ZombieVerter VCU Support

[alexbeatle]

Thank u for explaining that. I re-checked and think i know it.
The pin assignment works a bit different then expected. For the other PWM pins we do need timer 14 and 15 from memory oh and more AFIO assignments...

Johu; Can you elaborate on your LIN driver? and can you tell me if N channel PWM outputs are useful for normal PWM modes?
Or should we look for assignment of CHx pins only?

Have you had luck testing the pwm outputs? Thinking of using them for the cooling pumps controls, if possible.

[EV_Builder]

Nope; feel free to share your code when you have it
Yes; Tesla pumps are a nice candidate.

[nkiernan]

Just checking if some of the experienced Zombie & Codeblocks people might be able to guide me in the right direction on my query in the general section?

viewtopic.php?t=2367

[Kelju]

I have a problem with potnom value.
I have a dual pot throttle pedal and have calibrated it according to the instructions and seems to work according to the individual pot and pot2 readings.
I have the car on stands and only a few amps of current draw from the battery when accelerating, so no current limiting active.
Temperatures are low, so there should not be temperature derating active.
Despite of all this, the potnom value is limited and having the occasional maximum value of ~10 when the pedal is fully applied.
In fact while the pedal is fully pressed down, the potnom value is fluctuating somewhere between 5 and 10 and the maximum RPM of the motor is regulated to 2500. My rev limit is set at 6000.

I also tried disconnecting the throttle2 input and using the single pot mode, but the behavior did not change.

[Kelju]

I have a very old version of the source and I noticed that the code has changed quite a bit.
So I decided to clone the latest version and apply my changes.
It is always a good idea to first check that "make" works for the original unchanged code.
Turns out it does not. Probably something simple, but I got this error message:

Code: Select all

  CPP     obj/hwinit.o
make: *** No rule to make target 'obj/stm32scheduler.o', needed by 'stm32_vcu'. Stop.

edit: I checked the make file and everything seemed to be in order.
I went ahead and tried to "make" again without any changes and sure enough - IT WORKED!

So I guess it is time to break it with my mods...

[Jonr]

I have the transmission and inverter from an GS450h plus a zombieverter but no shunt. Does anyone have this combo running OK ideally in Kent/Surrey area of the UK? I've checked the wiring from the internal plug on the inverter through to the leg of the chip on the inverter and all good. I'm not getting comms between the zombie and inverter. I don't know if it's a config issue or hardware. Can you assist?

[dukbilt]

First : make sure you are wirring as per the wiki : https://openinverter.org/wiki/Lexus_GS450h_Inverter

Is it a GS450H inverter or some other variant?

I use Camry inverter G9200-33171 (2013 - newer that described in the GS450H wiki). Bassmobile said that he has not any issue with it and GS450H Drive unit in the Lexus GS450H VCU Support Thread.
About wiring: I've tried both variants of serial wiring. Both unsuccesful. What is correct? REQ+(vcu) <-> REQ+(inv)?
I've met this kind of serial protocol in the first time, so I'm not sure how tx and rx are calling.
Other wiring is correct I beleive.

chuux,

any progress with the G9200-33171 inverter?

[Kelju]

I have a very old version of the source and I noticed that the code has changed quite a bit.
So I decided to clone the latest version and apply my changes.

I applied my changes and was able to compile. However, after flashing, the web interface is not updating and the red "activity" led is not blinking.
I reverted to my original code and everything went back to normal.
Then I thought I'd try the newest code from the Git without applying my changes, but the behavior is the same. "Activity" led not blinking and no activity/coms via the web interface.

What might be the cause of this behavior and are others successfully running this code version?

[nkiernan]

Wondering what the main conditions are that need to be met to allow the Zombie pin 32 (inverter relay low side switch) to ground and so supply power to the inverter (GS450H set up)?

Doing to testing with a 60V bench supply, the '.bin' file Damian posted, on 10/30/2021, the ISA shunt and the HV and precharge contactors. Looks like mostly working just the inverter relay doesn't activate.

Dual channel throttle is calibrated and working. ISA is reporting all the correct voltages on 'udc', applying start from ignition, the negative and precharge contactors close, have to hold the key at start until the positive contactor then closes - few seconds later. Getting 60V through to the inverter but no 12V power to inverter from the inverter relay. Tried changing the gear position between P-R-N-D thinking it might only power on when drive or reverse are selected, no go.

I do need to get the web interface fired up again to check errors but wondered if there was a couple of key conditions that stop pin 32 grounding the relay?

[EV_Builder]

Did you asure the hardware part? IC/FET is there? Ground/Pos switched relay?

[nkiernan]

Did you asure the hardware part? IC/FET is there? Ground/Pos switched relay?

Good point!!! I believe it is there and ok but will need to get out the schematic again and recheck. Did another start up sequence so I could capture some web interface for confirmation there are no errors showing up. Started in Park and then I switched to Forward before taking the screenshots.

Here InvStat is 'off' as if it is not being allowed on rather than reading its status?

[Jack Bauer]

So I can't answer why its not switching in this case but I do know that 1)the design is correct as it works in two of my cars and 2)the software is correct at least in as far as I can tell. The track from the STM32 to the smartFET is direct so a continuity test would be a good starting point. Then check to see if the gate pin is at 3.3v when in run mode. The NCV8402 will shutdown if subjected to an overload so make sure you dont have a bad relay for example.

[EV_Builder]

The base class uses the virtual definition to allow a programmer who inherits to override.
I don't know the differences exactly in motors but i'm not sure if you are walking to many miles to get there.
(regarding the two engines).
Keeping them in seperate c++ files and sharing just the class for the statemachine would be the easyest solution for more people i feel.

To come back a bit to this discussion about StateMachines i have build my own, leaner / light, implementation of the famouse PackML standard.
I will upload the stuff to my github when i'm ready.

Not every state is needed and can be easily skipped.

[Kelju]

I thought I'd try the newest code from the Git without applying my changes, but the behavior is the same. "Activity" led not blinking and no activity/coms via the web interface.

What might be the cause of this behavior and are others successfully running this code version?

Any idea what might be causing this issue with the newest code?

[nkiernan]

So I can't answer why its not switching in this case but I do know that 1)the design is correct as it works in two of my cars and 2)the software is correct at least in as far as I can tell. The track from the STM32 to the smartFET is direct so a continuity test would be a good starting point. Then check to see if the gate pin is at 3.3v when in run mode. The NCV8402 will shutdown if subjected to an overload so make sure you dont have a bad relay for example.

Did some more testing this evening. So at ignition on and start, I can get the VCU 'opmode' into 'Run' every time and HV and pre-charge contactors work as expected with no error messages and 'udc' showing correct voltage (60V from bench supply).

I took a new relay, confirmed it was working and swapped it in for the inverter start relay, but no change. Have continuity between STM32 PA8/67 pin and the gate pin of IC6 (IC6 controls the inverter relay). Have continuity between IC6 drain pin and pin 32 of the vcu connector and all the way back to 12V, and the source pin has continuity to ground.

When I get the system into 'Run' mode, I get 11V to 12V at the Drain pin via pin 32 on the vcu connector as expected, but only getting 0.08V at the Gate pin.

So looks like hardware is all ok, and its the STM32 that is not sending 3.3V to the gate pin as expected? Are there certain conditions in the code that stop this gate signal? Wondered if something like the lower car battery voltage in the test is too low, but can't see that affecting it, or if brake signal needs to be present at start request or something like that? Previously had this GS450H gearbox and inverter running with the GS450H VCU, so should all be functional

[Jack Bauer]

Right. Looks like the software is doing something silly. I'll get it sorted as soon as I have a desktop pc back up and running as no point asking others to contribute. It would seem that those who can help won't and its unfair to ask people not familiar with sw to climb a very steep learning curve. After all the main stated aim of this project is to avoid the need for messing with sw.

[nkiernan]

Right. Looks like the software is doing something silly. I'll get it sorted as soon as I have a desktop pc back up and running as no point asking others to contribute. It would seem that those who can help won't and its unfair to ask people not familiar with sw to climb a very steep learning curve. After all the main stated aim of this project is to avoid the need for messing with sw.

Did a test to see if I could manually power the inverter. When I got the system in 'Run' mode and InvStat was showing as 'Off', I tried manually connecting 12V power to the inverter at 'Run'. Refreshing the web interface still showed InvStat as 'Off' which I was expecting as presume it is software in VCU that is setting that.

Didn't get any inverter whine to suggest it was powered on, and putting it in 'Forward' and pressing throttle didn't spin the motor. So couldn't bypass the inverter relay issue. Didn't expect it to, but wanted to make sure.

I will do some looking through the code as best I can to see if I can find anything that affects InvStat

[EV_Builder]

I can give it a shot but tell me where the source is from what you loaded. Sometimes people attach here on the forum new versions. And there is no way to know what and where the matching source is.

So either we only deploy to GitHub incl. Source and we can help or the poster of the attachment is the support resource.

NKierman please load the GitHub vanilla release. Retest and report.
I will tell you from source what the conditions are.

[nkiernan]

I can give it a shot but tell me where the source is from what you loaded. Sometimes people attach here on the forum new versions. And there is no way to know what and where the matching source is.

So either we only deploy to GitHub incl. Source and we can help or the poster of the attachment is the support resource.

NKierman please load the GitHub vanilla release. Retest and report.
I will tell you from source what the conditions are.

[EV_Builder]

I took a new relay, confirmed it was working and swapped it in for the inverter start relay, but no change. Have continuity between STM32 PA8/67 pin and the gate pin of IC6 (IC6 controls the inverter relay). Have continuity between IC6 drain pin and pin 32 of the vcu connector and all the way back to 12V, and the source pin has continuity to ground.

Yes i agree. PA8 / 67 = inverter power.

I think your relay should be wired to that output...

i think its just a matter of configuring the VCU correctly with the WEB GUI.

[nkiernan]

i think its just a matter of configuring the VCU correctly with the WEB GUI.

Thanks EV_Builder. So you're thinking its just my combination of web interface settings that's causing the issue? I can try changing round some things.

From the code snippet you attached, I should meet the conditions to get inv_out set ok. I start with opmode = off, get the start signal through ok, and not charging. Unless there's something about chargemode that's saying true for some reason, or invmodes is actually OpenI, and then inv_out won't get set.

I need to get compiling working and get a bin of the latest version.

[EV_Builder]

Thanks EV_Builder. So you're thinking its just my combination of web interface settings that's causing the issue? I can try changing round some things.

From the code snippet you attached, I should meet the conditions to get inv_out set ok. I start with opmode = off, get the start signal through ok, and not charging. Unless there's something about chargemode that's saying true for some reason, or invmodes is actually OpenI, and then inv_out won't get set.

I need to get compiling working and get a bin of the latest version.

yes; the invmodes must be set to GS450H

[nkiernan]

yes; the invmodes must be set to GS450H

Ok, will give this a check again tomorrow. It has been set to GS450H as shown below (Inv), but will try changing some settings to check

[nkiernan]

Some more testing this evening. From EV_Builders comments I changed around some more settings to do with the inverter and charging settings, and same result, inverter relay never gets gate signal to activate. Changed to a new .bin and back and no change. Things that are probably different from my set up to Damians working set up with the same .bin, I would have all charger inputs selected off, lower test voltage, and I don't have any vehicle CAN selected. As the inverter relay is governed in a way by charger selections, I'm being steered to a quirk around charging!


So using gitpod I compiled a new .bin from the latest stm32-vcu git repo. Uploaded this to the vcu and looks like the same result as kelju reported above, once uploaded the flashing activity LED stops and the parameters section of the web interface disappear leaving the first section for UART and SW uploads and the lower plot section.


Would anyone have a known working .bin with the GS450H setup I could test also maybe?

[johu]

Could it have something to do with my latest changes?
viewtopic.php?p=38975#p38975

You can go back to the version before these changes with

Code: Select all

git checkout b9c8f19

If it then works again, please someone find out where I killed it.
You can work for way forward through my various commits by using the hashes from here: https://github.com/damienmaguire/Stm32- ... its/master