Prius Gen 3 Inverter Logic Board Support Thread

[konstantin8818]

Interesting things I've found out about MGR (Sorry for the offtopic as this thread is not about MGR)
Max_motor_current for the MGR is ~110A according to this:

Also according to this and my pure knowledge about motor polepairs, if there are 8 coils, there is actually 4 pole pairs.
And in that thread it was a guess that resolver also has 4 pole pairs.
Will check it right now, with max manualid of 55A

[Jack Bauer]

So some guy measures the wire diameter, uses an online calculator designed for power distribution and bingo we have a max motor current. Oh and even better it fits in perfectly with the 650v "fact". Sweet. What could possibly be wrong.

Back in the real world it seems I need to provide another hint. Did I say the resolver settings were wrong? No I didn't. So if we have the right settings but the wrong behavour whats left? and no I won't answer this one.

[konstantin8818]

Sorry for the previous post.
I'm officially state that I'm now achieved confusion of the highest order: revolutions count changes whenever polepairs or respolepairs is changed.

What changes in terms of signals when we reverse motor direction?

Does direction depends on iq being positive or negative? Motor runs smooth both directions if I enter negative or positive manualiq, while manualid is 0.1
From now I'm not writing anything about this case. I'm reading now.

[Jack Bauer]

In order for the inverter to calculate the angle of the rotor it needs sine and cosine inputs. Each of these has two wires. Therefore there are 4 possible connections for the sine and cosine signals.

Johannes uses an arctan function of these two signals to calculate rotor angle.
https://www.rapidtables.com/math/trigon ... rctan.html

So, suppose you had one of those signals backwards. Guess what happens?

[Jack Bauer]

[konstantin8818]

So, suppose you had one of those signals backwards. Guess what happens?

Back in a days when I've graduated from uni, I've got like 4 points out of 10 for mathematics... Geometry however was 10 out of 10 at the middle of semester.... what I'm saying is I never was good at those sinus/cosinus stuff...
Anyway, thank you for the most obvious hint in history of hints, otherwise i would wondering about this for ages...

[Jack Bauer]

No worries:) I'm just trying to get people thinking. I was totally rubbish at maths also.

[Bigpie]

Are you able to post any hits to get me thinking?
I'm having issues with the buck/boost.

Setup:
I've got 83v connected to the right most side (DC Bus 2) and a volt meter on the left (DC Bus 1)

If I PWM pin 10 I get 83v on DC Bus 1, regardless of PWM value.
If I PWM pin 9 I get 0v on DC Bus 1, regardless of PWM value.

I've attached a scope on the correct pins of the connector on the Toyota board on the other side, so I'm assuming the 50 pin connector is not the issue.
BOOST_FAULT is held at 5v, I'm assuming it is pulled low for a fault.

I've got to be doing something stupid somewhere.

[bexander]

I just verified the MGR resolver polepair to 2. Set the motor to spin at ~1670rpm according to SW and then measured driveshaft flange speed with a optical tachometer to 244rpm. Result, 1670 / 244 = 6,84. Gear ratio is 6,86 so I call that a match!

[konstantin8818]

I just verified the MGR resolver polepair to 2. Set the motor to spin at ~1670rpm according to SW and then measured driveshaft flange speed with a optical tachometer to 244rpm. Result, 1670 / 244 = 6,84. Gear ratio is 6,86 so I call that a match!

Yup, already came to this conclusion as well "angle" parameter is calculated not only by "respolepairs", it also depends on "polepairs" as I've found out. I believe, Johu can tell exactly how it is calculated...
Anyway, for some reason, I had to swap sinus phases of resolver, for it to operate correctly. I really don't know why they were swapped in first place. Either toyota engineers can make it work correctly with software and they don't care what is the polarity, or someone already played with them before me.

[ZooKeeper]

Anyway, for some reason, I had to swap sinus phases of resolver, for it to operate correctly. I really don't know why they were swapped in first place. Either toyota engineers can make it work correctly with software and they don't care what is the polarity, or someone already played with them before me.

Apologies for being thick, but if I understand correctly, Resolver pin #5 is now MSN and #2 is MSNG? If so, that is very odd indeed.

[konstantin8818]

Apologies for being thick, but if I understand correctly, Resolver pin #5 is now MSN and #2 is MSNG? If so, that is very odd indeed.

Before swap sine, I've swapped cosine, it didn't helped. So I've switched it back. Red wire was connected to 5 and now it is connected to 2.

[ZooKeeper]

Apologies for being thick, but if I understand correctly, Resolver pin #5 is now MSN and #2 is MSNG? If so, that is very odd indeed.

Before swap sine, I've swapped cosine, it didn't helped. So I've switched it back. Red wire was connected to 5 and now it is connected to 2.

Whelp, tried MSNG <> MSN and MSN <> MCS and both were a total fail. What I learned was that in both cases, there are 65576 wrong syncoff settings and when MSN = MSN & MCS = MCS, at least it spins smoothly at low throttle input.

[konstantin8818]

Apologies for being thick, but if I understand correctly, Resolver pin #5 is now MSN and #2 is MSNG? If so, that is very odd indeed.

Before swap sine, I've swapped cosine, it didn't helped. So I've switched it back. Red wire was connected to 5 and now it is connected to 2.

65576

Is it a typo? 65536 is max gain 65536=0

[ZooKeeper]

Is it a typo? 65536 is max gain 65536=0

Yes, yes it is I ran from 0 to 65536 by 1k and got only growling/wiggling in either direction.

[Jack Bauer]

I'll get the instruction video made asap. Hopefully will clear up a lot of confusion. I've been a bit busy the past few days holding back floodwaters with 2 three phase sub pumps and a prius inverter so not entirely irrevalent:)

[konstantin8818]

I've been a bit busy the past few days holding back floodwaters with 2 three phase sub pumps and a prius inverter so not entirely irrevalent:)

And I was wondering where all our water gone... wells are dry... I'm digging water bore right now

[Bigpie]

I've blown up my power supply, this is my theory after looking at a diagram.

I've fed in to pin 9 and pin 10 complimentary PWM with a deadtime, verified this on oh-silly-scope using a Arduino UNO, all good. Signals as expected.

Code: Select all

int val = 80 ;
int dead = 20 ;
  
void setup() {
  Serial.begin(115200);//
  TCCR1B = TCCR1B & B11111000 | B00000010;    // set timer 1 divisor to  8 for PWM frequency of  3921.16 Hz
  TCCR1A = (TCCR1A & 0x0F) | 0xB0 ; // set pin 10 inverted

  
  analogWrite (9, val-dead) ;
  analogWrite (10, val+dead) ;
}

Run this on the board in the inverter and my power supply released all the magic. I noticed the diode on the high side on the diagram. My theory is I very briefly boosted the voltage above 80v and the diode lets it pass through to DC bus 2 and caused the boom?

With the upshot being that it cannot boost from DC Bus 2 to DC Bus 1 and therefore cannot be used as a charger on single phase for packs above 340VDC?

[konstantin8818]

Did you precharge main capacitor before connecting power supply?

[Bigpie]

I did, turns out I've blown an internal fuse on the power supply, My other thought was I boosted enough to enable the DC DC that then put a big load on the power supply.

Can it buck and boost is both directions?
I'm thinking maybe not because of the diode on the high side IGBT, but I'm far from an expert in these things, and trying to learn along the way.

[Jack Bauer]

It can only boost from the battery input to the hv bus and buck from the hv bus back to the battery input.

[Bigpie]

Thanks for clearing that up. I guess that means if you want to use single phase to charge the higher voltage pack you could use an external rectifier and boost from the battery input?

[celeron55]

That's one option. The other option is to use an external voltage doubling rectifier to feed the HV bus.

[johnspark]

Just a bit of info on the boost inductor: It is 258.5uH inductance at 10kHz, and 0.267 ohms resistance.

[johnspark]

just did a few calcs on the inductance and resistance. The resistance looks too high. Measured inductor wire, it is approx 1mm by 5mm or 5mm2. This equates to something more like 0.048 ohms.