ZombieVerter VCU Support

[Jack Bauer]

Yes will be required for gear sense on the 450h

[AdrianB]

Is there an exact reference for R14? There’s a bewildering array of 1K R0805 resistors listed on Mouser.

[Jack Bauer]

any 1k 0.1% 0805 will do.
e.g. https://www.mouser.ie/ProductDetail/Bou ... nxVA%3D%3D

[Jack Bauer]

Well folks here it is. The big release. Probably forgot a few things that I ended up fixing but here you go :
https://github.com/damienmaguire/Stm32- ... s/tag/1.01
-Fix non working CAN channels.
-Fix inverter menu defaulting to OI at boot.
-Fix inverter menu not refreshing on new selection
-Fix GS450h oil pump pwm output not working
-Calibrate GS450h MG1 and MG2 temp sensors.
-Correctly display GS450h selected gear based on valve body feedback switches
-Add GS300H inverter/transmission. Bench tested only. Some functions to be added.
-Car test of GS450h functions.
-E65 dash and CAN functions tested.
-CAN3 / Chademo tested.

In addition this code is now running in my E46 touring running the LIM and Leaf Gen1 system. It has been tested in the superyacht E65 with gs450h system by paa. I have bench tested the GS300h and vw instrument cluster. Its as tested as I can reasonably make it with what I have available right now. I would be grateful for some feedback from users with as much detail as possible here on the support thread.

[AdrianB]

Thanks for all your work Damien. I hope to have a couple of days next week to get my GS450h spinning on the bench. I’ll share any problems I find here.

[nkiernan]

Thank you Damien, was looking forward to this update

Gave it a quick run this morning. Testing with four 12V batteries for HV (approx. 49V). Initially have vehicle type set to 'none' and inv_can set to 'CAN1', the default. Is there supposed to be three CAN channels to select from the drop down? Only two showing currently.

Torque, potnom and pot1/pot2 plots below. Is torque scaling correct?

Pre-charge and contactor sequence works as expected and as before and opmode changes to 'Run' ok. But also as before, InvStat stays 'off' and no drive rotation in fwd or rev. InvUDC also reading 0V as before but I can measure 12V at the inverter power input (so Zombie has inverter power relay switched on) and 49V at the inverter HV input so inverter should be happy with HV and power (which is all it needed to run with the GS450H vcu). Attaching a few screenshots as I'm probably setting something incorrectly and someone might see it.

So changed inv_can to 'CAN2' and this time INVudc now shows the correct inverter HV (this comes from the Inverter to Zombie communication yes?) But still InvStat remains off and no rotation. Precharge error here was just first start attempt where didn't hold key to start long enough

My thoughts are that it must be something I have incorrect in my wiring (or settings) so will be taking another look at both again although the GS450H vcu still runs the setup (can still connect in either for testing). Tried changing vehicle type to classic and E65. E65 just shuts off contractors as expected as its CAN based. If InvUDC is reading ok, would that mean there is now communication between Zombie and inverter. So I still have some troubleshooting to do

[paaa]

Thank you Damien, was looking forward to this update

Gave it a quick run this morning. Testing with four 12V batteries for HV (approx. 49V). Initially have vehicle type set to 'none' and inv_can set to 'CAN1', the default. Is there supposed to be three CAN channels to select from the drop down? Only two showing currently.

Torque, potnom and pot1/pot2 plots below. Is torque scaling correct?

100922_TorquePlot.JPG

100922_Pot1Pot2Plot.JPG

100922_PotnomPlot.JPG


Pre-charge and contactor sequence works as expected and as before and opmode changes to 'Run' ok. But also as before, InvStat stays 'off' and no drive rotation in fwd or rev. InvUDC also reading 0V as before but I can measure 12V at the inverter power input (so Zombie has inverter power relay switched on) and 49V at the inverter HV input so inverter should be happy with HV and power (which is all it needed to run with the GS450H vcu). Attaching a few screenshots as I'm probably setting something incorrectly and someone might see it.

100922_Spot001.JPG

100922_Spot002.JPG

100922_Spot003.JPG

100922_Spot004.JPG


So changed inv_can to 'CAN2' and this time INVudc now shows the correct inverter HV (this comes from the Inverter to Zombie communication yes?) But still InvStat remains off and no rotation. Precharge error here was just first start attempt where didn't hold key to start long enough

100922_Spot005.JPG

100922_Spot006.JPG

100922_Spot007.JPG


My thoughts are that it must be something I have incorrect in my wiring (or settings) so will be taking another look at both again although the GS450H vcu still runs the setup (can still connect in either for testing). Tried changing vehicle type to classic and E65. E65 just shuts off contractors as expected as its CAN based. If InvUDC is reading ok, would that mean there is now communication between Zombie and inverter. So I still have some troubleshooting to do

ill check inv stat but I think inv stat and udc inv don't show anything but system is working otherwise. How are you loading new firmware ? through web page or st link ?

[nkiernan]

Loading through web page using phone. Laptop connection is less stable, then back to laptop for the testing. Looks to load OK with new temp parameters at the bottom of the list. Not set up for stlink

[Jack Bauer]

So invstat and invudc as well as other inverter params for 450h and 300h do work. The fact they dont in your setup indicates a comms problem. All I can say is that this firmware works so at least you can narrow down the problems. My first suspect would be a missing clk signal. Remember that the Toyota serial comms do not use any of the CAN channels. CAN3 is not in the menu items as of now as it is dedicated to Chademo.

[Domt177]

I’ve looked at the schematic and don’t think they are but are IC17 and IC12 needed in operation of is300h inverter & gearbox?

[Jack Bauer]

IC12 is used gear shifting in the 450h and IC17 is one of the CAN busses so neither would be required for test but IC17 would be needed if you intend to use more than one CAN bus.

[Domt177]

Ok thanks, just testing for now. Getting as far as run mode and torque commands but invstat stays off and no high pitched noise or spinning. Have checked all the comms and power wires placement and continuity

[nkiernan]

Ok thanks, just testing for now. Getting as far as run mode and torque commands but invstat stays off and no high pitched noise or spinning. Have checked all the comms and power wires placement and continuity

Sounds similar to what I was seeing today, run mode ok but no whine or spinning. Meant to do one more check but forgot, will try tomorrow, but previously I believe when I refreshed the web interface after applying ign on but before applying start, INVstat changed to On and INVudc read correct voltage. Then apply start and both changed to off/0V. Maybe I was dreaming, but will check again.

[nkiernan]

So invstat and invudc as well as other inverter params for 450h and 300h do work. The fact they dont in your setup indicates a comms problem. All I can say is that this firmware works so at least you can narrow down the problems. My first suspect would be a missing clk signal. Remember that the Toyota serial comms do not use any of the CAN channels. CAN3 is not in the menu items as of now as it is dedicated to Chademo.

Thank you for the reply. Yes, comms appears to be the issue (my bad on the CAN bus relationship, although thought it was interesting it shows correct voltage even if coincidental). So the GS450H VCU appears to work correctly which rules out the inverter/transmission side comms. If the firmware is now working, that means its the 8 comms wires (which I checked again but going to remove the GS450H Tee in case that might have some affect), or its Zombie hardware.

Is there a relatively straightforward troubleshooting option for the Zombie hardware to check the four comms lines ('For Dummies' type thing! )?

Could there be a debug/error log option in the web page at some stage similar to the new Tesla Model 3 PCS firmware to help confirm comms for the Toyota stuff is ok?

[Jack Bauer]

InvStat is designed to show active comms from any inverter toyota or otherwise. Built In Test (BIT) or loopback test of the hardware would indeed be possible if this were a better resourced project. Sadly it would seem to be just me and between limited time and software skill I wouldn't hold my breath. That said I would very much like to get to the bottom of this in case there is some deep bug still lurking thats not triggered by my bench setup or the setup in the E65. I'll mess about this morning and try to create the problem.

[Jack Bauer]

On the test bench this morning with 450h inverter and gearbox connected.
cold start : ok.
restarts : ok.
select a wrong inverter, attempt start, select correct inverter. Works as long as not another toyota inverter is selected.e.g. if trying to run a 450h with 300h selected it will not self restart and a vcu power cycle is required. Other selections ok no restart required. e.g. select Leaf when trying to run 450h.
Firmware load from : Linux laptop, win 7 laptop, android phone: ok.
select a wrong vehicle : ok but wont work with E65 as its expecting CAN. Reselecting the correct vehicle or just "none" or "classic" restores operation without vcu reboot required.
load old bugged firmware with win7 laptop, attempt start, doesnt work as expected. Load new firmware. Works as expected both from warm and cold reboots.
Not sure what else to try here. InvStat and inverter params display as they should.

[thornogson]

Stupid question , does using the IS300 with a zombieverter still need the Isabellahute current shunt ? if so does anyone know of a source of these, I've looked all over and in the video it seems to be on the bench (17.30) but not mentioned where the hookups are described

[Jack Bauer]

Yes the isa shunt is required and I have explained that in the video series and put links in the video description to :
https://www.evcreate.nl/shop/charging/i ... nt-sensor/

Over on the wiki page the videos are even indexed with the specifics of connecting the shunt and the why :
https://openinverter.org/wiki/ZombieVerter_VCU

[nkiernan]

Works as expected both from warm and cold reboots.
Not sure what else to try here. InvStat and inverter params display as they should.

Thank you for the detailed checks. It can only be a hardware issue at this stage, I'll have to recheck my soldering. Attaching json if this might help compare settings, maybe I'm messing that up, but don't think so.

GS450H_Bench.json

(1.08 KiB) Downloaded 35 times

[thornogson]

Yes the isa shunt is required and I have explained that in the video series and put links in the video description to :
https://www.evcreate.nl/shop/charging/i ... nt-sensor/

Over on the wiki page the videos are even indexed with the specifics of connecting the shunt and the why :
https://openinverter.org/wiki/ZombieVerter_VCU

My name is John , I have been a nob...and I should know better. I have been 'one of them '- exasperating people who ask a question rather than read a manual

Lesson here people ... Read the ..F*&£ing....WIKI , there's a clue in the name .

I have now watched all 3 videos and am a much better (informed) person. Thank you Damien for your patience .

[Jack Bauer]

No problem. I try to make the links as visible as possible.

[Jack Bauer]

Some folks had asked about the component situation regards the kits and fully built boards. As of an order today I'm pleased to say most parts are now fitted by JLC once again. See attached picture.

Components not placed but included in kit :IC27,29,29. IC19. Wifi module and socket.
Components not included in kits but fitted to fully built boards : IC10,IC12.

Missing 1k R14 is now placed and R35,R43 changed to 150k to reflect recent calibration of GS450h and GS300h temp sensors.

I do not make custom kits or one offs.

[Jack Bauer]

aaannnd just found a really major and deeply buried bug in the latest firmware which would cause the vcu to lockup and bootloop on the watchdog. Seems certain modules most notably the Leaf had not been updated to reflect the new web interface selection of CAN interfaces. This bug would need to following conditions to trigger :
Selection of a CAN driven inverter (e.g. Leaf),
Save params,
Cold start,
Bug triggered and quite impossible to exit without loading a modified source.

Now of course I run a Leaf drive train in the E46 so why didnt I catch it you may ask? well, because up to this morning I had not cold started the vcu The bug would NOT trigger if a CAN driven inverter were selected during normal operation e.g. chaging from a toyota to a Leaf and any other combination and of course I have been working on the toyota drive systems on the bench and no reason to save a CAN driven inverter and do a power cycle which would have triggered the bug.

So I have updated the release so anyone running this please head over and grab the update :
https://github.com/damienmaguire/Stm32- ... s/tag/1.01

[nkiernan]

Sounded like there were a number of people on this thread with the Zombieverter and a GS450H transmission, although some at the early bench test stage. Would be good to hear your feedback on how the new Zombie firmware is working with your setup?

[AdrianB]

Sounded like there were a number of people on this thread with the Zombieverter and a GS450H transmission, although some at the early bench test stage. Would be good to hear your feedback on how the new Zombie firmware is working with your setup?

Will do. Unfortunately I’m still waiting for my board to come back from my brother who’s done the SMD soldering for me. It might arrive today.