Tesla Model 3 Rear Drive Unit Hacking

[JaniK]

so tiny looks like you are getting it finalized. So I ask many questions:

How is the JLCPCB doing with deliveries at the moment?
Have you tested minimum voltages running the model3 inverter?

Can we see it on evbmw shop soon for preorder ?

Just suggesting:
"Support for a product, arrives when done, or a nice postcard from lanzarote"

[Jack Bauer]

JLCPCB are doing fine. No problems that I have encountered.
The rdu inverter seems to need about 75v on the hv side but I have no idea about the motor characteristics as yet.

Soooo...you want me to start selling a completely untested and unproven board? What could go wrong?

On a serious note I am happy to do that as long as the community accepts responsibility. This may take many versions and many iterations before it works properly. If the greater openinverter populus want to help this happen by means of buying alpha hardware then that's cool but that means accepting that things like blown inverters , build errors, etc will happen.

[Kevin Sharpe]

that means accepting that things like blown inverters , build errors, etc will happen.

This is an important point... please DO NOT buy this board unless you are willing to trash your motor/inverter, work with little/no documentation and NO PRODUCT SUPPORT.

The 'community' is not ready to hand hold people who are out of their depth and this is very much the bleeding edge of product development

[Jack Bauer]

Now that's what I call a product endorsement:)

[Kevin Sharpe]

Now that's what I call a product endorsement:)

Yeah, 45 years in the electronics industry helps

Unfortunately, we all know the NEWB's won't listen and will be trashing hardware and asking for help... at least we get to say "I told you so"

[JaniK]

OK, I understand, There are good points.

There is still much other things to do so I will patiently wait for something that is tested working. Since I am proud to say I am new on this stuff. I can solder trough hole stuff, but cannot design PCBs or program much.

Thanks for the info, happy hacking!

[Jack Bauer]

Now in production at JLCPCB. Changed the wifi header to smd type to avoid through hole pins poking down onto the oem pcb. Also added a 3v3 reg to allow for single 5v supply operation. May be deleted subject to testing.

Given recent developments the design files will not be released at this time.

[PacEmaker]

Respect

[cloudy]

Awesome, top work!

[JaniK]

Great work Damien!

[JaniK]

About the drive unit itself, can it be driven in reverse full time?
Or is there lubrication issues or something like the model S units?

[Jack Bauer]

Given that it uses an external electric oil pump I would see no problem with reverse. Of course we won't know until somone does it and shares the information.

[Jack Bauer]

In other news JLC have dispatched the prototype boards. Due into me this week with luck.

[JaniK]

Thank you for the answer, I will look into it some more.
I could run it in reverse in a car later if no issues arise.

[Jack Bauer]

Boards have arrived from JLC. First powerup test on the bench is a success:) Next up fitting one to the inverter and hopefully spin the motor. Lots to do.

[cloudy]

Looking good!

[JaniK]

It is a cute little board

JLC delivers quite fast? Do they come to you without the connectors?

[Jack Bauer]

[Jack Bauer]

Some info on the M3 rear du temp sensor and resolver wiring.

[Jack Bauer]

First attempt running the drive unit using FOC firmware in the prius inverter today. No luck as I think I have my current sensors mixed up. Going to do some investigation on this over in the Gen 3 prius thread then back here to see what happens:)

[Jack Bauer]

[Jack Bauer]

[arber333]

Damien just try giving it 10 poles and the same resolver. I suspect 10 poles should be much for such a small motor. It gives higher torque at lower rpm and good max RPM count regarding 360Vdc bus.

If it still doesnt work then try 8pole. I am sure they wont use odd pole numbers . I will look for some dissassembly videos.

[Kevin Sharpe]

Damien just try giving it 10 poles and the same resolver. I suspect 10 poles should be much for such a small motor. It gives higher torque at lower rpm and good max RPM count regarding 360Vdc bus.

If it still doesnt work then try 8pole. I am sure they wont use odd pole numbers . I will look for some dissassembly videos.

Two comments on the YouTube video;

"The prius gen2 motor is a 4 pole pair motor. It would pay to check the resolver too, pretty sure it is 4 pole pairs also. The inverter will not run properly if these parameters are wrong since FOC depends on correct rotor angle information to function correctly."

"Actual resolver may only have 2 pole pairs. In any case if you plot rotor angle you should get 4 0-360 degree transitions for every 1 turn of the rotor. Adjust resolver pole pairs until that occurs."

[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.

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.

In regards the M3 motor I have trawled some of the "Masked Maraduers" videos and think what we have is a 3 pole pair motor and a 3 pole resolver.