openinverter forum

** posts moved from thread "Communicating with the Lexus GS450h Inverter/Converter" **

Awesome work guys. I checked the source file gs450h_v3.ino to see what information is actually needed to drive it.
And it turns out, the only thing that is communicated towards the inverter are torque numbers.
Negative and positive depending on direction and which motor. Torque of 0 will be free/neutral I assume?

It looks like (at least from the source file) nobody is using the second planetary gear box? And shouldn't it be controlled by at least braking one of the wheels so it stays in one gear?

What I am wondering is how Lexus/Toyota was using torque in their application with the 2GR petrol engine in there.
I can see how you can control MG1 with torque, keep adjusting it as the engine keeps revving.
Neutral would be just to put MG1 with a torque of 0 to have it free running.
But I believe they also had a way to "lock it up", meaning that you can do engine braking on the petrol engine, when the battery pack was full, and you are still going of a hill.
I don't see how they would achieve that with having only torque numbers. Do they keep adjusting the torque numbers to keep MG1 at the same speed of MG2?
Probably I am overlooking the proper way to do this.

Swing wrote: ↑Mon Jan 27, 2020 3:56 pm Do they keep adjusting the torque numbers to keep MG1 at the same speed of MG2?

They're locked... have you seen this?

Would doing the same on a prius gearbox avoid having to weld the power split device?

Yes but the oil pump is driven off the engine, so the gearbox would receive and lubrication. Also mg1 would have a high ratio to mg2, so mg1 could over speed. Welding the gears would turn the oil pump.

Lexus has an external oil pump, so this isn’t an issue as such.

Kevin Sharpe wrote: ↑Mon Jan 27, 2020 6:02 pm

Swing wrote: ↑Mon Jan 27, 2020 3:56 pm Do they keep adjusting the torque numbers to keep MG1 at the same speed of MG2?

They're locked... have you seen this?

Sorry, I was very unclear.

I am aware that people are locking/fixing the center shaft with the petrol engine removed and pure EV driven.
I meant the situation where the petrol engine is still there.

As this car is coming down in price, I think the car as a whole is interesting to use.
That could be interesting in two ways. One where you chuck out the engine, saving 200KG and some space, good for placing batteries. Convert it to full EV with help of the board as discussed in this topic.
The other one is converting it to a plugin hybrid, and keeping many functions of the car. This has several advantages (and also some disadvantages)
It could do with a smaller battery, which is cheaper to purchase. And with that uprated battery (and cabling!) you could supply enough amps to keep MG1 and MG2 busy and add in the power of the 2GR petrol engine. That is some serious amount of power! Peak output could be over 600hp
And in most cases you can just drive short distances purely electric with more than enough power.

I am thinking to piggybacking the interface between hybrid ECU and inverter.
The first GS450h also has the advantage that the hybrid ECU and engine ECU are seperate. This gives the advantage of controlling or shutting off the petrol engine whenever you wish.

I am aware of what needs to be done on the battery side and the function of the original BMS that you will need to replace. That is a seperate topic.

What you are talking about is more like putting a different "skin" on a GS450h. Which is an interesting idea that would either be A) much simpler or B) much more complex than using just selected mix and match components.

Early on a few people looked at the feasibility of transplanting an entire Nissan Leaf drivetrain, battery and wiring loom into a different auto platform. Some got as far as pulling it all out and running it on the bench and the minimum requirements were pretty much every component. On the whole it's thought to be more complex than is worth the effort due to the interlocking of various safety and security systems. They seem too fully integrated to separate or selectively disable. This could be the case for the GS450h drivetrain also. The difference here is that the PCM is almost surely more tricky because of the synchronisation of both electric and ICE sides of the powertrain.

Definitely worth investigating though.

Swing wrote: ↑Mon Jan 27, 2020 9:10 pm One where you chuck out the engine, saving 200KG and some space, good for placing batteries. Convert it to full EV with help of the board as discussed in this topic.

This Lexus cost us £500 and the ICE, Exhaust, CATs, and fuel tank are worth £2000. With 250kW CCS rapid charging and >60kWh battery it will make a very simple and capable EV that can be converted in a few days

sfk wrote: ↑Mon Jan 27, 2020 9:27 pm What you are talking about is more like putting a different "skin" on a GS450h.

The difference here is that the PCM is almost surely more tricky because of the synchronisation of both electric and ICE sides of the powertrain.

I explicitly mean using the GS450h as a whole car, not transplanting anything. It is the lazy approach

However, when keeping the petrol engine, it becomes more complex.

Kevin Sharpe wrote: ↑Mon Jan 27, 2020 10:15 pm

Swing wrote: ↑Mon Jan 27, 2020 9:10 pm One where you chuck out the engine, saving 200KG and some space, good for placing batteries. Convert it to full EV with help of the board as discussed in this topic.

This Lexus cost us £500 and the ICE, Exhaust, CATs, and fuel tank are worth £2000. With 250kW CCS rapid charging and >60kWh battery it will make a very simple and capable EV that can be converted in a few days

Yes although I think it will be unrealistic to assume 500 quid as market value. If have seen them here for 5000 euro with broken battery light, left hand drive. Maybe someone will come by cheaper if you have time. Still, it is a lot of hardware for little money.
So that is plan A, where you spend some money on a nice battery pack. But all in all, it is pretty easy given that you can buy the developed board for this and just make some adjustments based on your battery pack.

However, I am very interested in plan B, because it could be more fun and cheaper.
I am thinking about a small battery pack, so much less costs, but which can deliver the amps. At the location of original hybrid battery pack.
And use maximum power of engine and motors at the same time.

However, I don't understand yet how to treat MG1 if the only input is torque. You have to give torque input to make it (for example) go as fast as MG2, so that you can lock the motor to MG2 speed (this would be for higher speeds).
I guess for the CVT mechanism they just keep adding negative torque to make MG1 generate power and keep the petrol engine at ideal revs. So it is probably constantly adjusted based on the engine speed input.

In general, I don't really understand if you could have, for example, MG2 lockup completely. I think if torque is the only measure, you can put in an enourmous amount of torque, but if it then still turns, you get a crazy amount of power back which you probably cannot handle. Or voltage may become too high, something like that.

Next to the discussion above, I still have two points. Is it correct you guys are not using the second planetary gearbox that would allow for a higher top speed? Shouldn't you at least fix one of the two gears upon setup? I don't see that in the code.

If I understood the car correctly, there is some kind of power measurement device on the end of the gearbox that measures the actual given power. It is probably also used to feed the dashboard gauge that is showing KW up to 250KW.
Did anybody use that? I think the info should be given somewhere on this message interface, but I could be wrong.
It is a great thing though, because you don't have to visit a dyno, it is like a built in dyno

Swing wrote: ↑Mon Jan 27, 2020 9:10 pm And with that uprated battery (and cabling!) you could supply enough amps to keep MG1 and MG2 busy and add in the power of the 2GR petrol engine. That is some serious amount of power! Peak output could be over 600hp

Bear in mind that the 288-650v converter is rated to around 30kw max, so for 600hp you would need a 650v pack to supply the motors directly.

Swing wrote: ↑Tue Jan 28, 2020 9:20 am Next to the discussion above, I still have two points. Is it correct you guys are not using the second planetary gearbox that would allow for a higher top speed? Shouldn't you at least fix one of the two gears upon setup? I don't see that in the code.

Default (neither shift solenoid energised) is high gear, seems suitable for EV use. I haven't looked at the shifting function yet, I doubt I will as I don't see a need - the torque output in high gear is already going to be putting strain on OEM differentials.

Swing wrote: ↑Tue Jan 28, 2020 9:20 am If I understood the car correctly, there is some kind of power measurement device on the end of the gearbox that measures the actual given power. It is probably also used to feed the dashboard gauge that is showing KW up to 250KW.
Did anybody use that? I think the info should be given somewhere on this message interface, but I could be wrong.
It is a great thing though, because you don't have to visit a dyno, it is like a built in dyno

There are current transformers on two phases of each motor. Feedback must be from these, as there is nothing else to perform that function.

xp677 wrote: ↑Thu Jan 30, 2020 2:18 am

Swing wrote: ↑Mon Jan 27, 2020 9:10 pm And with that uprated battery (and cabling!) you could supply enough amps to keep MG1 and MG2 busy and add in the power of the 2GR petrol engine. That is some serious amount of power! Peak output could be over 600hp

Bear in mind that the 288-650v converter is rated to around 30kw max, so for 600hp you would need a 650v pack to supply the motors directly.

Do you mean you would need 650V to get 600hp out of the drivetrain in pure electric alone?
I think the OP was implying that transplanting the whole hybrid system could yield 600hp if the ICE was involved in addition.

sfk wrote: ↑Mon Jan 27, 2020 9:27 pm They seem too fully integrated to separate or selectively disable. This could be the case for the GS450h drivetrain also. The difference here is that the PCM is almost surely more tricky because of the synchronisation of both electric and ICE sides of the powertrain.

Yes, this is the case for the GS. You could likely expand the battery pack in parallel without it noticing (you would need your own BMS), or remove all the control hardware from the Lexus and use your own PCM. And hope that dash data, etc is over CAN (I believe it is), and sniff that first.

sfk wrote: ↑Thu Jan 30, 2020 2:23 am Do you mean you would need 650V to get 600hp out of the drivetrain in pure electric alone?
I think the OP was implying that transplanting the whole hybrid system could yield 600hp if the ICE was involved in addition.

No, I'm saying that you can only get max 30kw from the motors unless you bypass the buck/boost converter. You can stick whatever voltage you like into the motors, they are rated for up to 650v.

In pure electric mode, the motors are rated for a combined 250kW (untested). So that's 335hp from the motors. Of course, you can probably exceed this.

Swing wrote: ↑Tue Jan 28, 2020 9:15 am I explicitly mean using the GS450h as a whole car, not transplanting anything. It is the lazy approach

I guess you could do some stuff to "enhance" a standard GS450h. Like swapping to lithium battery chemistry, giving up interior space to add extra battery packs in various parts of the car, adding a charging ability and "overclocking" the electric motors or forcing the car into pure EV mode more frequently.

Doesn't solve the problem of it being ugly though

The battery (and especially the BMS, which is flawed in some cars, leading to premature battery failure) is a weak point of the car. Adding some extra kWh wouldn't be a bad thing either.

However, the engine triggers at around 5-10mph, making an "EV" mode difficult to achieve on the "standard" Lexus hardware.

Thanks for the answers.

As a background: I am an embedded software engineer, unfortunately not so experienced in electronics.
I have built an electric scooter before using Mitsubishi cells and configuring a BMS.
And I have been studying the Prius/Lexus hybrid cars.

xp677 wrote: ↑Thu Jan 30, 2020 2:18 am Bear in mind that the 288-650v converter is rated to around 30kw max, so for 600hp you would need a 650v pack to supply the motors directly.

Ah yes, obviously it could not be that easy with only some controls.
But it makes sense, as MG1 feeding MG2 is where most of the power is going, and this is all on the ~600v plain, after the booster. I could have seen this coming. I will study pictures of the inverter to see what is needed.

But it also means my idea of using a BMW i3 battery is not suitable, unless I would fill the complete trunk, and I don't want that because of weight and safety. I need way more in series, so more compact cells.

xp677 wrote: ↑Thu Jan 30, 2020 2:37 am The battery (and especially the BMS, which is flawed in some cars, leading to premature battery failure) is a weak point of the car. Adding some extra kWh wouldn't be a bad thing either.

Yes that is my initial idea. Make it into a (powerful) plugin hybrid. Phase one would be replacing the pack+BMS, and keeping the rest of the car satisfied with BMS messages (SOC communication controls if the car will use battery or not)

However, the engine triggers at around 5-10mph, making an "EV" mode difficult to achieve on the "standard" Lexus hardware.

Once you control the inverter like you guys are doing (and yes that will require 600v battery pack to get the heavy car moving a bit, not being limited to 30kw, it should not be to hard.

The car has two ECUs. Later generations changed to 1 ECU, which is an issue to modify.
With 2 ECUs, one is the engine ECU and the other the hybrid ECU.
The Hybrid ECU is the boss, but you can fool it. For example: Saying that the engine is hot enough, means it will not run the engine if not needed.
Some more advanced stuff will be needed to run it at higher speed / power levels, and faking the turning engine.
And there is an out-of-gas-mode that you can use.

This would all be a bunch of man-in-the-middle attacks to get it running. While it will require time, you are working with a complete working car, so most of the work is already done. Should still be easier than building a car from scratch.
Making it into a full EV could be easier than PHEV, but it will require quite some money for batteries to give the car a usable range. Because of course, it is a big heavy car.
The PHEV way is also the way to a very powerful car.

Combining the forces is where the challenge is, especially in controlling MG1.
But with a working car you could learn from tracing what the ECU is doing with MG1 to control / put a load on the engine.

Swing wrote: ↑Thu Jan 30, 2020 8:41 am But it also means my idea of using a BMW i3 battery is not suitable, unless I would fill the complete trunk, and I don't want that because of weight and safety. I need way more in series, so more compact cells.

Damien is road testing BEXUS at ~480V Why not wait a couple of weeks and see how well that performs?

Swing wrote: ↑Mon Jan 27, 2020 9:10 pm The other one is converting it to a plugin hybrid, and keeping many functions of the car.

In theory you could replace sun gear and satellite carrier with solid piece compact enough to bearing being placed on it, wich will support outer piece replacing ring gear. This outer piece will tie together MG2, oil pump and output shaft.
So engine will be connected to MG1 only (you can replace engine with smaller one) and will work as generator only to provide power thrue the inverter to MG2 (same scheme as on chevy volt)
I bet a good lathe operator could fabricate such two pieces. The only thing you need is to find bearing that will fit.

Kevin Sharpe wrote: ↑Thu Jan 30, 2020 5:57 pm

Swing wrote: ↑Thu Jan 30, 2020 8:41 am I need way more in series, so more compact cells.

Damien is road testing BEXUS at ~480V Why not wait a couple of weeks and see how well that performs?

Yes it will be interesting. And I am not moving that fast anyways.
But what I meant is the amount of cells needed to reach the 600v. That is okay in a pure EV, but challenging in a PHEV, while.reusing cells from other EVs. And then preferably the ones that are liquid coolable. Anyway, we will see.
I am trying to identify the challenges so thats great.

konstantin8818 wrote: ↑Thu Jan 30, 2020 8:27 pm
In theory you could replace sun gear and satellite carrier with solid piece compact enough to bearing being placed on it, wich will support outer piece replacing ring gear. This outer piece will ...

I like the idea but it will not allow for all options like the factory setting does.
In your line of thinking, I could also weld the whole thing together, which will make it easier to control. Just enabling the petrol engine whenever it reaches above 1000rpm. But there will be lots of disadvantages too haha. Always dragging the engine along, those kind of losses.

To over come the 30kw limit of the buck boost converter. Could you take two inverters and parallel just the buck boost converters to give 60kw . Then the other mg1 output of the second inverter could be used as a charger. Therefore allowing lower voltage packs to deliver more power and have the convenience of charger there too?