Prius Gen 3 Inverter Logic Board Support Thread

[Bigpie]

Even though it looked like I'd done an OK job at soldering the pins for the DC DC converter I hadn't on the one to enable the DC DC converter. A bit of solder and now it's working.

[chrskly]

I noticed these exact same values today as well. Will go after it.

I'm seeing this as well on a dual-motor prius board running 4.97.R sine.

set ampnom 0
get ampnom -> 9.53

I've tried:
- Rolling the firmware back to 4.94.R sine
- Rolling back the loader to v3
- Wiping with stlink and re-installing everything

Any ideas?

[johu]

Oh I forgot. Yes I found this. At some point I started calling "SetTorque(0)" in stop mode and that results in given behaviour. You can set ampmin=0 for now, that will get least get rid of setting ampnom to a non-zero value. Will try to find a better fix in the next release.

[chrskly]

Oh I forgot. Yes I found this. At some point I started calling "SetTorque(0)" in stop mode and that results in given behaviour. You can set ampmin=0 for now, that will get least get rid of setting ampnom to a non-zero value. Will try to find a better fix in the next release.

Thanks @johu, can confirm that setting ampmin=0 allows setting ampnom=0. fslipspnt is still getting stuck on 0.53Hz though. Not a problem, but just thought I'd mention in case that's something that needs to be tracked down.

[mrmriza@msn.com]

Sorry very basic questions
I believe I can Prius gen 3 inverter to control leaf motor
Is there any complications ? Or is it easier to run leaf motor with leaf inverter?

What's the minimum battery volt required to run leaf mot+ Prius invt.?

I only need 50mph max

Thanks

[johu]

The Leaf inverter is mechanically a better fit, obviously. Voltage and current specs is all that matters and those match the Leaf motor for medium performance application quite well.

[bobby_come_lately]

I'm having some resolver issues. As soon as I put the inverter into forward, angle starts cycling from 0-360 - see plot. But the motor isn't spinning - it happens when it is not even connected. I've tested three sets of wiring and I even get the same thing when the inverter is only connected to 12V and GND.

I can't see any shorts on my board and I've checked for shorts between the exciter and sin/cos/encoders - all fine.

I've pretty much exhausted everything I can think of. Any suggestions would be very welcome.

[mjc506]

Sin/cos chip? If so, check the voltages you're getting on the EncA and EncB pins as you rotate the motor by hand - try to find min and max. Once you have these, you should be able to get the midpoint and amplitude - midpoint should be 1.65V and amplitude must be greater than 0.8V (more is better, so long as you don't exceed 0.0V - 3.3V)

If the midpoint is a long way off, I get the same symptoms. I actually made a pull request earlier with a work around.

[Isaac96]

I saw the same 0-360 angle movement when using an AB encoder. I don't think it's an issue, as my car drove fine.
Can you put it into manual mode and change fslipspnt, see if the angle changes faster or slower?

[johu]

Is that sine firmware?
Then what you see is slip. Set fslipspnt=0

[mjc506]

Ahh of course, electrical angle...

[bobby_come_lately]

Is that sine firmware?
Then what you see is slip. Set fslipspnt=0

LIGHTBULB MOMENT!

Of course, I can't set flslipspnt as zero because of the weird error of it defaulting to 0.53hz and not able to be changed down from that.

Thank you!

[midway]

I was looking for -5 volts for two hours, I got a board with replacing SGM3204 with LM2776DBVT. Please tell me the value of the resistor R113.

[Jack Bauer]

R113=10k 0805

[wacomg]

Hello folks !

I have bought 3 boards gen3 v1d , two of them are kits and one is partially build .
I salvaged all missing parts for partially build board from original gen3 invertor board , and now it works perfectly with sine and FOC firmwares and Toyota MGR

The question is :
Have we lost our external CAN bus (because on the external Ampseal connector we have only charger CAN bus ) ?

Can we be able to control all MG1, MG2 and Charger via CAN bus

[stibium80]

I am working to get an industrial ACIM running using a throttle pedal. I am able to successfully spin run my motor in manual mode, so now I'm on to getting my setup to work using an E46 hall effect throttle pedal. After spending time fiddling about with my inverter and searching this forum, I was able to get control the motor using the throttle pedal. Since I had this modicum of success, I wanted to share my progress with the forum in case anyone might find it useful.

Areas where I've made progress:

• I was getting an error message on the web interface of DERATE-CURRENTLIMIT whenever I started the inverter using my momentary "start" switch. I resolved this by updating my firmware from 4.87.R-sine to 5.00.R-sine.
• I am getting a "din_ocur" error, but this is an artifact of the prius board design and doesn't affect functionality (see viewtopic.php?p=19882#p19882
• I was able to successfully update the firmware over WiFi using the web interface. It's super easy to update. Compile the .bin file, put it on your iPad (or whatever) and using the web interface, select your .bin file and update the firmware. This was a matter of seconds for me. Again, super easy.
• Adjusting the following parameters was critical for me transitioning from manual mode to using throttle pedal:
  • udcmin (this had to be adjusted to get a response from potnom when using the throttle pedal; if you are getting a response from pot, but not potnom, this could be a culprit)
  • udcsw (this has to be adjusted for the HV contactor to engage)
  • I was using a 60V bench power supply as my HV source and I needed to adjust the parameters above in order to get the inverter to respond to the throttle pedal.
  • Set encmode = 0(single) to use channel A of the encoder. This worked for me with my 200 pulse A/B incremental optical encoder with numimp = 200.

Areas where I still lack understanding:
I'm still not clear on how to get the A/B mode of the inverter software to work properly with my A/B encoder. I'm using an optical incremental encoder with an A/B (90 deg. offset) and an index (Z). I do not anticipate using the index input, but wanted to have it just in case I might use it in future. I started off with "encmode" = 1(A/B), but I couldn't get the motor to work properly in this mode. My encoder has 200 ticks on it and so I used a value of 200. I decided to switch to manual model and set "ampnom" = 0 and "fslipspnt" = 0 and spun the motor by hand while plotting angle. Based on what I saw in the plot, the angle parameter went through two rotations for every one rotation of the motor shaft. I then updated "numimp" = 400, and restarted the inverter and while the motor responded to throttle input, the response was very weak. Also, the motor did not respond to any of the tuning parameters (like boost, etc.) with "encmode" = 1.

In a last ditch effort, I switched "encmode" = 0(single), set "numimp" = 200, and restarted the inverter. With these parameter values the motor finally responded as expected to the throttle pedal. In this mode, other parameters began to responding to changes in input, so I considered this a success.

However, I wanted to understand why the motor didn't run properly when "encmode" = 1(A/B). In order to understand what was happening, I plotted "angle" and "turns" with "encmode" = 0(single), set "numimp" = 200, and restarted the inverter. I switched to manual model and set "ampnom" = 0 and "fslipspnt" = 0 and spun the motor by hand while plotting "angle" and "turns". I observed that the angle went from 0 to 360 (as expected) and for every turn of the motor shaft, the "turns" variable incremented by 1. This was good and made sense.

I switched back to "encmode" = 1(A/B) from "encmode" = 0(single) and plotted "angle" and "turns". Again, I switched to manual model and set "ampnom" = 0 and "fslipspnt" = 0 and spun the motor by hand while plotting "angle" and "turns". However, instead of getting "angle" = 360 deg. corresponding to 1 "turns", I get something like an "angle" = 720 deg corresponding to 1 "turns". As I stated earlier, there's something I probably don't understand or a parameter I've set incorrectly, so I wanted to ask if anyone else has run into this and what they did to resolve it so they could use an A/B encoder.

Has any one else experienced this? I think there must be something I'm missing here. Anyone have any thoughts on how I can correct this and use the A/B mode for "encmode"?

I've made two tedious videos of my efforts and posted below:
1: Update firmware via WiFi, but motor doesn't respond to throttle because I didn't have "udcmin" parameter set properly.



2: Motor responds to throttle with "encmode" = 0 (single), but doesn't respond well with "encmode" = 1 (A/B).


My current setup:
power: NICE-POWER 60V; 5A bench power supply
throttle: E46 hall effect pedal (potmode = 0)
firmware version: updated to 5.00.R-sine (updated from 4.87.R-sine to fix DERATE-CURRENTLIMIT trip)
prius inverter number: G9200-47210
IGBT mode: parallel (I soldered the jumpers on my logic board so the IGBTs run in parallel)
prius DC-DC converter: updated resistor network using 210K resistors to enable converter to run at higher voltage levels
encoder: U.S. Digital E6-200-625-IE-S-D-D-3 (incremental); link: https://www.usdigital.com/products/enco ... tal/kit/E6

[arber333]

Hi i think enc channels A B on your board are linked together on main board so that you can only use channel A on input. Its hardware
Check schematic if you need to unsolder a jumper or resistor maybe.

I remember experimenting with AB encoder and ACIM motor. Generally motor would run on only A channel just fine. Difference shows only in little things such as reversing downhill in gear or a bit smoother running at really slow speed.

[Jack Bauer]

No it uses seperate lines for A and B there is no linking. Most likely you need to swap A and B.

[stibium80]

No it uses seperate lines for A and B there is no linking. Most likely you need to swap A and B.

I double checked my wiring for encoder channels A and B and everything was correct. As I was troubleshooting, I noticed that when I plotted the angle parameter and rotated the motor shaft by hand (in the anticlockwise direction), the angle parameter values increased for encmode = single and decreased when I set encmode = A/B. So, I was pretty well convinced I needed to swap the encoder channels.

When I swapped the A and B encoder channels, the motor spun up beautifully!

Note that when I switched back to encmode = single after swapping the channels, the motor would not spin.

Thank you Jack Bauer!

[180jacob]

The question is :
Have we lost our external CAN bus (because on the external Ampseal connector we have only charger CAN bus ) ?

Can we be able to control all MG1, MG2 and Charger via CAN bus

Correct, at the moment there is no external CAN access to the STM32s for MG1 and MG2on the v1d boards. I plan on experimenting with adding one CAN transceiver to the transceiver-less connection between the STMs but am in the middle of changing employment so all progress has stopped for the moment.

[Dylan Witt]

My electric mazda is driving with the gen three inverter and control board. Thank you all

[lumax]

Does the board is a 2 layers or a 4 layers ? JLC pcb recognize it as a 4 layers?

[UStas]

My electric mazda is driving with the gen three inverter and control board. Thank you all

Cool!

What is there motor?
Is there parallel connection of MG1 & MG2 power stages?
What maximum current did you get?

[m.art.y]

Yes I would be interested in that too. By the way has anybody run MG1 and MG2 in parallel for more power on a single motor on V1c board? How do we wire the motor and which jumpers we need to solder up - just the 6 by the 50 way connector or also the one next to R81? Thanks

[midway]

in parallel with six jumpers SJ1-6