GS450H gearbox with Camry inverter G9200-33171

[tom91]

Make a map again of your potnom and idc.

The measurements do not really make alot of sense to me. potnom and idc do not look to be linked as you expect.

[PetersonOctavius]

here you go

[tom91]

As I taught, have a look at your idcmin and max.

Bump these to like -300 and 300 and test again.

Also charge Throtramp to 1 or 2.

[PetersonOctavius]

I ll do that later when i get home and get back, thanks tom

[PetersonOctavius]

I tried with lower idcmin and higher idcmax, i did have less judder. But it didnt disappear. Throtramp was set to 1, default was 100

[tom91]

Logs...

We need data

[PetersonOctavius]

Is there a way to save the plots? I ve been taking screenshots while driving

[PetersonOctavius]

did couple of runs...judder getting less.. but still present
i also plotted torque this time around

negative torque cause the vcu is "in reverse" (wish i could change it)

[tom91]

You need to review why Potnom gets limited. Are you running V2.20A ? This should show you why it delimits if you do not key off.

Try looking at just Potnom and idc again. I do not know the ESP32 CAN enough to tell you how to save graphs/logs from it.

[PetersonOctavius]

Im running 2.17A
At the release of 2.20 the gearbox was not installed yet, i still tried it but its scary. No throttlre input still get gearbox spin.
Tried it once in the car… almost hit a wall with brakes applied.

[tom91]

Im running 2.17A
At the release of 2.20 the gearbox was not installed yet, i still tried it but its scary. No throttlre input still get gearbox spin.
Tried it once in the car… almost hit a wall with brakes applied.

This does not sound right please review all parameters before proceeding. Disconnect the HV to your inverter and review potnom and torque and adjust your parameters to correct.

You clearly have problems with your settings. Also note IS300h DOES NOT work with 2.20A release so its wierd you can get your inverter working using this selected inverter.

[PetersonOctavius]

Idcmax is at 300, during the runs i think i reach between 150-175Amps max. So not even near the 300max.
I feels like a derating, but im not even pulling that much amp.



upload the current .json also

[PetersonOctavius]

Is it weird that my tpmhs never reaches 30degrees?
I live on Curacao, in the caribbean. Those days outside temperature reaches 30deg but the highest ive seen from the inverter is 28 degrees.

[PetersonOctavius]

It looks like this sensor: viewtopic.php?p=9183&hilit=036906051G#p9183

Doesn't need any pull-up

Correct, its plug and play. What i didnt know is that the zombie has to be in run mode to show the value. I did buy a replacement sensor, plugged it in and now i have controlled brake boosting

[PetersonOctavius]

its been a while now, car was park waiting for some new wheel bearings.
i also installed a volvo S40 power steering pump, works right away, i use the cars fuel pump real to enable the pump when VCU runindicatioin i enabled.

went for some test drive today, still jittering...shakes the car pretty rough... i have to let go of the throttle pedal, then it drive fine again for a while. its unpredictable when the shaking will start.

did some data logging, this time with idc, potnom and INVudc.
i added the INVudc since it comes from the inverter, as a way to check if its a communication problem, i had jitters but INVudc was still there.

what i noticed in the src for the gs300h :
//mg1
htm_data[5]=(mg1_torque*-1)&0xFF; //negative is forward
htm_data[6]=((mg1_torque*-1)>>8);
htm_data[11]=htm_data[5];
htm_data[12]=htm_data[6];

//mg2
htm_data[31]=(mg2_torque)&0xFF; //positive is forward
htm_data[32]=((mg2_torque)>>8);
htm_data[37]=htm_data[26];
htm_data[38]=htm_data[27];

the last 2 lines of most lexus torque command for mg2 is most of the time equal to the firt 2 htm data, example:
for mg1:
htm_data_11 = htm_data_5
htm_data_12 = htm_data_6
but for mg2 data 37 and 38 are 26 and 27 and NOT 31 and 32.

the GS450H code is written as follow:
//mg1
htm_data[5]=(-mg1_torque) & 0xFF; //negative is forward
htm_data[6]=((-mg1_torque) >> 8);
htm_data[11]=htm_data[5];
htm_data[12]=htm_data[6];

//mg2
htm_data[26]=(mg2_torque) & 0xFF; //positive is forward
htm_data[27]=((mg2_torque)>>8) & 0xFF;
htm_data[32]=htm_data[26];
htm_data[33]=htm_data[27];

i have no idea how the Totoya Sync Serial works, but i just noticed this difference.

[tom91]

Different generations IS300h vs GS450h. People have working none juddering GS450h and IS300h setups.

So its parameters for your setup or wiring related issues.

[PetersonOctavius]

I used shielded utp ethernet from vcu to inverter, i grounded the shield on both sides.

Which params should i try to change?

[tom91]

Who knows. Compare your params to Damiens or params from others running a GS450h

Damiens E39 with a GS450h https://github.com/damienmaguire/Stm32- ... _220a.json

Please note you are running a setup that is "unique" IS300h inverter with GS450h transmission.

[PetersonOctavius]

i was on 2.17A, i switched to 2.20A, but a changed GS300h torque setting in the GS450H.cpp
in case 14 i change:
THIS
//mg2
htm_data[31]=(mg2_torque)&0xFF; //positive is forward
htm_data[32]=((mg2_torque)>>8);

TO THIS
//mg2
htm_data[31]=(mg2_torque*-1)&0xFF; //positive is forward
htm_data[32]=((mg2_torque*-1)>>8);

this is because the car used to roll forward when VCU pin REVERSE gets 12v and the opposite for drive.

i upgraded to 2.20 just now because of weird behaviours.
i put the rear wheels in the air...updated zombie to 2.20A and tested, it did fine. so i went for a test drive around the block.
happy to report i didnt get any jitters/juddering.

came home to write this update and go for another test run hopi for no jittering.
i ll get back with updates.

[PetersonOctavius]

i guess 2.20A fixed the jittering.
i did some slow and fast runs. reaching a max of around 130Amp (reported by ISA shunt)
didnt get jitters!!
FOR NOW i can say the camry inverter G9200-33171 works with the ZombieVCU using the GS300H settings. connected to the GS450H gearbox.

on to the next, make the gauge cluster display car speed

[Gregski]

excellent build and write up I wonder if it would have been smooth sailing using the 6th Generation (2007-11) Toyota Camry Hybrid inverter since the Zombie was developed using the 3rd Gen (2007-11) Lexus GS450h inverter / transmission combo

I am on a mission to do just that here: viewtopic.php?t=6239

[PetersonOctavius]

I think that s the one Arthur was testing back then for his jaguar.
I went for the 2012 cause @Zechy had some succes using it… and it was cheap

[PetersonOctavius]

its been a while, was pretty busy with life lately (i became a dad) and also waiting for new control arms for the audi.

recap on components being used :
- GS450H, only mg2 being used;
- Toyota camry inverter (cause its almost 3x cheaper than the GS450H inverter)
- Zombieverter VCU, firmware 2.22A, (1 small modification to the code was to invert the torque request for gs300h, as with my rear diff forward was reverse and vice versa.)
- Samsung SDI li-ion battery from batteryhookup.com (24kWh theoretically)
- Tesla gen 2 charger, with EVBMW board, firmware 1.17R
- Tesla gen 2 dcdc converter
- Volvo power steering pump.
- Original audi electric brake booster
- Flying ADC BMS from Johannes, installed but still need to finish setting up
- Tesla ac compressor, the one that uses pwm for speed control, installed but not wired up.

so far ive done quite some driving around, so i can confirm (again) that the VCU will control the camry inverter using the GS300H settings.
still need to try a public charging station.
max Amp reporter by the isa shunt was 129A, battery on idling was at about 370volts.

to do :
- connect and build a circuit for the AC compressor
- control the tesla charger via the VCU, i dont know how to set that up yet, CAN is conected, but i cannot find instruction on how to make the vcu control the gen 2 charger.
- configure the Flying ADC BMS to do some balancing and read amp draw from the ISA shunt.

below my latest vcu parameter json

vcu-222A-17april.json

(2.13 KiB) Downloaded 294 times

[PetersonOctavius]

Drove my first 100km on the electric frankeinstein, it all went well.

i also did some public charging, tesla gen 2 charger with the evbmw board works like a charm.

soon i will let an A/C shop clean and fill up the system, then i ll try to get the compressor working.

my fuel gauge has 2 input, both need the same resistance for the gauge to show fuel level. how do i make an electric variable resistor? can i use an arduino for that? i already have an ESP32 getting CAN messages and displaying it on a screen

[johu]

Great work!
It's probably the same fuel gauge as Touran. I use 2 BC547 transistors with 10k base resistor and 1uF low pass capacitor. I then PWM with 18 kHz.
You start out with both duty cycles at the same value until you reliably get the fuel gauge in middle position. Then as you increase one channel you decrease the other one by the same amount to move it around.

Once it detects invalid resistance it bottoms out and you have to power cycle the car