Paul and Sabrina's Cheap 3 Phase Inverter (AC Controller) with Field Oriented Control

[MPaulHolmes]

thingstodo, did you get the updated software I sent a couple weeks ago (april 11)? It's the updated code, but with the through-hole board incorporated into it.

Yes I would very much like to do nissan leaf board. I have a leaf board, and have been figuring out the pinouts, and am almost done. I could get a board working pretty quickly. I just need to do a prototype. I'll do that this month. Well, next month. it's the 30th haha.

[thingstodo]

Originally Posted by MPaulHolmes

thingstodo, did you get the updated software I sent a couple weeks ago (april 11)? It's the updated code, but with the through-hole board incorporated into it.

I read the email ... but didn't download or install the firmware or source code

I will do that! Likely tomorrow.

Still a bit confused by why the controller did not use the old settings to try starting the Polaris motor though. All I should need to do is power up and hit the throttle ... right? The encoder pulses rise to 3.9 VDC and fall to 0.2VDC .. that should be good for a 5V input?

I didn't get the logs loaded last night. Will get that part done tonight - hopefully

[MPaulHolmes]

Could you do:

stream-poscnt 1
data-stream-period 100

then slowly turn the motor shaft by hand to see if you get smoothly changing position. That will show that the encoder is working properly.

[thingstodo]

Originally Posted by MPaulHolmes

Could you do:

stream-poscnt 1
data-stream-period 100

then slowly turn the motor shaft by hand to see if you get smoothly changing position. That will show that the encoder is working properly.

I can do that! I could not remember which variable to turn on to check on the encoder. The mechanical rpm seems to be a random number generator when the output of the controller is not on

That will likely happen tonight as well

[02BMW02]

Originally Posted by MPaulHolmes

Yes I would very much like to do nissan leaf board. I have a leaf board, and have been figuring out the pinouts, and am almost done. I could get a board working pretty quickly. I just need to do a prototype. I'll do that this month. Well, next month. it's the 30th haha.

That would be great Paul, I will defiantly buy one if/when you produce them. here in the UK the breakers/dismantlers only sell leaf motor complete

If in the future you can hack the charging and the DC/DC convertor you will be onto a real winner

[thingstodo]

Originally Posted by thingstodo

I can do that! I could not remember which variable to turn on to check on the encoder. The mechanical rpm seems to be a random number generator when the output of the controller is not on

That will likely happen tonight as well

Results for April 28, 29, 30

I attached the raw logs. I typed a few notes while hyperterm was capturing. see if they make sense. I'll go through them a bit more tomorrow night, but I found out I have training tomorrow night so I only have about an hour.

Didn't get the firmware loaded in.

stream-poscnt 1 is a not a valid command, so I guess I need to load the firmware to get that command.

At the end of the April 30 log I rotated the wheels by hand. First back and forth several times. Lastly in the same direction for several seconds and it looks like I seem to have gotten a consistent direction and speed.

I think the encoder signal is coming through OK.

I still can't figure out why the controller is not putting out current when I move the throttle signal. Weird.

The base config from Andrew's siemens AC motor is in the logs in a couple of places. It should at least TRY to rotate! SIGH!

[wootwootman]

Hmmm I might be having a problem along the same lines as thingstodo

Originally Posted by MPaulHolmes

Could you do:

stream-poscnt 1
data-stream-period 100

then slowly turn the motor shaft by hand to see if you get smoothly changing position. That will show that the encoder is working properly.

I tried this but still only saw 0s. I was able to get my hands on a scope, so I hooked it up to the resolver and then encoder outputs, captures of which I've attached. From what I can tell, the encoder output seems to be fine (90deg shift between QEA-QEB, 0-5V, Index pulse every 1/4 turn, etc).



However, this was done without current sensors or throttle connected, so there were faults triggered. Are these faults preventing the "poscnt" variable from being updated/streamed properly? Or can there be another issue at play here?

[MPaulHolmes]

If the motor type is set to 1 (induction motor), streaming poscnt won't do any good. if the motor type is like 3 (leaf nissan leaf), then stream-poscnt should work. What motor do you have? I am sick today and can't think straight, and am still trying to help my job transition smoothly to being without me, so I won't be much use for a couple more days.

[wootwootman]

I verified that it's a 2013 Nissan Leaf motor (EM57), so it should be compatible. As far as the motor-type, I haven't looked at the config settings in a while so I'll take another look at them this evening after work.

Going over the schematic again, I don't think I populated C8,C9, or C10 (the LPF caps before the I/QEA/QEB inputs on the PIC). This would leave the signals basically unfiltered, but since I'm not running the motor or any voltage higher than board level (24VDC), I don't think this is the issue.

I'll be sure to populate them before I start powering the motor, just gotta choose the right values.

512ticks/rev * 10,000rpm (max EM57 rpm) = 5.12Mticks/min ~= 85kticks/sec

Since the encoder signals are square waves (which i believe need fast rise/fall times?), I'll raise the cutoff frequency like....four times? So 340kHz. Since there's already a 1k resistor, that leaves me with ~470pF for a cap value. Someone please correct me if I'm wrong

[MPaulHolmes]

I never use caps there but it wouldnt hurt if they were sufficiently small capacitance. The chip has a noise rejection feature built into the QEI module.