Inverter Firmware 5.35.R

[Zieg]

Okay, adjusting syncadv seems to have fixed the high speed unintended acceleration after full throttle, but I had to set it to 9, not 11. I don't understand the physics well enough, but I was wondering if maybe that's because the motor is mounted in reverse?

Well either way, I'm just happy It's fixed, thank you.

[johu]

I don't understand the physics well enough

Neither do I it seems. Unexpected result!

[glink]

Regarding the boat mode, I was investigating an issue with the motor keeping its pace even if pot is set to 0, so guess the boat mode will not be able to deal with this as it is regulating the pot, right? I'm unsure if my issue is due to wrong syncofs or something else. Seems to be related to the amount of resistance from the propeller, as if I test at the dock, motor is always following pot, but when running at speed motor is not obeying pot. When I disable field weakening, motor is always "behaving", both at dock and at speed, but with limited RPM.

[alexbeatle]

Is it possible to scale the regen from ibooster (CAN messaging) if using dual-channel wired throttle?

[johu]

Not directly but you can pick up the ibooster info and put it into the regenpreset field of the inverter control message

[alexbeatle]

Not directly but you can pick up the ibooster info and put it into the regenpreset field of the inverter control message

oh! That makes sense!
So when I started sending my auto-start via CAN control message, I set everything else in the CAN message as 0. This is probably why I lost my regen at the same time. I'll try to play with the regenpreset. Thank you!

[Ev8]

Does this new version allow can control of one oi board from another? If so can you give an example of what the tx and Rx settings would be? Thanks

[Zieg]

Okay, adjusting syncadv seems to have fixed the high speed unintended acceleration after full throttle, but I had to set it to 9, not 11. I don't understand the physics well enough, but I was wondering if maybe that's because the motor is mounted in reverse?

Follow up to this, I increased my fwcurmax after realizing it should probably be -500 instead of -400. That got be better acceleration at higher speeds (BMS reported 130-some kW and the battery wasn't even full!) but as soon as I let off the accelerator, it started accelerating on its own again.

I noticed the parameters wiki page says this about syncadv: "Must be set so that ud remains at 0 when coasting below field weakening speed". So I did some slow-speed coasting and ud was bouncing between 500 and almost -1500. How close to 0 SHOULD it be? Or is it more like it should be 'centered' around 0? I tried reducing syncadv from 9 to 6 and that seemed to shift the numbers to something closer to +/- 1000. I didn't dare drive faster at that setting though.

Is it possible that I should be setting syncadv to something like -10 (Or, I guess, 65525) because the motor is running in reverse?

[johu]

Or is it more like it should be 'centered' around 0?

yes

Is it possible that I should be setting syncadv to something like -10

No, that is handled in code

Code: Select all

advancedAngle = angle + dir * FP_TOINT(FP_MUL(Param::Get(Param::syncadv), frqFiltered));

Could you plot frequency and iq as well?

[Zieg]

Okay, perfect, thanks. I'll grab another log with frequency and iq as soon as I'm able. I'll have to do some more testing with different numbers but it's looking like I'll end up somewhere around 6 for syncadv. Very interesting, I'm glad you left it as a parameter rather than hard coding it!

[alexbeatle]

I know Tesla cars limit regen when battery >90%.
On my conversion I set the following. Is there perhaps a more elegant way of doing it? (BMS_SOC is in 0.1%, and I have 5% buffer)

Code: Select all

    if (BMS_SOC>=900) {Drive_Cmd_RegenPreset=0;} //apply no regen
    if (BMS_SOC<=850) {Drive_Cmd_RegenPreset=100;} //apply fully defined regen

[johu]

You could make it more gradual. Like

Code: Select all

RegenPreset=100 - (BMS_SOC - 850) * 2
RegenPreset = MAX(0, RegenPreset)
RegenPreset = MIN(100, RegenPreset)

[Boxster EV]

I've been running this Sine firmware in my Tesla LDU for sometime now with great success. However, I have noticed a couple of small quirks.

When changing direction, say reversing and then selecting FWD, the RPM or POT signal needs to be zero/resting otherwise the throttle doesn't engage. If you do start to apply throttle too early, nothing happens and the throttle needs to be fully released and then pressed again to work. Think of maneuvering in carparks frequently changing direction where it's most evident applying throttle/dir in the desired direction whilst travelling slowing in the opposite direction.

What's slightly more of a hindrance is in a similar situation, if the brake is used to slow the car reversing but not fully stop it, then FWD direction is engaged and throttle applied the car will attempt to move forward slightly but then register an overcurrent event, requiring a power cycle. Bear in mind the inputs during these scenarios are very mild.

Hopefully the feedback is useful to someone.

My file is here.

params.json

[Zieg]

Have you adjusted your DIRCHRPM at all? Looks like it was 100rpm from your parameters, not sure what speed that equates to though. Rough calc points to less than 1.5 kph?

[johu]

Unfortunately direction changes in the SINE variant were never tuned to work smoothly as my Polo only had a single channel encoder.

The behaviour is intentional but maybe not practical then?

Code: Select all

   /* Only change direction when below certain motor speed and throttle is not pressed */
   if (((int)Encoder::GetSpeed() < Param::GetInt(Param::dirchrpm) || userDirSelection == rotorDir) && !potPressed)
      selectedDir = userDirSelection;

[mane2]

What happens with Tesla motors if you up that value to like be like 10km/h? Factory EVs seem to have way higher limit.