openinverter forum

johu wrote: ↑Tue Aug 04, 2020 9:33 am In general it is hard to control motor speed of an unloaded motor with the throttle. Throttle input controls torque and of course very little torque will spin an unloaded motor up very fast. You can try using the idlespeed parameter for testing as well.

Thank you. Makes sense.

Something else; I've hooked up my board, and got the Sine firmware running without any error, motor "turns" but is shaking/jumping and dancing. after measuring, it seems that MWU is not active or providing a signal, I checked the darlington array, that is not receiving any signal as well from the STM32. Already tried resoldering the connections on the darlington array, but no change.
Is there something I can do to get this working again, or could this be an manufacturing error?

p.s. (my capacitor C13 also felt off the board shortly after receiving, soldered back on it and got it working)

If you are using an outlander motor then you need the foc firmware not the sine.

LRBen wrote: ↑Wed Aug 05, 2020 10:08 am If you are using an outlander motor then you need the foc firmware not the sine.

Thank you, correct.
I first try to get my induction motor on the sine software running, once successful I will hook up my outlander motor. want to avoid to destroy my outlander motor.

Bram wrote: ↑Wed Aug 05, 2020 10:05 am Something else; I've hooked up my board, and got the Sine firmware running without any error, motor "turns" but is shaking/jumping and dancing. after measuring, it seems that MWU is not active or providing a signal, I checked the darlington array, that is not receiving any signal as well from the STM32. Already tried resoldering the connections on the darlington array, but no change.
Is there something I can do to get this working again, or could this be an manufacturing error?

You should certainly see an INPUT to the array when fstat > 0 and amp > 0 e.g. by setting fslipspnt=1 and ampnom=1. Could it be that pwmpol is set to ActiveLow? If you can't see pwmpol type "flag pwmpol ~hidden" into the custom command field and refresh.
The darlington output of course only creates a visible signal when connected to the power stage AND the power stage is enabled via MSDN. So it should be resting at around 15V and pulses pull down to GND.

Bram wrote: ↑Wed Aug 05, 2020 10:05 am p.s. (my capacitor C13 also felt off the board shortly after receiving, soldered back on it and got it working)

Oh, that's embarrassing I've seen it happen on other boards, too. Maybe I should aim for an elcap-less design.

johu wrote: ↑Wed Aug 05, 2020 11:31 am You should certainly see an INPUT to the array when fstat > 0 and amp > 0 e.g. by setting fslipspnt=1 and ampnom=1. Could it be that pwmpol is set to ActiveLow? If you can't see pwmpol type "flag pwmpol ~hidden" into the custom command field and refresh.
The Darlington output of course only creates a visible signal when connected to the power stage AND the power stage is enabled via MSDN. So it should be resting at around 15V and pulses pull down to GND.

Thank you for your feedback Johu.
I've got it turning (on 2 phases), and then measure the signals from the Darlington array (in & output) 2 are active going down, while MWU stays flat on the input from the darlington compared to the other 2. Correct, Darlington output signal for MWU is always 15V, where the other 2 switching to GND.
PWMPOL is on ACTHIGH, MSDN on 12v, while running Fstat 5Hz and amp 37813(AMPnom 100%).


any other thoughts, or scenario's I can check?

It looks like you made no mistake here. Write a short email if you want a replacement board.
Meanwhile you can check whether the signals are present on the micro. It is the 3 parallel traces where it says "IC1"

EDIT: since the cap fell off it almost sounds like a mistake during reflow, maybe some pins just haven't made contact.

greetings members!, I am back after driving to Florida and back to pick up a spare inverter board that I own, and related prius work.
( including installing a leaf motor in a boat on another thread, and another tube vid)

in the meantime parts arrived for the original inverter board
uln2003 now replaced and board is turning a gen2 MG1 motor, and FYI the spare board is also giving the same results.

I dont see any PWM signal coming from the driver,so i assume my testing method is wrong and likely in the parameters

tuning commences tomorrow, here are the parameters I have so far.

(refresher., what is the goal for the currents with regard to phase and amplitude ?

Before I fried my adapter board, my Mg1 motor was running smooth as silk in manual mode from inverter. I didn't test pwm outputs, but definetly wasn't clonking like yours.

imwoody36 wrote: ↑Thu Aug 06, 2020 12:01 am I dont see any PWM signal coming from the driver,so i assume my testing method is wrong and likely in the parameters

What are the three PWM signals doing on the input to the ULN2003 driver chip?

Do you think this is the same issue?

Bram wrote: ↑Wed Aug 05, 2020 1:37 pm

johu wrote: ↑Wed Aug 05, 2020 11:31 am You should certainly see an INPUT to the array when fstat > 0 and amp > 0 e.g. by setting fslipspnt=1 and ampnom=1. Could it be that pwmpol is set to ActiveLow? If you can't see pwmpol type "flag pwmpol ~hidden" into the custom command field and refresh.
The Darlington output of course only creates a visible signal when connected to the power stage AND the power stage is enabled via MSDN. So it should be resting at around 15V and pulses pull down to GND.

Thank you for your feedback Johu.
I've got it turning (on 2 phases), and then measure the signals from the Darlington array (in & output) 2 are active going down, while MWU stays flat on the input from the darlington compared to the other 2. Correct, Darlington output signal for MWU is always 15V, where the other 2 switching to GND.
PWMPOL is on ACTHIGH, MSDN on 12v, while running Fstat 5Hz and amp 37813(AMPnom 100%).

Probably a good time to remind people that when you're stuck in a loop and the community can't solve your problems, both Johannes and Damien are available for hire**;

https://openinverter.org/shop/index.php ... duct_id=61

https://www.evbmw.com/index.php/evbmw-w ... evbmw-serv

**Note that both guys are extremely busy and therefore I'd recommend booking this help early if your development is on a critical path. Note also that New Electric will have some one day online courses in the future that focus on specific technical topics such as the Gen 2 adapter board

Well guys, I think I made a mistake in the documentation. I realized when testing one of my own boards. So in fact inverter PWM is on pins 28, 29 and 30 (per my documentation sheet) or 8, 9, 10 when looking at the 12-pin connector only. On the input of the ULN chip this would be pins 5, 6, 7.
Please try this and report back.

Likewise CPWM signal is found on Pin 31/11 and pin 4 of the ULN

Attached is a corrected documentation sheet.

Hi Johu,

it works!!
Thanks, I've got it running (on sine) smooth without any errors. with 3 similar amp, nothing abnormal or weird. Thank you for checking Johu

Champagne ! Thanks Johu

Nice, great relief

I have 3 pwm signals now as well.
now tuning commences... will check current again.

Just watched some of your earlier videos. Did you actually spin the Prius motor that violently on just 2 phases?

apparently, but I wouldn't say it was violent. ( well after I went down to 24 volts instead of 220vdc)

so, do I mess with any phase swaps now?

I would turn off all phase swaps and try the foc tuning procedure.

Yes I believe in the Prius inverter everything is lined up, but I haven't tested FOC software myself yet. Only resolver could be subject to swapping/checking if you get no movement

johu wrote: ↑Wed Jul 15, 2020 1:42 pm Just tested it with my Prius power stage and it works. If you install the 10k resistor you can measure down to about 0°C, otherwise it starts at 56°C. At startup it will check once for a voltage > 2.6V and if it sees that it will assume no 10k resistor is installed. It will of course fail to detect that if you start into a hot inverter, i.e. warmer than about 85°C.

Wanted to order 10K resistor for temp...Any specific wattage required for this resistor?

RetroZero wrote: ↑Fri Aug 07, 2020 2:58 pm

johu wrote: ↑Wed Jul 15, 2020 1:42 pm Just tested it with my Prius power stage and it works. If you install the 10k resistor you can measure down to about 0°C, otherwise it starts at 56°C. At startup it will check once for a voltage > 2.6V and if it sees that it will assume no 10k resistor is installed. It will of course fail to detect that if you start into a hot inverter, i.e. warmer than about 85°C.

Wanted to order 10K resistor for temp...Any specific wattage required for this resistor?

in my humble opinion any wattage will be okay down to 1/8th watt. this is a small signal.

look in an old piece of electronic scrap for one and harvest it.

Thanks imwoody26, I have 0,25w & 0,5w on its way... Preparing for new adapter board

Well I have intended, but mystery progress

- The resolver signal is back, I have absolutely zero clue how; then again I never figured out when or why I lost it.
- MGR "fights" me very slightly when I attempt to rotate it and measure angle with the inverter powered on
- All the STM power and grounds seem ok: 3v3 on 1, 13, 19, 32, 48, 64 and 0v0 (< 3mV) on 7, 12, 18, 31, 47, 63
- I have 4.4kHz on pin 54
- 28 = 80Hz
- Lastly, pins 57, 58 & 59 all have 9.5mV & no Hz

Does anybody knows why we have jumpers Sj1 and Sj2?
I'm struggling with getting a nice triangle out of my (outlander) resolver, and watched video's again from Damien, where he showed his prius transaxle with a nice triangle signal from his resolver. also tried it, with Sine and hand, but without success. also tried a few possible other combinations with resolver without result.
Then i had a look at the GEN3 encoder signal input and noticed that R1 and R2 are not active due to the open jumper, while its a hard line within the Gen 3 Schematic. Beside that the encoder scheme's are very alike.
my question: can I set Sj1 and SJ2 as another possible solution for my Resolver quest?

RetroZero wrote: ↑Fri Aug 07, 2020 8:30 pm Thanks imwoody26, I have 0,25w & 0,5w on its way... Preparing for new adapter board

Erm since I'm sending you the newest revision board the resistor is already fitted! Sorry about the confusion.

Bram wrote: ↑Sat Aug 08, 2020 7:42 amDoes anybody knows why we have jumpers Sj1 and Sj2?

SJ1 and SJ2 are for ENCODER use only! They activate pull-up to 5V. Do not close them for resolver. Set encmode to resolver.