GS450h with 3 motors?

[drprox]

This thread is aimed at understanding if it is possible to add a 3rd motor to the GS450h hybrid drive system.
• To achieve an overall power output of ~300kW using low cost OEM parts.
• Because even if I had a single 300kW motor it would still need a transmission, which in the case of an RWD car would likely be a stock manual transmission which may not handle the torque. I believe the Lexus system would.



I understand the basic principles:
• Ignoring the 2 stage reduction, MG2 is essentially direct drive to the output shaft.
• MG1 can provide power to the output shaft via a reduction of 2.6:1 if the engine is stationary.
• The Engine can provide power by spinning the carrier.
• MG1 can vary the output speed of the ring gear by spinning at different speeds.

Damiens setup:
• Engine input is welded ie. Stationary.

Motor specs:
• MG2 max power is ~110kW
• MG2 power will drop off at ??? rpm @ 650v
• MG2 max continuous power is ~25kW

• MG1 max rpm is ???
• MG2 max power is ???

Questions:
1. How much torque is required to hold the carrier? Surely it can’t be a lot else the engine would just start turning when in EV mode, unless it has a mechanical lock, or the motor itself acts as a lock by electrically ‘holding’?

2. If the Engine is replaced with a motor and was providing power in addition to MG1, would it essentially just add more power, or would MG1 require special control to ensure it doesn’t start generating instead and blow the inverter?

3. If the above is not possible, then could the whole ring / sun / carrier assembly be welded up and this would work but MG1 would be spinning slower than intended?

[slow67]

You would have to synchronize MG1 and the added motor. You are better off welding up the power split device. That’s my plan.

[Kevin Sharpe]

You are better off welding up the power split device. That’s my plan.

With the GS450H CVT you don't need to weld the PSD... just fit a splined collar over the input shaft and bolt that to the bell housing

[sfk]

Yes I understand your question and have thought about this myself.

What you are talking about is simply swapping the ICE for another electric motor, probably AC, and using all of them for drive output.

This would work but I don't understand the purpose unless you were building a drag car or something. The output figures for the L110 hybrid transmission published by Toyota are not necessarily the limits of the motors. Adding another motor would make things more complicated and you would lose the space in the engine bay that could be used for batteries or other components.

You would need another inverter too. The Lexus RX has 3 inverters (2 for front transaxle and 1 for rear) but not sure it would have the capacity for output you are after.

Without locking up (welding probably) the planetary gears I suspect the motor with the lowest torque would lose the fight and slip. It wouldn't be MG2 as is directly coupled to the wheels, it would be the weaker of MG1 and MGiceReplacement.

From what I understand in normal operation MG1 is never used to add torque directly to the driveline (except negative torque for regenerative braking). It only starts the ICE when stopped and takes torque from the ICE when running to generate power to send to MG2 for actual driving.

Looking at the diagram you can see there are no clutches or brakes to lock out the ICE or MG1 so would be hard to isolate their functions independently. ie stopping MG1 from spinning while using ICE without adding stationary torque to MG1 etc.

ICE can spin the carrier but only to spin MG1 which generates power to feed MG2.

In EV mode MG2 drives the wheels, ICE is stationary and MG1 freewheels backwards. MG1 doesn't contribute torque.

Max safe speed of the motors is around 10,000rpm I think. And welding the planetary gears would reduce the speed of MG1 which is a good thing.

The way the power split device works seems limited and illogical to me. Why wouldn't you want to be able to have MG1 and the engine provide torque directly? But I have read it works this way in many documents from Lexus and Toyota so have to accept it does.

[slow67]

I am building a drag car, so that’s why I wont lock the input shaft. I’ll weld the PSD in mine so MG1 and the 3rd motor wont fight.

[sfk]

Sounds awesome!

[MattsAwesomeStuff]

In EV mode MG2 drives the wheels, ICE is stationary and MG2 freewheels backwards. MG1 doesn't contribute torque.

"MG1 freewheels backwards" you meant. Just for archive's sake.

Max safe speed of the motors is around 10,000rpm I think

I haven't seen data on the Lexus, but for the Prius:

Gen 1 = 6,500 rpm
Gen 2 = 10,000 rpm
Gen 3 = 13,500 rpm
Gen 4 = 17,000 rpm

The numbers might be the same for the Lexus, or, they might at least follow a similar progression.

[sfk]

In EV mode MG2 drives the wheels, ICE is stationary and MG2 freewheels backwards. MG1 doesn't contribute torque.

"MG1 freewheels backwards" you meant. Just for archive's sake.

Max safe speed of the motors is around 10,000rpm I think

I haven't seen data on the Lexus, but for the Prius:

Gen 1 = 6,500 rpm
Gen 2 = 10,000 rpm
Gen 3 = 13,500 rpm
Gen 4 = 17,000 rpm

The numbers might be the same for the Lexus, or, they might at least follow a similar progression.

Yes. I meant MG1 freewheels backwards. Thanks. I edited my response to prevent confusion.

The max rpm depends on which motor on which gearbox you mean. But generally the max rpm seems to climb with each new iteration. 10,000rpm is a safe estimate for any Toyota/Lexus MG you might be considering.

[mdrobnak]

When you say freewheels backwards, in a stock (with ICE) situation, does that mean it's running the ICE backwards? Apologies if this is something elementary that I'm missing.

-Matt

[slow67]

When you say freewheels backwards, in a stock (with ICE) situation, does that mean it's running the ICE backwards? Apologies if this is something elementary that I'm missing.

-Matt

When the ICE is not running, MG1 is spinning backwards at 2.6x the speed of the output shaft.

[drprox]

Thanks for the replies. I think I am getting a better understanding now.

I'm going to assume that MG1 freewheels 'under inverter control' to ensure that 0Nm torque is applied to the ICE and the back EMF is balanced out? In which case I think the welding idea is probably best.

I've seen a few figures floating around for the power output of this transmission in the region of 250kW, but I don't know where that came from as I suspect it is more likely around 110kW for MG2 and maybe half that for MG1, so perhaps ~160kW.

I'm not building a drag car but I like performance cars so my goal is to push the boundaries a bit.
It will be a 5 series BMW, but I would like it to be able to match a high-end model (say a 535i).

Perhaps I am underestimating this transmission.

[Kevin Sharpe]

Perhaps I am underestimating this transmission.

BEXUS road, track, and dyno testing is about to start. I'd recommend waiting for the results of that before making any final decisions on the way forward

[Jack Bauer]

It will be a 5 series BMW, but I would like it to be able to match a high-end model (say a 535i).

Perhaps I am underestimating this transmission.

E39??

[drprox]

E61

[GaryClarke]

When you say freewheels backwards, in a stock (with ICE) situation, does that mean it's running the ICE backwards? Apologies if this is something elementary that I'm missing.

-Matt

As a hybrid, to an observer standing outside the transaxle, MG1 never spins in reverse. If MG2 spins clockwise then MG1 always spins clockwise. MG1 never spins anti clockwise.

MG1 and MG2 are not locked together. Both can speed up and slow down independently of each other. When MG1 slows down but MG2 does not, the change in relative speeds make it appear that MG1 is spinning in reverse. But it does not change its direction of rotation. It is slowing while MG2 does not. It is the differential gearing of the PSD that allows this to happen.

[mdrobnak]

When you say freewheels backwards, in a stock (with ICE) situation, does that mean it's running the ICE backwards? Apologies if this is something elementary that I'm missing.

-Matt

As a hybrid, to an observer standing outside the transaxle, MG1 never spins in reverse. If MG2 spins clockwise then MG1 always spins clockwise. MG1 never spins anti clockwise.

MG1 and MG2 are not locked together. Both can speed up and slow down independently of each other. When MG1 slows down but MG2 does not, the change in relative speeds make it appear that MG1 is spinning in reverse. But it does not change its direction of rotation. It is slowing while MG2 does not. It is the differential gearing of the PSD that allows this to happen.

Thank you for explaining this. My mechanical knowledge is significantly weaker than my electronics knowledge, and I sometimes have a hard time visualizing these things. My concern was that since the BMW S54 engine is an 'interference' style engine that spinning it backwards would be a big problem. As it seems like that is not actually the case, but simply an appearance due to the speed differential, then my idea is not unfeasible (uneconomical / impractical, perhaps).

Thanks,
-Matt

[sfk]

As a hybrid, to an observer standing outside the transaxle, MG1 never spins in reverse. If MG2 spins clockwise then MG1 always spins clockwise. MG1 never spins anti clockwise.

If you look at the way that the MG1, MG2 and ICE are mechanically connected together through the gearing, if the ICE was stationary (not running) and MG2 was turning forward (to drive the vehicle forward), it would be impossible for MG1 to not be spinning in the opposite direction to MG2.

[GaryClarke]

As a hybrid, to an observer standing outside the transaxle, MG1 never spins in reverse. If MG2 spins clockwise then MG1 always spins clockwise. MG1 never spins anti clockwise.

If you look at the way that the MG1, MG2 and ICE are mechanically connected together through the gearing, if the ICE was stationary (not running) and MG2 was turning forward (to drive the vehicle forward), it would be impossible for MG1 to not be spinning in the opposite direction to MG2.

Just when I thought I'd got my head around this. You are right of course, and I said the same myself in this post

Only in hybrid use could MG1 over spin. The gearing of the PSD means that MG1 would spin 2.61 times faster than MG1 - in the opposite direction.

However, I think the gen 3 transaxle, which I've been focusing on recently, with it's MG2 speed reduction gearing, is the exception. In this case MG1 never spins backwards. Which I think means that it can be used as a full time EV without any welding of PSD or immobilising of the ICE input shaft. Am I right on this?

[sfk]

I assume when you refer to Gen3 you mean Prius?
I'd have to look at the schematics for the layout of the Prius transaxle to be able to say how that works.
I was speaking about the L110 in the GS which is what this thread is about

[mdrobnak]

As a hybrid, to an observer standing outside the transaxle, MG1 never spins in reverse. If MG2 spins clockwise then MG1 always spins clockwise. MG1 never spins anti clockwise.

If you look at the way that the MG1, MG2 and ICE are mechanically connected together through the gearing, if the ICE was stationary (not running) and MG2 was turning forward (to drive the vehicle forward), it would be impossible for MG1 to not be spinning in the opposite direction to MG2.

I guess the appropriate follow-up question is, with how much torque? If connected to an ICE, you're going to need to overcome friction, which is going to be 50+Nm to get it to move.

-Matt

[sfk]

I don't understand your question.
How much torque in respect to what?

MG2 requires as much torque as it takes to get your car moving.
MG1 is not producing any torque and, if the bearings are any good, using nearly zero torque to reactively spin.
ICE is not running and not requiring torque.

Are you asking about torque required to start the ICE?

sk

[mdrobnak]

I don't understand your question.
How much torque in respect to what?

MG2 requires as much torque as it takes to get your car moving.
MG1 is not producing any torque and, if the bearings are any good, using nearly zero torque to reactively spin.
ICE is not running and not requiring torque.

Are you asking about torque required to start the ICE?

sk

If you're asking for torque out of MG2, you are saying MG1 will spin 2.6x the speed of MG2 in the opposite direction. My question is, if there is something attached that will require a force to turn (eg an engine), will MG1 be able to overcome that force and turn the engine? Or will the engine stay stationary?

-Matt

[drprox]

I am building a drag car, so that’s why I wont lock the input shaft. I’ll weld the PSD in mine so MG1 and the 3rd motor wont fight.

Do you have any motors in mind for MG3?

[sfk]

MG1 can and will overcome the torque required to turn the ICE. That is one of it's main functions - to start the ICE in place of a starter motor.

[mdrobnak]

MG1 can and will overcome the torque required to turn the ICE. That is one of it's main functions - to start the ICE in place of a starter motor.

Understood -if actively commanding MG1 to do so. I'm asking in the situation of when you're not trying to regen brake or start the engine.