openinverter forum

Hello Guys,

I recently got my self a Tesla SDU inverter and OI drop in board and it having some problem, here the list of question and problem that I have encountered.

1.) Is the OI drop in board for Tesla SDU can be program to drive PMSM with FoC
2.) I have tried using various kind of sensors, such single channel PWM encoder( not getting any response on graph) , sin cosine mlx90380( showing graph but missing 355- 2 degrees) , AB encoder. - no showing any on the angle graph.
3.) Power going to main or battery is always tripping as if it is deadshorted, but the on OI web interface shows negative values.
4.) Il 1 shows negative values even the inverter is on idle.
5.) Inverter always showing overcurrent if I adjust current limit to 20a.

Hope someone can help.

Best regards.

Iccie.

iccie2 wrote: ↑Fri Mar 27, 2026 4:29 pm Hello Guys,

I recently got my self a Tesla SDU inverter and OI drop in board and it having some problem, here the list of question and problem that I have encountered.

1.) Is the OI drop in board for Tesla SDU can be program to drive PMSM with FoC
2.) I have tried using various kind of sensors, such single channel PWM encoder( not getting any response on graph) , sin cosine mlx90380( showing graph but missing 355- 2 degrees) , AB encoder. - no showing any on the angle graph.
3.) Power going to main or battery is always tripping as if it is deadshorted, but the on OI web interface shows negative values.
4.) Il 1 shows negative values even the inverter is on idle.
5.) Inverter always showing overcurrent if I adjust current limit to 20a.

Hope someone can help.

Best regards.

Iccie.

I would suggest you go and research the shit out of posts on this blog as this is something no one has tried yet....
1. FOC should load but you need to know which settings/inputs you need to select in the web interface so it doesnt trip I dont know about running PMSM with tesla board, i know i did it with gen2 OI board
https://openinverter.org/forum/viewtopi ... 888#p25888
2. Could be that your sensor is saturating the input. Maybe you need a divider to lower voltage to 1.66V
https://openinverter.org/forum/viewtopi ... 979#p16979
3. there are various settings to swap the order of inputs and you have to get them correct
4. ditto
5. Overcurrent level is meant to be sensed for protection only so at 400A or more. This is not so much active value, rather taken as protection from OC.

We cant hold your hand here... you will need to research the SDU development threads here where there were those milestones you need to solve.

you are pioneering here as the SDU board is meant to drive the SDU motor with direct sine control.
You might be able to run a sin/cos chip on the AB inputs.
AB encoders are not suitable for PMSM control.

arber333 wrote: ↑Fri Mar 27, 2026 6:57 pm I would suggest you go and research the shit out of posts on this blog as this is something no one has tried yet....
1. FOC should load but you need to know which settings/inputs you need to select in the web interface so it doesnt trip I dont know about running PMSM with tesla board, i know i did it with gen2 OI board
https://openinverter.org/forum/viewtopi ... 888#p25888
2. Could be that your sensor is saturating the input. Maybe you need a divider to lower voltage to 1.66V
https://openinverter.org/forum/viewtopi ... 979#p16979
3. there are various settings to swap the order of inputs and you have to get them correct
4. ditto
5. Overcurrent level is meant to be sensed for protection only so at 400A or more. This is not so much active value, rather taken as protection from OC.

We cant hold your hand here... you will need to research the SDU development threads here where there were those milestones you need to solve.

Thanks,

I'm looking into it.

What I have done today are;

1.) Checked SDU main battery supply using diode mode. Got some reading on one way but not on other way.
2.) Checked SDU main battery suppli to each phase. Got initial deadshort for one second then disappear. Reverse leads I Got some readings.
3.) Tried switching using pinswaps,. But no changes on negative il1 readings.

I'm have not yet checked on board circuit level and schematics.

johu wrote: ↑Sat Mar 28, 2026 6:48 am you are pioneering here as the SDU board is meant to drive the SDU motor with direct sine control.
You might be able to run a sin/cos chip on the AB inputs.
AB encoders are not suitable for PMSM control.

Yes. I have little background on electronics specifically on board level. So it's seems like a uncovering a crime scene here to me.


Awhh I see. So therefore PMSM needs absolute position to control it FOC?. Now I get why some controller dont implement Ld and Lq designation.

I have tried mlx90380 , I checked both output pins but they don't read 3.3v on max and min it shows somewhere around 0.5v. And on the graph I am missing 0-5 degrees and 350-360 degrees using sin/cos encmode. As you mentioned the inverter can adjust minor discrepancy but I am not sure what it really mean. I tried using AB and Single channel encoder, but not showing any on the graph. I will check the board level soon.

Still currently learning about motor saliency.

For to days update,

I was able to run the inverter without tripping unlike before.,

What I have done was to correct current sensor power supply, now it's reading 2.5v on output pin using multi meter then set inverter pinswap to : 1 and 2.

The motor doesn't spin but rather vibrate after changing lq and Ld current value. I assume it needs absolute position to run the Foc.

I'll try to work with it sensor again using mlx90380. As I don't have working resolver the moment.

arber333 wrote: ↑Fri Mar 27, 2026 6:57 pm I would suggest you go and research the shit out of posts on this blog as this is something no one has tried yet....
1. FOC should load but you need to know which settings/inputs you need to select in the web interface so it doesnt trip I dont know about running PMSM with tesla board, i know i did it with gen2 OI board
https://openinverter.org/forum/viewtopi ... 888#p25888
2. Could be that your sensor is saturating the input. Maybe you need a divider to lower voltage to 1.66V
https://openinverter.org/forum/viewtopi ... 979#p16979
3. there are various settings to swap the order of inputs and you have to get them correct
4. ditto
5. Overcurrent level is meant to be sensed for protection only so at 400A or more. This is not so much active value, rather taken as protection from OC.

We cant hold your hand here... you will need to research the SDU development threads here where there were those milestones you need to solve.

1.)I actually saw someone who made PMSm running on version 9, but not on version 8 board ( the same board that I have)
What's the hardware difference on version 8 and 9?

2.) As far as the sensor I didn't use voltage divider but rather smd chip that converts 5v to 3.3v, then I used back converter. The smd seems like giving me a closer to perfect 360 position. While the buck converter have missing angles with suddent drop 0 degress then jump on some random angle depends on sensor and magnet position.

3.) I have done it. Now is working without tripping on pinswap 12.

4.) Working

5.) The board seem to work now.

At the moment I am running small motor rated on 20v, from portable chainsaw. It is also an ipm motor. With 2 pole pairs. Then attached a AB encoder into it which the inverter don't read. Converter it to sin cosine encoder.

johu wrote: ↑Sat Mar 28, 2026 6:48 am you are pioneering here as the SDU board is meant to drive the SDU motor with direct sine control.
You might be able to run a sin/cos chip on the AB inputs.
AB encoders are not suitable for PMSM control.

It seems like a and b channels is readding it but I am not getting full 0-360 rotor position using encoder . I checked encoder ground to pin 22 and 23 of MCU, pin22 is reading 1.3v while pin 23 reading 4.9v is it normal? Any guess what's causing this?


Best regards.

iccie2 wrote: ↑Tue Apr 07, 2026 2:13 am It seems like a and b channels is readding it but I am not getting full 0-360 rotor position using encoder . I checked encoder ground to pin 22 and 23 of MCU, pin22 is reading 1.3v while pin 23 reading 4.9v is it normal? Any guess what's causing this?


Best regards.

so are you reading A and B encoder or Sin Cos return encoder. there is a difference in pulse return. do you have a scope?
Check the.schematic for.resolver i showed you...there is resistor.that should be removed i think...

Normally there is but on the Tesla Board there isn't: @iccie Please stop complaining on YT that you didn't get replies here. You got plenty

arber333 wrote: ↑Tue Apr 07, 2026 5:22 am so are you reading A and B encoder or Sin Cos return encoder. there is a difference in pulse return. do you have a scope?
Check the.schematic for.resolver i showed you...there is resistor.that should be removed i think...

I have cheap mini handheld scope. But I am not sure how to. Use it,.

I have checked the resolver schematics, and tried to build on, but unfortunately some parts are not available locally, and may take some time to arrive.

Yes, I was able to get a signal from AB pin using mlx90380,
Originally I cannot get any reading on B pin ( I assume something messed up or grounded from voltage regulator) so I decided to direcly hook the sin and cosine output to c22 and c18 which leading to stm32 pins, directly bypassing the resistor from circuit, as expected it have bad judder without this resistors.

What I am planning to do is to install resistor directly to my encoder.

But problem still prevails, or atleast I am thinking it is a problem.

1.) The encoder gives twice (2x) 0-360degress on one single rotor turn.
2.) If I rotate the rotor in one direction I only get 0-360 climb but it doent do 0-360?

Are those two from above are normal? What should I do. ?

Best regards,

Iccie

Ah your chip could be set to different pole setting. Leaf motor is 8-pole, 3-phase, 48-slot machine. Could be that you have a 4 pole count set in chip and ...
Or you have FOC set incorrectly?

Read the datasheet if there is a possibility to set pole count on resolver chip.

arber333 wrote: ↑Tue Apr 07, 2026 8:49 am Ah your chip could be set to different pole setting. Leaf motor is 8-pole, 3-phase, 48-slot machine. Could be that you have a 4 pole count set in chip and ...
Or you have FOC set incorrectly?

Read the datasheet if there is a possibility to set pole count on resolver chip.

Hi,

Thanks for the reply,

I just turn it manually by hand , I have not yet made the motor run as i am still making the encoder mount.

I am actually using mlx90380 ( magnet encoder), i think its an evaluation board model. I have checked the datasheet but i cannot find any details on setting the pole count, i guess it doesnt have that .

I was wondering if it is normal or not, i try to mount different magnet to see it it change.


I might run resolver soon if this dont work and if the parts will arrive.

iccie2 wrote: ↑Tue Apr 07, 2026 9:33 am Hi,

Thanks for the reply,

I just turn it manually by hand , I have not yet made the motor run as i am still making the encoder mount.

I am actually using mlx90380 ( magnet encoder), i think its an evaluation board model. I have checked the datasheet but i cannot find any details on setting the pole count, i guess it doesnt have that .

I was wondering if it is normal or not, i try to mount different magnet to see it it change.


I might run resolver soon if this dont work and if the parts will arrive.

No, go in OI interface and see if you have correct pole pair set for your motor equivalent

arber333 wrote: ↑Tue Apr 07, 2026 10:05 am No, go in OI interface and see if you have correct pole pair set for your motor equivalent

Hi,

I checked the OI and doubled check if i put the correcr motor pole pairs , and its a correct motor. I also tried changing it but still i am getting 0-360 even if i just physically turn the motor 180 degrees .

But i was able to fix the unwanted noise and jitter. It is now playing below 1 degress with random spikes and almost not worth to fix.

Could it be software? Or firmware? Or code?

arber333 wrote: ↑Tue Apr 07, 2026 10:05 am No, go in OI interface and see if you have correct pole pair set for your motor equivalent

Hi ,

I was able to make the motor ( small 20v bldc motor from mini one hand chainsaw) run for around a second with some revolution and my low power supply ( 330w laptop charger) suddenlly, I did put resistance load to avoid tripping. The tried check all pinswap and it seems ps13 is the better but my dummy load still recieves some current. Enough to turn the grinder to spin at 20v and I am not sure what current ( but assume less than 15amps ) it is receiving. Then I played with different pole pairs and 2 pole pairs spin better along all pole pairs which happen to correlate with actual motor pole pairs.

Is it normal that it is pulling so much current ( assumed less than 15amps) from main battery side? Atleast to what I am driving?

Update : I am now playing with foc on manual id and IQ. I notices as the syncof change the current passing to dummy load is also getting lesser but the motor is spinning faster but lesser torque. I'm still trying to understand the FOC guide. I saw the video but I am trying to understand and digging it since I cannot really get to the point on how it works.

I'll try to put throttle and see how it goes.

Hi,

I was temporarily to fix the over current issue after playing with manual LD and lq then pinswap. The problem went back again after I fix the magnet and encoder. Not sure what happen, but the inverter is reading the correct encoder position. I was wondering if dead time has to do with it?

Some issues to fix;
1.) Minor erratic spike on encoder signal, planned to play with different resistoe value and placement.
2.) Current is detected as -8amps , I'm am not sure why.

Some questions neee to be answer,
1.) Is 360 degrees sensor reading while 180 degrees physical on rotor ok (2:1 ratio on sensor and actual) , if the motor is spinning? Could it affect efficient of the drive train? Can I see the code regarding sensor input?
2.) The power supply is not tripping anymore and the motor spins faster than before, but my current reading on OI is -8amp? What causes this?
3.) Do I still need to buy lcr meter to properly set the ld-lq?

Hope I can get some help.

arber333 wrote: ↑Tue Apr 07, 2026 10:05 am No, go in OI interface and see if you have correct pole pair set for your motor equivalent

Hi,

I was able to fix all issue above. Seems like the motor is happily running on above condition. I might install it soon on my electric motorcycle if time permits,

I was also wondering which a better inverter the model 3 RDU or this inverter?

I might up my plans and run 1700amps inverter soon. If I gain more knowledge. Over all I was kinda satisfied that with your help I was able to learn alot of things on single board.

Best regards,

Iccie.