You Wouldn’t Download A Nuclear Reactor, But Could You?

By pretty much any metric you care to use, the last couple of decades has been very good for the open source movement. There was plenty of pushback in the early days, back when the only people passionate about the idea were the Graybeards in the IT department. But as time went on, more and more developers and eventually companies saw the benefit of sharing what they were working on. Today, open source is effectively the law of the land in many fields, and you don’t have to look far to find the community openly denouncing groups who are keeping their source under lock and key.

The open source submarine that won 2017 Hackaday Prize.

In the last few years, we’ve even seen the idea gain traction in the hardware field. While it’s not nearly as prevalent as opening up the software side of things, today it’s not uncommon to see hardware schematics and PCB design files included in project documentation. So not only can you download an open source operating system, web browser, and office suite, but you can also pull down all the information you need to build everything from a handheld game system to an autonomous submarine.

With so many projects pulling back the curtain, it’s not unreasonable to wonder where the limits are. There’s understandably some concerns about the emerging field of biohacking, and anyone with a decent 3D printer can download the files necessary to produce a rudimentary firearm. Now that the open source genie is out of the bottle, it seems there’s precious little that you can’t download from your favorite repository.

Scratching an exceptionally surprising entry off that list is Transatomic, who late last year uploaded the design for their TAP-520 nuclear reactor to GitHub. That’s right, now anyone with git, some uranium, and a few billion dollars of seed money can have their very own Molten Salt Reactor (MSR). Well, that was the idea at least.

So six months after Transatomic dumped a little under 100 MB worth of reactor documentation on GitHub, is the world any closer to forkable nuclear power? Let’s find out.

An Imperfect Design

Founded by MIT alumni Dr. Leslie Dewan and Mark Massie in 2011, Transatomic originally intended to manufacture the reactors themselves. Their designs promised to be safer and more energy efficient than contemporary methods of nuclear fission by utilizing a concept from the 1960s known as the Waste Annihilating Molten Salt Reactor (WAMSR).

Dr. Leslie Dewan

The idea is that such a reactor could extract additional energy out of the “spent” nuclear fuel from other reactors. Additionally, by using molten salt rather than water as the reactor coolant, it eliminates the possibility of a Chernobyl-style steam explosion should the cooling pumps fail. Safer, cheaper, and more efficient, a perfected WAMSR could be a revolution in nuclear power generation.

Unfortunately for Transatomic, their design didn’t work. At least, not entirely. In 2016, the company realized they made mistakes in their original calculations which meant that the reactor couldn’t actually run on spent nuclear fuel as originally envisioned. Without the promised ability to consume what would otherwise be nuclear waste, interest in the design waned, and by 2018 Transatomic decided to suspend operations.

While the reactors wouldn’t be able to gobble up nuclear waste like they had hoped, Dr. Dewan still thought their core MSR design was worthy of consideration for other companies which might have the capital to actually build it. In a post on the company’s website, she announced that the company would begin looking into the logistics of releasing their designs to the public:

We’re extremely proud of the great scientific and engineering work we’ve done over the past seven years, and want to make sure that it can continue to further the development of the next generation of nuclear reactors. We will therefore be open-sourcing our intellectual property, making it available for any researchers – private, public, or non-profit – who want to continue the work we’ve started.

Expectation vs Reality

As promised, Transatomic did release a trove of documents to GitHub a few months after the announcement that they were shutting down their reactor development program. But it’s important to remember that the design was nowhere near the point where it could actually be constructed; it was still very much in the planning stages when they realized it couldn’t meet their goals.

So what the company eventually released is quite a bit different than what the average Hackaday reader might be expecting when they hear about a nuclear reactor design going open source. If you were hoping for circuit diagrams or some source code, you’ll want to temper your expectations.

The closest thing you’ll get in the repository is a series of PDF “blueprints” that in many cases appear to be only a step above block diagrams. There are a couple illustrative cutaways in the various white papers which make up the information dump, and the details on the theoretical design are fascinating enough if you’re into that sort of thing. But don’t think you’ll be coming away from the experience any closer to building a functioning nuclear reactor than you are currently.

Arguably, one of the most interesting aspects of the release is the GitHub community’s response. With more than 570 stars and nearly 60 forks at the time of this writing, the aptly named “reactor” repository has certainly gained some attention. Unfortunately, none of those forks actually have any activity, not that we think anyone at Transatomic was holding out hope that one of them would generate a Pull Request to fix their fuel consumption issues.

The repository is also subject to the same nagging logistical issues that we’ve seen time and time again on GitHub. Someone had to remind Transatomic to add a license file, and there’s currently an open Issue that asks if editable CAD files can be added in addition to the PDFs. It’s somehow comforting to know that even nuclear physicists have trouble putting together a decent README.

Half-Life Goes On

The last commit to the “reactor” repository was in December of 2018, right around the time Transatomic’s Twitter account went dark. At this point, it’s probably safe to assume the world has gotten all the information it’s going to get out of the company. As promised they released the data they’d accumulated after seven years of research under an open source license, but from an outsider’s perspective, the biggest surprise is perhaps how little material that actually amounted to.

Realistically, there’s very little that even the most driven hacker could glean from the Transatomic files. It’s the same thing that keeps the 3D models NASA has released of their Mars rovers from helping you jump-start your own mission to the Red Planet; the prerequisites required to capitalize on the technology are simply too high. While it’s likely that some nuclear energy whitepaper of the future will reference these documents as part of their research, nothing contained within them will help the next “Nuclear Boy Scout” on their mission. Which is probably a good thing, given how such endeavours have gone in the past.

In the end, one has to wonder if things would have gone differently had Transatomic operated with this level of transparency from the beginning. Rather than wasting years working on a fundamentally flawed concept, opening up their designs for peer review from the start could have allowed them to address the issue before it was too late. Unfortunately for Transatomic, and perhaps the world, that’s a question we’ll never know the answer to.

43 thoughts on “You Wouldn’t Download A Nuclear Reactor, But Could You?

    1. Considering the dozens of articles they have published on 3D printed guns, I think they’ve got a handle on it.

      Seems more likely your understanding of the situation is the one that’s flawed. Especially if you think the reports in the news are in any way representative of what a desktop 3D printer is capable of.

      1. The hoopla about printable firearms died down when they realized that the plastic guns don’t work/aren’t better than plumbing store zip guns, and that even the metal printed ones suffer from a number of flaws and practical issues that haven’t been solved.

        Such as, you’ve painstakingly made a gun. So what? Will you also print ammunition?

        1. Thinking that the poor choices made by CW still define what’s currently out there is marvelously Luddite. The hybrid designs are quite nice. As for printed ammo? That was done *years* ago. The “pure” designs were less than impressive, but functioned. That rabbit hole is one I’ve not followed, but given that nothing stands still….

          1. > As for printed ammo? That was done *years* ago.

            In what sense? Do you stuff cordite down the extruder of your desktop 3D printer? How exactly do you print a primer and a brass casing?

            Of course it’s possible to print a bullet, but unless you make a working cartridge for it, your gun is practically useless, or about as useful as a flint lock pistol that has to be loaded with the bullet, the powder, and then the ignition mechanism separately.

          2. >The hybrid designs are quite nice.

            But they’re also missing the point of what the moral panic over 3D printed guns was about.

            No. You can’t “download a gun”. You can download (and print) some parts of a gun. All the actual functional parts that are difficult to obtain, difficult to manufacture, and/or illegal are still just as difficult and/or illegal as ever.

            For example, 3D printing a gun barrel, with rifling, for anything else than low pressure pistol rounds is going to fail, because it has to survive a proof test and SLM/SLS printed parts contain so many little defects that they just shatter under the stress. Real barrels are work-hardened and heat treated to achieve the correct properties.

            E.g. .45 ACP: 140 MPa test pressure. 9 mm NATO: 240 MPa. Ordinary steel has a stress fatigue limit around 130 MPa assuming the material has no defects or irregularities, which come from being sintered or melted together from powder. An analogous situation is with welding, where the stress fatigue limit can drop to as low as 30 – 80 MPa thanks to the changes in the grain/crystal structure and all the gaps and tiny gas pockets, inclusions etc.

            This means your 3D printed 1911 has a good chance of shattering on you even with the mildest loads – and this is why the company that made the news a couple years back by printing that very gun declined to put it through an actual proof load test. They opted instead to fire many regular rounds, which doesn’t prove anything. If there is a defect, it can go on the first shot, or it can go after 10,000 – the point of the proof load is to break it on the first if it is going to break at all.

          3. “How exactly do you print a primer and a brass casing?”

            I think it would be fairly easy to come up with an electrically fired, muzzle-loading pepperbox:


            with the barrels being store-bought pipe, just as used in “zip guns,” just keep the chamber pressure low enough. No need to rotate the barrels as with the old school firearm, just rotate the firing voltage. Heck, it could even have a mode to rapid-fire like a fully automatic weapon. The barrel supports and grip would be 3D printed with the electronics and battery charging port in the grip. With no rifling, it would not be a really long range shooter, but the DIY cast slugs could be aerodynamically designed to remain stable just like the slugs used with smooth bore shotguns.

          4. > an electrically fired, muzzle-loading pepperbox

            The question is, “Should we be worried about 3D printed guns?”, and the answer seems to be: “Someone might make a early 19th century style firearm with no accuracy or power, a great chance of detonating in your hands, which takes half an hour of fiddling to reload, and requires 9 Volt batteries to fire.”. Somehow I’m not convinced.

            Electric firing is also difficult to get right. If the charge doesn’t ignite properly at once, the powder grains may break down and turn partially to gas instead of burning evenly, and you may have a detonation in the barrel.

            This is because the pressure rise in the barrel causes the broken down and partially gasified bits of gunpowder to all ignite with a delay – kinda like knocking in an engine – which causes a massive pressure spike. This is the reason why black powder guns were very limited in chamber pressure – you had to make sure that the gunpowder will not detonate when you fire it, and with low pressure you only get low muzzle velocities (or ridiculously long barrels) so you have to compensate with a big lead ball for a bullet.

        2. I thinke you should widen your horizon a little. Somebody wanting to make a gun, is probably more than just a computer nerd whose only understanding of constructive work is limited to a 3 D printer. Though that would help a lot. There are millions of metalworkers with various degrees of skill out there. Semi literate tribesmen in Pakistan manufacture functioning guns. The skills are aquired over a few years, but the instructions are “downloadable”. High grade steels are available from scrapped industrial – car parts. So If I wanted I could knock out a lethal gun in my home workshop within a day, and a pretty good one having the resources of a maintenance workshop shop at work available within a week. Primers a bit tricky but brass ammo can be turned and if I make the the effort, I could even draw and form the cases. And who says it has to be a solo effort anyway.

    1. Who does Tom think he is talking to normal people?
      In my opinion the kind people who read Hackaday are more likely to be the kind who would download a reactor design just for curiosity.

      1. Perhaps it isn’t as common place as Tom assumed, but the “You wouldn’t download a __” part is based on the RIAA fud campaign argument “You wouldn’t download a car, would you?” while trying to convince the public copying mp3’s was immoral.
        That had the unintended effect of cementing the difference between a physical item, where taking it means depriving the current owner of it, vs information, where taking it means the current holder still keeps it.

        It’s a phrase pretty much at legendary meme status now.
        He is in effect saying with much sarcasm that of course most people would download one if such a thing was possible – which is of course the twist, as it apparently is, when before I suspect most people would assume (me too) that it wasn’t, just like downloading a physical car

        1. Funny enough they were sued buy the composer of the music for using it illegally as the rights were issued for a different ad entirely. Although they made it right and paid the composer the royalties, it’s still damn funny.

  1. I thought the MSR was to use thorium in it’s reactor. There is a ton of information on that type of reactor. The US actual did test on that, and it survived an actual msubseltdown. It experiment was shut down because the US GOV wanted water type reactors for their submarines.

  2. >anyone with a decent 3D printer can download the files necessary to produce a rudimentary firearm.

    It may surprise even many on this site that AR15/M16 rifle blueprints have been in the public domain for decades, and the 1911 handgun for over 100 years. You could even get books from the library with many plans for homemade gats.

    Though it might not resemble modern open source too closely, gunsmithing has been reasonably “open source” for as long as guns have existed.

    1. All one needs is some basic machine tools, and actually it can be done with hand tools. Also contrary to what the press would have you believe, it is legal to manufacture a firearm. (as long as you are legally allowed to posses one.)

  3. I doubt that what has been released is everything that was created, but much of the information on nuclear reactors will probably be restricted under ITAR regulations and can probably not be released to the general public.

  4. What’s so special about a nuclear reactor? I know of a guy who has reproduced Fermi’s Via Panisperna work with neutron amplification, by grabbing a block of natural Uranium, which causes the Hydrogen in his hand to moderate the ambient neutrons, such that they produce a very slightly enhanced fission in the natural Uranium block. No, it’s not recommended, but it works and is measurable.

    As for natural Uranium, it will undergo fission. Look up the CANDU reactor design, which is another approach to a nuclear reactor, which uses natural Uranium (U-238), with Heavy Water for a moderator. It’s claimed to be much safer against meltdowns, since, if the water leaks out of the reactor, there’s no longer a moderator for the neutrons, so the reaction shuts down.

    Plus, the trick with designing a nuclear reactor is the safety features, not so much the raw design. After all, the Chicago Pile-1 was, essentially, a stack of ultra-pure Graphite blocks, with natural Uranium, both in the form of Uranium Oxide, and Uranium metal. For that matter, consider Oak Ridge’s X-10 Graphite Reactor, which operated safely for 20 years (but, which was disturbingly way too much like the design of Chernobyl!).

    As for the safety features, consider the disastrous run of Westinghouse Electric Company, and their screw-ups in nuclear reactor design, which plunged them into bankruptcy.

  5. With ready availability of 3D printers with metal output, the open sourcing of reactor designs and blueprints can be quite risky and may prove to be disasterous if accessed by unIslamic Muslims and/or anti-Islamic non-Muslims bent on building a radiological device on a shoe-string !

    Are the cash-greedy MICs hatching these potential WMD open sourcings in the hope of kick-starting drooping non-Nuke Club nuke demands ?

  6. Sure…anyone who wants to mimic the first meltdown, get impaled and hang from the ceiling. And better yet…use the irradiated ambulance at fairs for years after for the kiddies. (Idaho falls 1961). “Share the pain” (Fukushima 2011).

  7. Real plans for fission weapons have been on the net decades. It’s fissionable material that will get you shot before questioned by the CIA, Secret Service, MI6, FSB etc..

    Which is interesting considering a sniper rifle or certain types of xray are far more effective as weapons… Even unrefined uranium or equipment to do it or even researching any of itwill get governments hunting you..

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.