GS450h Cuts Out on Heavy Throttle Application SOLVED

[dadiowe]

image.png (35.21 KiB) Viewed 2344 times

The above is my trace when the throttle was pressed and held at the point the cut occured. It does look very different to your trace.

Just a few thoughts on your parameters. I am not saying any of the following will work but its worth a try.

I see you are using single pot mode. Damien mentions a "gotcha" in his zombieverter video part 3, 36 minutes in saying that if you use single pot mode do not have a second channel wired in.

Also on the throttle, after update 1.10A there was a different method for calibrating the throttle which says to exceed the min and max values slightly rather than having the parameters inside the min and max values.

I see your UDCmin is set at 512 perhaps try reducing that to say 350 as an experiment.

Perhaps try dropping your idcmax to 500 or less. Damiens last video shows varying values up to 500 demontrating that the higher the number the higher the Amps

Also your BATTcap is set at 22Kw is that correct?

Is your wiring loom to the inverter original or home made? I had awful trouble modifying the connector to the inverter and making the connections reliable.

I hope one of these ideas helps.

[royhen99]

These plots might be more useful for diagnosis if the current is plotted on the right hand axis and at a higher resolution. More like this from the wiki,
https://openinverter.org/wiki/File:Plot.png

[Jackk]

Update....

So I replaced the encoder wires with 6 core shielded wire. Funnily enough the original cables I had installed were twisted and had copper tape wrapped around the whole length but was not grounded at the inverter end.

With the new shielded cable installed it still has the terrible cut out. In fact I would say it is worse now

I moved the encoder wires as far away as possible from the HV cables going to MG1 and MG2 ... The inverter is mounted about as close as it can be to the gs450h.

Please see pics for reference...

One thing I would say when it does cut out if I keep my foot in the same position it continues to almost switch the power on and off violently. Obviously I do not like doing this as it puts much strain on the drivetrain.

The only thing that has improved my issue so far is removing the current limit from the code.

Shall I get another plot of torque and IDC whilst trying to keep my foot still? Before my plots show me letting off the power.

[m.art.y]

How is your throttle pedal calibrated are you leaving enough room for noise? If this happens when pedal is fully pressed could it be that the pedal is in the territory where it gets outside of the calibrated area due to noise/wrong calibration?

[Jackk]

How is your throttle pedal calibrated are you leaving enough room for noise? If this happens when pedal is fully pressed could it be that the pedal is in the territory where it gets outside of the calibrated area due to noise/wrong calibration?

Ill re calibrate it but fyi its nowere near flat out when it cuts power... Probably half way

[evMacGyver]

The only thing that has improved my issue so far is removing the current limit from the code.

"6)if idc shows very high spikey values in time with the cutout then the vcu is receiving erroneous data from the isa shunt."

So is your CAN bus shielded where isa shunt is? If removal of current limit improves (fully?) your issue must be what Damien suggested, problem with current sensing.

[dadiowe]

I see from your parameters thsat all your CAN devices are set to 0 is it worth setting them up as per damiens settings as above?

[Jackk]

UPDATE.

Thanks to everyone for the help. I took the car out for another test just now while playing around with some parameters. My ISA seems to be reporting sensible values. I saw 350 amps max.

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! . I must thank Dadiowe for this as he suggested it!!

Could this be due to voltage sag under load going below UDC min? My Udcsw is still set to 512v so would it be ok to leave the UDCmin to 400v?

Interestingly my pack voltage 552v but the INV-udc is reporting 456v which seems strange

Thoughts?

[evMacGyver]

My ISA seems to be reporting sensible values. I saw 350 amps max.
..
Interestingly my pack voltage 552v but the INV-udc is reporting 456v which seems strange

Congratulation solving your mystery! Voltage sag is most likely the cause.

It's still strange that when you removed Idc limit from the code and so far it had most improvement. If you did set current limit high enough before current limit should not affect at all.

My experience on inverter voltage reporting is that my inverter reports less voltage while precharging, I don't remember exact difference but it was like hundred volts or even more. After precharge is complete and contactor closed inverter will show voltage correctly. Not really sure what happened to inverter as it was not always like this, but now I trust only ISA shunt voltage reading.

[royhen99]

From looking at the code the voltage derating control has a fixed gain and starts working within 20V of UDCmin. With a higher pack voltage this is a smaller percentage of UDC and so is possibly harsher than with a lower nominal voltage. Not sure if this is a cause of your problem but a lower gain might allow for more gradual derating before hitting UDCmin. Plots with torque, IDC and UDC may help to find a solution.

[PatrcioEV-ATX]

Funny, related story. I have also been experiencing the torque cutout under heavy throttle issue, but I'm using the GS450h VCU, not the Zombieverter. After reading through this thread, I realized I had the same solution as well because.... I had written in an extra bit of code with a parameter called DeadVoltage (thinking I'd be clever and protect the batteries).

My get_torque starts off with "if ((readyVal==true)&&(Sensor.Voltage>DeadVoltage))". Needless to say, as soon as I dropped the DeadVoltage parameter, my torque cutout immediately went away. Completely a voltage sag problem. Good times!