As many of you know, we’ve been working on open-sourcing a drivetrain from a 2011 Smart ForTwo Electric Drive. We can control the DC-DC converter, the AC unit, the charger and make the motor turn. It was time to move from our complete drivetrain test assembly to a workbench assembly where each of the components are set up and wired both independently from, and to each other, verifying we can run each of the items successfully independent of the vehicle’s wiring harness and together as if they have been placed into a custom EV build.
After moving everything to a bench and chasing down lots of wiring issues, setting up liquid cooling and a high voltage power supply, etc., we realized that even though we had all of the components working (to some extent), we still had to:
- map the motor data streams for some semblance of speed control over the motor (all we can do now is make it spin at one speed)
- build a control board
- write software to pull everything together into a single package and make it more than less usable.
All of this would be done by trying to control the motor through the existing control board via the existing software and whatever surprises that may have in store for us.
We were batting around different ways of making all this work, noting that none of us currently working on the project are what you’d call “programmers.” After much discussion and debate, we decided to take a page out of EV West’s book and build a new board that would control both the IGBT driver board and the DC to DC converter board and bolt into the existing control board’s location inside the motor housing. Now we can take advantage of work that has already been completed, and move the project along more quickly. Our board design is patterned after Damien Maguire’s Tesla build.
The new approach is to replace the inverter’s control board and software; this will make the motor easier to control, as we have access to the software that runs it, but slightly less efficient due to the type of control strategy being used. However, the control strategy we’ll be employing will work on almost any synchronous-wound AC motor and is approximately only 1-2% less efficient – a huge advantage for the small loss in efficiency.
Ideally, after we’ve built this board and have it working, we’ll be able to use it as a template for almost any other AC synchronous motor produced (i.e. Tesla, GM, Nissan), and quickly adapt it to run any of these motors. Of course the plan is to open source everything so you can build one, too. We may also make a small run of the boards for the Smart; if we’re going to build 4 or 5 for ourselves, we may as well have 50 made.
I’d like to thank Jarrod, Damien, Collin, the Software Guy, and all the other open source folks out there making this possible. Also thanks to EVWest who’ve been instrumental in making this happen and supplying the parts we need.
Here’s a link to Damien Maquire’s Tesla Project so you can see how he does it: https://www.youtube.com/playlist?list=PLPHK4T9kKEyblTISV7u-PrSgSV6VD4YYz
4 Responses
Hello Kevin
Would be really interesting to discuss with you the challenges you faced, trying to replicate your researches here.
Have you used the research of Charlie’s miller on how to control an ECU over the CAN bus? And replicating and faking ECUs? I would bet some checks are done at the software level that disable some functionnality, but would like to know how far you went on this path (basically to decide whether i should or no buy a smart fortwo that now not working because of safety interlocks of the airbag)
let me know if you could have a short free slot to discuss over skype,
Thanks,
Louis
Hey Louis
Thanks for your interest in our project and sorry for the late reply…always busy it seems.
There weren’t really any software Locks on functionality for our model other than top speed, at least that we ran into.
But we were also not trying to get a whole car working, just the motor/controller, ac, and charger. What you’re talking about sounds like a whole other can of worms.
Additionally we’re working with a 2011 model year and it seems almost every single model year of this vehicle has some significant differences between years.
I’d be wary, not even the dealers around here seem to be able to change out a battery pack and get the car going again…I imagine airbags may be more difficult.
Unfortunately I cannot tell you if it is possible or not to get the car running again after the airbags have deployed.
Have you had any luck with asking local dealers if they can unlock it for you?
Best of luck – Kevin Smith
Hi Kevin & Illuminati Team,
I met you out at the recent Battery Show in Michigan this past September, and wasn’t aware of your work with the Smart car. I’m a graduate student and obtained a 2013 Smart EV from a salvage yard, and have taken it mostly apart. Have you made further progress with the motor controller? I’ve been able to read battery diagnostics and information from Arduino via CAN bus connection, but haven’t yet tried to write any commands or get the motor to spin. I’ve tested my battery and even though cells are low voltage, by testing a few they appear like they’ll come back. My airbag sensors and fuses blew during the accident, but I understand you have to actually reset the controller rather than just replace fuses.
Thanks,
Wes
hi @Wesley Connor and ilimanti team, I own a 2013 smart fortwo Electri with P01851c error, I charge up to a level 3.7v each cell before was 3.1v the error cannot be erase with MB star c4 (xentry) anyone with same progress