DIY Electric Car Forums banner

1991 MR2 + 2004 Prius Transaxle

64K views 67 replies 17 participants last post by  Johnny5 
#1 ·
I got a new donor car yesterday and I wanted to post to mark the beginning of this conversion. It is a 1991 Toyota MR2. I have a 2004 electric transaxle and inverter out of a Toyota Prius that I am going to put in it. I have been experimenting with the control of the electric transaxle for some time now with partial success. I was trying to fit it into a 1966 VW squareback but I got overwhelmed with the amount of rust and restoration that was needed on the donor body. I should have listened to the advice of others that cautioned about what makes a good donor car. I have to learn the hard way apparently. I may still be in the hard lesson mode since there are not many people trying to use the Prius motors but time will tell.

This car needs nothing but a paint job and the interior is perfect so I can focus solely on the electro mechanical aspects of the project. The mid engine design looks like it will work well for my transaxle and conversion plans.
Land vehicle Vehicle Car Coupé Sports car
Land vehicle Vehicle Car Steering wheel Center console

I will post some pictures of the engine removal once that happens but the wife has declared that there will be no dis-assembly until the VW shell is gone.

The first phase of my plan is to get the car to a grocery getter stage with the small 48 cell 40Ah pack that I already have. In this stage I am not going to bypass the DC boost within the inverter and I am only going to engage the main traction motor (MG2). In this configuration I will be limited by the boost circuit of a 100 battery amps or so.

Once this is functioning I will attempt to increase the power. This is where the fun will start. In addition to adding more cells I have a couple of ideas how to get a considerable amount of power with the Prius setup. I can either start to include the other motor/generator (MG1) and bypass the boost circuit or I can use the MG1 inverter stage IGBTs and some additional inductors to increase the DC boost circuit capacity driving MG2.

Thanks
Jeff
 
See less See more
2
#3 ·
its nice to see someone attempting to reuse a prius synergy inverter.

im waiting for the inverters to go down in price on ebay, before i grab one to tinker with. theere's many of them for sale, but the idiot sellers dont realise that no one changes a faulty inverter, since its bullet proof, and most times, covered by waranty. very soon, ebay will be flooded with this inverters from wrecked priuses,

i dont mean to bust your bubble but that inverter is not "intelligent". cannot operate by its own, it still needs the main ecu to control the boost converter, mg1, and mg2.

this guy tore one down: http://techno-fandom.org/~hobbit/cars/ginv/

from your post, seems that you have already figured that out. looking forward to your conversion, and success.
 
#4 ·
Thanks for the encouragement.

I am making my own ECU. I have a couple of versions that are a ways from optimal but do sort of work. I got screwed up with my first donor car which was holding me back quite a bit. I hope I am able to mount the transaxle in this car in relatively short order and get on with the inverter/motor control aspects of the project.

I bought the Prius inverter, transaxle, and half shafts off ebay for $550.

I consider this a low budget conversion project.

Jeff
 
#5 ·
I got the ICE out this weekend.
Vehicle Automobile repair shop Car Auto mechanic Auto part

Degreased the engine bay.
Vehicle Car Auto part City car Engine

And starting to line up the new motor mounting.
Automobile repair shop Vehicle Car Bumper Automotive exterior


I looks like the MR2 axles might work with only light modifications. The inboard CV joint shaft has the same spline as the prius so should plug right into the differential, but the outer diameter of the CV cup hits the transmission housing. I hope with a little grinding I can make it work. "Better to be lucky than smart".

I have decided to use the 2001 transaxle with MG1 and MG2 welded together in instead of my 2004 transaxle for this first trial. My idea is that this motor will spin at a higher rpm with lower battery voltage without using a battery boost converter. I may have less torque but hopefully good enough and a better match for my smaller battery pack for now.

Now I am working on my motor mounting options. I might build a frame to fill the void left from the ICE to connect to the far motor mount and fill it with batteries to help balance out the weight side to side.

My body is not used to this kind of manual labor. Hopefully I can get past this part of the job without too much injury.

Regards
Jeff
 
#6 ·
I forgot to mention that I also removed the gas tank from underneath the car.

I have no experience with AirCon compressor and refrigerant. You can the see the compressor dangling in the engine bay picture. I would like to purge the refrigerant and remove it for reattachment to the transmission ICE shaft later in the project.

If anyone has experience or suggestions with this to help me avoid any pitfalls I would appreciate it.

Jeff
 
#9 ·
I bought myself a transmission jack and lifted the Prius electric tranny in place. Fits like a glove.
Vehicle Engine Auto part Car City car

I had a clearance problem at first but noticed that the MR2 inbound CV cups on the two axles were non-symmetrical. Swapping the left and right gave about an inch extra position shift in my favor.:)

I also thought that I had to grind down the MR2 CV cup for clearance but this was not case so the only modification was swapping the left and right cups.

I am now going to start framing up the mounting. My rough plan is to build a battery support structure where the ICE used to be.

Long term goal is a preassemble transmission + inverter+ battery unit that can be swapped out for the MR2 form factor in a single day with minimal donor modifications. This is my plan even though I don't expect my prototype to resemble this goal at first attempt.

Jeff
 
#11 ·
I made an offer on the MR2 Spyder convertible but the person didn't take it. I then decided it might not be so bad to use a less expensive version to get the kinks out. If it works out my plan is to have a kit for various mid engine and front wheel drive models.

Thanks for your interest. I am sorry if progress will seem slow. I don't get a whole lot time to work on it.

This weekend I spent more time verifying clearance and motor mounting options. I am sure as soon as start cutting and welding I will think of a better design.
 
#12 ·
Seeing the conversions at EVCCON raised the bar on me quite a bit. I was very impressed with the workmanship.

I was trying to mount my prius electric transaxle in place of the original MR2 transaxle all the way against the driver side and then put batteries in place of the ICE to counter balance it. This design would work but I was not thrilled with the complexity of the framing and weight management.
Vehicle Engine Auto part Car City car


I then realized that by swapping the long and short axles it would shift the transmission towards the passenger side over 10inches and basically center the weight it in the engine bay. It is very heavy since the electric motors are inside. I would guess the transmission and inverter are over 300lbs combined.

Vehicle Car Engine Auto part Technology


This picture is just a moch up of what it will look like and I like it much better. I now need to build a beefy cross member and brackets that the transaxle will hang from. I also decided not to use any of the original rubber motor mounts. I figure the cross member will act as a rigid subframe for the motor. I may need to build in some compliance or extra strength in the bracketing to the frame for flexing.

I also only have a portion of my batteries at this time and they will easily fit in the trunk with a very short wire run. This will save me from having to build elaborate boxes to get started on the control electronics. Once I get the thing operational with this pack size I can add more battery boxes in the engine bay on either side of the tranny.
 
#13 ·
I am working on my plans to hack part of my inverter for my battery charger solution. Chargers are expensive and it will be cool if it works out. Here is a picture of my inverter.
Electronics Auto part Technology Engine


I am using the GenII Prius inverter that has the bi directional boost/buck converter between the inverter and the battery pack. The purpose of the boost converter is to supply up to a 500VDC bus for the inverters to overcome the back emf of the motors at higher rpms. This Bi-Directional Boost Converter also bucks down large back emf voltages to put current back into the battery during regenerative braking. Here is a diagram of the Toyota Prius hybrid power distribution.
Text Floor plan Technical drawing Plan Diagram


My car will not have the ICE of course and the final configuration of my power setup is TBD pending some experiments and testing but this is what I have to work with.
My first charging configuration test is to tap one of the phases of the 30kW Motor/Generator with AC line voltage. Mainly because I don't have another inductor to use for the test. This puts the AC source in series with the motor inductance and the inverter acts a full bridge rectifier without actively switching it via the IGBT body diodes. I am using a couple 500W flood lights as a load in place of the battery in this experiment.
Auto part Engine Technology Electronics Automotive engine part


Here is a picture of the non-power factor corrected line current using the built in inverter phase current sensors. These signals were very ugly before I made a circuit to clean them up. The phase current peaks are when the line voltage peak exceeds the DC bus voltage as expected.
Oscilloscope Screen Technology Electronic device Electronics


Here is a picture of the DC bus voltage of about 150V peak and 25V ripple with the flood lights loaded on it.
Oscilloscope Screen Electronic device Technology Electronics


Now the hard work. I will attempt to write the control algorithm that can switch the inverter phases in such a way to produce a power factor corrected boosted +400VDC that the converter can buck down to my battery pack voltage whatever that ends up being. In this experiment I hope to be able to demonstrate controlling the battery voltage output by dimming the flood lights.
I have no experience with Power Factor correcting algorithms but here we go.
Thanks Jeff
 
#16 ·
I have simulated my Bridgeless Power Factor corrected boost battery charger that I am planning using the Prius inverter switches and built in buck/boost converter.

The green line is the single phase line current and the pink line is the ac voltage. The dark blue line is the dc current going into the battery pack. I normalized everything so it would graph together. They are in phase thus the Power Factor Correction.

Text Screenshot Technology Font Line


The bridgeless part of this is that I don't use a full bridge rectifier like traditional ac to dc boost circuits. Instead it uses the bidirectional inverter half bridges as both boost and rectification at the same time.

If this works the way I am thinking I will basically get a free battery charger since i am using the existing motor inverter. I think I can also use the motor inductance but for now I am using the inductor that is for the Prius dc boost (373 uH).

I have not gone live with the AC part yet but my experiments and simulations with dc to dc boost/buck moving charge between battery packs works as expected.
 
#18 ·
I plugged some things together today and did a video of tires spinning on the jack stands. I hope this helps motivate me. I haven't been spending enough time on this project lately.



I have the transaxle rigidly mounted to the frame. This may have been a mistake. I can feel and hear quite a bit of the vibrations when I am sitting in the car during this test.

Thanks
Jeff
 
#20 ·
This is my favorite thread on this site.
There is so much potential (pun!?) in dealing with used Prius parts to make
a simple conversion.
Jeff, please consider this encouragement to get this project on the road.
It is really a ground breaking project.
Thanks!
tom in maine
 
#21 ·
Thanks Guys for the encouragement

I only have a 6kwh pack (48X40Ah) to put in it at the moment. I hope it is enough to do some neighborhood driving and work out some kinks before spending more.

I have a lot to do but would like to be on the road early next year with this grocery getter phase of the project.

Jeff
 
#22 ·
I like having this holiday time to get some important things done like working on my car:)

I just got my FOC (Field Oriented Control) working for the main traction motor. This is a milestone for me. I was not sure it was going to work. So far I am pleased.

Previously, I was just modulating the phase voltages based on throttle command and rotor position. Now the throttle commands the motor current (torque) and the voltages go to whatever they need to in order to meet the command. This type of control is what I will need in real world driving.

My algorithm tries to maintain a current vector of -30deg off of what traditionally one uses for BLDC motors. For this type of motor I don't want the Id=0 like other BLDC motors might.

Id = I*sin(-30)
Iq = I*cos(-30)
I am using this Id Iq relationship based on the stall torque graph of this motor published in one of the ORNL document.

This motor is a IPM (Interior Permanent Magnet) motor which has a significant reluctance torque component that overlays the permanent magnet torque component. Id=0 will not take advantage of the reluctance torque from my understanding.

Here is a video of it in action this morning.

 
#25 ·
Offtopic: good camera work, as you where not shaking the camera over the place, also a good camera voice & no rubbish info on the side.

I am very impressed you have got the Prius transaxle to spin .
Protection against overloads, the motor tuning etc may all still involve a lot of work still.. Not sure though, I never played with IPMs.

Eitherway, keep up the work, you may be able to pull something off that enables the re-usage of the Toyota drive train for the people :)

What power levels you expect the motor/inverter can go upto? I assume higher than with the ICE still attached, as the coolant used to be shared with the ICE, or was this separate for the traction inverter?


//steven
 
#26 ·
Are you making fun of my camera work:)

Lots of tuning ahead and lots of unknowns still. You are right about there being many exceptions and overloading handling yet to be implemented. I have already run into some of them.

I hope I don't try and push things too far too soon without the proper controls. However knowing myself and my tendency to cut corners I will probably learn the hard way.

The motor/inverter coolant is separate from the ICE normally with two radiator sections in the Prius. In my MR2 conversion I will be pumping coolant through the main radiator in the front of the car. This should give me good heat transfer and perhaps raise my continuous power ceiling.

(Example of cutting corners) I didn't take the time to figure out and wire in the motor and inverter temperature sensors.

The expected power is a good question. I have several different configurations to explore.

1st option: (approx 40kw) Double the battery voltage. MG2 only and DC boost. I read that MG2 is rated at 50kw and the DC boost is rated at 20kw. Since the Prius battery is around 200V nominal and boosted to 500V, I take these ratings to mean 100 battery amps and 100 motor amps sounds reasonable. The DC boost circuit is the bottle neck. In this configuration I will need more battery volts to increase power. I am thinking of fitting 128 cells for a nominal voltage of 422V which would be 42.2 kw at 100 battery amps through the boost circuit. Boosting the voltage to 500 is helpful for reaching the higher motor rpms. I am not sure if the built in 12V DC to DC can handle the higher battery voltage but this is not a deal breaker.

2nd option: (approx 40kw) Double the DC boost. Still using MG2 only but using the unused MG1 igbts and an extra boost inductor to parallel the built in 20kw DC boost switching. This would allow for more battery amps and my pack voltage can remain lower perhaps 64 cells and 200 battery amps in this case.

3rd option: (approx 70kw) Double the battery voltage and motors and bypass the DC boost circuit. Rigidly connect MG1 and MG2 by welding the planetary gear internally. 160 battery amps and 128 cells. The addition of MG1 will be interesting I think. It spins twice as fast as MG2 for the same voltage so I expect it to add some significant torque at the high end of the rpm range. At highway speeds I expect MG1 will be doing most of the work and MG2 will be deeper in the field weakening region.

I am not sure which route I am going. Perhaps I will try all of them and others as I learn.

I am very open to suggestions and perhaps I can exceed these expectations.

Note: I have bench tested the dc boost circuit at 60amps continuous without any liquid cooling. It gets warm but seems to handle it fine. At 80 amps I think I was hitting a thermal fault limit after some time.

Regards
Jeff
 
#27 ·
Hi Jeff.

What operation is the Prius inverter doing in your setup please? Have you totally bypassed the inverter control system and are controlling the IGBTs themselves? I'm a little confused about what your controller does to 'control' output if the Prius inverter is still a major factor.

Is it at all possible to hack and 'remap' the Prius inverter and keep the protection and control algorithms as Toyota decided originally and simply tweak key parameters like voltage and Max current? This would be ideal but I suspect there is reason you have not done this. Please share.

Thank you.
 
#28 ·
The prius inverter has no intelligence. It is only the high voltage high power switches. It provides access to the IGBT gate drivers and will turn on whichever transistor you tell it to without asking if you are sure or not. It also provides some analog signals that represent bus voltage and phase currents.

The Prius ECU is where the inverter plugs in on the Prius. This is where all the Toyota synergy drive decisions and motor control are made. I am not using the Prius ECU. I have never considered trying to use it.

My controller is decoding the rotor angle, sampling the phase currents, and modulating the IGBTs accordingly using the signals available from the inverter.

Jeff
 
#29 ·
Ah right so the hardware is within the inverter and the software in the ecu. That is useful. So all that is needed is a good interface module and control algorithm to run the hardware. I know I'm over simplifying it but that is the basics right?

May I ask what options you considered for controlling the inverter before going this route please? Thanks
 
#30 ·
You are correct. I am working on the interface and control algorithm. Some inverters already have this built in and respond to a command set via CAN bus for example. The Prius inverter does not.

I am not sure I understand your second question.
I enjoy DIY projects. Doing both hardware and software would be a bit too much for me. I am content with learning the software side of things for now and this is a low cost way of getting that experience.

Regards
Jeff
 
#31 ·
4th Option: Re-wind MG2 stator to a lower voltage, or switch from Y to Delta winding configuration (the tap point is relatively easily accessed).

I am hoping to get my hands on a Prius transaxle to do something similar to what you've done here. My plan would be to not use MG1 at all, rewind MG2 to a lower voltage, and run my own controller (which is nearly complete, just need to fine tune the software). Winding the stator to 24-26vac (assuming a 50Hz "base frequency" and a total reduction between MG2 and the diff of 8:1 from memory) would take you all the way to 110kph before field weakening starts. With over 300Nm of torque all the way up there, it would make for a quick car. You'd need to handle a load of current and go up to 400Hz though, hence the custom controller. It could be a sub 7s 0-100kph car.
 
#32 ·
That would be something to see. I have taken a peek at the MG2 stator and rewinding that amount of copper would be :eek:. The gear reduction in my 2004 transaxle is 4.11:1 I think.


Yesterday I went ahead and welded the planetary gear between MG1 and MG2 as stated in option 3. As I listed my options this one seemed to make the most sense especially for testing purposes. I should be able to run the motors against each other (One drive and one regen) and test out some higher currents.

Regards
Jeff
 
#33 ·
Jeff
What I was asking is how did you choose the arduino control for your project?
Also, in your research have you seen much difference between generation 1 and 2 Prius drives? I wonder if it would be easier to find a can bus controlled system for interface purposes where the manufacturer has already programmed control algorithms etc.

Not knocking your project, I think your progress is excellent and I'd love to do something similar with a Lexus RX/highlander rear transaxle or Camry transaxle for a larger vehicle, just wanting the simplest option to begin as I'm not as skilled as you in electronics or software and not got much time to learn at the moment.

Thank you for sharing your progress and success.
 
#37 ·
Arduino just because I am using it for other hobbies.

I have a Gen1 also. Gen 1 might be better for lower voltage EV than the Gen 2. The big motor, MG2, on the Gen1 spins faster at lower voltage. It has heavier gage windings so it can probably handle more current. It is rated for lower power because of the lower voltage.

I picked the Gen 2 because there are a lot more of them on Ebay.

Basically they are all the same, including the Camry, Highlander, and Lexus.

Jeff
 
This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.
Top