Heavy torque cut in low gear GS450h & Zombieverter

[dadiowe]

I have a massive torque cut in low gear or auto setting. I enclose my parameters for the current settings.

You will note that I am running 2.20 craflag. This is because I can't get a reliable connection of my wemos mini to succesfully read parameters and spot values. I have with, a lot of help from a forum member, managed to get a fairly good connection with tcan but only with the above.
It appears that there is a problem with the libopeninverter version, it has an SDO response bug which stops 2.17A firmware.bin connecting to the tcan device.

The torque cut error occurs with any version of the firmware.bin that has gear change enabled.
In High gear with the throtmax set at 80% I have a reasonable take up, with the throttle any higher % I can get some torque cutting at "pull away" throttle.
In Low gear it is impossible to pull away without massive torque cutting.
If I set throtmax to 30% I can pull away but, if I apply hard throttle I get torque cutting but not as extreme as at 80% throtmax. This is the plot below from today.
I have had this problem in the past when only high gear was available and updating to version 1.11 seemed to help, all be it with a throtmax setting of 80%.
I recall that another forum member "Jackk" had about 600v on his car. He found "I then played around with UDCmin. This was set to the correct value of 512v. I put this down to 400v and BOOM no torque cut. I have not tested about 50mph yet but crikey this thing is fast!"
The settings for the plot are what Damien suggested I tried the last time I had this problem.

Thanks for looking

[Bigpie]

UDC would be helpful too, but it looks like you're hitting the current limit set "idcmax": 350,

[tom91]

Please note a bunch of "saftey" checks have been correctly reinstated once I got started on fixing the Zombie coding.

Voltage limits and current limits were not working as they should. So your behaviour with the same settings will vary when going to newer code. However this is good as its now hitting limits. If your PotNom changes from 100% when flooring you are derating for some reason, check UDC, IDC and Temps.

[Bigpie]

Side note, you're running my development fork, was only really intended for bench testing / verifying your SDO comms issue. You should switch back to a release version, while I do run it on my car, its not intended for others to use.

EDIT, Tom has verified the can SDO bug exists in 2.17 build on github.

[dadiowe]

udclim 1000
Idcmax 5000
Ist picture at 80% massive torque cut
2nd picture at 30% moderate acceleration
3rd picture at higher revs 30% still torque cutting

[dadiowe]

Will go back to 2.17 A 1st thing 2moro

[tom91]

picture

No pictures.

[tom91]

You IDC is going negative...

Guess changing IDCmax does nothing as that would be a positive value.

Your settings

Code: Select all

  "idcmax": 350,
  "idcmin": -350,

[dadiowe]

Sorry just noticed the lack of pictures. Also now have udc and idc set to max settings. I will post the photos tomorrow as there is a nice bottle of shiraz calling to me

[tom91]

Do you have anything to manage your battery voltages? How are you charging?

From your sale you have a 7 x VW I.D. 3 batteries 12s- 84s

Code: Select all

  "udcmin": 240,
  "udclim": 380,

This yields a max allowed voltage of 380/84 = 4.52V, This is fire territory reduce this to a max of 4.2V per cell.
Low voltage of 240/84 =2.85V, which is very low but fine under load.

The ID batteries are prone to heating with high loads, so youneed to monitor this too.

[tom91]

Also now have udc and idc set to max settings.

NO WHY ARE YOU DOING THIS DO YOU WANT TO BREAK SOMETHING

Stop dicking around and just follow our guidance.

[dadiowe]

My batteries are 12s so 4.2 x 12 x7 = 352.8 volts this is what my charge is normally set to. Charging with outlander and zombie. BMS thunderstruck, probably not the best but I bought them 3 years ago when there were fewer options. Does UDC max and min only have an influence over charging?

I promise from now on I will do exactly as asked Tom.

Picture titles say what they refer to.

[tom91]

Cannot see anything on the pictures. Please makes sure to get the whole screen. Better yet grab screenshot. You also need to say which values are on the left scale and which are on the right.

UDCmin and UDClim work on limiting your Torque request based on reported UDC.

[dadiowe]

Thanks Tom, exactly which values would you like me to show and how do I know if they are using the left or right scale?

Is voltspnt what controls the max charge voltage then?

[tom91]

Is voltspnt what controls the max charge voltage then?

Yes

Your previous pictures do not show the full plot views.

[dadiowe]

I understand about the photos I was just asking what parameters you need on the plot. This will be with 2.17A of course also what would you suggest would be Safe settings for udc and idc?

[tom91]

Leave them as they were at start of thread and change just one of them. UDCmin UDClim and IDCmin IDCmax, please only change one. Start with IDCmax as per BigPie's suggest. Raise it to lets say 500A.

Plot the following:
Potnom on left
UDC and IDC on right

[dadiowe]

Really struggling to get a plot using the wemos mini, as soon as I pull away I lose the wifi connection. This was the reason I tried to get the T-can working. Even with the T-can it was difficult but as I understand it the T-can won't work with 2.17A. Is there a way to connect via the serial connections on the zombie and read parameters that way?

[tom91]

Where is your zombie mounted? I know too little about your build nor is there very little info on your build on the forum.

Possibly an external wifi antenna for the laptop fixes it.

[uhi22]

For these kind of issues, would it be possible to add a spot value "limitationalReason" to the software, which shows the reason why the torque is limited? This would save a lot of analysis effort from my point of view.

[tom91]

For these kind of issues, would it be possible to add a spot value "limitationalReason" to the software, which shows the reason why the torque is limited? This would save a lot of analysis effort from my point of view.

Not really as alot of the base parameters need to be set right, but yes it is being worked onn to give a basic indication where to start looking.

[manny]

Looking at this plot

What is going on! under regen the UDC goes to 800V (if I read the plot correctly)


Is the switching value between UDC and IDC because of communication problem or something else?

[uhi22]

I guess the udc has a different scale on the right hand side of the diagram.

[dadiowe]

"Where is your zombie mounted? I know too little about your build nor is there very little info on your build on the forum."

I will probably post some pictures when I am happy that everything is where I want it to be and when I feel it is the finished article.


"Possibly an external wifi antenna for the laptop fixes it".

Interestingly I can connect to the zombie via my home network and the connection is solid when I am sitting in the car. The connection only seems to be "flaky" when I log in using 192.168.4.1 with my laptop or phone.

Zombie is sitting in the battery box at the moment. However the lid is not on the box, again only for testing. Bigpie has compiled a fixed 2.17A fixed version so I will be trying the Tcan asap. If that does not give a good connection I will move the wemos to inside the car with wired connection to the VCU. I will then try and get some good screen captures with both scales in view.

Side note, I had not noticed that the 2 plot buttons had a little l & r next to them. Funny how you can't see something in front of you sometimes.
Hopefully I can some logs over the weekend, subject to domestic duties from the boss.

[tom91]

Looking at this plot

What is going on! under regen the UDC goes to 800V (if I read the plot correctly)


Is the switching value between UDC and IDC because of communication problem or something else?

Guys please stay out of this, my self and BigPie are actively working on this. I do not want to cloud this thread up with guessing.