[jddcircuit]

I acquired a 2001 Prius transaxle for $0 some time ago. I finally got some time to work on it. I completely took it apart and can’t find anything wrong with it. I don’t know what error code it produced to make the transmission shop swap it out.



I got both electric motors (MG1 and MG2) spinning using a Zilog Motor Control eval board that I already had and an Analog Devices Resolver to Digital eval board that I bought. The commutation software and resolver decoding was a little tricky but I got it working. With this success I am pretty sure that I can design and build my own control board with enough trial and error. I only used a 24V 3A supply which was fine for proof of control but at 24Volts my car would only go about 3 mph.


I just ordered a 2002 prius high voltage inverter ($100) that I want to hack into. I should be getting it in about a week. I am thinking that this will be easier and cheaper than me building my own. It has a voltage boost circuit, inverters for each motor with hall current sensors, etc. Even if it doesn’t work I should learn more from taking it apart and reverse engineering it than it cost me.



I learned that MG1 spins a few times faster than MG2 at the same voltage. This was expected, MG2 is the bigger of the two. MG1 and MG2 are connected to each other via a planetary gearset. MG2 drives the ring gear which is connected to the axle differential gearset at a 4.1 to 1 reduction. I have seen MG2 referred to as the traction motor. MG1 drives the sun gear and the planet carrier protrudes on a shaft where the ICE would connect if I had one.



I have a plan on how I want to configure the drive system that I haven’t quite seen yet so I thought I should get some feedback. I have seen where someone welded the planet gears rigid which allowed MG1 torque to add to MG2 torque driving the wheels.

My plan is to add a disc brake or clutch plate to the planet carrier output shaft where the ICE connected. By holding the shaft stationary and driving MG1 in the opposite direction of MG2 I should be able to produce a wheel torque of ( 4.1 X ( MG2 +( 3 X MG1))). I am trying to preserve the torque multiplying of MG1 through the planetary gear during low speeds. If this is correct I should be able to accelerate very quickly. At around 40mph, MG1 will be spinning at it’s max speed and I will release the brake holding the planet carrier. This should allow me to continue to accelerate using MG2 only driving the ring gear and to cruise. I will then set the MG1 speed to whatever speed is appropriate to drive the planet carrier shaft which is connected to the oil pump and to any other pump system (ie air conditioning) that I might attach to it. This configuration would also allow me the option of still using MG1 coupled with a small ICE to charge my batteries while the car is parked if needed.

My question is whether or not I am correct in my assumption about transferring MG1 torque to the wheels in this configuration and preserving the torque multiplying of MG1? Any recommendations experience on other configurations using this transaxle is welcomed?

Has anyone already made their own control unit for the prius inverter or have any wiring documentation?

Thank you
Jeff

 

[dimitri]

Have you seen this thread on PriusChat forum?

http://priuschat.com/forums/prius-phev-plug-in-modifications/85924-all-electric-prius.html

My understanding is that DC-DC voltage booster is the weakest link in Prius's electric drivetrain, requiring ICE assistance when peak power is required. This is why Prius in pure EV mode is anemic.

 

[Coulomb]

I have a plan on how I want to configure the drive system that I haven’t quite seen yet so I thought I should get some feedback.

Sounds interesting. I've mused over this a bit.

My plan is to add a disc brake or clutch plate to the planet carrier output shaft where the ICE connected. By holding the shaft stationary and driving MG1 in the opposite direction of MG2 I should be able to produce a wheel torque of ( 4.1 X ( MG2 +( 3 X MG1))).

That would appear right to me, from the equations I've seen, except that the 3 is actually 2.6 according to a web site I checked. (That's for the Generation II prius; the Gen I may be different).

However, these things are easy to get wrong, and I'm suspicious since MG2 is supposed to be the "torque" motor and MG1 is supposed to be the "speed" motor. But maybe that's only because MG1 can add to the speed of the ICE (by spinning backwards).

With the speed reduction (and therefore torque multiplication) of 2.6, the 10,000 RPM maximum of MG1 would run out at about 3800 RPM (at MG2), which is about 64 mph (Gen II). If MG1 is limited to 6500 RPM (as it is by default when the ICE is not running), then this limit becomes 2500 RPM at MG2, or 42 mph.

I am trying to preserve the torque multiplying of MG1 through the planetary gear during low speeds. If this is correct I should be able to accelerate very quickly. At around 40mph, MG1 will be spinning at its max speed and I will release the brake holding the planet carrier. This should allow me to continue to accelerate using MG2 only driving the ring gear and to cruise. I will then set the MG1 speed to whatever speed is appropriate to drive the planet carrier shaft which is connected to the oil pump and to any other pump system (ie air conditioning) that I might attach to it.

Interesting. One problem is that you would not have air conditioning or other mechanical loads running during low speed operation (or if you did, you'd miss out on most (?) of MG1's additional torque).

To do this with the existing Prius inverter, you'd probably have to have a 500 VDC pack (Gen II; 273 V for Gen I) to bypass the 20 kW limit of the DC-DC boost converter.

This configuration would also allow me the option of still using MG1 coupled with a small ICE to charge my batteries while the car is parked if needed.

Err, where would the ICE go then? Connected to the same place as on a regular Prius (ICE input of transaxle), but with the clutch across the ICE output?

 

[jddcircuit]

Thank you for the replies.

I will plan on a higher voltage battery pack than the 200 volt pack in the prius. Could it be that the 20kw rating on the boost circuit is due to the current limit of 100amps. If this is the case then 300V gives me 30kw and so on. If I realize that I can't produce enough torque to reach higher speeds then more batteries to the rescue I would think.

I probably won't invest in batteries or a donor car until I get control of the inverter that should be here in a few days. Hopefully by then I will have learned more details about the inverter and it's control interface.

 

[jddcircuit]

Sounds interesting. I've mused over this a bit.

That would appear right to me, from the equations I've seen, except that the 3 is actually 2.6 according to a web site I checked. (That's for the Generation II prius; the Gen I may be different).

However, these things are easy to get wrong, and I'm suspicious since MG2 is supposed to be the "torque" motor and MG1 is supposed to be the "speed" motor. But maybe that's only because MG1 can add to the speed of the ICE (by spinning backwards).

With the speed reduction (and therefore torque multiplication) of 2.6, the 10,000 RPM maximum of MG1 would run out at about 3800 RPM (at MG2), which is about 64 mph (Gen II). If MG1 is limited to 6500 RPM (as it is by default when the ICE is not running), then this limit becomes 2500 RPM at MG2, or 42 mph.

Interesting. One problem is that you would not have air conditioning or other mechanical loads running during low speed operation (or if you did, you'd miss out on most (?) of MG1's additional torque).

To do this with the existing Prius inverter, you'd probably have to have a 500 VDC pack (Gen II; 273 V for Gen I) to bypass the 20 kW limit of the DC-DC boost converter.

Err, where would the ICE go then? Connected to the same place as on a regular Prius (ICE input of transaxle), but with the clutch across the ICE output?

I think 2.6 is correct. I was approximating. This thing is still a bit conceptual.

I plan on having the shaft spinning in a positive direction during low speeds and even when MG2 is fully stopped. What I picture happening is that during high command torque situations MG1 will spin negative more quickly than MG2 will spin positive due to less load. This will cause the planet carrier shaft to slow to zero speed and then go negative unless something stops it. I can apply a brake to hold it at zero speed or use some type of screw in screw out brake plate that will apply itself with a few negative rotations and then release with a few positive rotations. With the shaft being held then MG1 torque will be available at the wheel providing a boost torque.

However The screw in screw out approach will not allow the MG1 to participate in regen braking I think.

The items that may also be connected to the shaft like airconditioning, hydrolics, external ICE would be done with belts and pulleys. These could coexist with the shaft braking mechanism in theory.

 

[Coulomb]

I will plan on a higher voltage battery pack than the 200 volt pack in the prius. Could it be that the 20kw rating on the boost circuit is due to the current limit of 100amps. If this is the case then 300V gives me 30kw and so on.

Well, there is a current limit certainly, but a voltage limit as well. Above a certain input voltage, the boost converter will just go phut.

I've just realised that if your inverter is for the Gen I, then it won't have a boost converter anyway; the battery and motors run at the same voltage (nominally 273 V, but it will take well over 300 V I'm sure).

If your inverter is for the Gen II but the motors are for Gen I, then the inverter may be capable of higher voltage than the motors can withstand. What might work OK is about a 312 V pack, the Gen II inverter, and Gen I motors, and you somehow disable or ignore the boost converter. With a larger pack than the stock Prius, hopefully the voltage swings will be lower, so you can get away with a higher average voltage to the inverter.

 

[jddcircuit]

http://www.youtube.com/watch?v=2olyxIhN_HE

Based on the speed of the differential output using a 24VDC supply (see video) I calculate that my car could have a top speed of 68mph with a 240VDC battery pack and a 24inch tire.

The motors inside the prius transaxle are 3 phase brushless dc type and their back emf limits their speed. In order to go faster I would need to supply a higher voltage. This should be fine for my first grocery getter prototype.

I can't wait to start hacking the prius inverter I ordered. I will also start looking for a higher DC power source.

Cheers

 

[Coulomb]

The motors inside the prius transaxle are 3 phase brushless dc type and their back emf limits their speed. In order to go faster I would need to supply a higher voltage.

Not necessarily true; I post about that here on your other thread.

 

[jddcircuit]

My donor car has been selected. I am very excited.

 

[jddcircuit]

Here is my motor and high voltage inverter from a Prius that I plan to use.

 

[nimblemotors]

Jeff, this is a very exciting project!

Are you planning to use the prius transaxle in the rear and eliminate the VW transaxle?

I also have a junkyard genI prius transaxle and inverter. I haven't had time to mess with it yet, actually wanted to use it as an ICE engine dyno as my first use of it. Might get on that project soon, and would love to have someone else to work with on it.

I have built a BLDC controller for the 16hp Mars motor, that powered my Electric Jetski (project not finished). I could pull out that controller and try to get it working with the MG, but it needs bigger power switches (inverter) to handle more than 16hp. That is what I hope the prius inverter would allow. Seems the worse case is just to use the IGBT's in it, or really my thought is even just the water-cooling technology is worth the $75 I paid for the inverter.

For an EV, I think you want the MG1 and MG2 tied together so it operates as a single electric motor. It would be very complicated to do anything. KISS. I wouldn't weld the planetary, but pin or spline the planetary gears so they don't spin.

I'm not sure what effect it has if MG1 windings are different so it spins faster at the same voltage than MG2. This might be a problem with tying them together?

 

[Mike D EV]

Hi guys,
Happy to see that someone is finally getting this concept to the operating stage.
I put this up over three years ago.
http://99mpg.com/Projectcars/evinsight/

I will be getting back into this very soon, as I have finally cleared the decks in my workshop.
While using the unmodified Synergy drive would allow many interesting possibilities, it seems to me that combining of the motors via the welded planetary may be the most efficient way to use the drive.
I have two second gen drives, so I will try it both ways.
The ICE shaft also runs the transmission oil pump, so from a lubrication standpoint, it is better for that shaft to always be turning to fully lube all the internal bearings.

This combined motor system also allows removing all the final drive gearing, to drop over 100 lbs of weight and the use of the ICE input shaft as an output shaft to drive any transmission, thus allowing this to power large vehicles.
The combined motors should be able to push my light weight Insight with no other transmission to well over 80 MPH.
You mention using an analog devices resolver decoder evaluation board for feedback?
What board did you use?
Mike

 

[jddcircuit]

Jeff, this is a very exciting project!

Are you planning to use the prius transaxle in the rear and eliminate the VW transaxle?

I also have a junkyard genI prius transaxle and inverter. I haven't had time to mess with it yet, actually wanted to use it as an ICE engine dyno as my first use of it. Might get on that project soon, and would love to have someone else to work with on it.

I have built a BLDC controller for the 16hp Mars motor, that powered my Electric Jetski (project not finished). I could pull out that controller and try to get it working with the MG, but it needs bigger power switches (inverter) to handle more than 16hp. That is what I hope the prius inverter would allow. Seems the worse case is just to use the IGBT's in it, or really my thought is even just the water-cooling technology is worth the $75 I paid for the inverter.

For an EV, I think you want the MG1 and MG2 tied together so it operates as a single electric motor. It would be very complicated to do anything. KISS. I wouldn't weld the planetary, but pin or spline the planetary gears so they don't spin.

I'm not sure what effect it has if MG1 windings are different so it spins faster at the same voltage than MG2. This might be a problem with tying them together?

Thanks for the interest. I am very excited. My plan is to not use the vw transmission. My donor car suspension needs some modification so it can accept a CV joint. It is a 66 which has a swing axle arrangement that I need to change. Then I need to make a Prius to VW axle. I am learning as I go but I think it might work.

I am not going to weld the planetary gear together yet. I want to get car rolling using MG2 first. I can keep the carrier shaft spinning using MG1 at any speed I want to keep lubrication pump going.

I then have some crazy ideas about how to use MG1. If I hold the carrier shaft stationary with a brake or gear then it will transfer 2.6X MG1 torque to the wheel through the planetary gear. I can use this for a low speed acceleration. I can then release the shaft to spin under the control of MG1 to lubricate and to run an air condition compressor or such during cruising. I could also use MG1 as a generator for my batteries when the car is parked with a small ICE and some type of pulley arrangement on the shaft if I am camping out somewhere. It would also be a cool campsite generator.

 

[jddcircuit]

I used the AD2S1210 resolver to digital eval board. I plan to layout my own board with the motor control, resolver and current sensing integrated together eventually. The two eval boards I had were for me to learn how the commutation works. I still have quite a bit of work ahead of me to get a well syncronized commutation at high motor speeds.

 

[Mike D EV]

thanks for the AD2S1210 info.
I was looking into the resolver interface chipset that is used in the prius, but this looks easier.
It was great to see the video of the unit actually running off the eval board.
Gets me motivated to get back on the project.

Have you tried running the motor open loop?

The dev board I purchaced for the task:
http://www.microchip.com/stellent/idcplg?IdcService=SS_GET_PAGE&nodeId=1406&dDocName=en537020
Good luck

 

[jddcircuit]

thanks for the AD2S1210 info.
I was looking into the resolver interface chipset that is used in the prius, but this looks easier.
It was great to see the video of the unit actually running off the eval board.
Gets me motivated to get back on the project.

Have you tried running the motor open loop?

The dev board I purchaced for the task:
http://www.microchip.com/stellent/idcplg?IdcService=SS_GET_PAGE&nodeId=1406&dDocName=en537020
Good luck

I got it spinning open loop first before I got the resolver board. I don't have a lot of current capacity so it was tough to get any torque during start up.

I need to build a high voltage dc supply to test out the prius inverter that I got. I want to see how fast I can get it spinning.

 

[Mike D EV]

You can get a prius battery pack with BCM and contractors for $400-$600 from most junkyards as well as e-bay.
We plan on using them for another project:
http://99mpg.com/Projectcars/phevinsightiigetti/

A simple HV CC charger can be built from off the shelf switching supplies to charge it up.
http://99mpg.com/Data/resources/downloads/relateddocuments/dual_stage_grid_charger1.pdf

Of course it will not get us far in an actual car but it is a good test power source capable of > 100A.

 

[jddcircuit]

You can get a prius battery pack with BCM and contractors for $400-$600 from most junkyards as well as e-bay.
We plan on using them for another project:
http://99mpg.com/Projectcars/phevinsightiigetti/

A simple HV CC charger can be built from off the shelf switching supplies to charge it up.
http://99mpg.com/Data/resources/downloads/relateddocuments/dual_stage_grid_charger1.pdf

Of course it will not get us far in an actual car but it is a good test power source capable of > 100A.

Thanks Mike
I will seriously consider these options. I definitely want to try and capitalize on junkyard parts as much as possible. Junkyard EV is the theme of my project. I am willing to try buck and boost type of arrangements in order to make existing battery packs compatible with my system.

 

[jddcircuit]

I got the engine out. If I am going to use the entire prius transaxle inplace of the VW transaxle then I am going to have to convert my swing axle suspension over to Indepenent Rear Suspension. I didn't know what this was until I started taking things apart. It turns out that if I bought a model later than 1973 it would already have the suspension I wanted.

 

[jddcircuit]

I removed the swing axle transmission and torsion spring frame. I got a later model suspension frame and trailing arms from various different make and models. I painted and assembled the new suspension.