openinverter forum

Hi fellow enthusiasts,

after almost driving my conversion for two years now i would like to open a topic on a subject that i think is something that will push all conversions a big step towards that factory feeling:

There seems to be something wrong with the calculation of the basic inverter parameters that spin the motor. It is not consistent through the RPM range. In the low RPMs there is less power than in an original Leaf, then from 2000 till 3500 there is full power and then it gets less. Of course it cant maintain the power on the upper end of the RPM, but still it feels rougher than in an OEM vehicle. Next subject would be regen. I called Johannes about this with the idea to instead set a certain percentage of regeneration we rather should set a target value in KW or Amp. So we would maintain the same amount of regen while decelerating through a certain range of RPM. He told me that's the way it already is. But also here it is obviousely not as good as it could be. He already told me it's about the motor, the resolver and especially the math. I hope Johannes will explain this more detailed to get the people on board of this that we actually need! I also hope that we can puplicly discuss that subject and hopefully get the information to put in the inverter source code.

My intension here is that hopefully someone will come up with the solution for this.
Also i'm happy to invest some money in the research of this, like i did before with the more meaningful Ampere-Limits! if you can't solve this problem on your own like me, maybe you would also like to invest some money in the research by Johannes. He's never let me down when i had a problem with something, so i'm sure if he can gather the right information he'll solve this, too! And please don't get me wrong - i love to drive my conversion. But compared to my OEM EV (Ioniq) it just doesn't feel right and i think it could be!

Cheers, Philip

I think the root cause to this problem is "syncadv" i.e. the increasing angular error between measuring and processing. It has never been properly tuned and I don't have a good tuning method for it besides just trying different values. Maybe there is even some non-linearity in the resolver, I don't know.

Regen in FOC is specified as a percentage of throtcur, so essentially it is specifying a motor current. If everything were linear then regen would be linear throughout the non-field-weakened rev range.

The way to calculate regen current is simply throtcur*brkmax or throtcur*brknompedal (and possible scaling by pot2). So say brkmax=30 and throtcur=6 than off-throttle regen would be 30*6=180A motor current (not DC current!).

Since I have a tesla LDU, I don't think it affects me. Nevertheless it should be possible to set Syncadv to 0 and to change the value for Syncofs externally via CAN depending on the speed. With a map, non-linear courses would also be possible.

To throw more confusion into the mix pm (leaf motor) and induction (Tesla) use different firmware and control methods.
To my understanding it is the pm motors that seem to be having this issue.

Edit: apologies missed the probably not affected bit

rstevens81 wrote: ↑Wed Sep 29, 2021 11:27 am To throw more confusion into the mix pm (leaf motor) and induction (Tesla) use different firmware and control methods.
To my understanding it is the pm motors that seem to be having this issue.

that's exactly what I wrote! I am not affected.

So Zapatero experimented with syncadv a bit before a special incident which he will soon present

Apparently the problem can be cured by carefully increasing syncadv. A good spot is somewhere between 10 and 20 dig/Hz. Maybe 13 or so.

I will do some testing myself soon.

Yes, syncadv does the job and also the linear increase of fweak instead of a steady value. Johannes sent me a beta firmware that i tested yesterday - succesfully! The car had no more torque loss after 4000 rpm and pulled strong up to 7000 rpm. Power output was now constant high throughout the rpm range. Regeneration was still an issue that need so be solved!

Incident: Unfortunately the first firmware had no RPM-lilimter. That and the fact the syncadv acts strange the higher you set it according to how the engine spins led to a fried motor. The rpm went for a second around 25.000 or so - now i have sparks between the motor and the chassis, i think the rotor didn't like it. The car doesn't drive anymore, i'm just able to rev it once and then when the rpm is zero again i can rev it again. Motor is definately fried Bought another one that should arrive next week.

I sacrificed a motor for the cummunity
Reminds me of some of Damiens videos though

johu wrote: ↑Thu Sep 30, 2021 9:54 am So Zapatero experimented with syncadv a bit before a special incident which he will soon present

Apparently the problem can be cured by carefully increasing syncadv. A good spot is somewhere between 10 and 20 dig/Hz. Maybe 13 or so.

I will do some testing myself soon.

Very interesting! Maybe this can help to cure some of the odd behavior of the MGR-motor as well, specially together with the 5.11 SW?

Possibly yes. Want to put it back in?

Yes, if it can be made to work, sure. Much more sensible motor in my car...

A thought regarding regen. The syncadv should teoretically be a stationary value but in practice it is not then maybe when changing sign on the torque (regen) requires some change to syncadv as well?

Well, that's what you get when the thirst for power is too big

The whole story here: viewtopic.php?f=9&t=1960

.