Electric vehicles and structuring and sustaining a FOSS community

The Hangar Off Topic
#1

If there ever were a more appropriate off-topic label… .

I’ve been doing some posts on solar estimations using data from OpenDroneMap, i.e. here and here. In the first of those, I made reference to acquiring an EV and looking for ways to power that. Some folks have asked what I acquired and it’s an odd-ball choice – it is a 2017 Mercedes Benz B250e. It’s kind of a box-of-parts car, like if Mercedes-Benz made the mistake of financially bailing out Tesla and decided to learn some lessons from that mistake by putting together some Tesla parts into a bclass MPV, selling as few as possible to comply with California codes, and then bail on Tesla investments after/while building their own EV team. Oh, never mind. It’s not like that: it is that.

Anyway, it’s got a tiny battery (28-31kwh usable), not fantastic efficiency, no rapid charging, but is super comfortable and kind of the nicest commuter EV at its price. And it almost immediately died due to a coolant leak in the motor coils and is waiting on a (thank goodness warranteed) motor replacement.

But that’s not what I came here to talk about. I came here to talk about this really cool FOSS forum and wiki I discovered called openinverter.org:
image

It’s a bunch of folks building, reverse engineering and then building, and open sourcing all the info needed to build EVs, from motor control, charger development, battery controls, etc… It’s really cool.

And the reason I bring it up (other than I am looking to add fast charging to my B250e if/when it ever comes back to me) is that I am a super outsider there. I’ve only read stuff, not even posted yet. I know not much about cars, even less about HV electronics, and I am reading through their documentation and forum discussion, and youtube videos, their cottage industry of suppliers and support contracts. Getting to see a FOSS community from the outside for the first time in a while, and it is a super useful exercise for me for understanding what we do well and what we might improve.

Oh, and if you get the urge to buy a Mercedes-Benz B250e or RAV4 EV, go ahead and double check the motor before purchase using the directions here (youtube link):
image

Those Tesla parts weren’t so reliable, as it occurs, but it is easy to check for damage.

4 Likes
#2

Not to derail, but I hope to see a similar scene grow to sUAS, where open firmware rules the space, from ESCs on up to control logic (I have a reliable source that tells me DJI white-labels OSS code).

Part of the EV space should be an international standards body to keep components, code, etc all reasonably interchangeable and shared… Think about how awful personal computers would be if you couldn’t just grab parts and expect a certain interface or bus to behave a certain way.

1 Like
#3

That doesn’t derail at all, or maybe it just derails in a way that strikes my fancy. I have taken a pretty deep dive into the ecosystem in question and understand better their intents and purposes. They basically are building this ecosystem out on 3 (now) intertwined tracks:

  • True from first principles FOSS controls of electric vehicle systems
  • Backwards engineering of existing mass-produced parts to create similar FOSS versions
  • Backwards engineering of existing mass-produced parts communication protocols.

That last one is pretty interesting and compelling to what you state your desire is for the hardware ecosystem. Cars (as you may know) handle component communication on a set of CAN buses (controller area network buses). The hardware for this is really quite simple to minimize the weight and cost of the network. The folx at openinverter have been working through the problem of reverse engineering the communication protocols of the components so they can use a common and useful component and place it in another context and have it still operate.

For example, while in-car CHAdeMO fast DC-adapters are relatively easy to build, fast DC CCS style adapters (which have higher practical charge rates and are becoming the North American and European standards) are quite hard to build. So, why not take a controller for CCS charging from a trashed BMW i3 and put it in your custom build electric vehicle. Well, you don’t do that because the controller will know it’s not in an i3 and shut itself down. That is unless you know how to speak the same language and convince it is is still in its parent vehicle.

Bringing it back to drones, the lack of industry standard CAN bus equivalent aside, it would be really cool to be able to swap a DJI camera gimbal or similar into a DIY drone and have enough info about the protocols make it think it was still in the appropriate parent vehicle.

1 Like
#4

Should have been MAVLink, but I gather UAVCan is gaining some pretty nice adoption.

Damn 3DR for giving up when they did.

1 Like
#5

True side note: I think if I can manage to install a CCS charger on my box of parts Merc, with a 28kwh 400v nominal battery, I’ll be able to do a 10-80% charge in 20 minutes which is only 2 minutes slower than an IONIQ5, a Porsche, Lucid Air, etc… Now, that 80% charge won’t take me nearly as far, but we’ll ignore that little fact for now. :laughing:

Edit: it might be faster: something close to 14 minutes.

1 Like
#6

You’re making me want to grab a Smart EV and go live my dream of slapping Harbor Freight panels all over it and using it as my commuter… Granted the Aptera will fill this role much nicer, but… Tinker :smiley:

1 Like
#7

You could always pick one of these up:

A used place here in town has one they’d love to just unload for probably most any offer at or above their cost:

image
image1920×2560 338 KB

A smart car might be a better bet as it can probably reach a higher max speed, but I find the thing darn charming.

1 Like