FOC manual mode, large iq spikes

[Bigpie]

I've put the inverter in manual mode with the physical gearbox in neutral, plotted id and iq, set both to 5 amps. iq has a large spike to 40 amps. Is this a problem? I've got a lack of umph when it comes to steep hills that I'm trying to tweak away if possible with this motor.

[Bigpie]

So further testing in manual mode, 0 manual id and manual iq

il1 and il2 seem very noisy. It's a gen 3 prius inverter with a board from Damien btw.

[johu]

Crap in, crap out. Do you get the same noise with sensors unplugged? Gotta be some hardware issues. Scoping ahead. Both power supplies (26V and -5V) and the actual output of the sensors.

[Bigpie]

This is the board on the desk with 12v power supply and the sensors plugged in.
MG2 IL2

MG2 IL1

26v

-5v

3.3v

5v

I'm not an expert, but nothing jumps out. 26v seems a little noisy but I've no idea.

[johu]

Can you vertically zoom il1/2? There is some noise but the 4V scale makes it hard to judge. 1 mV jitter would result in about 0.2A jitter I think. Or in other words your 20A jitter would require +-100mV on the scope.

[Bigpie]

Sensors plugged in
IL1

IL2

Sensors unplugged

[johu]

Ok with sensors plugged in it looks good, just about 15 mV jitter.
If signals are ok it could we the reference voltage. Maybe if you take a longer term (100 ms) shot of 3V3. Also vertically zoomed.

[bexander]

Sensors plugged in
IL1
Screenshot 2021-08-22 at 3.10.22 pm.png
IL2
Screenshot 2021-08-22 at 3.11.55 pm 1.png

Sensors unplugged
Screenshot 2021-08-22 at 3.15.35 pm.pngScreenshot 2021-08-22 at 3.15.47 pm.png

On the plots of IL1 and IL2.
These plots don't show the whole truth due to how they are produced. When zooming in on a large scale plot you will get very low resolution.
I suggest setting the scope to measure AC instead and then set channel sensitivity down to 100mV or so. Also probes at x1 if they aren't already.
Then set a "auto" trigger at say 20mV.
Expect to measure close to 0V with some large noise spikes.

[johu]

oh, good catch

[Bigpie]

Thanks for taking a look.
Here's IL1

IL2

3.3v

3.3 measured as AC

[johu]

Ooooh, 3V3 is totally out of shape. I assume any other analog inputs look similar. Does it improve if you remove the Wifi? There was a problem with 3V3 reg on some earlier Prius boards, not sure which one you have. They needed swapping in a stronger regulator and an LC filter for the Wifi.

[bexander]

Also the CAN transceiver might create disturbances on the 3V3. Looks like close to 10us pulse repetition. Might be CAN if nothing else is connected on the CAN bus. The transceiver get no response and tries to resend repeatedly.

When doing the 3V3 measurement, try using trigger auto to get a more steady state plot. You are using single trigger in posted plot. Decrease trigger level if needed to trigger in auto.

[bexander]

Looked in my old thread about similar disturbances and it might be a good idea to look at a larger time scale like 500us/div or 1ms/div.
My old thread:
viewtopic.php?p=18288#p18288

[Bigpie]

It does seem to improve when the wifi is unplugged. Wifi powered with an external 3.3v.

Looks like that's the issue, thanks

[johu]

Alright! That looks like the usual jitter now.