openinverter forum

Thought it was time to draft a new release. Tested in Polo and working. This is still scalar control software, FOC release will come later.

Changes:

• Added error message for stuck PWM pins
• Added possibility to selectively delete CAN mappings
• Made real use of GPIO bit set/reset register
• Do not calculate slip in synchronous mode
• Added support for Blue Pill board (reduced and remapped pins)

To use the CAN map deletion feature you also need to upgrade your web interface (I think index.js suffices)
https://github.com/jsphuebner/stm32-sin ... ag/v4.53.R

Hi J did you implement some PI softening of throttle zero transition because i still feel quite a surge when i quickly release throttle through 0 and then press it back to positive torque.

arber333 wrote: ↑Wed Aug 28, 2019 11:40 am Hi J did you implement some PI softening of throttle zero transition because i still feel quite a surge when i quickly release throttle through 0 and then press it back to positive torque.

I’m also curious about this as well. I’ve gotten fairly good at driving around it, but now that I’m starting to push the car harder in corners, it would be nice to have it transition smoother on it’s own.

In the upcoming version I was planning to have *one* throttle ramp for all transitions. I.e. stepping ON the accelerator but also coming OFF. And also touching the brake pedal. You guys could try it out as I can't really reproduce the problem with my car.

johu wrote: ↑Thu Sep 05, 2019 12:09 pm In the upcoming version I was planning to have *one* throttle ramp for all transitions. I.e. stepping ON the accelerator but also coming OFF. And also touching the brake pedal. You guys could try it out as I can't really reproduce the problem with my car.

I think this is not the fault of your motor, but rather mine. You use 1Hz for start slip, I hve the best efficiency with 3Hz and 21% amps at 7000pt voltage. This makes the car very lively, but torque is humongous.
Now that i thought more about this i really dont have a problem when starting or regenning. Really vibrations show mostly when on level road with throttle near or at 0. Then inverter transitions from drive and regen very often and i can feel this as shaking on chassis. If i apply a little more throttle or brake this subsides immediately. BUT results are most evident on clutch splines and coupler after one year in use splines are mostly gone!
By splines i mean center of the clutch plate that has softer splines than gearbox shaft.

I think wide (or configurable) 0 throttle field would help for start, later i can help you test PI loop also.

BTW, did you consider that probably FOC will solve this anyways. Just because you need PI to govern torque response.

Well i tried the 4.53R firmware.
For me there are no drastic changes noted. I found some issues though.

1. I couldnt get cruise mode to work. And i had a feeling if i would hold throttle pedal at or around 0 position then cruise would start to pull, but after normal throttle would be in control.
2. I got severe vibrations in regen mode under 2500Rpm. My regen setting was 50% and Regen pot was some 60% open.
3. I got tolerable vibration in drive mode around 2500Rpm but only if i kept throttle from 0% to 10%. Torque would set and then release in frequency about 2Hz or maybe 3Hz. If i applied more torque this would get dampened. I tried to play with P and I filter factors but no result.

In the end i returned to the old code with your elaborate filtering method which at least dampens drive mode vibrations and regen vibrations seem to be less.

arber333 wrote: ↑Wed Aug 28, 2019 11:40 am Hi J did you implement some PI softening of throttle zero transition because i still feel quite a surge when i quickly release throttle through 0 and then press it back to positive torque.

Late response (been away for a while, back now) I also feel the same behavior. When the throttle is near the zero point, the slip jumps from -1*slipmin to +slipmin instantly. While driving the car with speeds near say 5-10 km/h, releasing and pushing the throttle (or driving near zero point) makes the car very jumpy. fslipmin is 1.8 Hz, boost is 3000. We need something to ease in the transition between regen and torque.

The slip and ampnom is IIRFILTER'ed in the CalcAmpAndSlip() method actually, but it's not smoothing out the slip enough (at least for my case). I tried using rampup and rampdown functions on fslip and ampnom to make it even smoother but didn't work as I expected.

nailgg wrote: ↑Fri Sep 13, 2019 8:17 pm

arber333 wrote: ↑Wed Aug 28, 2019 11:40 am Hi J did you implement some PI softening of throttle zero transition because i still feel quite a surge when i quickly release throttle through 0 and then press it back to positive torque.

Late response (been away for a while, back now) I also feel the same behavior. When the throttle is near the zero point, the slip jumps from -1*slipmin to +slipmin instantly. While driving the car with speeds near say 5-10 km/h, releasing and pushing the throttle (or driving near zero point) makes the car very jumpy. fslipmin is 1.8 Hz, boost is 3000. We need something to ease in the transition between regen and torque.

The slip and ampnom is IIRFILTER'ed in the CalcAmpAndSlip() method actually, but it's not smoothing out the slip enough (at least for my case). I tried using rampup and rampdown functions on fslip and ampnom to make it even smoother but didn't work as I expected.

How did you determine you need boost 3000? I was driving for 2 years with slip 4000 and ampmin 12. Motor was a bit like rubber cord. When i started it needed quite a lot of throttle and then i had to release it since it was too much for driving. Regen was also a bit jumpy. I also had at least one IGBT blown per year.
When i bought semiron drivers with desat at first i had a lot of desat events on driveoff, which told me parameters were off. First i increased boost untill desat wouldnt happen and i kept increasing untill desat happened again. Then i reduced boost to a round 7000!
Next step i increased ampmin up to when i could takeoff into a small hill with very little throttle and kept climbing. No bungee there....
Try if similar happens with your design.
I am attaching a firmware Johannes did for me to experiment with. It involves some funky math but vibration is largely gone. Regen is still weird though. I have regen reduced to 30% and my regen pot is closed to 50%...

Is the throttle jitter caused by trying to implement 1 pedal driving?
Wouldn't this problem go away if Zero throttle = Zero torque?
ie. take your foot of the accelerator and car coasts 100%?
Then put a pot on the brake pedal so that first part of pedal movement introduces regen until you have 100% regen plus physical brakes starting to be applied?

That's how I would set this up anyway. I want to be able to completely freewheel at will.

sfk wrote: ↑Sun Sep 15, 2019 8:43 pm Is the throttle jitter caused by trying to implement 1 pedal driving?
Wouldn't this problem go away if Zero throttle = Zero torque?
ie. take your foot of the accelerator and car coasts 100%?
Then put a pot on the brake pedal so that first part of pedal movement introduces regen until you have 100% regen plus physical brakes starting to be applied?

That's how I would set this up anyway. I want to be able to completely freewheel at will.

Yes i have it setup exactly like that. I just had to remove my Colvern pot from brake. It developed some jitter at 1/3 of motion... I replaced it with normal twist linear pot under the dash. It works but it still doesnt clear jumpy regen. It is a lot better with old code.

Hm... i just may try the dual pot concept since i have prius pedal with two channels. Maybe this will be better.

arber333 wrote: ↑Sun Sep 15, 2019 10:05 pm It is a lot better with old code.

Hi Arber,

which version was it better with?