LDU Tune Issue - WOT Mystery

[P.S.Mangelsdorf]

As many of you know, I have an LDU in a 1940 Chevy, powered by two Chevy Volt (aka Opel Ampera) batteries. This is a very similar setup to Jon Volk's and others'.

I have been chasing a WOT tune problem for a while. This weekend I had my first chance to test on a track, and was able to determine a few things. My base parameter file is attached.

If I hold the throttle just shy of WOT, it will pull all the way through the 1/4 mile.

If I hold WOT at any point for more than about 2 seconds, it will hit what I believe is an overcurrent trip and shut off. Since I was on track, I couldn't pull up the error message before power cycling. (once I power cycle it starts right up and will run fine to get off track.) I have GoPro footage of the runs I haven't gone through yet to get some more accurate timing and see if that helps at all. I believe that it was at least 2 seconds because my 1.9 sec 60-foot time with a WOT launch was better than my 60-foot time when holding just shy of WOT.

I tried launching by holding throttle and brake, then dumping brake, and by rolling into WOT both quickly and slowly. All three ways, any sustained WOT caused a trip, and seemingly regardless of speed.

I tried the following, with no success:

• Raise fweak

• lower fslipconstmax

• lower fslipmax

• lower both fslipmax and fslipconstmax

On the street I previously tried lowering boost.

In theory, my tune is "softer" than Jon Volk's or BoxsterEV's. I'm at a loss here, I cannot figure out what is going on.


I do not have a data log of the runs. My phone won't reliably connect to the LDU when I'm not physically holding the phone up to the LDU case, and I don't have a way to secure a laptop in the car. I'm working on a plan for the later, or a cheap tablet to do the same.

If anyone has some ideas, I'm all ears. The only possible idea I have that I haven't checked yet is maybe the potmax value is too low and WOT goes out of range? But I'm not sure why that wouldn't be instantaneous. Is there some other parameter that only comes into play at WOT that I'm missing?

[Boxster EV]

Strange issue. The main differences from my tune is that I run a lower throtramp and ampmin setting.

I am however running one Chevy Volt pack and it's becoming a little saggy in its old age.

[P.S.Mangelsdorf]

I'm working on a data logging setup at the moment. I was able to check while driving in my neighborhood, and the error code is in fact OVERCURRENT.

My current hypothesis is that the current gain is incorrect (others have said that the gain for the LDU seems incorrect), and that is causing the LDU to hit ocurlim. My first test will be to log current on my ISA shunt and LDU via CAN, and compare the two. I was setting up CAN from the LDU anyways, because my new Speedhut EV gauges arrived last week.

[P.S.Mangelsdorf]

I have some updates, but they are only more confusing.

Based on a discussion with @Clanky , I learned that this issue is affecting others, as well as my setup. Based on that, the assumption was that this was an issue in the firmware. I recalled this issue springing up with some updates I made in early 2022, and being confounded at the problem then. I eventually assumed it was connected to my axle damage, but now know that wasn't the case.

On to the update:

Today I tired backdating the firmware as far back as 4.87. I know when I was back on a firmware in that time frame, I didn't have this issue. When I tested today, I was still hitting an error and shutting off.

This leads me to two possible conclusions, but neither make much sense.
1) There is some sort of damage to the inverter that is only showing itself under full power, and that damage is also present on @clanky's car.
2) There is some hidden setting that was changed, and so even when I back dated it didn't get updated.

Any setting that would seem to lead to this issue matches known working parameter sets, or is softer than them. I'm really at a loss here.

Current parameter set is attached. This set is absolutely wonderful to drive, with the exception of this full throttle problem, which is quite annoying because the point of building an EV like this is the face-melting pulls.

[johu]

Just wondering, if you set fslipmax=fslipmin essentially nothing should happen above slipstart=62%. Does the issue then go away? Of course not meant as a solution, just a test

[P.S.Mangelsdorf]

Just wondering, if you set fslipmax=fslipmin essentially nothing should happen above slipstart=62%. Does the issue then go away? Of course not meant as a solution, just a test

Just tested this, and yes, the issue went away, it did not trip.

This was with fslipmax=fslipmin=2.37 and battery at about 50-60% SOC. I held the throttle down to about 80mph and then ran out of road testing space.

[johu]

So could it be that simple, those 3 Hz are more than the power stage can handle? Now you can slowly increase fslipmax to see when the issue returns

[P.S.Mangelsdorf]

So could it be that simple, those 3 Hz are more than the power stage can handle? Now you can slowly increase fslipmax to see when the issue returns

I don't think so. This gets to my point about other known working parameters. JonVolk was running the same battery set up and the same inverter, with fslipmax of 3.18. BoxsterEV is running 5, though with a less powerful battery. The set in the parameter database uses 3.09.

[P.S.Mangelsdorf]

I may have solved it. Lowering fslipconstmax seems to have addressed the issue. However, I was testing at mid-low SOC, and I could swear I tried lowering fslipconstmax before, so I need to try again on a full battery to confirm. It also doesn't answer the issue of why it needs softer settings than similar setups, but its progress.

[mane2]

In my Sport LDU setup (240V-300V pack), when using fslipconstmax of 9 it makes the motor shake / vibrate. Only when having it the same as fmax it seems to be fine, but I feel like I'm missing some power that that setting could bring. It even shakes when set to 5 or even 4 a bit. I don't understand how someone can have it up to 9..

[johu]

It must be seen together with fconst. fconst must be a good deal higher than fweak to make physical sense. It's the frequency at which fslipconstmax is applied at WOT

[mane2]

It must be seen together with fconst. fconst must be a good deal higher than fweak to make physical sense. It's the frequency at which fslipconstmax is applied at WOT

Yeah, I had it set to up to 800hz even. But I think in my case, the problem seems to be voltage sag. With WOT voltage might drop down to 230V which causes Sport LDU to go into some mode that it won't put out much power anymore. Only with almost 100% SOC I can do WOT accelerations (because the voltage won't drop below 240V - which seems to be the low end limit to LDU).

I guess I need to add module or two.. Or is there some tricks to prevent voltage sag?

[P.S.Mangelsdorf]

My car is off the road right now due to broken motor mounts, but I did remember something from when I was playing around to try to solve this that did not make sense.

I tired once to set fslipconstmax=fslipmax=3. It almost immediately after launch had an OC trip. If fslipconstmax was really the cause of the other trips, this shouldn't have happened, right? Because it should have been disabled?

I need to play around with this more but its going to be a few more weeks before I have the car back up and running. The (sort of) funny thing is that I broke the motor mounts trying to solve this in the pits on day 5 of Drag Week.

[P.S.Mangelsdorf]

I believe I have a rough tune that functions. It still needs work, but this did let me get a good pull in.

I took the car to a large parking lot at a local community college where I could just keep making changes without disruption.

Turns out fslipconstmax of 9 is perfectly fine. At least in my car, it seemed rough when set lower.

I ended up playing with fslipmin and fslipmax. I backed both way down, and then started moving fslipmax up first in increments of 0.1 and then 0.05.
Once I got a trip again, I went back to the last setting without a trip, and worked up by 0.01. At approx 85-90% SOC, I was able to get a long pull in without tripping an OC/Desat error at fslipmax=2.63. At a higher SOC, it may need to be a bit lower. It seemed to still pull well at that setting. fslipmin at 1.4 is probably a little too low, but worked ok.

There's still a lot to fine tune and probably more power in it. I think there is more fine tuning of fslipmax, and there's definitely room to play with fweak. It definitely needs more street driving to confirm it works at all speeds and loads. I needed to stop with my testing and was planning to do more during the week.

Unfortunately, I damaged the inverter on the drive home. I don't believe it was due to the tune, the 12V system was having issues and I'm fairly certain I dropped a contactor while already decelerating. Now as soon as I trigger start, I get and OC error.

Anyways, I've attached my parameter set here, you are all welcome to give it a shot.