Prius Gen2 adapter board development thread

[johu]

Haha
Shouldn't you see your typed password in the auto-complete list? I didn't make it a "password" field...

[ZooKeeper]

It shows one digit at a time as you type it in, as is normal. The second I hit "ok", I knew I had fat-fingered something. Once I figure out how to get it on serial comms, I will change the PW again with AT commands and be back up and running!

This is 100% my fault, I should have taken more care in entering the new PW digits, but I was too cavalier about it and screwed myself.

[imwoody36]

So I couldn't complete my mission to test the battery charging feature. Anyone have a left over buck/boost module (without inductor)? Yes I could just buy another inverter but my space is limited so I'd rather avoid it.

EDIT: whats the temperature curve for the heatsink sensor, anyone got that?
[/quote]

Yes Johu..I have many gen2 prius inverters laying around....want a care package? I am willing to ship for free just tell me what you want

[johu]

I have a Prius inverter as well, but I'm too lazy to find the temp curve myself. It's supposed to be in the service manual but I couldn't find that either. So if someone could just post a screenshot that would help.

EDIT: I checked some old mails and found it. Small problem: I forgot the voltage divider so there will only be valid values starting at like 40°C

EDIT again: looks like -30mV per °C. Or -37 digit/°C. So T=167°C - ADC / 37 . Maybe I'll just add a "linear" option for tmphs

More editing: So it seems like the lowest measurable inverter temperature without voltage divider would be 56°C. So maybe it would be a good idea to add a 10k resistor in parallel to C37 to form a 1:2 voltage divider. I will add a hard coded formula in the next firmware that is always used in Prius hardware, i.e. "snshs" becomes meaningless, just like for the Tesla boards.

[RetroZero]

Thanks for checking all that Johu.
So for testing purposes, keep inverter cooled below 75°C (real) as to not create temp error.
Whilst waiting on next firmware update, a 10k resistor in parallel with C37 will solve the problem.

Me personally, I'll wait until firmware, even if I need to put iceblocks inside the inverter. Adapter board looks too good to go soldering resistors in addition

[johu]

The firmware is not going to increase the input voltage of the ADC It will just get you a valid reading in the first place. Without the resistor you only see temperature changing above 56°C.

[johu]

Just tested it with my Prius power stage and it works. If you install the 10k resistor you can measure down to about 0°C, otherwise it starts at 56°C. At startup it will check once for a voltage > 2.6V and if it sees that it will assume no 10k resistor is installed. It will of course fail to detect that if you start into a hot inverter, i.e. warmer than about 85°C.

[RetroZero]

I wired up everything and got preharge and contact relays working.
I have 4 throttle faults (even with the pedal not connected)... Set potmin and potmin2 to values that are in plot values (when connected) but still the same faults.
Step by step

[ZooKeeper]

WHEW! That was not fun, but I got back in

As we can see, I have messed about some with the parameters trying to get spin. Screenshot is obviously not attempting any output as am not in manual mode and test params are zero.

[RetroZero]

I started with 12v supply.
Boost 37000
Mode Permanent Forward
Fslipspnt 8
Ampnom 75%.
Got light bulbs in delta form to flash, then hooked up MG1

[johu]

Cliffhanger
Then what?

[ZooKeeper]

I started with 12v supply.
Boost 37000
Mode Permanent Forward
Fslipspnt 8
Ampnom 75%.
Got light bulbs in delta form to flash, then hooked up MG1

Thanks, I will try that as well and verify MSDN on #25.

[imwoody36]

this is how I added a 10k resistor to c37

[RetroZero]

Nice.
I still have 4 throttle faults. But when I press the pedal, at about 3/4, potnom jumps to 85%.(I set pot values to min and max found in actual values when looking at spot values section) .
And light bulb on hv line starts flashing... Like dead short....
I'm only running on 12 v, I should get some motor spin on that, since it worked in manual mode?

[ZooKeeper]

IT WORKS! Flashing lights AND Jiggling MGR

MORE POWER!!!!!! "I'm Givin' her all she's got Capt'n"

Answer was provided by JH, I had pin 25 of the Prius Inverter marked for shifter I/O, needs 12v "Run" (Ignition switch, relay, etc) signal.

<edit: Adding settings, hope this helps others>

[LRBen]

Got the delta lights to flash using manual mode. Used the settings that Retrozero used as well as connecting inverter MSDN pin to the 12v supply.

I was expecting each light to flash in series. What I got was all lights flashing at once but at different levels of brightness. Is that the correct outcome?

[arber333]

Got the delta lights to flash using manual mode. Used the settings that Retrozero used as well as connecting inverter MSDN pin to the 12v supply.

I was expecting each light to flash in series. What I got was all lights flashing at once but at different levels of brightness. Is that the correct outcome?

Hm... that would be consistent... you are using incandesent bulbs. I used LEDs so i guess filament doesnt cool down very fast. Maybe you could command it 0.3Hz slip? That should make each light flash 1 second apart. What deadtime are you using?

A

[RetroZero]

Yeah, that's what I got. Good to see another 12v start up in progress.
I didn't touch deadtime. Left it as I found it. Don't have values with me, but I think it was 26...

[LRBen]

Hm... that would be consistent... you are using incandesent bulbs. I used LEDs so i guess filament doesnt cool down very fast. Maybe you could command it 0.3Hz slip? That should make each light flash 1 second apart. What deadtime are you using?

A

Using 0 deadtime. Tried it again at 0.3Hz slip. Now they all flash at the same time in one second intervals.

[arber333]

Yeah, that's what I got. Good to see another 12v start up in progress.
I didn't touch deadtime. Left it as I found it. Don't have values with me, but I think it was 26...

You may want to use Deadtime 160 to start with. Remember Damiens gen3 destruction story?
It is just empty space between pulses. If it is too wide it doesnt hurt. If it is too short on other hand...

A

[LRBen]

I thought the dead time was only needed on the Gen 3 inverter?

[RetroZero]

Yeah, I remember those 2 videos like it was yesterday. From what I understand, gen 2 only switches one side, but I'll give it a go when I can get back to my project...off to a wedding for the weekend . I want spinning MG1!

[arber333]

I thought the dead time was only needed on the Gen 3 inverter?

Ah... i am not sure really. Maybe not. But what does it hurt to try?
Hm... on the other hand i remember gen2 has all low transistors switched on and all high transistors off when at standstill? Then your current would really flow directly through the delta connection. Maybe what you need is a Y connection. Then connect all bulbs like that...

But that would mean any 2 bulbs would be on at a time...
Better try with a simple ACIM motor...

[LRBen]

I'll hook up an ACIM over the weekend. Seems like progress anyway.

[johu]

You may want to use Deadtime 160 to start with. Remember Damiens gen3 destruction story?
It is just empty space between pulses. If it is too wide it doesnt hurt. If it is too short on other hand...

No we are on a gen2 here. The lower channels remain unused, the power stage creates its own deadtime. Leave it at 0 or it will distort your waveform.