter · #20 · (Edited)
Thanks for all the tips, Jeff! It really helps to hear from someone who's been there. :)
Thanks also PSTechPaul - those last two links combined w/ the ORNL papers should lead somewhere...:)

Right now I'm trying to resolve what seem to be some really big motor output issues. Something in their specs just refuses to make sense. I can make things work by fudging numbers (like the base speed of the motor.)

Let's assume this motor has a torque-speed and power-speed relationship similar to the ones I posted for the 2004 Prius. Obviously, with the rotor being a completely different diameter and length, this curve will be different, but follow a similar pattern.

The Highlander Hybrid has a "governer limited" top speed of 114 mph, according to Car&Driver:
http://www.caranddriver.com/reviews/2011-toyota-highlander-hybrid-road-test-review

With stock P245/65 R17 (Base) tires, the tire circumference would be 93 inches, and would make 681.5 revolutions/mile. Thus, the rear axle would be spinning @ 1294.9 rpm. With a gear ratio of 6.86:1, the motor would be spinning @ 8882.7 rpm. It would probably be safe to say the rotor is balanced to spin at 10,000 rpm, although a gernading rotor would make a very exiting stop. :p

IF the 50kW specification is correct, the peak horsepower (assuming a curve similar to the 2004 Prius) occurs right around or slightly after it's base speed. Further assuming the 96 ft-lb torque specification is corrrect, then the corresponding rotor speed would be:
rpm = HP*5252/Torque = 68hp * 5252 / 96ft-lb = 3720 rpm

This is nearly double the Prius' base speed, but corresponds roughly with the base speed spec of 610 RPM if the 6.86:1 gearing is included. Perhaps they were referring to the AXLE speed? :confused:

So, given that the rotor can physically handle high rpm, is it reasonable to think the base speed is 3720 rpm?

Because the diameter is smaller, would the back EMF generated be smaller than the Prius? Otherwise, the voltage is going to be much higher than 650V. In fact according to the ORNL test, at 6000 rpm, the Prius back EMF was 850V peak.

I'm still trying to wrap my head around the fact that this significantly smaller rotor could put out the same power. I've seen it frequently in examples like geared starter motors, but this situation has some serious back-EMF limits. I mean, I really don't want to have a 1000V battery pack! :eek:

Would any motor experts out there have any advice for practical limits that can be put on these rough analyses?:confused:
 
1 - 20 of 134 Posts
Top