Prius Gen 3 Inverter Logic Board Support Thread

[konstantin8818]

Wrong way=) groove must contain a red sealant and face towards inverter=)

Also socket must go thrue plate like this:

[bobby_come_lately]

Aha! I did wonder. Reinstalled the right way around, with the seal.

[konstantin8818]

Aha! I did wonder. Reinstalled the right way around, with the seal.

So how it sits? Does it fall off or hangs in? It is pretty soft so I believe when you tighten all bolts there is no gap left between plate and casing?

[Jack Bauer]

That's amazing guys:)

[Martin1775]

Many thanks to Damien,

the prius inverter now can be used with his board without changes, when using his customized Firmware.
viewtopic.php?f=14&t=488&start=60#p11328

I use the MGR connected straight connected pin1 to pin1 etc. and with syncofs=22100 it run's fine. I think the offset is necessary because the Prius defined U at pin2, that will be L1 at the motor, and from the inverter hardware ist is assigned to L2.
Up to now only in idle tested with low voltage (36V), in forward direction up to 4000 1/min and smooth in both directions.

MGR and MG2 wired the same, but with MG1 the phases are different.

Designation at the Prius inverter in the Toyota documentation is: pin1=W, pin2=U and pin3=V
Designation at the RX450h inverter in the Toyota documentation is: pin1=W, pin2=V and pin3=U
Designation at the MG2 and MGR in the Toyota documentation is: pin1=W, pin2=U and pin3=V
Designation at the MG1 in the Prius documentation is: pin1=V, pin2=U, pin3=W
Designation at the MG1 in the RX450h documentation is: pin1=U, pin2=V, pin3=W

[bobby_come_lately]

Aha! I did wonder. Reinstalled the right way around, with the seal.

So how it sits? Does it fall off or hangs in? It is pretty soft so I believe when you tighten all bolts there is no gap left between plate and casing?

Fits fairly well. But printed in PLA there's a bit of bowing around the middle from the pressure of the sealant ring - see picture. I wonder if it would be better in ABS/PETG? Or whether it's worth stretching the mounting holes out a fraction of a millimetre to provide some tension?

[bobby_come_lately]

Done a quick blanking plate for the old battery connector hole. Not fully tested yet - waiting on some 47mm o/d o-rings that should provide a snug fit.

[konstantin8818]

Fits fairly well. But printed in PLA there's a bit of bowing around the middle from the pressure of the sealant ring - see picture. I wonder if it would be better in ABS/PETG? Or whether it's worth stretching the mounting holes out a fraction of a millimetre to provide some tension?

Maybe when ampseal socket will be screwed to plate, it will straighten it. Or maybe opposite
Anyways, I've made second version: with 10mm offset to make it more sturdy.

gen3 plate V2.zip

(23.08 KiB) Downloaded 137 times

Looks almost OEM

[Jack Bauer]

Some foc testing :

[arber333]

That is just ridiculous. I expected it would start in manual and wouldnt be stopped.
I would recommend you calculate your syncoffset + 45deg, just because it is an IPM.

It is strange, i havent yet seen motor offset higher than 15000. I think them engineers setup resolvers to first quadrant and then use the offset. I think if you find offset higher than that you need to swap your phases and resolver pinout.

Nice work D!

EDIT: Maybe Elon poles can be counted? Did anyone open the Model 3 motor yet? Pole count should be visible on stator. Then automakers use the same pole count on resolver. They are lazy not like industrial motors which have the strangest of resolvers....

[Bassmobile]

Fits fairly well. But printed in PLA there's a bit of bowing around the middle from the pressure of the sealant ring - see picture. I wonder if it would be better in ABS/PETG? Or whether it's worth stretching the mounting holes out a fraction of a millimetre to provide some tension?

I would suggest Nylon if your printer can do it.
Going to download and print in nylon. I would be happy to print one for you and send it off. Been using Nylon here for most 3d printed parts because of it's ability to withstand temperatures once printed, and it's high tensile strength.

[Kevin Sharpe]

Did anyone open the Model 3 motor yet?

Lots videos by Munro in the Model 3 threads

[bobby_come_lately]

I would suggest Nylon if your printer can do it.
Going to download and print in nylon. I would be happy to print one for you and send it off. Been using Nylon here for most 3d printed parts because of it's ability to withstand temperatures once printed, and it's high tensile strength.

Interesting - thanks. This may be a different thread, but I was planning to use PETG for parts on the car because of its water resistance. I thought printed nylon struggled with that? Though they say that about PLA and I printed an RC boat that seems to be fine after a few tests. I'd need to upgrade my printer for nylon too, unless I used one of the low temperature ones.

[MattsAwesomeStuff]

EDIT: Maybe Elon poles can be counted

https://cleantechnica.com/2018/03/11/te ... -in-depth/

Two year old article discussion speculation about the new motors, but it mentions that the stickers given to dealers say it's a 6-pole.

[Bassmobile]

...PETG for parts on the car because of its water resistance. I thought printed nylon struggled with that...

In terms of water absorption, Nylon PA 12 printed parts absorb very little moisture. PA-12 is highly water-resistant but not completely waterproof. PA-12 printed parts should not remain submerged in water for extended periods of time. But the same goes for most of the other parts on your hacked Prius Inverter.

[Jack Bauer]

Thanks to everyones help I have now got the prius gen 2 mg2 spinning beautifully with foc using the V1c board in a gen3 inverter. Attached parameter file.

Some things I learned :

1)I wasnt pumping in enough manual Id to find the right syncofs value. ie i put in a syncofs, tried 5a then if the motor didnt spin I figured this was the point of no spin. I just needed to man up and pump in some real Id:)

2)Once I had found the point of no spin I tried run mode but got runaway condition. A brain cell fired up and I added 32768 to the no spin syncofs value. Result perfect throttle control but running in the opposite direction.

Hope that helps someone.

[konstantin8818]

Thank you for sharing parameters. I hope Lexus MGR has the same parameters as prius MG2, so it will save me a lot of time. Maybe only need to tweak syncofs as Martin wrote earlier.

[Martin1775]

Here is my parameter list from testing with the MGR. I use a start button, a forward and reverse switch and a poti for acceleration and Damiens FW.

From the first test's I changed the connection to the motor because the phases from the Motor don't fit. I changed it so, that U is L1, W is L2 and V is L3. I think that's right, because of the exchange from L2 and L3 in the software, the controller see it in the right assignment. Now it runs without offset (theoretically it should be 255 from the former value 22100).
The resolver is connected as in the original motor description MRFG=GND, MSNG=SINA, MCNG=COSA, MRF=EXCA, MSN=ENCA and MCS=ENCB. The motortemp sensor is actual unsupported, the 47k NTC is only for the inverter temperature selectable.

With 36V it was fine in both directions, I've increased the voltage to 60V and the motor can still be steered perfectly in the forward direction up to 6000 rpm, but now he jumps at around 500-1000 rpm when I decelerate in reverse, accelerate is OK, there is the maximun speed 2000 rpm.

I've tested it fast with different syncofs +/- 1000 but it get not better, so I think I must get some time to testing it.

[konstantin8818]

Much appreciated! Will test it on johu's new board hopefully. I've bought gen3 inverter dirt cheap, 15 euro or so, and as I expected, boost IGBTs are blown(and I can't see deep inside to find out how's mg1 and mg2 IGBTs are feeling), so unfortunately no full efficiency from Damien's board for me, unless I'll use it as dual motor controller with separated charger.
Did anyone printed second version of the plate? Maybe improvements are needed?

[bobby_come_lately]

It's on the printer now Konstantin - I'll let you know how it comes out tonight.

[bobby_come_lately]

Much appreciated! Will test it on johu's new board hopefully. I've bought gen3 inverter dirt cheap, 15 euro or so, and as I expected, boost IGBTs are blown(and I can't see deep inside to find out how's mg1 and mg2 IGBTs are feeling), so unfortunately no full efficiency from Damien's board for me, unless I'll use it as dual motor controller with separated charger.
Did anyone printed second version of the plate? Maybe improvements are needed?

So, printed revised plate. See photos. Still a little bowing but much, much better. As you can see from the other pictures, that rubber gasket is both fairly thick and quite stiff. May need a stiffer plastic to get a really tight fit, or a deeper channel for it?

Bigger issue is DFM now - I printed without support as I didn't want to dig material out of the channel for the gasket but that leaves a lot of spaghetti with the overhang.

What are you designing in? If it's Fusion360 would you be happy to share the raw design and I can do some iterations?

[bobby_come_lately]

PS - Damien: when can we buy the little boards for the Ampseal connector? My board just arrived (thanks!) but would love to get one of these too.

[konstantin8818]

What are you designing in? If it's Fusion360 would you be happy to share the raw design and I can do some iterations?

I'm working in Kompas 3D. Ain't Fusion360 capable of working with .stl? I have it but don't like to use it because it is heavy and it rely on internet connection, and my internet is prehistoric...
As for the part itself, I'll make groove a little bit deeper, but it would be great if you measure how far sealant is poking out, so I won't overkill

[bobby_come_lately]

I can pull in STL but F360 allows cloud sharing of raw files - all the sketches and features - which means it's easier to make tweaks. But as you say, it is *very* dependent on internet connection.

My cheap calipers say the gasket is currently protruding 2mm from the face.

Thanks for all your work on this btw - much appreciated!

[konstantin8818]

My cheap calipers say the gasket is currently protruding 2mm from the face.

In original plate gasket pokes out like 1,5mm and from your photos I can't really tell if it fits perfectly inside the groove.
So I've widen the groove 0,3mm inside and made it 0,5mm deeper. I hope now it should fit like a glove

gen3 plate V2_1.zip

(22.92 KiB) Downloaded 94 times