December 24, 2008
Space Station Cargo Delivery Still Expensive

Orbital Sciences Corp has won a $1.9 billion contract to carry 20 metric tons of cargo to the International Space Station in 8 flights. Think about those numbers. That's $95 million per metric ton to move cargo from ground level to low orbit. Those deliveries start in 2011 and run through 2015. A metric ton is 1000 kilograms or 2204.6 lbs. So the cost of putting stuff into low Earth orbit in 2015 is still going to be around $43k per lb or $95k per kg. At these prices large scale human colonization of space still seems a very distant prospect.

Those costs will come down a lot if a beanstalk into space built using nanotubes becomes possible. A bigger cost reduction for a Mars mission will come from nanotech advances. A bunch of nanodevices that can transform Mars landscape and produce needed supplies for a colony would reduce the size of the payload needed for setting up an initial colony.

The idea of moving to Mars remains unattractive to me at least until it becomes possible to massively terraform the planet. Move massive amounts of oxygen and nitrogen from Titan and Triton and suddenly a Mars where we could walk around outside and breath its air becomes a lot more appealing.

Nanotech assemblers will be the enabling technology for large scale colonization because they will lower the cost of creating the spaceships needed for terraforming.

Share |      Randall Parker, 2008 December 24 01:26 PM  Space Exploration

James Bowery said at December 26, 2008 12:24 PM:

OSC is little more the latest (early 1990s) small-sat component of the old boy launch service oligopoly of GD, MD and MM. Their price figures are interesting for historical comparison to the new breed of launch companies coming up through the ranks. The contract to SpaceX may be an attempt to buy off that breed -- and it may be successful in corrupting it given its nascent state -- but NASA isn't going to be able to do that to all of the launch companies unless it becomes a launch service monospony.

PS: Please get off this kick about terraforming. Planetary surfaces aren't the right place.

Randall Parker said at December 26, 2008 1:00 PM:


We can not safely live off the surface until we can achieve quality of both design and manufacture far in excess if six sigmas. How we going to do that?

Engineer-Poet said at December 26, 2008 5:35 PM:

How are we going to do that?  Redundancy and over-engineering, as always.  E.g., it doesn't matter if you need 3 cooling units running if you've got 6 on tap and can do without any of them for a period of hours before matters get critical.

Some of the details of maintaining life support look fairly easy to me at the scale of an Island 2 or 3.  For instance, cooling the interior takes a heat pump.  You have a solar mirror to supply high-pressure steam from a (redundant) set of boilers down manifolds along the axial spine.  The steam runs steam turbines which drive air-cycle chillers; the heat from the compressed air boils water which is exhausted down the low-pressure manifolds to the condensers.  Electric drive is another option, so steam is only used for heat rejection and not power.  If too many of the boilers or condensers fail, you go to twilight conditions to reduce heat load until they're fixed.

Of course, this isn't going to happen at $43,000/lb.  I'm waiting for some smart cookie to find a way to have something like a Space Ship 1 meet up with a rotating tether to attach cargo or swap feedstocks for products; SS1's successors are already a heck of a lot cheaper than $43k/lb.  Products which might be made better in microgravity today include ultra-clear aerogels (no bluish cast) for high-performance windows.

James Bowery said at December 26, 2008 6:34 PM:

At more than 20 fatalities per 100,000 licensed drivers in the US, I think we can live with something less than 6 sigma -- especially if it gives us options like many thousands of times the land area of earth in habitats where communities are of a size and character that don't require cars.

Randall Parker said at December 26, 2008 10:02 PM:


Some failure modes in space habitats kill everyone on-board who can't make it into a spacesuit on time. Ditto for Mars colonies which have enclosures.

I was watching a TV show about an enclosed Mars habitat and one of the engineering problems is the need to control air pressure. If, say, the farm plants emit too much gas (e.g. a bunch of them die for some reason and start breaking down) the pressure increase could burst the dome. A small system is stressed more heavily by what goes wrong.


Redundancy and over-engineering still probably won't get us to 8 or 9 sigma.

Suppose we get nanoassemblers that make manufacturing so cheap and rockets so lightweight that launch becomes affordable I think reliability and design quality will become the biggest problems for L5 habitats and Mars habitats. If we can terraform then our margin for mistakes and errors becomes way larger.

To make Mars suitable for human habitation I would want to use nanoassemblers to build massive space-faring automated ships to go harvest Triton and Titan. Bring the N, O, H from these places to Mars. Build up a big atmosphere. Also, build a massive fleet of satellites to reflect light toward Mars so it gets Earth-level insolation. Then bring in massive amounts of plants and animals to populate the place. Then show up maybe 50 years later once robots have built up lots of infrastructure.

James Bowery said at December 27, 2008 7:24 AM:

Some failure modes in space habitats kill everyone on-board who can't make it into a spacesuit on time. Ditto for Mars colonies which have enclosures.

Ditto for airliners and even oceanliners.

You posit things like nanoassemblers but reject the industrial learning curves that have an established track record of providing acceptable risk to benefit ratios in the recent past.

Engineer-Poet said at December 27, 2008 7:27 AM:

Do you really think that a habitat which has to contend with temperature variations won't either have air reservoirs with pumps or be engineered for pressure variations?

Small habitats can be protected against pressure loss.  If WWII submarines can have independent compartments with pressure doors, so can habitats.  Larger habitats would take hours or days for even meter-size holes to cause dangerous pressure loss, so I question what kind of failures you're talking about.

If we get nanoassemblers we can have them doing continuous maintenance and rebuilding on any habitat we want.

James Bowery said at December 27, 2008 8:39 AM:

Since "He who refuses to do arithmetic is doomed to talk nonsense."

Let's assume nothing but Island One type habitats as posited by O'Neill where, as Engineer Poet states, even meter-sized holes allow _days_ for repair/evacuation. An Island One habitat supports 10,000 people. There were 41,059 traffic fatalities in the US during 2007. This is equivalent to the catastrophic loss of 4 Island One habitats per year. The population of the US is 300,000,000. It would require 300,000,000/10,000 = 30,000 Island One habitats to equal that population. This implies an "acceptable" mean time between catastrophic failure of 30,000/4 years or nearly 10,000 years.

What catastrophic failure modes do you posit and how do you derive your estimates for their mean times?

James Bowery said at December 27, 2008 11:01 AM:

BTW, Engineer-Poet's idea that the suborbital launch companies spawned by the Ansari X-Prize might feed a Mach 12 100km altitude rotovator arm is a good one that occurred to me some time ago.

It may turn out to be even better now that the economic crisis has deprived so many potential "space tourists" of the means to patronize those suborbital launch companies.

Maybe it is time to form a company. I used to attend weekly seminars at the California Space Institute at Scripps in La Jolla with Joe Carrol but I don't know Hans Moravec.

Randall Parker said at December 27, 2008 7:33 PM:


We do no spend that much time in airliners. We can decide to drive. We can decide to stay home. Living in a habitat is 24x7 all weeks of the year for years on end.

Yes, some people die in airplanes. Typically not complete communities though. Also, airplane risks of plummeted.

Mean times between failures for habitats: The thing about statistics is that you need a base of data to use to derive failure rates into the future. We do not have this for the habitats.

For airplanes we had failure rates in the 1930s that were far lower than the rate at which rockets fail today. The progress rate in airplane safety has cut the failure rate by orders of magnitude. Rockets remain very high risk after hundreds of launches. I did some calculations about rocket and airplane safety in a previous post. I dug up crash rates from the late 30s. My impression is that airplanes were quite possibly never as dangerous as rockets are today.

So we need safe rockets to get up into space (or a beanstalk) before we can even think about worrying about habitat failure rates.

Will habitat safety be easier to achieve than rocket safety?

Engineer-Poet said at December 27, 2008 10:00 PM:

We do have historical experience with large static tensile structures.  How many suspension bridges were built up to 1958, and how many of them have failed due to non-dynamic factors?

Randall Parker said at December 27, 2008 10:37 PM:


Those static structures had far fewer functions than a space habitat.

If an airplane crashes we die of deceleration. In a habitat in orbit lots of ways to fail can kill us in lots of ways. We need a temperature range, an atmospheric pressure range, presence of some stuff in the air, absence of a much larger list of gases, and still other needs. Each of those needs is a way the habitat can fail and kill us. Most of those needs require active and complex control systems.

James Bowery said at December 28, 2008 8:40 AM:

The fact that there is only one dominant proximate cause of death in an airplane crash does not tell us the probability structure of ultimate causes of death: pilot error, control system malfunction, improper turbine blade inspection, etc. You need to look at the entire Bayesian structure.

Moreover, active control is not, in itself, a show stopper. It is, in fact, pervasive in life: homeostasis. The only question is whether the control system, including humans actively participating, is, in aggregate, reliable enough to keep the system within bounds over long stretches of time.

The comparison of space habitats with rockets is pretty much nonsense. Rockets operate at the very limits of material and chemical bond strengths.

The comparison of rockets with airplanes would make sense were it not for the fact that there was nothing even approaching the Kelly Act of 1925 in the area of rocketry. If airplanes had remained under the control of Langley and the US Army, the progress in cost and reliability would have been as abysmal as it has been in rocketry up until the recent competition arising from information technology multi-millionaires. Things are changing more rapidly than you think.

Finally, think back to the death rates of the first settlers from the UK in the US. 1 in 4 were dead within a year for some time. It wasn't until centuries later that one could arrive "penniless at Ellis Island" and make a go of it. You underestimate the desperation with which people long to escape the tyranny of the majority restrained by a vague laundry list of selectively enforced "rights".

Engineer-Poet said at December 28, 2008 9:24 PM:

Quoth Randall:

those static structures had far fewer functions than a space habitat.
You specifically mentioned failure modes which could kill the inhabitants before they could don spacesuits.  The only failure modes which allow this are sudden loss of pressure (minutes at most) or complete mechanical failure.  What, other than structural failures, can cause this?

Temperature-control problems would take hours to become critical, and have multiple corrective responses.  Life-support and other failures would cause gradually worsening problems over days to years.  They're not the kind of thing that can be addressed with spacesuits.

Each of those needs is a way the habitat can fail and kill us. Most of those needs require active and complex control systems.
I'm with Bowery here.  We're not working with transmutation or radioactive fission products, so problematic trace gases would have to be chemical compounds derived from biological or industrial activity.  Industries which produce such gases would probably be isolated outside the habitat proper, leaving the biological problems.  Biogenic gases can either be burned to CO2+H2O in catalytic systems or eaten by some other biological agent.  If the control system stops working right, you have a while to recognize and correct the problem before it becomes critical.  Again, not an issue for space suits.

I think that Biosphere II is a useful example.  The system was loaded with way too much organic carbon, which decayed with the production of CO2 and the consumption of oxygen.  This did not happen quickly (no "space suit" issue); it took weeks.  The corrective would have been to convert some rapidly-growing biomass to charcoal and setting it aside, removing carbon from the biological cycle and boosting oxygen levels.  Once this is understood as a problem it would not be difficult to handle.

Randall Parker said at December 28, 2008 9:33 PM:


There are tons of ways to make toxic gases in a hurry. Fire is one way. Two pipes bursting is another way.

Yes, you can assume various chemicals that are capable of being dangerous aren't handled in the habitat where people breathe. So a chemical plant that would be in atmosphere on Earth would instead be outside of atmosphere of a habitat. People would either work in the plant in spacesuits or there'd be tubes with atmospheric pressure and pressure doors that run around the chemical plants.

Some habitat failure modes would happen slowly enough to give you time to get into a space suit. But then what? Some of the failure modes would leave a messed up habitat for an extended period of time. The amount of capital equipment needed per person to support such a habitat would be far greater than the amount of capital equipment to support each human life in the United States down here in atmosphere of a planet. What's the economic justification for living in an orbital habitat?

This greater need for capital equipment (e.g. to seal in atmosphere and to provide assorted back-ups for failure modes) makes habitat life more expensive. Also, getting up there in the first place is still extremely dangerous and the rate of progress in making the rockets safe seems pretty slow to me.

Roderick Reilly said at December 29, 2008 12:17 PM:

Excuse me, but why didn't you mention SpaceX in this same post? SpaceX's prices are much lower than OSC's. I'm not going to get into the mistery of the different price structures for each contractor, but the bright side is that SpaceX is going to be delivering larger payloads for less than OSC.

Also, please spare us the "Space Elevator" nonsense -- I hear that silliness all over the place. Why is it that mature, intelligent, knowledgeable adults immediately connect "carbon nanotubes" with "space elevators?" Try connecting nanotubes to better tennis rackets first, then car bodies and airplane fuselages.

THEN space elevators, yes?

NO. THEN launch vehicle fuselages for SSTO rocket-powered vehicles. Space elevators are a long, l-o-o-o-o-ooong way off.

Remember, between stes "A" and step "Z" are a whole bunch of other letters.

Randall Parker said at December 29, 2008 12:32 PM:

Roderick Reilly,

I saw the SpaceX announcement too. But I was focused on cost per ton and I couldn't find a source for how many tons the SpaceX rocket would lift and so I could not calculate a cost per ton or cost per pound.

If you want to show data on payloads of the two rockets with a URL then please do. Just asserting a difference really does not help readers who do not know you are or whether your claims can be trusted.

SSTO: Yes, I could have gone there as an intermediate step. Not clear why they'll be more efficient though. Rocket 1st stages that fall back to Earth do not have to be carried all the way into orbit.

Could have brought up hypersonic ramjets too as a way to reduce fuel weight. But I have no idea whether they'll become practical before the nanotube space elevators.

Car bodies and airplane fuselages: Well, I'm only comparing tech for putting stuff into orbit.

Remember, condescension doesn't make arguments more persuasive.

Luke said at December 30, 2008 7:36 AM:

Musk said "Long-term plans call for development of a heavy lift product and even a super-heavy, if there is customer demand. We expect that each size increase would result in a meaningful decrease in cost per pound to orbit. For example, dollar cost per pound to orbit dropped from $4,000 to $1,300 ($8,800/kg to $2,900/kg) between Falcon 1 and Falcon 5. Ultimately, I believe $500 per pound ($1,100/kg) or less is very achievable." [11]

Post a comment
Name (not anon or anonymous):
Email Address:
Remember info?

Go Read More Posts On FuturePundit
Site Traffic Info
The contents of this site are copyright