I don't see an option to measure offsets in the mobile version of the tool. Is it possible there? I see it on the pc version but my stormcore is in a silicone beaded enclosure that I was hoping to avoid opening for a while. No problem if it is PC only, I can re-seal it.
Just found something a bit weird, not sure if its because of the new firmware or not.
My hardware is 2 low side shunt, with current and phase filters.
If I set 200A motor current and 20A battery current, this weird limiting happens and the motor slows down. I have a 20A power supply which is not limiting the current and the input voltage is solid.
If I then up the battery limit to 30A it reduces the affect.
Is that foc mode? I quickly checked the code in the foc algorithm and I did not see anything suspicious. If you are using foc mode, I will have a deeper look.
Is this going to get pushed into master at some point and made the main branch everyone uses?
I made some VESC compatible hardware, and for months struggled with motor screaming, random stalls resulting in violent braking and all manner of problems even though everything checks out perfectly on the hardware (loosely 100/250 based; I don't have that hardware so guessed what it was like based on the config file)... I put this branch on it a few days ago and suddenly... everything works perfectly... Even HFI works properly on a 5uH (1.5uH?) CA120-150kV motor.
There are substantial differences I have observed - the inductance measured on several motors I have tested varies by a factor of 2+, with the later code probably giving more credible results. The detection is far more reliable (previously I used to frequently fail detections on my custom hardware often and fail with FSESC 6.6 with skateboard size motors) but now it just detects with no issues.
Not sure if this is the panacea it seems or if you have other issues buried underneath this new one I am unaware of, but having seen the old firmware repeatedly put a 100+V spike back onto my 48V lab PSU while instantly locking the motor, only to have the same hardware perform very well with the new firmware... I feel like it might be time to make a new release widely available, and the updated tool referenced from the project with the ability to upload new firmware from it?
Update, having noticed that since I started writing the last post TechAmNu has been posting...
I also saw a strange effect at 200A - I set FOC openloop 200A and 300 speed whatsits... The motor span fine, board got hot very fast... But... Setting 250A only ever went to ~215A. Like TechAmNu, my PSU was not limiting, draw was about 300W @30V and the PSU is capable of about 500. I had set a battery limit of 20A. I thought not much of it, assumed it was PSU oddity related... but now reading someone else saw it as well...
What is the best way to measure and reduce the torque ripple I'm experiencing? Is there a way I can observe if this is instability in the current control. Is adjusting Kp and Ki the best way to reduce?
By the way, thanks for all the work that has gone into the 5.3 beta. I'm running beta 59 and the recent braking fix has eliminated a brake cogging problem I'd been experiencing in 5.2. This is great progress.
What beta version number does the firmware page show at the bottom?
5.3 Beta 60
What are the values on the foc -> offsets page? Can you try to run measure offsets and disable the calibration at boot?
I don't see an option to measure offsets in the mobile version of the tool. Is it possible there? I see it on the pc version but my stormcore is in a silicone beaded enclosure that I was hoping to avoid opening for a while. No problem if it is PC only, I can re-seal it.
Primary Side:
Current Offset 0: 1990.49
Current Offset 1: 1958.90
Current Offset 2: 2019.50
Voltage Offset 0: -0.0021 V
Voltage Offset 1: -0.0015 V
Voltage Offset 2: 0.0036 V
All Voltage Offsets Undriven are 0.0000V
Secondary Side:
Current Offset 0: 2020.35
Current Offset 1: 2009.76
Current Offset 2: 2013.79
Voltage Offset 0: 0.0047 V
Voltage Offset 1: -0.0023 V
Voltage Offset 2: -0.0023 V
All Voltage Offsets Undriven are 0.0000V
You can run foc_dc_cal from the terminal in the mobile version and read back the motor configuration a few seconds later.
Hi Benjamin
I set you a PM message with function I am looking for to be add to VESC.
Would you please look that and check how easy this can be add ?
Thx
Martin
Martin
Just found something a bit weird, not sure if its because of the new firmware or not.
My hardware is 2 low side shunt, with current and phase filters.
If I set 200A motor current and 20A battery current, this weird limiting happens and the motor slows down. I have a 20A power supply which is not limiting the current and the input voltage is solid.
If I then up the battery limit to 30A it reduces the affect.
50A battery limit removes the affect completely.
Do you want to tell something about that large speed controller A500S ?
Hi TechAUmNu,
Is that foc mode? I quickly checked the code in the foc algorithm and I did not see anything suspicious. If you are using foc mode, I will have a deeper look.
Yes its FOC, I haven't seen this before with older versions.
@Markula its A400S, 100v fets
I tried again with heavy load and it's actually starting to go backwards!
200A motor 10A battery with load
200A motor 20A battery with load
200A motor 50A battery with load. Limited a bit at the end by the psu current limiting and reducing the voltage.
That is crazy. Can you also make some pictures of the FOC graph?
Do you use any feedback? Position/velocity pid?
10A battery
20A battery
50A battery
You don't use a BMS, do you?
And what do you use to generate that setpoint?
No just straight out of power supply. No other limits should be affecting it but the input current limit.
The sweep is done by the experiments tab, duty cycle sweep from 5% to 25% at 0.25s per 1%.
Checked all places where that parameter is used, but found nothing strange. I do not understand how that parameter can cause that behavior.
Hi,
Is this going to get pushed into master at some point and made the main branch everyone uses?
I made some VESC compatible hardware, and for months struggled with motor screaming, random stalls resulting in violent braking and all manner of problems even though everything checks out perfectly on the hardware (loosely 100/250 based; I don't have that hardware so guessed what it was like based on the config file)... I put this branch on it a few days ago and suddenly... everything works perfectly... Even HFI works properly on a 5uH (1.5uH?) CA120-150kV motor.
There are substantial differences I have observed - the inductance measured on several motors I have tested varies by a factor of 2+, with the later code probably giving more credible results. The detection is far more reliable (previously I used to frequently fail detections on my custom hardware often and fail with FSESC 6.6 with skateboard size motors) but now it just detects with no issues.
Not sure if this is the panacea it seems or if you have other issues buried underneath this new one I am unaware of, but having seen the old firmware repeatedly put a 100+V spike back onto my 48V lab PSU while instantly locking the motor, only to have the same hardware perform very well with the new firmware... I feel like it might be time to make a new release widely available, and the updated tool referenced from the project with the ability to upload new firmware from it?
Update, having noticed that since I started writing the last post TechAmNu has been posting...
I also saw a strange effect at 200A - I set FOC openloop 200A and 300 speed whatsits... The motor span fine, board got hot very fast... But... Setting 250A only ever went to ~215A. Like TechAmNu, my PSU was not limiting, draw was about 300W @30V and the PSU is capable of about 500. I had set a battery limit of 20A. I thought not much of it, assumed it was PSU oddity related... but now reading someone else saw it as well...
What is the best way to measure and reduce the torque ripple I'm experiencing? Is there a way I can observe if this is instability in the current control. Is adjusting Kp and Ki the best way to reduce?
By the way, thanks for all the work that has gone into the 5.3 beta. I'm running beta 59 and the recent braking fix has eliminated a brake cogging problem I'd been experiencing in 5.2. This is great progress.
Man, I think I finally found something. There is a feedback behavior that kicks in with the current battery limiting implementation.
Hacked the firmware a bit to trace mod_q and iq_target values. Since iq_target makes mod_q and mod_q is used in iq_target limitation, this occurs:
total current = iq_target
FOC phase = mod_q
zoomed in:
You cannot see this behavior in the other traces, they are way too slow for this.
Solved the problem by adding a tiny bit of low pass: