I see 3 circuit boards and a GitHub repo that has an .ISO of the diagnostic software. They include DBC files that define the CAN interface, but that's it.
There's basically nothing here? Or am I missing something? No firmware, no CAD, no design files, no engineering data. It's like they released the service manuals and a couple PCB release files they scrounged up from someone's e-mail, but there's very little in here beyond what you'd get from service manuals or from having a PCB in your hand and following traces around. The PCBs are extremely basic.
I don't get it. Was this just a trick to get headlines from journalists who don't know how to interpret the files on the website? Or was there supposed to be more and it's just not up on the website yet?
Also, it's the Roadster -- a $100k+ supercar which sold less than 2500 units. Not the sort of thing an owner would typically be doing DIY repairs on, if they're even driving it. (I wouldn't be surprised if a substantial number of the remaining Roadsters are held by collectors.)
Yes it's a stunt, but c'mon, those "special tools" are something every manufacturer has, and they're usually not that special, and they're easily replaced.
One of the Tesla ones is literally a storage cart. Another is a sawed-off 26mm wrench head, the "clip tool" looks like it could be cut out of plywood or a sheet of metal in a few minutes.
I guess, but if you ever reversed engineered any 80's and 90's automotive electronics any bit of info you can get is amazing. Currently dealing with reverse engineering a wire protocol / command structure on a discontinued 90's automaker using logic probes its not really fun
Well, I agree any information is helpful. At the same time, this is so blatantly not "all design and engineering files" it is absurd. That statement is a clear lie.
>Was this just a trick to get headlines from journalists who don't know how to interpret the files on the website?
You're discussing a Twitter post by the same guy who knowingly committed securities fraud ("funding secured.") in order to protect the share price of his company.
One of the greatest grifters in history.
Do you expect him to be honest and upfront in his PR dealings?
"All design and engineering" -- even for Elon that's a ridiculous statement. Not only does this come with the disclaimer "It ... may not accurately reflect the actual production models or parts sold.", it doesn't include any software source at all.
I really don't understand why he is so willing to post such obviously verifiable lies. He also said he open sourced the twitter algorithm and what we really got were a few files which did not make it possible to understand the algorithm.
And then even when he posts obvious, verifiable lies, there is a crowd of people who believe him anyway.
I just wonder why is he so willing to sacrifice credibility with people who understand the world just because there's a load of suckers who won't see the lie? I guess he considers having more suckers on his side more valuable at this point, as he can hire people with credibility as needed. But to me it's a weird choice.
right but suckers and informed people offer different things. it is notable to me that elon is selecting for manipulating suckers here instead of appealing to informed people. not surprising to me per se, but a useful example in discussion with other people about what to believe in the press.
For the same reason cult leaders say outlandish to their followers, they only want those who are willing and able to be led astray to follow them versus those who might question or defy them.
It also allows for opportunities for the cult as a whole to get their two minutes of hate in and to punish an enemy/scapegoat/turncoat/etc, reinforcing conformity with what the leader preaches.
It's also a power thing. They have so much power over others that they can say blatant lies and followers will trip over themselves to believe and defend it.
The rest of the files were not theirs to release. The Roadster was based on a Lotus vehicle. A license to use a design doesn't grant you permission to release that IP to the world!
I don't think anyone is complaining that Tesla is not releasing Lotus design material. I think people are complaining that obviously Tesla-designed bits aren't in the materials despite Elon saying "All design & engineering of the original @Tesla Roadster is now fully open source".
> To accommodate the battery pack, motor and transmission, Tesla designed a new rear subframe. In order to handle the increased weight in the rear, new rear wishbones were designed.
I don't see any design materials for Tesla's rear subframe or rear wishbones in the materials that Elon linked to though.
Honestly, releasing what they did is commendable. Typically automakers (including Tesla at times) fight right-to-repair type stuff. It's just irritating to read the headline and think "holy crap they open sourced a car!" and then realize it's really "oh, they released some above-average technical documents about the car".
If you legally can't release something as open source, maybe don't tweet that you're releasing it as open source. Or at least say something like "we're open sourcing the parts we can" (and actually do that).
I'd seriously consider paying double that for one which was genuinely open-hardware.
I'd never consider Tesla since the way it works, you buy the physical hardware, but in practice, Tesla controls your car. It's not true ownership. This seems like a 180 degree turn. At least for me, it would seriously increase the odds of buying a car and the price I'd be willing to pay.
It'd be interesting to see a company pull this off with an in-production car. Knowing the crowd, ideal audience would either be a pickup truck or a basic sports car.
> This seems like a 180 degree turn. At least for me, it would seriously increase the odds of buying a car and the price I'd be willing to pay.
Take a look at the linked files before you get too excited. They didn't release much more than the service manuals.
There's very little source material on the page. 3 simple circuit boards and some definitions for the CAN interfaces. They released an .ISO of the diagnostic software, but that's generally floating around on the internet for most cars if you know where to look anyway.
An old American V8 has about a million options for everything, can get parts everywhere for cheap (I got a brake rotor for $8 recently).
It seems that it would be ideal to have the open source cad design files so you could make your own rotor when it breaks, until you realize it would cost you $1k to make one.
If this is your goal get an American car from the 60s. The problem is they're super unreliable, because the things that make internal combustion engines reliable are complicated and not easily replicatable by the average hobbies engineer.
I design this stuff for a living and I just buy a car, don't touch it, get it serviced at the dealer, and get a new one when I reach the right side of the bathtub curve and it's service life is up and is stops being reliable.
> We keep reading that people are willing to pay for a "Dumb TV"
They aren't though. People might be willing to chose a dumb tv, all else being equal. But not pay _more_ for a dumb TV. With subsidies from the smart tv services, it isn't cheaper.
First off, I think it’s important to recognize the intention here before getting into areas for improvement. There is a lot of material in here for the 2008-2012 Roadster. If I was an owner, I’d appreciate this level of access.
I’m curious to get a lawyers take on their brief terms on this service page combined with these patent and “open source” terms [0]. Do these amount to OSD-compatible terms? I don’t see any restrictions to immediately disqualify, but I also don’t see explicit licenses everywhere I’d expect to see them. Also, I’m seeing this on my phone, so I haven’t checked diagnostics software ISO file for source code.
A nit, it’s good practice to remove confusing confidentiality notices on schematics and hardware documents contained in their ZIP files.
Is this textbook open source / open hardware? Not sure.. is it something published without restrictions other than “no warranty”? It appears that way.
I do appreciate you found a webpage containing the actual term "open source," unlike <https://service.tesla.com/roadster#:~:text=Disclosed%20Resea...>, which is linked to from the readme, but its actual content appears to be closer to "gl;hf" than any licensing terms. Also, furthermore the open source link you found doesn't contain the word Roadster, so it's further unclear whether the roadster has any such upstream repos to link to
Either from the Tesla factory safety issues you mentioned, or from the MediaMatters / Twitter allegations and lawsuit. Plenty of reasons for him to want people to look elsewhere.
I was kinda hoping to see actual drawings for e.g. the panels or the seats or whatever. So if I really did want to build one myself or create replacement parts I could do so.
It would be really great if they'd release all the mechanical diagrams, there is an infinity of software out there already, and a lot of PCB cad designs. But if you want to study professionally designed mechanical parts there is a lot less material out there.
Yep, the original Tesla Roadster was based on the Lotus Elise, but they stretched the wheelbase a couple of inches, and from a glance you can tell that the body panels were changed.
I didn't realize there was a tracker for the launcher Roadster https://www.whereisroadster.com/ - I clearly misremembered that they sent it into the Sun
The opposite, really - the chassis and basic suspension design, as well as most of the interior parts, are from the Elise. The rear sub-frame, hubs (5-lug vs 4-lug), complete drivetrain, brakes, and all body components besides the windshield are specific to the Roadster, though.
The Elise itself was sold in several forms with a ton of different ICE drivetrains (Rover 4-cylinder, Toyota 4-cylinder, Toyota 6-cylinder, GM Ecotec) and is popular for these kinds of custom drivetrain builds, since it was a highly modular platform - fundamentally the Elise is an aluminum tub with a front composite crash structure glued and a rear tubular subframe bolted on.
> The information provided here is being provided as a courtesy to Roadster enthusiasts and was created during the design phase of the Roadster for research and development. It isn't manufacturer reference or repair and maintenance material, and may not accurately reflect the actual production models or parts sold. If you use this information, it's your responsibility to ensure that you follow all laws and safety protocols as we don't provide any warranties on any work done by non-Tesla personnel whether or not you use the information provided here. You also understand that if you make or design parts or create new repairs or procedures based on this information, we won't be responsible for them and they won't be genuine Tesla parts or accessories or Tesla-approved procedures.
that lengthy disclaimer sure seems to disagree with el(on) jefe.
Would love to see any early plans for the 2 speed gearbox. :)
Given that this is so heavily based on a Lotus Elise, it feels like the engineering scope is going to be quite different from a normal vehicle. It'll be interesting to see how much of the Elise's design will incidentally also be visible here.
I mean, if you have the firmware binary and the relevant arguments for the ISP (that is, In-System-Programmer), that's sufficient to make some PCBs do their jobs. Maybe not these, but you don't need source code as such to program a thing.
Yes, but it does not allow your to modify it, replicate it, or make it your own. The entire philosophy behind "free" software and open source, etc.
In this case, it's highly likely that one or more components on those PCBs is unavailable now, so you need this source code to update what may need to be updated to be able to keep things alive.
Also the "Parts Manual (EPC)" [0] is interesting since it contains the part numbers of a couple of models (maybe all, I don't know what models they offer) and is very well organized. I'm not sure I'd find this for my VW.
What's missing from https://parts.vw.com/ ? The Tesla UI is a little better, I guess. It's also not particularly difficult to find the VW parts catalog software, EKTA.
There's an industry-standard set of interchange formats for spare part number listings called PIES and ACES. Unfortunately no vendor I'm aware of openly shares the underlying PIES/ACES data.
I never really figured out why there's not a corresponding non-US market official site, it might be a regulation thing indeed.
https://volkswagen.7zap.com/en/rdw/ has what I think is a scraped version of EKTA (it's supplied by https://levam.net ) and is absolutely infested with horrible ads, but has pretty complete catalogs for ROW cars.
To save others the bother of finding the relevant link: https://epc.tesla.com/en-US/catalogs/301 is the Tesla Roadster but if one drills into (e.g.) "10.15 - Door Shells and Beams
- Door Shells and Beams - 1" <https://epc.tesla.com/en-US/catalogs/301/categories/22077/su...> then it does seem to enumerate the parts but does not provide any details other than an internal part number and conceptually the assembly order of them
I believe that's what the flag option under the submission is designed to communicate, but it's just one vote in a sea of them, and given the current 141 upvotes it's been my experience that unless you're a high roller it's not going to do anything
Also, pedantically, this site isn't designed to be Hacker Truths, otherwise the submissions would be way different; it's just news, and it is newsworthy that such a repo was created and populated, even if it's not (currently?) what the tweet claims it is
I wonder what the motivation for this is. I can’t imagine this would be a priority for a company currently going through hell with their new model. Are the dropping some servicing support?
The more Elon just blatantly lies the more I feel like he’s a grifter. I used to have a ton of respect for him after Lee grew Tesla and SpaceX. I still kind of do but since the Twitter acquisition I’m finding myself asking what is this guy’s problem more often than not
What you're proposing is how patents are supposed to work. The original proposition is that you get time-limited legal protection in exchange for putting your invention in the commons. Of course this has been subverted e.g. by trash patents esp. in the software area where the description just says the equivalent of "magic happens here" on step 17.
It doesn't require trash patents. It just require engineers turned entrepreneurs wasting their time in court trying to enforce claim or testifying on behalf of their investors to enforce claim. Time spent in court is time not spent on engineering.
Even if these entrepreneurs were successful, now you have a 20 years monopoly. That's not without cost to the economy at large.
That's even assuming patents are necessary to disclose secrets. We have companies that specialize in benchmarking and reverse engineering products. Then why are the hell we need patents as soon as the product is available on the market and someone can duplicate it?
If you have trade secrets to protect, why the hell would you leak it and hope that the arduous process of enforcing your patent claims would pay itself back? Lawyers are expensive.
To patent something, you must disclose it publicly.
Trade secrets are kept secret, but as a result can't be patented. At least not directly.
IP can be copyrighted, but it doesn't have to be patented. Copyright lasts much longer than patent protection.
There's no practical way of determining which trade secrets "go along with" patents because that could describe literally everything the company does. It wouldn't make sense to force companies to reveal all of their secrets or surrender their copyrights as soon as their first patent expires.
Yeah, but the detail is that pragmatically in pur current system the patents, as they are written, do not generally provide enough information or detail to replicate many complex systems. You could have a patent and still be a decade away from making something that works, if you have the design data, the cad files, the assembly instructions, the travelers and manufacturing documentation, that all makes is considerably shorter. That's the gap the original commenter and I are highlighting.
I see 3 circuit boards and a GitHub repo that has an .ISO of the diagnostic software. They include DBC files that define the CAN interface, but that's it.
There's basically nothing here? Or am I missing something? No firmware, no CAD, no design files, no engineering data. It's like they released the service manuals and a couple PCB release files they scrounged up from someone's e-mail, but there's very little in here beyond what you'd get from service manuals or from having a PCB in your hand and following traces around. The PCBs are extremely basic.
I don't get it. Was this just a trick to get headlines from journalists who don't know how to interpret the files on the website? Or was there supposed to be more and it's just not up on the website yet?