anybody work on the LS600h inverter?

[Kevin Sharpe]

I wouldn't attribute this to malice.

I think the intension is clear;

With Bigmouse his help we got the transmission / inverter running with our own control module.

[Isaac96]

Right, it's not exactly the nicest when people don't share what they've discovered or built.
However, it's a universal tendency -- once something's done, you no longer need help with it, so there is no real incentive to post the solution. This shows up on every forum I've ever been on (all 4 of them).

I hope we can encourage people to share solutions more often, but I wouldn't go so far as to suggest banning those who haven't gotten around to it yet.

-Isaac

[Kevin Sharpe]

I hope we can encourage people to share solutions more often, but I wouldn't go so far as to suggest banning those who haven't gotten around to it yet.

We started this forum with the intension of doing things differently. If you don't want to share then my advice is to go somewhere else. Feel free to start a thread to discuss this if you wish.

[Lwerewolf]

Just as a heads-up, I finally ordered an ls600h inverter, should be able to provide info in... 2-3 weeks, hopefully. Sorry for the delay, other things took precedence

A bit about the part numbers - g9200-500(11/21) - resp. (lhd/rhd) don't show up on toyodiy or the EPC app, I assume they're a very early production model or something. I ordered a g9200-50011. Now to secure wiring before it arrives.

[storm4710]

Just as a heads-up, I finally ordered an ls600h inverter, should be able to provide info in... 2-3 weeks, hopefully. Sorry for the delay, other things took precedence

A bit about the part numbers - g9200-500(11/21) - resp. (lhd/rhd) don't show up on toyodiy or the EPC app, I assume they're a very early production model or something. I ordered a g9200-50011. Now to secure wiring before it arrives.

nice looking forward to it

[Lwerewolf]

Arrived today (well, yesterday), still haven't gotten around to wire it up. I think I'll end up rearranging the terminals on the stock wire harness, hopefully without breaking the harness-side inverter connector housing, as it's unobtainium, as others here have found out. Then it's a matter of connecting the MGs to the inverter, since the 450h lances "slide in", whereas the 600h (and most other Toyota hybrids) are a bolt-together affair. I have my old blown inverter laying around, so I can use the, uhh... connector and resolver box to get something to bolt on to, but I haven't decided on (and found out) where to get cables & crimp terminals from. At any rate, it'll be a "does it run" setup, but that does include highway speeds and full throttle applications, in order to check whether the engine maintains revs as before - the ls600h keeps the engine below redline until >200kph, whereas the GS hits redline at around 140 - yes, this is a 4th gen gs, but the 3rd gen displays pretty much the same behaviour - other than the redline being 6400rpm and being reached a few kms above - backwards rotation limit on MG1... so if I have to guess, the l110(f) transmission might not have enough power to transmit the ls600h engine (2ur-fse)'s output at lower MG2 RPMs.

Just for the record, the GS wiring (I posted the LS wiring in one of my first posts in this thread):

[Lwerewolf]

Opened the inverter in order to repin the 40pin connector, as the wire harness-side connector is, well... let's say that I uhh "broke" it without really breaking it, and I gained a glimpse of the insides . You have to insert your "SST" in the depin hole and push "downwards" to free a terminal from the retainer - this is after you've gotten the retainer at "half lock" (i.e. per-pin individual lock), of course. Repinning the inverter makes switching between the GS and LS inverters way easier and safer (no rearranging 37-ish terminals every time), anyways.

At any rate, put the inverter on top of the original inverter, connected only the 40pin connector (i.e. no MG/Battery/AC connected), put the car in "accessory" mode (everything but the HV system running) - "check hybrid system". Reader - "p0a0d". There's an interlock plug in the battery connector, whou would've thought to check . At any rate - shorted that, cleared the code, car stopped complaining. My laptop needs to charge so I used hybridassistant to check whatever I could - basically the IGBT temperatures for mg1 and mg2 - and they were at 20 and 21 deg. C respectively - same as my house's room temperature. Outside was 11deg. C, engine was 16deg. c, HV (the battery) was reported around that range too. I have "motordata - hybrid" as well, which can read out everything from the HV ECU, but I forgot to use that... ah well.

Next up - up-sizing some fuses (to match the ls600h wiring diagram), connecting the MG harness to the inverter somehow, plumbing coolant, connecting AC & battery, testing.

So far, I'm very hopeful

[storm4710]

Opened the inverter in order to repin the 40pin connector, as the wire harness-side connector is, well... let's say that I uhh "broke" it without really breaking it, and I gained a glimpse of the insides . You have to insert your "SST" in the depin hole and push "downwards" to free a terminal from the retainer - this is after you've gotten the retainer at "half lock" (i.e. per-pin individual lock), of course. Repinning the inverter makes switching between the GS and LS inverters way easier and safer (no rearranging 37-ish terminals every time), anyways.

At any rate, put the inverter on top of the original inverter, connected only the 40pin connector (i.e. no MG/Battery/AC connected), put the car in "accessory" mode (everything but the HV system running) - "check hybrid system". Reader - "p0a0d". There's an interlock plug in the battery connector, whou would've thought to check . At any rate - shorted that, cleared the code, car stopped complaining. My laptop needs to charge so I used hybridassistant to check whatever I could - basically the IGBT temperatures for mg1 and mg2 - and they were at 20 and 21 deg. C respectively - same as my house's room temperature. Outside was 11deg. C, engine was 16deg. c, HV (the battery) was reported around that range too. I have "motordata - hybrid" as well, which can read out everything from the HV ECU, but I forgot to use that... ah well.

Next up - up-sizing some fuses (to match the ls600h wiring diagram), connecting the MG harness to the inverter somehow, plumbing coolant, connecting AC & battery, testing.

So far, I'm very hopeful

nice work buddy

[Lwerewolf]

I guess this is the better place to post this (i.e. not the gs450h inverter temperature thread thingy).

At any rate, first try was a failure.

The inverter fits somewhat. Following notes are for LHD cars:

I used my old blown inverter's water inlet/outlet pipes (they're pressed into the main casting - just... "wiggle" them out) as 90deg. coolant hose joints, in order to use the otherwise stock coolant plumbing. The gs450h's inverter "cage" (the thing to which the power steering ECU/inverter is mounted to) obviously won't let the ls600h inverter fit (since it's a bit "wider"), but removing it and mounting the EPS ECU straight to the car's body does work (the ECU bolts to the chassis anyways - the bracket just has holes and a bit of support on one side). The fuel pipe heading into the HPFP gets pushed against said HPFP by the battery connector - solvable by either removing the noise insulation covers off the HPFP (they're stock on the ls600h, not stock but present on some 2GR-FSE-equipped cars - there's a TSB for "ticking" noise), or by rotating it - it's the rubber portion that is curved to almost a full circle, can be fitted to "curve" to the other side.

Inverter outputs when viewed head-on are W-U-V on both the ls600h and gs450h inverter - checked via techinfo.toyota.com, I also have an EU-spec ls600h manual. The difference is that the generator output is on the inside (relative to the center of the car) for the ls600h,and on the outside for the gs450h.

Battery and AC connectors are the same. The wiring diagram indicates that the AC's positive and negative poles are swapped, but I simply do not see how that can be the case, given that the assembly that has the battery & AC connectors, the AC fuse (450v 40a) and the connecting bus bars is the same between both inverters (checked) - either I'm missing something, or the LS (or maybe it's the GS?) wiring diagram is wrong on that account (EDIT: the GS diagram & DTC troubleshooting pages appear to be wrong in this regard - AC pinouts are reversed).

I used H07V-K 35mm2 cables with crimped copper terminals - 1 meter in length - to connect the inverter's terminals to the old blown inverter's MG connectors housing - I use it as an interface between the slide-in-style gs450h connectors and the bolt-on ls600h terminals. No shielding implemented.

What happpened - p0a78-306 and p0a7a-344 are set when the car attempts to "start the engine" (which is all the time when engine coolant temperature is <40deg. C - engine warmup, catalyst lightoff, etc). The "ready" light does light up for a second, then the DTCs are set right as the car attempts to crank the engine - the cranking itself sounds/looks just like when the car is set in "compression check mode" (i.e. disabled engine fuel supply and ignition).

p0a78-306 (for mg2) and p0a7a-344 (for mg1) both indicate "Motor(mg2)/Generator(mg1) torque execution monitoring malfunction" - "If the difference between the requested (MG1/MG2) torque and the actual (MG1/MG2) torque exceeds a predetermined value, the MG ECU determines that there is a malfunction in the execution or monitoring of the (MG1/MG2) torque. Then, the hybrid vehicle control ECU will illuminate the MIL and set a DTC."

The MG ECU is built into the inverter, from what I've read from the few internal presentations on the evolution of Toyota's hybrid system that I could find - there are a few more on scribd but they don't seem to want to take my money... ah well. My guess is that the MG ECU is the thing with which the HV ECU (the hybrid vehicle ECU - basically the orchestrator) is communicating via the HTM/MTH USART lines.

Unfortunately, neither DTCs' freeze frame data includes MGx exec/actual torque values, or RPMs... actually, there's just about nothing inverter related in the freeze frame data. Attempting to log the event manually wasn't much of a success - from what I gather, techstream's "snapshot" recording function doesn't run on the ECU itself, so I'm left with the relatively atrocious "few frames per second" sampling rate that I get with an adapter. Good news is that what I saw from that was promising - proper pre- and post-boost voltages are reported, motor/generator RPMs seem in line with what was expected (the engine doing maybe a single revolution and then stalling back-and-forth - tiny RPM fluctuations in mg2, somewhat larger in MG1), torque request/exec values seem "decent" too...

Not sure what to make of this. One possibility is the lack of shielding on the MG extension harness thing that I made - might be fooling the CTs, given that MG1 & MG2 cables are pretty much mixed with each other due to the way that the old inverter breakout box is arranged:
Top (original MG harness side): (MG2-W MG2-U MG2-V MG1-W MG1-U MG1-V)
Bottom (inverter internal side): (MG2-W MG1-W MG2-U MG1-U MG2-V MG1-V)
Another possibility is that I didn't repin something on the inverter connector properly - will recheck. Given the amount of care I gave that, I kind of doubt it, but we'll see.
Yet another possibility is being mislead by the wiring diagram (the AC pinout still comes to mind).
Could also be a different motor magnet/winding/etc configuration leading to "unexpected" (to the inverter) torque values.

Re: whether it works or not with the controller board - I'd say that there's a high probability that it does, given that it did at least attempt to spin the engine in the right direction

I think that's all I have to report for now, questions and insight are always welcome. Plan moving forward is to attempt to depin the MG 3 phase harness (seems possible) and convert it to a bolt terminal (nondestructively), check the wiring diagrams and the various DTC pages for every little thing again, etc, etc.

[Jack Bauer]

Thanks for the hard work. And for sharing the results.

[xp677]

The MG ECU is built into the inverter, from what I've read from the few internal presentations on the evolution of Toyota's hybrid system that I could find - there are a few more on scribd but they don't seem to want to take my money... ah well. My guess is that the MG ECU is the thing with which the HV ECU (the hybrid vehicle ECU - basically the orchestrator) is communicating via the HTM/MTH USART lines.

Correct. The "MG ECU" is the microcontroller on the control board within the inverter casing.

Unfortunately, neither DTCs' freeze frame data includes MGx exec/actual torque values, or RPMs... actually, there's just about nothing inverter related in the freeze frame data. Attempting to log the event manually wasn't much of a success - from what I gather, techstream's "snapshot" recording function doesn't run on the ECU itself

Techstream/TIT pulls data from the HVECU. At least on the GS450h, there is no "outside" communication possible via the synchronous serial port, the only communication allowed to the inverter is via the microcontroller in the HVECU.

The MTH packet for the GS includes torque, rpm, position, inverter temp, bus voltage, and some other data. The HVECU passes this on.


If you're looking to reverse engineer the serial data (to allow the inverter unit to be used in electric vehicles) and you have a complete car, the easiest way is to put a logic analyzer onto HTM-, MTH-, CLK-, REQ- wires, leave everything else alone, and drive the car as normal, while recording the data. I found it helpful to record a video of me doing this, so you can "line up" various events (accelerating, etc) with the data.

It's very likely that the data is near identical to the GS450h, so by using the code from the GS450h VCU, you can likely solve this problem from both ends!

I'm off work for the next few weeks, if you are able to send me a (understandably huge) data dump, I might be able to help out.

Edit: not sure if it's much help, or if it's been posted before, but here's a teardown/evaluation of this unit: https://www.osti.gov/servlets/purl/947393

[Lwerewolf]

I know that there's no (official ) way to get data from the serial lines externally, I thought that the "snapshot" feature was a time-limited (well, memory-limited) internal recording mode of the ECUs that can be communicated with - engine, HV, abs/vsc/trac, etc - basically internal logging, so that you don't rely on the obd2 adapter readout speed - the best one that I have (some OTC thing that can't run half the utilities but reads HV ECU data fast) has a refresh rate of about 4hz on HV ECU data, which is very insufficient for catching this condition

I've seen all the ORNL teardowns, the original diyelectriccar thread, I think all the gs450h discussions here (communicating, support thread, buck/boost, connectors, etc, etc), the service manuals, so on.

For context, I'm not trying to build an EV (yet), I'm trying to replace the inverter in a stock EU-spec LHD 3rd gen gs450h (gws191l) with an ls600h unit - they share the same transmission (well, one has the LF1A transfer case at the rear). Reasons are listed here:
viewtopic.php?f=14&t=776
tl:dr - anything without IGBTs with double-sided cooling seems to be blowing up a lot (3rd gen gs450h, rx400h, ct200h, hs250h, pre-gen4 prius, etc), whereas I've only found somewhat obscure evicence of one (rx450h) inverter failure of the newer design (started with ls600h, then 450h, then 4th gen gs450h/gs/is/rc300h, etc). More details on why mine is not in a great shape - in the thread linked above, I've gone way too off-topic

At any rate, next up is connecting the MG cables to the 450h inverter the same way that I did to the 600h - need to fabricate "blade to screw terminal" copper adapters for that. Hopefully that will net me the same codes. After that, I guess I'll have to fabricate a shielded "adapter" from the 450h to the ls600h inverter - the blades are 13mm x 2mm, haven't found anything like that for sale anywhere - Yazaki does list something mostly identical, but the blades are 9.5mm in width. Guess I'll have to fabricate female connectors out of bent copper pipes or something.

[xp677]

So you'll be using the HVECU from the LS600h as well, or the original GS450h unit? Either way, you're in for a headache. Let me know your plans, it sounds interesting.

[Lwerewolf]

gs450h HV ECU. There are DTCs for Engine and HV ECU mismatch in the manual, plus the ls600h's engine has a different powerband and is resp. kept at different revs, so I'm guessing that an ls600h HV ECU won't work - this assumes that the ECU takes care of orchestrating the CVT aspects.

Plans are as I noted above - make a proper "adapter" for the GS harness to connect to the LS inverter terminals, see if it still throws the "throttle monitoring" DTCs - I imagine the CTs don't really like the two motors' cables coiling around each other without any kind of shielding... or something like that.

Other than that, I've been meaning to ask this for some time:
https://github.com/damienmaguire/Lexus- ... 50h_v7.ino

Code: Select all

    //speed feedback
    speedSum=mg2_speed+mg1_speed;
    speedSum/=113;
    htm_data[0]=(byte)speedSum;
    htm_data[75]=(mg1_torque*4)&0xFF;
    htm_data[76]=((mg1_torque*4)>>8);

I don't recall seeing this mentioned anywhere. Any clue as to what it is? I'm guessing it's something calculated off of driveshaft/wheel speed.

[xp677]

It will be interesting to see if the two ECUs from different cars are able to communicate with each other. I doubt it, but would love for it to be so!

In the code you quoted, the first 3 lines are just data fed back from the MTH packet. I guess as a sort of safety check, to ensure that the HVECU has received the speed information somewhat correctly.

The last two lines should be in the section below, for the MG1 torque command.

The source of the speed signals is the motor resolver, via the MGECU (inverter) as part of the MTH packet. You can see this earlier in the code:

Code: Select all

      mg1_speed=mth_data[6]|mth_data[7]<<8;
      mg2_speed=mth_data[31]|mth_data[32]<<8;

MG2 speed is an appropriate place to get vehicle speed from. I wrote some extra code for this for my project, but I'm not sure if I mentioned it:

Code: Select all

//for 255/40/R17 tyre
#define TYRE_WIDTH 255
#define TYRE_RATIO 40
#define WHEEL_SIZE 17 //inches
#define DIFF_RATIO 4.1 //x:1

#define TYRE_CIRCUMFERENCE (WHEEL_SIZE*25.4+TYRE_WIDTH*TYRE_RATIO/50)*3.14159265359/1000 //metres travelled in one wheel revolution
#define DIST_PER_MG2_REV TYRE_CIRCUMFERENCE/3.9/DIFF_RATIO/1000 //km car travels in 1x MG2 revolution
#define LOWGEAR_RATIO 2.05263157895 //how much lower the low gear is compared to high

      veh_speed_kph = mg2_speed * 60 * DIST_PER_MG2_REV;
      if (veh_speed_kph < 0)veh_speed kph *= -1; //for reverse
      if (prnd_gear == GEAR_LOW)veh_speed_kph /= LOWGEAR_RATIO; //low gear

Stick the defines at the top, and the 3 lines of code within the if(mth_good).

[Lwerewolf]

Great, I was worried that it had something to do with engine RPM control. Still wondering why the inverter has a camshaft input sensor... other than to compare it to the value that the HV ECU is sending and set a DTC if it differs.

Re: different HV ECUs talking with different engine ECUs - not sure how much use that'll be, given the nature of the THS system.

[mdrobnak]

That is super odd. I know my gearbox has a traditional 12V tach signal + CAN data and will error if they mismatch, but I can't see why the inverter would need it.

[Lwerewolf]

Just a heads-up for anybody that might be waiting on this, I've been, ehh... busy with other projects and the weather hasn't been too great outside... and now that I cleaned up the garage somewhat, an mr2 aw11 ended up in it..

I estimate it'll be at least a month more until I come back to this.

[PacEmaker]

Just a heads-up for anybody that might be waiting on this, I've been, ehh... busy with other projects and the weather hasn't been too great outside... and now that I cleaned up the garage somewhat, an mr2 aw11 ended up in it..

I estimate it'll be at least a month more until I come back to this.

Lwerewolf I just want to say thanks very much for your work up to this point, also thanks to Tom as well.

I'm looking at an LS600h inverter as an option too but I don't expect I'll be able to contribute on the coding side of things, not a software guy me, at least not yet! Once mine arrives I will definitely want to try some bench testing, so I'm keen to see the fruits of your research.

[Xenith]

Im very interested in this as i have just blowen my firsg inverter (this is my 3rd gs450h) and would like to upgrade to the twin cooled inverter as i drive this car to work 1000km each way with sprited open road use and im heavy on the foot around town

[PacEmaker]

Just as a heads-up, I finally ordered an ls600h inverter, should be able to provide info in... 2-3 weeks, hopefully. Sorry for the delay, other things took precedence

A bit about the part numbers - g9200-500(11/21) - resp. (lhd/rhd) don't show up on toyodiy or the EPC app, I assume they're a very early production model or something. I ordered a g9200-50011. Now to secure wiring before it arrives.

Looking at them on Yahoo Japan I can confirm G9200-50021 (RHD version) is from 2007/2008. The 2009 ones I've seen have part number G9200-50070 (RHD). Later part numbers are G9200-50072 in 2011 and G9200-50074 from 2012. All these are RHD versions, because Japan. I have no idea what the differences are or whether the later inverters can be retrofitted to the earlier cars, so I'm playing it safe and assuming the earlier ones will be most likely to work with the GS450h VCU but we'll see.

I've ordered a late 2007 that should be the same as yours (except RHD). Thanks for your work Lwerewolf and looking forward to seeing how it goes!

[Xenith]

Just an fyi i am currently running the igbts and boards out of a 2011 camry in my 2006 gs450h im using all the gs450h wiring and capacitors worksgreat

[Eflyers]

By all means: show us and tell us how you are doing this!

[Lwerewolf]

The casing is pretty much the same, the IPM (igbt module) and control board(s) have the same physical dimensions and will bolt right in - no clue whether the wiring from the external socket to the on-PCB sockets is the same but it very well could be. The camry modules do have less IGBTs though - check the ORNL article, the "unpopulated IGBT slots" are populated on the gs450h boards, probably for a good reason given the power output differences.

The weather here is pretty bad and the garage is taken by the mr2 (which I can't leave outside for a multitude of reasons), so it's still on hold. Re: yours - be aware that you'll need the ls600h-specific wiring harness (for the two 3-phase motor connections). The connectors are the same on the transmission-side, not the case on the inverter-side. As I noted previously, during my first attempt I "extended" the harness without extending its shielding, which I suspect is the reason for the "torque execution monitoring malfunction" codes that I got - the car attempted to crank the engine just fine... for the split second that it did.

[Xenith]

The casing is pretty much the same, the IPM (igbt module) and control board(s) have the same physical dimensions and will bolt right in - no clue whether the wiring from the external socket to the on-PCB sockets is the same but it very well could be. The camry modules do have less IGBTs though - check the ORNL article, the "unpopulated IGBT slots" are populated on the gs450h boards, probably for a good reason given the power output differences.

Thanks for the info Lwerewolf
The use of the camry igbts was due to the only 450h iverter i could find in aussie was $1500 and would have to wait for it to be posted . i was able to get the camry one for $145 at the local pullapart making the gamble worth a try
the plug sockets were the same as i used all the gs450h everything just the igbt,s from the camry .
So with the less igbt's will i just get less power or will i be driving more current through them .
Were do i find this article "check the ORNL article"
Looking forward to ur replay