Follow the reluctant adventures in the life of a Welsh astrophysicist sent around the world for some reason, wherein I photograph potatoes and destroy galaxies in the name of science. And don't forget about my website,

Saturday, 21 June 2014

Learning To Love The Bomb

One of the first spaceships I ever modelled was the Discovery from 2001 : A Space Odyssey. This was back in 2002 when I first started learning Blender, for one of those absurdly ambitious, doomed-to-failure projects that all 3D hobbyists go through. I wanted to render all of the space scenes in the novel which weren't included in the film. I failed miserably, of course, but I did learn how to model spaceships. Not very well, you understand, but it was a beginning.

More ancient historical images can be seen here.
OK, it's a crappy render, but it's still recognizable as the classic design featured in the film - a long boom connects the obvious habitation module and compact drive section, with some standard-looking rocket motors at the back. The boom turns out to be to keep the crew well-separated from the nuclear drive unit. The whole design is very practical - everything serves a purpose, but it's also elegant (well, the real movie version is, at any rate). The only major concession to form over function is the lack of waste heat radiators (we'll get to them a bit later).

But there was another, quite different design that was originally considered. This one wouldn't have used rockets. It would have used bombs. This would have been the first mass-popularisation of the Orion* drive, where nuclear bombs are used to blast a ship forwards. The whole thing is made survivable for the crew by a pusher plate connected to the main ship by huge pistons, reducing the acceleration to survivable (actually quite modest) levels of a few g.

* Absolutely nothing to do with the modern NASA vehicle, this was a conceptual study from the 1950's.

Image credit : me.
I've written about and animated Orion before quite extensively, of course (have a look at the first link if you're not familiar with it). Briefly, it was a 1950's plan to sidestep useless chemical rockets altogether and use the massively greater power of nuclear explosions (which were quite in vogue at the time) to launch truly stupendous payloads (thousands of tonnes) into space. It would also be pretty good for nipping around the Solar System - journeys could be shortened from months to weeks. So, naturally, an Orion drive was considered as an early concept for the Jupiter-bound Discovery in the 1968 masterpiece 2001 : A Space Odyssey.

Image courtesy Winchell Chung.
Looking at the design, though, it's clear why it was rejected for the film. It looks more like something constructed in Kerbal Space Program than anything that Kubrick, with his notorious use of visual clues to manipulate the audience, would ever have allowed. You can't have this sort of ship convey any kind of subtle message, or any message at all except perhaps, "YEEEEEE-HAAAAAAAH !".

No ! That's another Kubrick film entirely !
Only the spherical habitation module is recognizable from the final movie version, and it's rather crudely bolted on to the main ship by what is essentially a big stick (it was only concept art, after all). There also appear to be two, grossly asymmetrical cooling fins, which don't help matters. The rear of the ship looks far more similar to something you'd find in a James Cameron movie - a plausible, practical design that would look just fine on its own, but which is completely inconsistent with the front end.

We can deduce a few things from this drawing. Its overall ungainly appearance suggests an orbital construction instead of a ground launch. Given the size of the habitat module it appears to be about the same size as the movie version, 10-20m diameter. That's pretty small for an Orion though it would be on the long side - about the same as the movie version at 140m. The big cylindrical tanks are presumably for storing fuel (i.e. bombs), with their bulk suggesting this is a ship built for speed.

That would seriously impact the movies's depiction of a remote, isolated crew, and there'd be absolutely no need for most of them to hibernate. I also imagine that convincingly animating the movement of the pusher plate and pistons would have been extremely difficult with 1968 special effects. For all these reasons, I reluctantly conclude that it was rejected for the best. But Orion and 2001 are two of my favourite things, so I can't not model this. I would lose the same amount of self-respect as if I suddenly decided to become a naturist, join the BNP and work for a bank*.

* Though I'm not quite sure all of those are morally equivalent.

I like to think this is an improvement on my 2002 version.
I deliberately set out with the aim not to re-create the reference picture in exact detail, but to make something with all its major features in a way that Kubrick wouldn't have spat on in disgust. 2001 paid exceptional attention to detail in terms of... well, everything really, but perhaps most unusually in terms of making the spaceships realistic. Not merely believable, but almost to the point of being NASA-worthy design studies.

Nonetheless, Kubrick was not adverse to breaking realism's house and burning its legs down (or possibly the other way around) if it helped make a better movie. I've already mentioned the lack of cooling fins in the movie-version Discovery; another noticeable example is the massive size discrepancy of Space Station V in different shots (which frankly is almost as bad as in certain Godzilla movies). And rightly so : there's absolutely no point in making a realistic spaceship for a feature film if it leaves the audience cold.

So, the main change was to make the propellant magazines quite a lot smaller, no wider than the habitat module. That in turn also makes the pusher plate smaller, giving the ship a much sleeker profile and looking less like it's butt-heavy. It's still an ungainly thing, but hopefully looks more like a respectable spaceship and less like a a bunch of flying grain silos on pogo sticks.

Other changes are subtler. I don't like the crude join of the habitat and drive sections in the sketch, so I used something based on the movie version of the ship. Potentially this could also be another piston, reducing the shock that the crew experience still further.

I considered ditching the giant cooling fins and replacing them with some much smaller, recessed panels. These might actually be more accurate. Most nuclear engines produce huge amounts (gigawatts) of waste heat from their reactors, which they have to get rid of to stop the ship from melting. But not Orion - its waste heat goes straight out into space*. The ship will probably still need a nuclear reactor for powering other systems, but only a few megawatts, and even that might be a stretch. The only major power requirements would be for a powerful radar transmitter (via the AE-35 unit) or possibly a high-tech magnetic shield to deflect the solar wind. 

* Some people criticize Orion for being inefficient, since not all of the material from the bomb impacts the pusher. This is like saying solar power is inefficient since the Earth only receives a minute fraction of the Sun's energy - true, but totally irrelevant.

In the end I liked the big cooling fins too much. So they stayed, but now they're symmetrical and rather heavily braced to deal with high accelerations. They're still pretty substantial, probably overkill for a ship like this, but meh. Perhaps it uses its radar to vaporise passing asteroids instead of studying them. Maybe its RCS thrusters use nuclear power so that it can... umm... turn around... really, really quickly. Yeah.

The main thrusters are big, obvious rocket nozzles, but there are also other smaller exhausts scattered about the whole ship. So it can orient itself however is needed, albeit in a slow and stately manner. The exhaust from the main nozzles doesn't quite intersect the cooling fins, so they're safe, but it will scour the hull of the magazines. That could be unsightly, so I added some protective teardrop-shaped panels, based on Apollo design drawings.

I made a small modification to the mounts of the pistons to the pusher plate. Here I used hexagonal prisms in imitation of the rockets on the movie Discovery. I wanted to pay as much homage to the movie as possible, while still largely following the concept art.

Orion aficionados will have spotted that this version does not have a protective plasma deflection sheath around the pistons, as the original design did. The purpose of the sheath was to divert any plasma that made it through the hole in the pusher plate around the shock-absorbing pistons. A trapdoor was supposed to open and close to let each bomb through - it's never been clear to me whether this trapdoor was supposed to be on the pusher plate itself, or on the end of the ejection tube. Anyway this version doesn't have a sheath, so presumably it relies on a trapdoor on the pusher plate.

Finally, the AE-35 communications antenna. Lots of people have pointed out that the pulse unit magazines would block its field of view if it points backward. And they would, but this is a non-issue. I'm assuming that the antenna is fully steerable (as is the ship), so it can still point up or sideways. Given the movement of the planets along their orbits, there's no guarantee that Earth would always be directly behind the magazines... but more importantly, spaceships don't have to point in the direction they're travelling. And wherever the unit is placed, the ship will block part of the antenna's field of view.  

Finally finally, as you'll already have noticed, I also added the ships' name, American flag and NASA logo, none of which are visible in either the drawing or the final movie version. There's no doubt they would be if the ship were ever built though. Since another concept ship recently drew some astonishingly hostile criticism by large parts of the internet, I'd better state it in no uncertain terms : NASA isn't planning to build a ten thousand tonne spaceship propelled by enough nukes to devastate a small country. Just in case you were wondering.

On then, to the artwork. A nice thing - well, one nice thing - about 2001 is that there are different accepted versions of the story. In the movie, the ship goes to Jupiter, discovering an alien monolith in orbit.

In the novel, the monolith is on Japetus (or Iapetus, whatever), which turns out to be a frickin' awesome moon of Saturn. Saturn was rejected for the film because it was difficult to render in the 1960's, but that, of course, is no longer an issue.

The moon here is Enceladus, a proposed target for the original Orion project. Fewer images exist of Iapetus that were suitable for making a nice composition.
The story never describes Discovery in Earth orbit, but of course it was bound to have been there at some point.

Animation ? Duh. Winchell Chung says it best :

"Wild horses couldn't prevent Rhys from animating an Orion drive Discovery from 2001. It is just too much fun for an ordinary person, but infinitely too much fun for an Orion fanatic like Rhys."

Quite right. The main task was to get the firing sequence right. In my first video I deliberately rendered the explosions in an unrealistic way because I wanted to clearly illustrate a basic principle of the ships' operation : the shaped nuclear charges. Most of the mass of the bomb can be directed towards the ship, which is a good deal more efficient than if the debris expands in a sphere. Shaped charges can be extremely good at collimating the plasma into a respectably narrow cone of 20 degrees or less (the exact figure is still classified).

The main problem with this early rendition is that the speed of the plasma is much too low - absurdly low. So in subsequent videos I've tried to make things much more realistic. According to George Dyson's "Project Orion" book, the whole nuclear event - from detonation to impacting the pusher plate and re-expansion - takes about 1 millisecond : less than 1 frame in standard 25 fps video. Rendering only a single frame is not really a viable option (I want the audience to see the explosion, not get a subliminal impression of it), but I restricted myself to 5 frames for the whole event. That's still fast enough that it appears instantaneous to the eye (at least at 25fps, probably not in the GIF though).

Unlike my previous efforts, this version features both the spherical detonation and the plasma jet from the shaped charge.
But is this really what it would look like ? Perhaps not. Project scientist Freeman Dyson says (in his son's book) : 
"The debris goes out from the bomb essentially invisible. You don't see anything until the stuff is stopped... that won't produce anything very spectacular in the way of a flash until it hits the ship. Then all its energy is converted into heat and so you get about a millisecond or so of intense white flash. And very little else."

My naive understanding says that anything which is hot will radiate. And the atomic fire is very hot indeed (10,000 K in the jet, over 100,000 K when it hits the plate) so presumably something would be seen in visible light, however faintly. But I am not a plasma physcisit, so if you have more information, let me know. As for the exact shape, colour and density of the plasma, that is of course a complete guess. I chose to make it look like a gun-muzzle flash because I thought it would look cool.

The firing rate here is about 1 bomb per second, which is about the typical design spec for a ground-launched Orion. It doesn't need to be so fast for obrital manoeuvres because it's not trying to avoid crashing, but it's an aesthetically nice rate to depict. Interestingly, early on in the real-life Orion project a much higher firing rate was considered : 4 (smaller) bombs per second. Since the detonation point is something like 50-100m behind the ship, you need a powerful gas gun to shoot the 1-tonne bombs out the back. But you can't reload a gun like that four times per second. There are two solutions, both of which ought to worry even the most steely-eyed of astronauts.

Option 1 was not to shoot the bombs through the pusher plate at all. Instead, the bombs would be ejected through the side of the hull, guided along rails, and then either shot by a catapult system or propelled by rockets to their detonation point. Presumably the rockets would also adjust the orientation of the bombs to ensure they were pointed toward the pusher plate - in some designs this meant the bomb would do a full 180-degree flip. Well, what could possibly go wrong ?

Option 1 being rejected on grounds of sanity, option 2 was scarcely less dramatic. From Dyson's book again :

"... a gun a metre in diameter... 10 metres long, weighing 2.5 tonnes to project a 1.5 tonne projectile at 200g's. Obviously this can't be reloaded every quarter of a second so you need maybe 10 of them... this will probably wind up as a battery of Gatling gun-type gadgets."

That's right - they wanted a Gatling cannon that fired nukes, because science. Whether you'd need such a gobsmackingly terrifying contraption for the more leisurely rate of one bomb per second, I don't know.

Before I shut the hell up and let you watch the video, Orion always begs the question : "would it have worked ?" The answer, I think, is probably yes - with a catch. Experiments like Operation Plumbob do seem to indicate that a pusher plate (and therefore the ship) could survive the explosions, but there are plenty of unanswered questions. 

For instance, could a system be engineered to reliably eject one-tonne nukes at 200 mph, with an absolute guarantee that they wouldn't detonate too close to the ship ? What would happen if the ship veered off course and had to be destroyed ? What about if a bomb did detonate early - would it risk the others detonating too ? Even if everything worked perfectly, would the fallout from the bombs be anything to worry about ? Well, the answer to that last one is no, but - and this is the catch of course - it doesn't matter.

Orion is a scheme so monumentally audacious that barring the threat of an asteroid stike, it isn't ever going to fly, assuming it would work at all. The total explosive yield for a 10,000 tonne ship for a Mars-bound mission is something like half a megaton of TNT - enough to destroy Hiroshima thirty times over. Does anyone really believe, deep down, that it would be perfectly safe to launch that kind of devastating firepower, or that doing so wouldn't cause massive public outrage, however misguided that outrage might be ?

Of course not. Perhaps Orion could open the road to the stars, but it is, and will only ever be, a dream. Let it go, people, let it go. 

But it still makes for good science fiction. On that note, I suggest you set the volume to "deafen", watch the video and forget about crappy old reality for a couple of minutes.

... if you're one of those people who are chronically unable to forget about reality, you might be worried about the EMP. Well, don't. Unless specifically designed to cause it, a nuke at these orbital altitudes wouldn't cause enough of an EMP on the ground to do any damage unless it had a yield of about a megaton or more. These blasts are much smaller (few kilotons) and would cause problems only for nearby satellites. Or so I'm told.


  1. Before I watched the video, I'd thought the soundtrack might be "The Blue Danube:"

    Dum dum daa da da DA--


  2. Rhysy:

    Good to find you! You may recall me as Makkikomi from the Orion board that was run by that nutcase from Australia. We had some early discussions on the board related to your Orion to Mars project, which turned out great. Anyhow, its good to see you still in "the business" so to speak. I'll always be sorry that board went down. There was a rare combination of really smart folks there who were nice to talk to. I'll always be curious about some of them and where they are at this point. I recall one guy (I think he went by "Doc") who was a retired nuke designer who had his own site extolling the Orion project, but all were interesting characters (which is no surprise as they were all Orion nuts--a limited and elite group of eccentrics). Anyhow, I'm glad to find someone I recall from that period. Keep up the good work!


    1. Makkikomi ! Good to hear from you. Yeah that board did indeed go pretty far south. It was a nice little community until things fell off the cliff.

      If I'm not wrong, the retired nuke designer was called "maddoccanis". He has a facebook page and a, uh, interesting website.

  3. Won't let it go. Love the project and love your work on it. This is the way humanity can have the inner solar system. When enough of a reflexively irrational fear of anything nuclear can be behind, maybe... at least for those in power.

    The first one, I think you just gotta take the plunge and launch it in one piece. Anything else is too laborious and poorly cost effective. The first guy could get to the moon, and if not act as one, help construct a decently sized space station at L1 to gain the expertise needed to learn to assemble really big things in microgravity. Then perhaps the second and following ships could be built to move on to Mars, Ceres, Saturn, etc.

    Unless we're willing to take a couple hundred years of careful husbanding of chemical shaving ounces NASA style like we're used to.

  4. Oh yeah it would seem like the AE-35 would just be installed on one of the booms holding on the radiator fins so it could clear the propellant magazine. Rolling in one axis seems preferable to multiple necessary to get the antenna towards the broadcast destination (which would probably almost always be directly on Earth for this mission), and launched on a parabolic course to keep it that way, at least while Hal is driving, but brings to mind that at least one more antenna would be desirable as a backup on the other side, perhaps. Especially on the first mission, near constant telemetry would seem almost a prerequisite.

    Perhaps multiple folding booms for phased array antennas on Discovery II.

    Are you going to do the Leonov next? ha ha

  5. The only problem is, it looks rather stupid.

    1. Thank you ever so much for that thoughtful and insightful comment.

  6. You've probably already read Vernor Vinge's _Marooned In Realtime_. But if you haven't - you should check it out. Granted, using a 'stasis field' is cheating, kinda.