Tesla Model 3 Rear Drive Unit Hacking

[AMP3R]

Who have seen this solution? https://www.ev-controls.com/product/ev-controls-t2c/

[Jack Bauer]

This topic is relating to open source solutions for running the model 3 drive units not commercial closed systems.

[vwbrady]

This topic is relating to open source solutions for running the model 3 drive units not commercial closed systems.

Where's the like button on this post?

[JockE28]

Following with great interest

I hope I can contribute to get this motor spinning in my own project EV.

Do we know that all 980 motors are the one we want?

Mine has p/n 1120980-00-F

[davefiddes]

The 980 variant has 4 MOSFET pairs per phase vs the 990's 3 pairs. Info on the wiki came from a Tesla Reddit that indicates this boosts the current capacity from 600 to 800 amps. Bit academic at this stage of the project. That's the only difference I'm aware of.

[JockE28]

A little o.t. but does anyone have a DU that they can take a few measurements on? I would need the outer diameter of the "motor" part to check if my mockup 3d-printing scaling is correct


picture på svenska

[davefiddes]

There's a high quality 3D scan here: https://grabcad.com/library/tesla-model ... vshop-fr-1

I found that measurements taken from it in Autodesk Fusion 360 and DesignSpark Mechanical 5 match up well with measurements taken in the real world on my drive unit.

I don't think there is a simplified CAD model shared anywhere. It sounds like you are building one?

[EV_Builder]

How is SW progress going?

[davefiddes]

It's still going. Progress is slow but forward. Took a while in the last few months to build a PC based tool to allow configuration of openinverter systems via CAN. Now working on getting CAN communications working reliably on the C2000 now I have a client. Will provide a more in depth update when I have a minimum viable inverter running (i.e. something others can help with development of).

[JockE28]

Overall dimensions seams to be correct in that scan and my model. Just want to be double checking. This will take a "few" days to print so just confirming the first parts will make me a little calmer.

[JockE28]

It's still going. Progress is slow but forward. Took a while in the last few months to build a PC based tool to allow configuration of openinverter systems via CAN. Now working on getting CAN communications working reliably on the C2000 now I have a client. Will provide a more in depth update when I have a minimum viable inverter running (i.e. something others can help with development of).

Nice work

[EV_Builder]

It's still going. Progress is slow but forward. Took a while in the last few months to build a PC based tool to allow configuration of openinverter systems via CAN. Now working on getting CAN communications working reliably on the C2000 now I have a client. Will provide a more in depth update when I have a minimum viable inverter running (i.e. something others can help with development of).

Did you use an RTOS?

[davefiddes]

Did you use an RTOS?

No. As you can see from the code published so far I'm sticking closely to the existing architecture of Johannes' codebase.

[davefiddes]

Took a couple of basic drive unit measurements:
Max Width (from outside of mounting lug to mounting lug): 675mm
Max Depth (back of DU to front edge of mounting lug): 565mm
Measurements probably +/- 5mm at best.

Checking against CAD model in DS Mechanical shows the width as 6752mm. So out by x10. IIRC Fusion360 it was spot on but I don't have that installed any more.

[JockE28]

Took a couple of basic drive unit measurements:
Max Width (from outside of mounting lug to mounting lug): 675mm
Max Depth (back of DU to front edge of mounting lug): 565mm
Measurements probably +/- 5mm at best.

Checking against CAD model in DS Mechanical shows the width as 6752mm. So out by x10. IIRC Fusion360 it was spot on but I don't have that installed any more.

Seems to be accurate. Thanks a lot

[JockE28]

Just found another super exiting cat video on YouTube. Nice work Damien

[Roschii]

Hi guys,
I'm new to the project and I'm having some issues with the software.

We want to install the Tesla motor in an Evum (a small commercial vehicle from Germany). In addition, he should then be converted to the off-road version with a roll cage, etc. This is not a problem, but none of us are software professionals...
We bought the JTAG adapter board from the evbmw shop and are now having problems copying the chip. I'm working my way through davefiddes' guide on github. However, we are currently already stuck on the build process.

I downloaded the github files and saved them in the created c2000 folder. However, when you run

Code: Select all

cmake --preset c2000 ../..

always the following error message (image 1).


Before that I had compiled the libopencm3 file for STM32 in VS code (at least I hope the output is in picture 2)

I would be really grateful if I could possibly help me on the jumps.
Thanks very much

[davefiddes]

My c2000-inverter project as it currently stands is not a functional inverter. If you attempt to run it on a Tesla Model 3 drive unit it is trivial to brick the drive unit and render it completely inoperable. Once bricked it is impossible to fix at this point in time.

Damien is working on an alternate logic board based on the existing STM32 inverter software which may possibly produce a working solution sooner. This project is still in its very early stages.

If you must use a Tesla Model 3 drive unit I would highly recommend using one of the commercial solutions from Ingenext or EV Controls.

An Open Source hardware and software solution will emerge eventually.

[Roschii]

Thanks for the answer, good to know.
So Mr. Huebener's software on https://github.com/jsphuebner/stm32-sine doesn't work either for the drive unit?

[johu]

No, because as the name suggests that is written for the STM32 processor, not for the TI. Dave is trying to port stm32-sine to ti-sine and Damien is trying to fit an STM32 to the M3 board that runs stm32-sine

[Roschii]

Ah ok, thank you very much

[someguy]

My c2000-inverter project as it currently stands is not a functional inverter. If you attempt to run it on a Tesla Model 3 drive unit it is trivial to brick the drive unit and render it completely inoperable. Once bricked it is impossible to fix at this point in time.

How exactly is the drive unit getting bricked ? Does the micro controller lock itself ? Is it frying gate drives ? Is there a security chip on the board ?

[someguy]

Damien and Dave: have you seen this ?

https://static.rainfocus.com/rsac/us21/ ... 01bujx.pdf

[davefiddes]

How exactly is the drive unit getting bricked ? Does the micro controller lock itself ? Is it frying gate drives ? Is there a security chip on the board ?

It is trivial with TI Code Composer Studio and a suitable JTAG adapter to erase all the flash sectors in the TMS320F28377D containing Tesla software. Tesla have used the security features of the MCU to lock access to the flash preventing readout so there's no way for anyone outside of Tesla to restore it. I know it is easy to do because I've done it and rendered my £1500 drive unit worthless.

[someguy]

It is trivial with TI Code Composer Studio and a suitable JTAG adapter to erase all the flash sectors in the TMS320F28377D containing Tesla software.

OK.

Tesla have used the security features of the MCU to lock access to the flash preventing readout

To confirm, there is no way to read Tesla's firmware ?

so there's no way for anyone outside of Tesla to restore it.

Can you write your own firmware into the processor or is the processor locked from writing new firmware into the flash area ?