Zombieverter & Gen2 Leafstack - RPM surging in Forward direction but stable in Reverse

[Iceman96]

Hi Brainstrust - I need some assistance with my full 2014 Leafstack that runs smoothly in Reverse direction, but is unstable and surges when spun up in the forward direction. I have searched the forum but have not found anything yet that matches my problems. This is my setup:

Full 2014 Leafstack ( with PDM divorced so it will fit in the vehicle)
Complete Original 2014 model 24 KWH BATTERY (approx 73% SOH)
Zombieverter running firmware 2.20A

I took a couple of sound recordings which I will upload, along with a copy of my paramaters file. I have it configured with the PDM, Inverter, and BMS all on CAN1, and where possible, I have copied the parameter settings Damien is using for his L200. I am assuming my problem is related to regen, as the surging is not there when I spin the motor in reverse. I have experimented with different settings for that, but without success.

I have wired everything up as per the “ Leafstack with a Zombieverter “ wiki, and something thing strange I have noticed is that once the 12 volt battery is connected and I then log in to the zombieverter that the spot value for “Invstat” shows up as “On”, even though I have not supplied the ignition or start signals to the zombie yet. In the wiki it says that I should get “Invstat on” after I turn on the T15 ignition input? When I do turn on the ignition the “Invstat” remains on, yet as soon as I send the start signal to the zombie the “ invstat” changes to off, even though the contactors do their thing, and the zombie goes into “Run” mode???
Another strange thing is that when I was setting up my min and max value’s for the 2 throttle pots my “potnom” shows random values, often negative, and I have never seen a value higher that 1.5, regardless of the throttle position?? I was expecting to see this change from zero up to 100% dependant on throttle position?

So I possibly have more than 1 issue going on here, and if anyone has any ideas your advice will be greatly appreciated.

Thanks - Andrew

params 16.json

(2.13 KiB) Downloaded 22 times

[tom91]

Let me guess, you are running the motor on the bench unloaded.

Turn the regen values to 0 it will go away.

[Iceman96]

Hi Tom,

Thanks for your reply. The stack is in the vehicle when I run it, but yes, the gearbox is in neutral, so the only load onthe motor is the input shaft to the gearbox. But i have also tried jacking up the car and putting the car in gear so it is spinning the rear wheels. No real change -if i select reverse the motor spins up gradually and smoothly - no problems at all. But if I select forward it is unstable, starts suddenly from a gradual throttle input, and is generally unstable.
I took your advice and I set the “regenmax” to zero and tried that. Reverse is still fine, and Forward is slightly better, but still not very good. I tried it with this setting with the rear wheels spinning, but again , it is unstable in the forward direction. I even tried applying a bit of handbrake to place some load on the motor but it made no improvement.

From my reading of the Parameters page that you have populated I see that “regenmax” is calculated based on the “potnom” value. On my system I definitely seem to have a problem with this value, as I have never seen a realistic value in this field regardless of the throttle position. Plus the only time I see any value in that field is when the Zombieverter is in Run mode.

I just dont understand why it is so good in reverse, and yet is so bad in forward - even with regenmax set to zero?

[Bratitude]

I just dont understand why it is so good in reverse, and yet is so bad in forward - even with regenmax set to zero?

in the web interface, when you select forward with the switch, dose the din values say forward?

iv never had effective reverse from the zombie. had almost no power. but that was on 2.17

[tom91]

Please make a plot of potnom and speed so we can review what is happening.

[Iceman96]

in the web interface, when you select forward with the switch, dose the din values say forward?

iv never had effective reverse from the zombie. had almost no power. but that was on 2.17

Hi Angus - Thanks for your input. Yes the Digital inputs for both Forward and Reverse both turn on as expected when I select either direction. And the “dir” spot value shows Neutral when neither input is selected, Drive when The Foward input is turned on, and Reverse when the Reverse input is selected. So everything seems to be ok there.
However The “ Invstat” spot value behaves oddly I believe. It shows up as “on” as soon as I log into the zombie ( IE when 12 volts is connected to the Zombie, and to pins 46 & 48 on the inverter, and pin 18 on the PDM). It remains as on when I turn on the ignition and the zombie digital input T15 turns on, then it goes to “off” as soon as I give the Zomie its “start” input.

Is that normal behaviour? as from my reading of the wiki it does not seem right?

[Iceman96]

Please make a plot of potnom and speed so we can review what is happening.

Hi Tom - wasn’t sure how to export the plots you requested, so I just took screenshots. I hope that is sufficient. One of the plots is just me gradually applying the throttle from 0 to 100% with the car turned off. Feedback as expected from both pots, but no feedback at all from potnom. The next plot is with the zombie in Run mode and Forward direction selected, I only applied partial throttle as the car was in neutral. Again, expected feedback from both pots, but almost nothing back on potnom. And the last one is in reverse, with similar results. I hope this is what you are after, and please let me know if there is anything else you need

[Iceman96]

Please make a plot of potnom and speed so we can review what is happening.

Hi Tom - I just realised you asked for potnom and speed. I will go and get those plots for you

[Iceman96]

Hi Tom, - this time I got the plots for speed and potnom as requested. First plot in Foward and 2nd in reverse. As you can see, no readings for potnom in either direction. Also, I gradually applied the throttle pretty much the same in both directions, but you can see from the plot,that in the forward direction that the motor gets up to the revlimeter ( 3000 rpm at present) far more aggresively. While in reverse it is a far smoother and gradual build up to speed. I backed off the throttle straight to zero the same in both forward and reverse, and in both instances the rundown to zero rpm is pretty much the same. I have regen set to zero at the moment.

[Iceman96]

Hi again Brainstrust,

I have an update on my throttle problems, after following Tom’s advice of turning off regen, it definitely improved my no load performance. It is still very smooth in reverse, and is a lot better in Forward, though it still seems to come on suddenly compared to reverse. I still seem to have issues with potnom and invStat however.

So I decided to take the car for a drive on the road to see if that made any difference. 1st problem I encountered was I had no torque in reverse. Unless the car was on dead flat ground, it really struggled to get it to move in reverse. I also noticed that the Zombie showed a “Throttle12diff” error message. My car is a manual, so I got arround this by selecting reverse gear and running the motor in forwards. I then tried driving the car forwards and now it performs much better. I thought the Performance and throttle responce while driving was very good. My only complaint would be that the initial start off is still a bit sudden. But once it is moving it behaves well. I will attach a copyof the current parameters in the Zombie, as well as a picture of the error message, and a plot i took of the vehicle taking off from a standing start on flat ground and accelerating up to approximately 60kph. ( car was in 3rd gear, which gives me a final drive ratio a little bit taller than the standard Leaf ratio). Potnom still does not show meaningfull values, and InvStat is still showing as off, even though the zombie is in Run mode, and drives the vehicle OK??

Thoughts anyone?

params 19.json

(2.13 KiB) Downloaded 11 times

[J0hannes]

After filling the coolant system, just before closing shop for the day, I did a short back and forth test. I still have 2.17 in my ZV and haven't updated to 2.2x.
Similar thing going on as in my first test "drive", forward direction I can hold against the brake and it really wants to go. Reverse direction I can literally floor it and it barely has any chooch. I'll do some CAN logging next to get a clearer picture on what's happening.

Posted this in my built topic. Haven't done any CAN logging or updated to 2.2 yet, but yeah, seems that we have similar troubles. Can you do CAN logging on your setup?

[tom91]

“Throttle12diff”

this will create a judder

Other thing to try drop your "throtramp": 10 to lets say 2.

Potnom still does not show meaningfull values

this is a problem, can you plot potnom, speed and torque

I had no torque in reverse

"throtmaxRev": 50 is what limits reverse torque, you can try changing this.

[Iceman96]

Posted this in my built topic. Haven't done any CAN logging or updated to 2.2 yet, but yeah, seems that we have similar troubles. Can you do CAN logging on your setup?

Hi Johannes - thanks for your reply. I will try and get some can logs today. I am currently using the throttle pedal that was in the car originaly, ( car is a Holden Rodeo, so not sure if the throttle pedal is an Izuzu or GM part), but I also have the original Leaf pedal from the donor car, so I am planning to try that today to see if it makes a difference ( I dont expect it will, but it is an easy thing to try). I will also try doing a plot like yours that includes torque to see if mine is similar to yours.

[Iceman96]

this will create a judder

OK - if I changed my throttle to single channel would that eliminate it as a cause of the judder at takeoff?

Other thing to try drop your "throtramp": 10 to lets say 2.

i will give it a try and report back

this is a problem, can you plot potnom, speed and torque

Will do

"throtmaxRev": 50 is what limits reverse torque, you can try changing this.

I had increased it to 50 from the default value of 30 that Damien uses, but it made no difference. But I will increase it to 100 and see if that does

[tom91]

OK - if I changed my throttle to single channel would that eliminate it as a cause of the judder at takeoff?

Set it to single channel and it will eliminate the error being called, which can be source of your juddder.

I had increased it to 50 from the default value of 30 that Damien uses

Note the Nissan leaf motor is limited in the the stock inverter firmware.

[Iceman96]

Further update:

As per my reply to JOhannes - I tried the original Leaf throttlepedal. I still get the “Throttle12Diff” error, however it is more responsive than the throttle pedal I had been using. It also made no difference to the lack of grunt in reverse.

I have not been able to get any CAN logs yet, as I am having trouble loading the GVRet sketch on to my CanDue for some reason. But I will persevere.

I have taken some plots which I will include below, but I can clearly see the negative torque requests when I select reverse, yet unless the car is on dead flat ground ( where it will creep backwards), it just cant move the car. I even tried applying the handbrake, and in reverse there is nothing, but in forwards it can easily overcome the handbrake.

For Tom - I did the plots you requested, and I thought I was still getting nothing on the Potnom channel. But I realised that due to the scale required for the RPM plot, that the changes in the Potnom channel probably would not show up. So to check my theory I did a plot of just the Potnom value. And sure enough, it now goes from 0 to 100 and back as I applied and removed the throttle. So maybe I need to start another thread to find out how to make the best use of the Plot feature? I have searched the Forum and the Wiki, but have not found anything helpful yet.

Also Tom - I noted your point that the Leaf inverter firmware has some inherent limits, but this leaf motor had plenty of torque in reverse when it was in the Leaf, so I dont understand why it doesn’t now. From my plots it appears that the Zombie is sending the negative torgue request ok, yet for some reasong the inverter is not applying it to the motor.

For JOhannes - I tried doing the same Plot as you had done, but due to the problems I am having with the Plot scaling I mentioned above, it does not show the change in direction like your plot did. But you can clearly see the negative torque request when the car was in reverse and would not move regardless of the amount of throttle applied.

This plot is taking off from a standing start on flat ground, accelerating to 60kph, slowing for a corner then accelerating to 60 again before backing off.

This plot is of just the Potnom value showing that it does go from zero to 100

This plot is an attempt at recreating JOhannes’s plot. Starting off, accelerating to 60kph, then coming to a stop and selexting Reverse, apply the throttle again but car would not move due to a very slight gradient.

[Iceman96]

And another update:

I realised today that I could not here my contactors opening on shutdown. I opened up the battery pack and checked them individually, and both the positive and negative contacts are stuck closed. The precharge contactor is still ok. I am not sure when they failed. I have ordered some replacements, but they wont get here for 3 weeks, so I wont be able to perform any more testing until they arrive. Very frustrating!

[tom91]

Get a CAN log please shows more detailed information.

The plots look fine to me.

[Iceman96]

Get a CAN log please shows more detailed information.

The plots look fine to me.

Thanks Tom - once my new contactors are in I will get some CAN logs for you.

[J0hannes]

Looked at parameters for the L200 Damien has shared on Github
He's using a single channel pot throttle setup. I'll try this also to see if it makes a difference on my setup.

[Iceman96]

Looked at parameters for the L200 Damien has shared on Github
He's using a single channel pot throttle setup. I'll try this also to see if it makes a difference on my setup.

Hi Johannes - I like what you have done with your excel comparison workbook. I have also made a comparison workbook, but i set mine up so it shows me what the parameter value is, rather than highlighting any differences. But after seeing yours I am going to copy your idea and add it to my workbook.

And yes, I have had a good look at Damiens parameters for all of his cars, but with particular interest in the L200. I had noticed that he was using single channel for the throttle. I changed mine, and while it stopped me getting the “throttle12diff” errors, it made no difference at all to my lack of torque in reverse. I even rewatched his video on the first drive of the L200 last night to see if he drove his in reverse. He definitely does, but I cant tell for certain if he is actually reversing the motor direction, or if he is using reverse gear in the gearbox. I suspect that he is a actually using the gearbox, so it is possible that he has not tried running his motor under load in reverse yet.

I also finaly managed to get the gvret sketch loaded on to my Arduno Due yesterday, but as you no doubt read in my earlier post, I also managed to weld the contacts closed on my Positive & Negative contactors ( I suspect I may have done this when unplugging the charging cable while it was under load). So while I can now get some CAN logs from the zombie, there is probably not much point until I can get some when it is trying to move the car in reverse.

I will be interested to hear how you get on with yours- Good Luck!

[J0hannes]

I also managed to weld the contacts closed on my Positive & Negative contactors ( I suspect I may have done this when unplugging the charging cable while it was under load). So while I can now get some CAN logs from the zombie, there is probably not much point until I can get some when it is trying to move the car in reverse.

I will be interested to hear how you get on with yours- Good Luck!

Yeah, I didn't look at Damien's params too closely earlier. I've been working on mechanical stuff on my build recently, so haven't put much focus in this yet.

Since you are cracking open the battery anyways to replace the contactors, you could in the meantime see if you can get them moving by tapping them to continue testing and diagnosing your setup. If you get them moving, I'd still replace them once you get new contactors.

I'm sure we'll get these running properly eventually

[Iceman96]

Hi JOhannes - good idea. I will see if I can free up the contactors to allow me to keep testing as you suggest. I have ordered some Kilovacs like the ones Damien put in his truck, so hopefully I wont have the same problem in the future. I will also pay more attention when charging the car too ! I am actually really happy with how my car performs with the 80KW Leaf motor, so your Suzuki should be an absolute rocket! Like you - I am confident that we will be able to get these issues sorted. I was at the stage where I was going to repackage my battery anyway, so even if I can’t resurrect the old contactors, i will put the time to use building some battery boxes and getting them installed in the car.

[Iceman96]

Well I went with JOhannes’ suggestion and pulled out the contactors and gave them a few good bumps on the workbench. It did the trick and they both came unstuck. I power cycled them a few times on the bench with an ohmeter across the contactors and both seemed to work ok. I reassembled the contactor block and put it back in the battery pack and reconnected everything.

So now with 12VDC going to the zombieverter, and to the correct inputs on the PDM, Inverter, and the Battery pack, but no “Ignition on” signal, or “Start” signal going to the Zombie, I logged in to the zombie to see what state it was showing. Everything looked ok to me except for the fact that the “InvStat” feedback was showing as on. But no volts were showing on udc, and the DC-DC converter was still off. I then did a download of the parameters which is below:

params No1.json

(2.13 KiB) Downloaded 9 times

I then connected up my arduino Due and started savvycan logging on my CAN1 bus, which has the zombie, pdm, inverter, and bms all on it. I continued logging for a bit as it was, and then turned the ignition on. I saw no appreciable difference on savvycan ( IE no new channels appered in the list of what was being logged). I then gave the zombie a “Start” signal. I-heard the contactors close, the zombie went into run mode, and udc1, udc2, and Invudc were all indicating approx 384volts, the DC-DC converter powered up as my aux volts wentup to 14.8v and everything seemed okother than my “InvStat” was now showing as off. I checked Savvycan and there was now extra channels being monitored.
I left it like that for a few more seconds, before turning off the ignition. It sounded like the contactors dropped out, but i was not sure. The zombie went into “Off” mode ok, “InvStat” went back to on mode, i think the dc-dc turned off as i am fairly sure the aux volts dropped back to approx 13volts. After a short wait i stopped the savvycan logging and the file is below:

rodeo-leaf log.csv

(3.32 MiB) Downloaded 10 times

Something strange that I noticed was that it appears that the value I had saved for udcsw of 360volts had been ovverwritten by the actions of starting up the inverter, as it was now showing a value of 364volts. (Not sure if it is a coincidence or not, but that was a value I had been using previouslyfor udcswI ). I tried doing a refresh, but the value remainned at 364 volts. So I did another parameter download which is attached below where you can see the value has changed:

params No2.json

(2.13 KiB) Downloaded 8 times

I then cycled power to the zombie and checked the parameters again, and udcsw was back to 360volts again???

I then tried powering up the inverter again to see if it did the same thing (it did), but on powerdown the contactors sounded funny again, so after isolating eveything again and checking, i found the contactors stuck closed again. So even though I did not run the main traction motor at all, the power draw of the dc-dc converter was enough to reweld the contacts, so I wont take any more risks with these contactors, and will have to wait until my kilovacs arrive and are installed before I can get any savvycan logs of the “no torque in reverse” issue. But hopefully the one log that I did manage to get may give some insight into why my “InvStat” is behaving the way that it does???

[tom91]

“InvStat” is behaving the way that it does???

Does nothing with the Leaf inverter. As indicated by the spot value description on wiki. https://openinverter.org/wiki/Zombiever ... pot_Values