Prius Gen2 adapter board development thread

[RetroZero]

I assume : VCCIO = IO = Ov reference. (0.2v on my multimeter) , and VCC = -3.3V
STM Pin 60 = GND = OK
STM Pin 01 = VBAT = 0.2v - OK? - (VBAT = IO = 0v?)
STM Pin 12 = VSSA = -3.3v - OK
STM Pin 13 = VDDA = 0.2v - OK?

Would that seem correct so far?

[jap]

I think those values are all over the place, you could look at what pins should receive from STM32 datasheet:

https://www.st.com/en/microcontrollers- ... 03rb.html/

[imwoody36]

thank you for answering kiwifiat,
you are right the three coils in the resolver are 14 ohms, 14 ohms and 8 ohms.
I have photographed the toyota schematic with my oem harness laying on top.
yes the colors match, with black and white being the excitation.
the openinverter board is producing this almost sinusoidal signal unloaded. the distortion shows up when connected to the 8 ohm resolver

I am baffled why the vector signal entering the openinverter board will not show a valid angle when plotted with the webinterface
and I would argue this is a problem that could be fixed in firmware.

[Jack Bauer]

Ok, lesson time considering I have just been doing this with the Model 3 drive unit. In order to work the FOC software needs the exact angular position of the rotor. It does this by feeding one end of the sine and cosine from the resolver to two adc inputs. The other ends are held and about 1.65v. Why 1.65v? well its 3.3v/2 Now, in order to get the most accuracy we want the sine and cos amplitudes to be close to 0v at minimum and close to 3.3v at maximum. We can tolerate less but more is very bad as we loose angle information. ok so far?

Now we end up with two numbers. Feed those into an arctan function and presto an angle in degrees. That's how it works.

So lets scope the sin and cos pins on the stm32 and what should we see?

If it clips, thats bad.

If it is very low amplitude, thats bad.

[jap]

And those pretty waves we see only with a spinning rotor, right? With rotor/resolver standing you have only 2 steady values for sin/cos.

[ZooKeeper]

Unless I am mistaken, you will not see pretty waves in the VCU GUI, but something more like a sawtooth. The pretty ones are on only on scopes.

[RetroZero]

I think those values are all over the place, you could look at what pins should receive from STM32 datasheet:

https://www.st.com/en/microcontrollers- ... 03rb.html/

Thanks jap,
So VBAT on Pin 01 of STM is 0.2v (16v supply and GRND cable of multimeter on Negative battery pole) - Should be between 1.8v and 3.6v
Went back to beginning - Power supply: 12v (multimeter GRND to Negative Pole)
D1 = +16v
Power supply Pin 4 +4.5V Pin 1 +8.14V
Power supply Pin 5 +15.9V Pin 2 -2.97V
Power supply Pin 6 +2.53V Pin 3 -2.15V

Voltage Regulator Pin 1 + 2.5V
Voltage Regulator Pin 2 +0.34V
Voltage Regulator Pin 3 -2.93V

I deduce from these values that Voltage Regulator Pin 2 (VCCIO) is the supply to STM Pin 1 (VBAT)
If so, the actual voltage of +0.34v is incorrect - should be between 1.8V and 3.6v

However, I do not know if the problem is before voltage regulator as the power supply values don't make sense.

[RetroZero]

More internet digging later....ground is not where GRND of multimeter should be used for reference. Capacitors!!!
So start again...
Voltage Regulator
Pin 1 = Ground reference
Pin 2 = VCCIO 3.3V
Pin 3 = VCC 5.0V
Makes more sense. And I have 3.3V at VBAT Pin 01 on STM. I have 3.3V when testing across Pin 01 and Pin 60 of STM.
So STM has power, I think..
I can confirm 3.3V at Alive LED and 1.95V between STM and LED

[RetroZero]

VDD 1,2,3,4 all have 3.3V at STM32
Between VBAT and VSSA I have 3.3V
Between VBAT and VDDA I have 0.0V
It seems like all my supplies and grounds are OK....that means I fried my nice new STM32?

[imwoody36]

thanks " elon" while trying to figure out what pins are the sin and cos in to the STM32, I remembered this was one of the boards Johu had to rotate the chip.
so I reflowed it with hot air. and my symptoms are different... and maybe a little better. some circular jerky motor movement.



so, now I will go back and retest the polarity of the sin and cosine connections. you dont happen to know what pins I should scope do you?
I can likely go find a point to probe nearby.

[ZooKeeper]

Woody, diagram Johu posted earlier in this thread on 29-April on Pg. 29 may help.

Schematic done, does anyone feel like routing this out in KiCAD?
I've included a footprint for the ModICE connector but I believe there is a better one in KiCAD.

EDIT: I have started layout, therefor removed the Eagle file here.

[imwoody36]

update!
I can now plot a sawtooth, just like damiens in his youtube. the only change was reflowing the STM.
now when I attempt motor spin , I get a few moments before the contacters cut out and they wont restart until I power cycle.
no errors but Ive been up for 10 hours and maybe I should take damiens advice and ponder a bit.

is there a way to restart without power cycling?

Im assuming I need to play with W.C.U phases or perhaps sin cos polarity
does anyone want to see a tube video of the reflow under a microscope?

[RetroZero]

I would like to see that as I am going to attempt an STM replacement....

[kiwifiat]

update!
I can now plot a sawtooth, just like damiens in his youtube. the only change was reflowing the STM.
now when I attempt motor spin , I get a few moments before the contacters cut out and they wont restart until I power cycle.
no errors but Ive been up for 10 hours and maybe I should take damiens advice and ponder a bit.

is there a way to restart without power cycling?

Im assuming I need to play with W.C.U phases or perhaps sin cos polarity
does anyone want to see a tube video of the reflow under a microscope?

That is good news, did reflowing the stm32 fix the clipping present on the resolver drive signal?

[Jack Bauer]

I would like to see that as I am going to attempt an STM replacement....

NOOOOOOOOOOOO. Have you tried to access the mcu via stlink?

[Jack Bauer]

update!
I can now plot a sawtooth, just like damiens in his youtube. the only change was reflowing the STM.
now when I attempt motor spin , I get a few moments before the contacters cut out and they wont restart until I power cycle.
no errors but Ive been up for 10 hours and maybe I should take damiens advice and ponder a bit.

is there a way to restart without power cycling?

Im assuming I need to play with W.C.U phases or perhaps sin cos polarity
does anyone want to see a tube video of the reflow under a microscope?

Ok now we're talking. Lets ask ourselves some questions first before doing anything. why would the the contactors drop?
.....because we hit overcurrent limit........
.........why would we do that given that FOC firmware uses current control......must mean we are either not receiving the correct current readings into the stm32, or we are reading them incorrectly (not calibrated) OR we are reading the wrong current for that phase leg. Remember my fun with the Gen3 inverter where Johannes expects to be able to read the current on phase A and B but the Gen 3 inverter has current sensors on phases A and C!

Even if you have resolver incorrect (motor won't spin) it should still be possible to control the current into the stator via throttle. Once you get that happening THEN we mess with the resolver.

[RetroZero]

I would like to see that as I am going to attempt an STM replacement....

NOOOOOOOOOOOO. Have you tried to access the mcu via stlink?

Sorry, I don't have one. Tried through wifi to restore default parameters. Looking at what other alternatives I might be able to come up with trolling the forum.
I saw somewhere a home made adapter connected to USB port....but before doing anything, I would like to understand the whole software reload and reset process.....trolling again. I really don't want to touch the STM if it can be avoided..

[RetroZero]

For info, I did create the problem by supplying 12V to Pin16 (precharge) with the board functioning.
I heard a little pop and then Alive LED went dead.
I tried testing the Mosfets in Diode mode with multimeter, but no readings in iether of them.From what I understand, they are usually tested out of circuit.

[Kevin Sharpe]

For info, I did create the problem by supplying 12V to Pin16 (precharge) with the board functioning.
I heard a little pop and then Alive LED went dead.

I'm not following this thread closely but when you say "supplying 12V to Pin16 (precharge)" do you mean you put 12V on an output pin?

[RetroZero]

Yes was trying to tidy up wiring and test precharge circuit when I 'dropped' output wire on 12 v supply. Poof. Very upsetting and frustrating . Easy to say that there's a lesson to be learnt, and taking a step back can sometimes be the best way forward, but it's a huge setback I created. Pissed off with what I did.

[Kevin Sharpe]

Yes

Welcome to the world of RND

Assuming the precharge driver chip is not current limited then it (or the tracks supplying power) probably failed. In the first instance check signals and grounds around the driver chip to see whether it's still functioning. Also try and work out where all the current from the 12V supply went. It's possible the driver chip failed and the CPU pin had 12V applied which would be bad news. This is speculation... I haven't looked at the schematic or driver datasheet.

At this point it might just be easier to buy another board rather than spend time trying to debug the faulty board.

[Jack Bauer]

You will need an st link to program the boot loader if you replace the stm32.

[jap]

Before changing STM, I would check for other failed components. You practically created a short with +12v directly to ground through board.

[RetroZero]

I would buy another board in 2 seconds, problem is Johu is on well deserved holiday. As this is a professional project, time is of the essence at the moment. Mid August really complicates things for me.
As for st link, yes I would need to get one to program the stm...if I go that route.
All my VSS have 0V and all my VDD have 3.3v . VBAT also 3.3v
I notice Output Precharge and DCSW are connected in series to ground. Could it be possible that another element tied to common ground caused the short?

[RetroZero]

Before changing STM, I would check for other failed components. You practically created a short with +12v directly to ground through board.

Started tracing and found different resistance values for the 2 X 100uF ECAP. C13 and C23
C13 In one direction 3.3K Ohms for about a second and then Open Circuit . 3K Ohms in other direction
C23 0.6k Ohms in one direction . 225 Ohms in the other.....
Could I have cooked C13 capacitor with +12v?
Anyone able to give me rough values? Since they are identicle capacitors, they should give me identicle values?