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 11 September 2010

Project Orion : How To Nuke A Spaceship Without Killing Anyone

See more here and a spin-off here.

As the creator of the "Project Orion" video that's floating around the internet, I thought it was about time to clear up some misconceptions that have been circulating. And give, hopefully, a basic introduction to Orion - though there are many more detailed articles easily found through Google. But mine has pictures ! And some shiny new video clips  For those who have wandered in from cyberspace, the video to which I refer can be viewed here (a specially uploaded higher-quality version than the standard YouTube video) :

(I'm also not going to respond to critiques of the animation itself  because the project is now a scary 7 years old)

The basic idea is to propel a spaceship forwards by detonating nukes ~100 metres behind it. The bombs are shot into position from the ship by a powerful gas cannon (they have to be detonated about once per second, too fast to simply drop them into position). A huge, ~800 tonne plate (usually called the pusher plate), is between the bomb and the ship proper, which are joined together by gigantic pistons. As the pusher moves upward from the impact of the explosion, gas in the cylinders compresses and momentum is transferred relatively slowly from the pusher to the vehicle. The pusher is actually two plates separated by giant gas bags - these absorb the initial massive wallop of acceleration (hundreds of g or more) on the plate, while the pistons slow things down to human-survivable levels. Actually accelerations can be surprisingly modest in the ship, as low as 1-2g.

One particular comment should be specially addressed : "No ofense ment, but that? roket looks rediculis." Ah, well, it isn't a rocket. If anything it's the opposite, having stuff thrown AT it to make it go, not thrown AWAY like a rocket. Incidentally the designs are based very closely on those in Dyson's book - probably too closely. I doubt there's any need for it to be quite so streamlined.

When the pistons compress enough, their internal gas pressure is enough to drive the plate backward. The next bomb prevents the plate from continuing backward off into space. Bombs are either delivered through a hole in the center of the plate, or even around the side (which is just too wacky for me). And make no mistake - the thing is bloomin' massive. Almost as big as Liechtenstein.

Here we come to the first big question - won't a bomb just vapourise or blast apart the ship ? The answer is a resounding "no" to vapourisation, but still a very big "maybe" to blasting it apart. Vapourisation can be dismissed thanks to test such as Lew Allen's balls (insert joke here), in which two graphite-coated spheres survived a close-proximity (of order meters) nuclear explosion. Some of the early nuclear tests involved the bomb at the top of a tall metal tower, which was not vapourised but was blown to smithereens (fragments of the towers being found after the test).

Answering whether the pusher can survive the blast is - in my opinion - the big unknown. The original engineers thought that it could, but never got an actual nuclear test to prove it. The bombs aren't island-smashing H-bombs or even city-killing Hiroshima yield devices, they're much smaller - except when out of the atmosphere. But even here the massive shockwave from compressed air is absent, considerably reducing the impact force (in fact some propellant mass has to be added to the bombs). Still, assuming the explosion doesn't smash the plate to bits, the hot plasma might still severely ablate the pusher.

One of the weirder possibilities for protecting the plate was to have it sprayed with oil between each blast. The bomb's plasma is only in contact with the plate for a microsecond, so it simply doesn't have enough time to impart much heat. The oil is used to absorb that heat, evaporating away before any damage can be done. Again, no-one really knows if this will work. Certainly it can't guard against any failed bombs, which could send horrible fragments at high speed into the plate doing untold damage. Here's a clip from another (stalled) project.

If you doubt that bombs can be stable enough for any kind of useful propulsion, you have but to look at the videos of the "hot rod" the team built to prove just that. I think this is quite wonderful. It's one of those moments where all your assumptions are dispelled and all you can do is look back and say, "Oh." Of course, whether you could be confident that the debris from the nuke wouldn't be so unbalanced as to twist the shock absorbers is another matter (not one that seems often addressed - either it's a non-issue for some reason or no-one wants to know).

At this point I should respond to comments that in my video you can see most of the plasma moving toward the pusher. While you might expect a spherical blast - particularly in space - this is not quite true. A substantial amount of propellant is contained within the bomb in a flat disc facing toward the ship (as I mentioned, in space there's no air so you need some extra mass within the bomb). According to Dyson, this gets reshaped by the explosion into a cylinder which impacts the pusher. I wanted the video to convey this, albeit at the expense photorealism - plasma moving very much faster than in the video. What you'd actually see, apparently, is the blinding flash of the explosion and then almost simultaneously a glow on the pusher as the plasma compresses and heats up - but not much more (at least in space). Here's another clip.

So let's assume then that the idea is sound in principle. Personally, my first three thoughts on the project were : 1) It'll just blow up 2) If it doesn't, it will be wildly unstable 3) Fallout will be horrific. (1) now seems unlikely, but with a big if, (2) I think is also debatable, and then we come to the far more emotionally charged (3). Before addressing this, I want to clear up a few more video-specific points.

Firstly, the rocket boosters (based on the Shuttle's existing SRB's) are NOT used to get the thing into orbit, only to ~100 km altitude (i.e. above the atmosphere, for reasons which will become clear). Lifting things to this height and orbiting them at this height are massively different prospects. To lift something that high you have to get it to a speed of ~1400 m/s (it will then immediately begin to fall back), whereas to orbit you need speeds more like ~8,000 m/s. That's why the X-prize was won with a much smaller budget than NASA's. It's also why I show the SRB's detaching at a high altitude. It coasts upwards for many kilometers until the bombs kick in.

Occasionally, people remark that it would be better to just build the thing in space using conventional rockets. No. That is wasting the benefit of the vast energies available from A-bombs, which give you an incomparable heavy-lift capability. And although space-borne Orions still perform very well indeed, if space is ever really going to matter to people we need to send up more than a few dozen elite astronauts.

(but for God's sake don't think I'm suggesting that Orion is the way to do this)

Another point of confusion is the deployment of structures attached to the ship on long tethers. These are to serve as habitat modules. Rotation of the ship provides ~1g at distance ~300m without causing nausea. These are only intended as representations; I did no calculations to determine how massive/large the habitats could safely be. Here's an earlier concept image I did where the tethers also serve as supports for solar panels. Then I realised that these would be nearly pointless since the ship could have a nuclear reactor onboard.

A few people have said that I "skipped over the landing sequence", which confuses me because I didn't think I did. Still, I certainly don't know if landing the thing is at all feasible. Whether aerobraking would have any useful effect on several thousand tonnes of steel in the thin Martian atmosphere... hmm. To say nothing of the difficulties of landing with a giant flat plate on rocky terrain (but maybe there are more suitable areas of Mars), let alone taking off again.... The ship might not necessarily need to land safely, of course... it could just be transporting resources / heavily protected bulk equipment to a colony, with astronauts landing in a more controlled fashion. So even if it's a 1-way trip, that's not necessarily a show-stopper.

And so we come to fallout. I've no idea how dangerous radiation actually is, but it seems to be intensely controversial. Pro-nuclear and anti-nuclear complainers will bandy about completely different numbers. As a moderately pro-nuclear but cautious observer, it's very difficult for me to judge who's right. So I'm going to largely ignore things like total yield or number of bombs detonated. All I'll say is that in the animation, all of the bombs are detonated outside the atmosphere where they can't irradiate ground dust - for which there seems to be a consensus as the major source of fallout.

However, I WILL get on my hobby-horse about one unfortunate statistic - the idea that each launch would kill about 10 people. This was calculated by the original team, assuming bombs were being detonated in the atmosphere. However, not only is it the worst marketing ever, it's a self-evidently flawed statistic. About 7.5 million people die of cancer worldwide every year. You couldn't even measure a change of 10 ! There'd be absolutely no way you could attribute a single case to an Orion launch - natural variation in the death rate is vastly higher.

Moreover, it should not be shocking that a method of transport causes casualties. Car-crash casualties are about 42,000 per year for the US alone. And don't bother saying that you can avoid this by not getting in a car - nearly 5,000 of those killed are pedestrians, to say nothing of cyclists and motorcyclists. It also seems quite inconceivable to me that the inventors of the motor-car could have been so naive as to envisage their invention would never kill anyone, but that didn't stop the car becoming rather popular. However, if they'd pulled out a PowerPoint presentation demonstrating that the wonderful new invention would, over about a century, kill tens of millions of people, one suspects they would not be held in particularly high regard.

If that still all sounds a bit harsh - no rational person, after all, wants to kill anyone at all - let's put that into perspective and play a little game. View the photos below and order them by number of yearly casualties.

The answer (from highest to lowest) is :
H) Cars
E) Lightning
G) Stairs
I) Lions
F) Dogs

With A,B,C,D all about the same. Cars kill about 1.2 million per year worldwide, lightning (surprisingly) about 10,000. Stairs are death traps, killing over 1,000 people in the UK alone every year. Data for animal attacks is, admittedly, more iffy, but people do get trampled by cows and savaged by dogs every year.

Now, fair enough, relatively few people are at risk from crocodiles or sharks, but pretty much everyone is at risk from dogs, lightning, and stairs. And that's the point, of course, Orion causes a marginal extra risk. Just like putting more cars on the road, it is NOT the same as lining up 10 people for a firing squad. So let's stop worrying about fallout from a successful launch - I'd be far more concerned about what would happen if a bomb detonated while still inside the ship.

Which about rounds up this little tirade. It remains only for me to say that there's no point getting all misty-eyed about Orion, as it remains a madcap, entirely untested and unproven technology. For all we know, the pusher plate might survive but buckle horribly under the stress of the first explosion. In any case, if it had gone ahead in the 1960's... well, launching giant nuclear spaceships at the height of the cold war doesn't spell missed-opportunity-for-utopia to me. Nevertheless, its potential is too high to dismiss out of hand - but I suspect it probably will be.

(Finally, for some reason on YouTube there are a number of comments about wormholes as an alternative. GOOD BLOODY GRIEF ! WAKE UP PEOPLE ! You've all be watching too much Farscape ! Consider yourselves slapped very hard indeed for stupidity)


  1. Interesting article, and probably the most complete idea I've seen on the subject of nuclear propulsion systems. I always wondered about the following in a system such as this:

    1. How efficient is this mechanism of blowing up bombs behind a spacecraft? Could the charge be shaped such that it only directs the force in one direction (especially in space)?

    2. By slowing down the explosion with the use of pistons, is that reducing the efficacy of the propellant? (I have no knowledge of propulsion systems, so I genuinely am curious)

    3. Since the bomb is dropped through the hole, wouldn't its tendency (in space) be just to keep going away from the ship, thus minimizing the impact of the propellant?

    4. The whole through which the bombs drop seems to be somewhat of a weak point of the pusher plate. Is there risk of bomb fragments or radiation flying through the hole and damaging the ship or the occupants?

    1. 1) How efficient ? Not very, but that's actually not a problem, because the sheer amount of available energy is huge.
      Yes, shaping the charges to project most of the mass is an important part of the design (in the video, I tried to show this). A flattened disc of material at the top of the bomb will expand into a narrow cone (though don't ask me how it does this).

      2) No, the pistons do not affect the overall performance. The amount of momentum absorbed is still the same - the pistons simply spread the transfer of this momentum over a longer time to the main section. So, instead of the momentum being absorbed in a tiny fraction of a second, it can be spread out over about a second or two, which massively reduces the acceleration - but NOT the final speed attained.

      3) Not sure I understand the question correctly. Yes, the bomb will tend to just keep going when it's dropped through the hole, but this won't have any affect on the propellant - the detonation distance can be controlled.

      4) Yes, the hole is a weak point. Alternative designs featured side-mounted launching systems, but this requires the bombs to very precisely orient themselves at the moment of detonation. Which seems like far too much to go wrong, to me.
      From what I remember reading, if the explosives if the bomb (for whatever reason) failed to cause a significant amount of fission, the impact of the fragments on the ship could be very serious. The pusher plate can withstand (it's thought) repeated brief bursts of hot plasma, but surviving impacts of high-velocity projectiles is another matter (a few inches thick isn't all that substantial - think of how warships can be sunk by conventional torpedoes).
      The ship includes a "plasma deflection cone" (see schematic) to prevent any plasma which enters through the hole from doing much damage. Radiation isn't much of a concern because the hull can be thick enough to provide shielding (also, in space at least, the duration of contact with any radioactive material will be extremely short).

      Thanks for reading !

  2. It's interesting to see technology evolve. I remember looking at 19th century 'flying machines.' Technically they had what they needed in the late 19th century(to fly), but the theory and combination of tech necessary to fly was started by the wright bros. The ridiculous machines we see in old video and pictures attempting flight is pretty much what we see now attempting space travel. Although its barely conceivable now how we can possibly breach the depths of space with human passengers, the theory and tech haven't arrived- so we get these great ideas of dropping nukes behind us for a push! But it's ideas like the Orion we need- big, scary, and slightly obsurd that'll get us where we need to go.

  3. I read Dyson's original treatise, and his (well-written) later book, co-written with his son. In it, he comes across as a very moral man.

    He calculated (using a standard method of the time) that slightly more than one death would occur per ground launch of the small Orion. As you point out, many more people die in road accidents - but his personal moral code prevented him from permitting this to happen as a result of his work, and he withdrew. Around this time, the atmospheric test ban treaty was in the works (following the detonation of the USSR's BSD-1 50Mt) and the project was shelved. I'm simplifying here, please be tolerant!

    The ejected nukes were shaped charges, designed to be "clean", with the plasma generated from the vapourisation being directed by the shape of the charge, container, "neck", and propellant plate. Essentially, it generated a tube of plasma with defined characteristics pointing back at the ejection port in the pusher plate. As you note, there would be a risk of asymmetric thrust resulting from a plasma pulse being offset from the central axis of the ship. To counter this, the charges were meant to be spin-stabilised during ejection. Oddly, the Coca-Cola corp was consulted about the design of the storage magazines for the charges and the delivery mechanism to the firing port. A "wobbling" charge would not detonate as this would disable the detonator.

    The ship was never meant to land - once launched it would remain in space: landings were meant to be handled by shuttles. Re-fuelling was meant to be achieved by a "canning plant" in the ship to manufacture more charges from materials gathered at the destination (I forget the details of this). However, the team was postulating 30-day out-and-back jaunts (with their families taken along - larks!) to Jupiter, in a non-stop return journey.

    General Atomic owned the Electric Boat company, who were manufacturing nuclear subs, and postulatied the yards would build large Orions in dry docks next to the subs, using the same workforce and construction methods, at a similar or larger scale!

    Finally, your chemical launch to above Karmann was in one of the original studies but was not extensively developed; however, this does seem like the most realistic way to employ a ground-launched Orion - the charge composition and yield was meant to be controllable by the delivery mechanism changing magazines and delivering charges of varying type to the firing stream chosen to suit the flight profile, so the risk of both radiated and dust-based hazard could be controlled.

    Stephen Baxter's "Flood" and "Ark" SciFi novels involve descriptions of an Orion launch during global catastrophy, and the SyFy miniseries "Ascension" covers what life aboard an Orion generation ship might be like.


Due to a small but consistent influx of spam, comments will now be checked before publishing. Only egregious spam/illegal/racist crap will be disapproved, everything else will be published.