Communicating with the Lexus GS450h Inverter/Converter

[arturk]

Most of the pin-outs are gathered in the support thread:
viewtopic.php?f=14&t=396

I will try to come up with detailed diagram very soon.

The easiest way by far would be to use OEM wire loom and customize it to only use necessary wires. This is approach Damien is taking.

I do not have OEM wire loom so I am creating my own from scratch. My goal is to have all OEM (or compatible aftermarket) plugs for transmission connections. All plugs except for the gear shifter have been identified and are easy to obtain.
As far as inverter side, it is probably best to replace it, few have done it already. I am going to look for alternatives and post my findings.
For bench testing I connected to inverter using generic 3-way connectors. I will post some pictures tonight.

[arturk]

Does anyone have "07-11 Lexus GS450h Service & Repair Manual" PDFs to share?

In particular I am looking for the following chapters:
- HYBRID VEHICLE CONTROL
- HYBRID BATTERY CONTROL
- HYBRID TRANSMISSION / TRANSAXLE
- CAN COMMUNICATION
- AIR CONDITIONING

I have complete "2007 Toyota Camry Hybrid (AHV 40 series) Service & Repair Manual" if anyone is interested.

[jon volk]

What’s the effective combined Kw of both motors? It seems like only mg2 values are published at 147kw?

[sfk]

The MG1 is not used for propulsion in normal GS450H operation. Only as electrical generator taking power from the petrol engine or regenerator taking power from the drivetrain.

Perhaps it didn't make sense to Lexus to publish a power rating for MG1 for this reason?

[arturk]

Agree, I think they did not publish rating because it was not being used for propelling the car thus somewhat irrelevant for GS450h user.
Anyways, MG1 it is rated at 134 kW (180 HP).

Combined: 134 kW +147 kW = 281 kW (376 HP)

[Kevin Sharpe]

What’s the effective combined Kw of both motors?

iirc Damien believes 220kW (300HP) is possible. If everything goes well we will put the Super Yacht on the dyno in the Spring

[sfk]

That's really quite a substantial drivetrain if you work them together. It's covered in other posts on this forum in more detail but in short; you can lock the input shaft and use MG1 for propulsion but because of the ratios it spins at a higher RPM than MG2.

[jon volk]

That’s a pretty impressive power per dollar value.

[Dre.esky]

Agree, I think they did not publish rating because it was not being used for propelling the car thus somewhat irrelevant for GS450h user.
Anyways, MG1 it is rated at 134 kW (180 HP).

Combined: 134 kW +147 kW = 281 kW (376 HP)

That’s impressive but do we have to run the transmission at 650V to get that power out of it

[Kevin Sharpe]

That’s impressive but do we have to run the transmission at 650V to get that power out of it

The Super Yacht will be tested with a ~450V battery initially. If we then need/want to use ~600V the proposal is to use the converter in the Lexus inverter to buck/boost as required by the charger, DCDC, etc.

Obviously, a ~600V battery could potentially support 120kW CCS charging

[arturk]

Short video with overview of connections between components.
Apologize for annoying inverter whine noise, it is worst than I anticipated.

I know Damian was going to make one in the future but since he is so busy I figured I will take a crack at it.

[sfk]

Thanks for this video. All of these discoveries are a great resource for us all. Please keep making them!

[Dre.esky]

Great video good luck with the progress

[Kevin Sharpe]

I know Damian was going to make one in the future but since he is so busy I figured I will take a crack at it.

I'm sure Damien will be thrilled that you're doing this work... thank you

[Jack Bauer]

Damien is indeed delighted. Thank you very much for the contribution and It's great to see another controller up and running. I'll be rolling in the pull down resistor mod into the next version.

[arturk]

I'm sure Damien will be thrilled that you're doing this work... thank you

Damien is indeed delighted. Thank you very much for the contribution and It's great to see another controller up and running. I'll be rolling in the pull down resistor mod into the next version.

Guys, I am trying to give back to this fantastic community.
Thank you for your encouragement!

[johu]

So this also means inverter and motor don't have to be matched, right?

[arturk]

I hope this is the case.

Both Camry and GS450h inverters are almost identical including their control boards which are exactly the same (down to Toyota Part Number).

Based on my testing basic functionality that is currently implemented in Damien's VCU works.
However when I was trying to take it to implement REVERSE gear functionality, I was unsuccessful with that.

Now, the question is whether it is due to us not fully understanding how to control those inverters (not having all the HTM bytes figured out).
Or is it because Camry inverter is using slightly different HTM codes due to its firmware.

This is why I am anxiously waiting for someone with Lexus inverter to confirm additional capabilities (such as working REVERSE gear functinality).
Also, I noticed that ones motors start sniping down as result of letting off the throttle, I should not be pressing on the throttle until motors stop completely. Otherwise entire transmissions starts shaking really badly util motors stop and then they will start sniping back up.

Hopefully Damien or xp677 will be able report their results very soon.

Update: Corrected grammar

[jnsaff]

iirc Damien believes 220kW (300HP) is possible. If everything goes well we will put the Super Yacht on the dyno in the Spring

According to the tests here https://www.osti.gov/servlets/purl/947393

Motor efficiencies are above 90% for a great portion of the operation range and the 18 second power capability of the primary LS 600h motor is about 110 kW at 4,500 rpm.

The setup to run at high power levels probably need some serious cooling as with stock cooling the 50kw steady power level overheated the motor in about 12 minutes for 25kw it was about an hour. They did not test MG1 but there are a few interesting data points to consider.

[Kevin Sharpe]

They did not test MG1 but there are a few interesting data points to consider.

One of the great things about converting a car, driving it on the road, and testing on the dyno is that you learn what happens in the real world

Lots of unknowns about reusing hybrid components but I have no doubt Damien would be a very rich man if he received €1 from everyone who told him "you're doing it wrong"

[jnsaff]

They did not test MG1 but there are a few interesting data points to consider.

One of the great things about converting a car, driving it on the road, and testing on the dyno is that you learn what happens in the real world

Lots of unknowns about reusing hybrid components but I have no doubt Damien would be a very rich man if he received €1 from everyone who told him "you're doing it wrong"

Ha, I defos don't want to be one of those guys but I found this study a fascinating read.

[arturk]

The setup to run at high power levels probably need some serious cooling as with stock cooling the 50kw steady power level overheated the motor in about 12 minutes for 25kw it was about an hour. They did not test MG1 but there are a few interesting data points to consider.

Jnsaff, very good point. I am familiar wit this paper and it is great source of information.
Definitely, cooling has to be adequate to application.

We need to keep in mind that very high load will be experienced only under heavy acceleration for brief period of time.
As far as highway cursing demand is much smaller.

Here is an example of Tesla Model 3 Standard Range (I know I am comparing it with most efficient EV on the planet).
Battery: 50 kWH
Range: 220 miles
Crusing @60 mph it will take roughly 3.6 hour
Which means average power needed is ~13.6 kW

This is just rough estimate but I think it gives us good idea.

[xp677]

Hopefully Damien or xp677 will be able report their results very soon.

My system is currently back in pieces for mounting into the car, I'm also revising my control PCB. My previous work was just a proof of concept. Hoping to have the motor running at 600v in the car within 3 months.

Reverse should work, make sure your accelerator pedal function is returning negative torque when in reverse. I also had a line to disable MG1 in reverse.

Damiens code differs from mine in this regard, my code has a load of accelerator pedal checking and mapping, as well as gear selection, engine braking, creep, and cruise control, which is specific to my project, so did not make it into Damiens version of the code.

I'm currently working on a digital dash to complement this controller, based on cheap ebay LCD screens.

And also loads of other side projects including an intelligent cooling system, AC controller, BCM for the car, etc.

[arturk]

I appreciate response.

Reverse should work, make sure your accelerator pedal function is returning negative torque when in reverse. I also had a line to disable MG1 in reverse.

I did try returning negative torque and I could not get motors sniping in the other direction (they were making noise however). I tried to disable MG1 as well, and when I do, MG2 stops responding really weird.
Perhaps I am doing something wrong although it does not seem to be complicated.
I fear that there could be an issue with my Camry Inverter however my understanding is that reverse engineering was done using Camry, RAV4 and GS450/600 so I am hoping this is just silly mistake on my part.

Anyways, some of the features you describing (gear selection, engine braking, creep, and cruise control) sound interesting to me as well.
Would you share some of your code with me so I could try to isolate problem?
I do realize more tweaking is need to get everything running smoothly but I really need to be able to move my project car in an out of garage since it is too crowded. For this I need reverse

Your dash looks really great, nice work on all of those icons!

[xp677]

All of that code is currently untested, I wrote it after disassembling my test setup, using the Arduino serial monitor to preview the results, so I have no idea if it works in reality. My practical results are the same as shown in your video, the code which I shared with Damien for his controller was based on that.

Once I have my code tested, I'll be releasing the updates. As you can see from the serial readout, I was still in the middle of deciphering the MTH stream when I shared the code.

Reverse should work with negative torque values, that's how it is in my GS450h. Maybe there's an issue with signed integers being converted to the HTM data.