≡ Menu

Spacecoach on the Stage

I’m glad to see that Brian McConnell will be speaking at the International Space Development Conference in Toronto this week. McConnell, you’ll recall, has been working with Centauri Dreams regular Alex Tolley on a model the duo call ‘Spacecoach.’ It’s a crewed spacecraft using solar electric propulsion, one built around the idea of water as propellant. The beauty of the concept is that we normally treat water as ‘dead weight’ in spacecraft life support systems. It has a single use, critical but heavy and demanding a high toll in propellant.

The spacecoach, on the other hand, can use the water it carries for radiation shielding and climate control within the vessel, while crew comfort is drastically enhanced in an environment where water is plentiful and space agriculture a serious option. Along with numerous other benefits that Brian discusses in his recent article A Stagecoach to the Stars, mission costs are sharply reduced by constructing a spaceship that is mostly water. McConnell and Tolley believe that cost reductions of one or two orders of magnitude are possible. Have a look, if you haven’t already seen it, at Alex’s Spaceward Ho! for an imaginative look at what a spacecoach can be.

ISDC is a good place to get this model before an audience of scientists, engineers, business contacts and educators from the military, civilian, commercial and entrepreneurial sectors. ISDC 2014 brought over 1000 attendees into the four-day event, and this year’s conference brings plenary talks and speakers from top names in the field: Buzz Aldrin, Charles Bolden, Neil deGrasse Tyson, Peter Diamandis, Lori Garver, Richard Garriott, Bill Nye, Elon Musk and more. My hope is that a concept as novel but also as feasible as the spacecoach will resonate.


Image: Ernst Stuhlinger’s concept for a solar powered ship using ion propulsion, a notion now upgraded and highly modified in the spacecoach concept, which realizes huge cost savings by its use of water as reaction mass. This illustration, which Alex Tolley found as part of a magazine advertisement, dates from the 1950s.

Towards Building an Infrastructure

We have to make the transition from expensive, highly targeted missions with dedicated spacecraft to missions that can be flown with adaptable, low-cost technologies like the spacecoach. Long-duration missions to Mars and the asteroid belt will be rendered far more workable once we can offer a measure of crew safety and comfort not available today, with all the benefits of in situ refueling and upgradable modularity. Building up a Solar System infrastructure that can one day begin the long expansion beyond demands vehicles that can carry humans on deep space journeys that will eventually become routine.

The response to the two spacecoach articles here on Centauri Dreams has been strong, and I’ll be tracking the idea as it continues to develop. McConnell and Tolley are currently working on a book for Springer that should be out by late summer or early fall. You can follow the progress of the idea as well on the Spacecoach.org site, where the two discuss a round-trip mission from Earth-Moon Lagrange point 2 (EML-2) to Ceres, a high delta-v mission in which between 80 and 90 percent of the mission cost is the cost of delivering water to EML-2.

The idea in this and other missions is to use a SpaceX Falcon 9 Heavy to launch material to low-Earth orbit, with a solar-electric propulsion spiral out to EML-2 (the crew will later take a direct chemical propulsion trajectory to EML-2 to minimize exposure time in the Van Allen belts). The water cost is about $3000 per kilogram. The Falcon 9 Heavy should be able to deliver 53,000 kilograms to low-Earth orbit per launch. McConnell and Tolley figure about 40,000 kilograms of this will be water, while the remainder will be other equipment including the module engines and solar arrays. From EML-2, various destinations can be modeled, with values adjustable within the model so you can see how costs change with different parameters.

The online parametric model has just been updated to calculate mission costs as a function of the number of Falcon Heavy 9 launches required. You can see the new graph below (click on it to enlarge). At a specific impulse of 2000s or better for the solar-electric power engines, only two launches are required for most missions, one taking the crew direct to EML-2, the other carrying the water and durable equipment on a spiral orbit out from LEO. It is only the most ambitious destinations like Ceres that require three launches. At $100 million per launch, even that mission is cheap by today’s spaceflight standards.


Brian notes in a recent email that the launches do not need to be closely spaced, because the spiral transfer from LEO to EML-2 takes months to complete. The crew only goes when everything else is in place at EML-2. For more on this model, see spacecoach.org. I’ll be interested to hear how the idea is received at ISDC, and how the upcoming publication of the spacecoach book helps to put this innovative design for interplanetary transport on the map.


Comments on this entry are closed.

  • James Stilwell May 18, 2015, 13:41

    Awareness is the game now!
    You have the genius to get it done!
    Keep going!

  • david lewis May 18, 2015, 18:23

    If the price is 100 million, then there would be a market. With an ever increasing number of billionaires in the world (proving the theory of trickle-up-capitalism far outweighs that of trickle-down) I would imagine no few of those would be more than willing to purchase one.

    For one, it would be a way to boost their fame, and would have the potential to pay itself back despite the cost. Landing on an asteroid, planting a flag of some sort, and naming it, would be a sure attention getter. If selling a product, that would be a great way to advertise its uniqueness. Plus there would be the ego factor.

    Not just individuals but companies would be jumping for a chance to purchase one, for much the same reason. Sure, special affects could duplicated conditions on say Ceres, but they wouldn’t be able to claim an ad was shot there. There’s a money-making proposition for someone should these ships become reality – a business of shooting advertisements in space, actually on-site, to garner attention for some product.

    Then, there would the groups that can fund such a purchase for research purpose: major universities and research organizations. Even smaller groups could get together and share the cost of one.

    Then there would need to be some sort of search-and-rescue group, aka coastguard, for when those ships run into trouble. More sales there.

    With sales potentially in the thousands that would certainly jump-start things. Hotels in orbit, and ‘cheap’ trips to the moon. And once the life support become good enough, the perfect get-a-way for those who want to escape justice.

    Oh well, even if the reason is greed, at least it would be a step in the right direction – a firm step taken to get humanity out of Earth’s cradle.

    ….. but I wonder what happens if the water on Ceres is found to have microscopic life? That might complicate things when trying to fill up the gas tank since there would be fears of contaminating the Earth with it with unknown consequences.

  • J. Jason Wentworth May 19, 2015, 1:10

    This technology is a very good “fit” with human patience, politics, and economics. Electrothermal (arc-jet and resisto-jet) rocket engines have adequate-to-good specific impulse, coupled with sufficiently high thrust (especially when such engines are clustered) to enable tolerable mission durations. They aren’t as propellant-efficient as ion thrusters and Hall Effect thrusters, but when your working fluid (propellant) is *water*, propellant availability and cost aren’t pacing items. Also:

    Electrothermal engines raise none of the objections that nuclear thermal rockets face, and they are simple, safe, long-lived, and easy to produce and service. These engines have always been considered a “poor third” in performance to electrostatic and electromagnetic thrusters (except for a few attitude control thrusters, relatively few electrothermal engines have been flown in space), and consequently they haven’t received the level of development work that their “more favored kin” have enjoyed. But just as turboprop engines are widely used in aviation despite not being the fastest or most fuel-efficient aircraft powerplants, so too will electrothermal engines prove the most favorable for space travel applications that “mesh well” with their characteristics.

  • DCM May 19, 2015, 4:29

    If there’s microscopic life don’t bother it. Leave it alone.

  • Alex Tolley May 19, 2015, 12:46

    Thank you for posting a note about Brian’s talk in TO this week. We are looking forward to the feedback we get. Any obvious holes in the approach should be exposed, as well as further questions that we need to address. If we get any industry interest, that would be a plus.

    As I monitor what is being said at various conferences, it seems to me that the argument for electric engines is becoming increasingly evident. Coincidentally the USAF’s X-37B is to be launched on Wednesday with an array of Hall Effect thrusters, using Argon as the propellant, for maneuvering in orbit. I’m also noting the advantages of inflatable hulls is being increasingly touted as a solution for crewed deep space missions, as they offer more living volume and the plastic material offer some radiation protection from protons as well as reduced secondaries for GCRs. It seems that the last major remaining idea is the use of water as propellant, allowing all the benefits we have outlined to be gained.

    Brian and I are anticipating his talk to generate interest and further feasibility questions that we can address in the future for this approach to spaceflight.

  • david lewis May 19, 2015, 13:13

    “If there’s microscopic life don’t bother it. Leave it alone.”

    I don’t see why we would/should leave it alone.


    The place could have upwards of hundreds of millions of cubic kilometers of water – more so than all the fresh water on Earth. It would be like saying we couldn’t use water from the Great Lakes which only has a few tens of thousands of cubic kilometers because it has algae in it.


    At 1 ton a cubic meter, if each ship was to take 1000 tons from Ceres each trip, that would be 1,000,000 trips per cubic kilometer of water.

    The problem as I see it is contamination, whether it be Ceres by Earth, or Earth by Ceres.

  • Brian McConnell May 19, 2015, 13:54


    We’re open to using whatever works best. Besides electrothermal engines, we know electrodeless Lorentz force engines will work with water and CO2 at an Isp of 1,700s or better. What we’d really like to know is if Hall Effect thrusters will work as is, or with minimal adaptations, will work with water and carbon rich gases. If the latter proves to work, then we have an off the shelf engine that’s already been used extensively with satellites.

    In any case, we view this as a systems integration problem, not unlike building the best possible computer using currently available components. All we care about is finding an electric propulsion technology that works with water (and ideally also carbon rich gases), and runs at an Isp of 2000-3000s (that’s the sweet spot for spacecoaches).

    Then upgrade and replace engines as needed or as improved units come along to increase spacecoach operating range or further improve economics.

  • J. Jason Wentworth May 20, 2015, 0:35

    Thank you, Brian. Hall Effect thrusters produce greater thrust than comparable ion thrusters, and they are compact and grid-less (which should make them longer-lived than gridded ion engines). They also seem to work well on propellants other than xenon (tests using argon [and other noble gases, if memory serves] and even bismuth have demonstrated good performance). Plus:

    A “flex-fuel” Hall Effect (or other type of) electric thruster like this would be very useful for general-purpose spacecraft and spaceships. Perhaps a colloidal water solution would be sufficiently easy to ionize for such thrusters (colloid thrusters, incidentally, would be good for attitude control for these vehicles). As an alternative, mildly salty asteroidal and/or cometary water (these bodies are suspected to contain it) might work well in such “flex-fuel” electric thrusters; if they were made of corrosion-resistant materials, using this brackish water in the thrusters shouldn’t be a problem.

  • Harold Daughety May 20, 2015, 1:34

    For me, the best thing about the spacecoach concept is that is is doable. A voyages to the stars is not doable at present and may never be.There currently are no warp engines or worm holes to get us there and nothing remotely in the pipeline. But all components for the spacecoach are available. Just a little engineering and a lot of money and away we go! Once in space, untethered from the earth, humanity will maintain a presence there forever. Then, maybe the stars. . . .

  • Rob Henry May 20, 2015, 5:36

    Am I the only one troubled by the measuring of Isp in units of seconds. That is as wrong as saying the spin rate of Earth is 15, when in fact it is 15 (arc)seconds per second. The point is that these are different units that share nothing but the name of their measure.

    Likewise the imperial measure of Isp is not really seconds but pound per pound seconds. These completely different lb units can’t cancel out, and even if you try to do so you get g. Thus 1700gs is quite acceptable but 1700s for Isp is just gobbledygook. Just because every other American engineer has made the same mistake for half a century doesn’t mean the rest of you have to continue it forever! (or does it?)

  • John Freeman May 20, 2015, 8:42

    I don’t suppose you need an engineer/blogger? :)
    In seriousness, what would be the developmental road to seeing one of these actually happen?

  • Alex Tolley May 20, 2015, 11:40

    Rob Henry. It depends on the units, you picked the one that gets you to gs.
    See here for explanation:
    “If mass (kilogram or slug) is used as the unit of propellant, then specific impulse has units of velocity. If weight (newton or pound force) is used instead, then specific impulse has units of time (seconds). The conversion constant between these two versions is the standard gravitational acceleration constant (g0)”

    The rocket equation is:
    V = Ve*ln(Mo/M1)
    V = g*Isp ln(Mo/M1)

    Ve units must be the same as g*Isp units

    V or Ve = m/s
    g = m/s^2

    Isp = Ve/g = (m/s)/(m/s^2) = s

    I hope this helps clarify.

    Another way to think about this is to equate the Isp to mean how long does the engine have to run to reach the final velocity? If we set ln(m0/m1) = 1, i.e. (m0 = 2.718, and m1 = 1), then V = Ve = g * Isp. Then Isp means how many seconds will the engine run, if it was accelerating at g, until the final velocity of the rocket is the same as the exhaust velocity. Trivial, I know, and potentially confusing as high Isp engines accelerate very slowly in practice, but I think it helps to put the units into context.

    I hope I’m not teaching a grandmother to suck eggs with this explanation. ;)

  • Brendan May 20, 2015, 13:29

    My reaction to this is that while inventive, water is a precious resource. Granted, there is a lot of water in the solar system. Shooting it out the back of a spacecraft seems wasteful.

  • Brian McConnell May 20, 2015, 14:27


    exhaust velocity = g x Isp, g is measured in meters/second^2, Isp in seconds, so simply multiplying units of measure gives meters/second, which is what you’d expect for exhaust velocity.

    maybe it’s done differently in imperial units, but we did all of our analysis in metric, and all of the literature we reviewed was also in metric.

  • ljk May 20, 2015, 17:02

    Brendan said on May 20, 2015 at 13:29:

    “My reaction to this is that while inventive, water is a precious resource. Granted, there is a lot of water in the solar system. Shooting it out the back of a spacecraft seems wasteful.”

    There is a LOT of water in the Sol system. Europa’s global ocean alone has twice the volume of that molecule in liquid form than all the oceans of Earth combined. There are literally billions of comets circling our yellow dwarf star. We also know of a number of other solar systems with comet belts of their own, some like Tau Ceti being even more voluminous with those ancient ice balls than our own.

    While I do not believe in being wasteful with resources of Earth, I think we would have to launch a LOT of water-using spacecraft in order to make any kind of a dent in the water supply for our celestial neighborhood.

    BTW, this is also why I get a rueful kick out of subpar science fiction plots which involve aliens coming to Earth to steal our water: There are so many other places to go to get the same substance with a lot less hassle in terms of much more shallow gravitational wells and no angry natives to contend with.

  • Rob Henry May 20, 2015, 19:55

    Thank you for the reply on Isp unit usage. Wikipedia makes the best of this mistake that it can, but taking Isp in seconds fails dimensional analysis. Interesting that you work in metrics yet still use Isp in seconds, as I recall a movement to made kg(force) a unit of weight. It failed long ago and preventing the confusion of weight and mass became one of metrics strong points. Assuming, for the moment that logic continues to fails and we are stuck with this archaic unit, what on Earth is g(0) on a planet where surface gravity varies from 9.8337 m/s2 to 9.7639 m/s2 ? Is the reference the base of the Kaaba, or street level outside the New York stock exchange, or is every culture free to pick its own g(0) in a sort of ‘don’t worry, be happy’ arrangement?

  • Mark Zambelli May 21, 2015, 10:13

    ljk said…

    “…science fiction plots which involve aliens coming to Earth to steal our water…”

    What… do you mean that the late Robert Urich movie ‘Ice Pirates’ isn’t a documentary!? :)

    Cerean microbes would throw a spanner in the works for astrobiological reasons and until we worked around that sticky point we’d still have megaliters of water on the moon etc, as ljk mentions.

    This Spacecoach idea really deserves as much press as possible to garner the interest it deserves. The potential for becoming a game-changing idea with minimal (as I see it) requirement for new technology (if any) is really exciting. Goodluck gents.

  • Rob Henry May 21, 2015, 17:03

    On a more practical note, why the need for chemical thrusters at all for crew? Why not use that water initially to continually increase apogee of the stagecoach, while keeping perigee at LEO level, and only transfer crew when the orbit can be broken by only one last pass through the Van Allen belts

  • Brian McConnell May 21, 2015, 17:21

    One thing to keep in mind. I understand why people want to focus on ISRU, but (puts on systems engineer / finance hat), what matters is which source of water is cheapest.

    We know we can deliver water by the metric ton to LEO for a cost of $1,700/kg via F9H. There are lots of unknowns around ISRU, from basic feasibility, to the fully loaded cost of water delivered via ISRU. Unknowns = risk = higher costs.

    If the F9R becomes real, water will be its ideal payload (zero replacement cost if a heavily used booster explodes), so it wouldn’t surprise us if the cost of delivering low value payloads to LEO declines even further with the F9R.

    That’s why we don’t include ISRU in our models for the most part. Of course, we’ll want to explore how to do it, but for the foreseeable future, the cheapest, most reliable source of water for these ships will be Earth. It’s difficult to model something that’s almost all unknowns.

  • Alex Tolley May 22, 2015, 10:55

    @ Rob Henry what on Earth is g(0) on a planet where surface gravity varies from 9.8337 m/s2 to 9.7639 m/s2 ?

    varying g doesn’t help you either. In your original example you had 1700gs. G will vary there too.

    In practice, the variance of g is so small compared to other factors, such as where you launch from on Earth, that it doesn’t really matter in determining performance. Even in space other factors will probably impact performance more than the “constant” g. Bottom line for me is that Specific Impulse measured in seconds works fine, but as long as we use the correct form of the equation, one can use either units for Isp.

  • Eniac May 22, 2015, 20:53

    Rob is right, Isp in seconds is a flawed measurement because its definition is dependent on a constant (g) that has no meaning in space. Exhaust velocity is the clean quantity to use for comparing rocket engines.

    I am curious how feasible it actually is to operate ion engines with water. What ion are we talking about? OH-? OH3+? What do we do with the extra/missing hydrogen? Does anyone have more than handwaving on this?

  • Alex Tolley May 22, 2015, 22:53

    @Eniac – the baseline Spacecoach uses microwave electrothermal engines. A microwave turns the water to a high temperature plasma that operates just like a rocket engine after that. Isp is around 800-900 seconds, nearly twice that of a chemical rocket.

    Ideally we would want something with an Isp of 2000-3000 seconds. To do that we need something more like a hall Effect thruster. Those engines need propellants that are easily ionizable. Water with some additives may work, but that is speculative and needs development, which is why we envision running a competition to see what emerges. There are other approaches that will allow an Isp in the 1500s range.

    If we have the power supply, then one fallback is to simply electrolyze the water to hydrogen and use something more in line with the VASIMR technology but with lower Isp, again demonstrated technology. However we have not as yet calculated the water requirements for such a ship where we are discarding the oxygen (ie most of the water mass) for use as a propellant. That doesn’t seem to make a lot of sense.

    Another fallback is to use a mass driver and eject water ice pellets, but the performance of such an engine is unknown and may require enormous amounts of water if the “exhaust velocity is more comparable to a chemical rocket, with associated higher costs.

    So there are engine technologies that work at the lower end of our performance range that gets the ship started, and we are expecting that the higher performance electric engines will emerge in due course.

    Rather than aim for the best engine technology before we start, the idea is to use the most economic technology available that meets teh requirements, and to upgrade as newer, higher performance technology becomes available at a good price. Affordable utility, rather than aiming for the “top of the line”, best performance.

  • Alex Tolley May 24, 2015, 11:04

    I should say that a mass driver is going to be very massive and/or power hungry, especially as we want the ejection velocity to be be of the order of 8-30 km/s. Without some clever engineering, this isn’t a good approach. A rotating arm that ejects the masses might be better, but there is little information on the performance of such devices and their suitability as engines.

  • Eniac May 25, 2015, 1:00

    Alex, as you say, not using the oxygen as propellant is not a good option. I may be wrong, but it seems to me that “additives” are not going to do it, either. You need to ionize all of the water in a controlled fashion (i.e. a beam rather than plasma), so the ions can be accelerated. Perhaps the most feasible option for high Isp would be twin O– and H+ ion engines. That way you could use all of the water, and do away with neutralizing e-beams at the same time. Efficient water ionization and separation into O– (or OH-) and H+ streams would be a formidable engineering challenge, I suspect, but not inherently infeasible. Perhaps the peroxide can be of help here, too.

    Mass drivers, preferably the lightweight rotating tether sling type, are great for launching inert payloads from low gravity bodies on slow journeys, but not so good at propelling fast, manned ships. We know quite well that the “Isp” performance of rotating tethers will not get much beyond 2 km/s. That is roughly the tip speed beyond which even the strongest tethers will break. Magnetic mass drivers have similar limitations. and are much heavier.

    This is an extremely exciting concept, I would love to get involved with it.

  • Alex Tolley May 25, 2015, 9:47

    @Eniac We know quite well that the “Isp” performance of rotating tethers will not get much beyond 2 km/s.

    Thank you. I think I saw a figure like than mentioned in the Landis paper on lunar slings: Journal of the British Interplanetary Society, Vol. 58, No. 9/10, pp. 294-297 (2005). “Analysis of a Lunar Sling Launcher”

    So that definitely rules out that approach as the performance is too low.

    Accelerating the plasma from RF heated water might also be an option. The helicon plasma thruster concept uses nitrogen and gets an exhaust velocity of 10x that of chemical engines. Maybe that approach can be adapted?

  • Eniac May 25, 2015, 22:52

    The helicon thruster is a plasma rocket, not an ion engine. Akin to VASIMR, really, only simpler. It is indeed much easier to adapt to arbitrary propellants. Probably the way to go, but can it get up to the Isp we want? Probably. Make the plasma hot and the field strong enough.

    Apparently they built one from a soda can and soda bottle. Fits well with the stage coach concept, I’ll say!

    See also this: http://en.wikipedia.org/wiki/Helicon_double-layer_thruster

  • Eniac May 25, 2015, 23:09

    This is a study of various propellants in a double layer helicon:


    Right out of the box, this one turns out an Isp very close to the ideal range Alex mentions here, earlier.

    Ion energy distribution functions are measured using a retarding field energy analyser located 7.5 cm downstream of a helicon double layer plasma source, respectively, operating with four molecular gases: nitrogen (N2), methane (CH4), ammonia (NH3) and nitrous oxide (N2O). For radiofrequency powers of a few hundred watts, and a magnetic field diverging from about 0.013 T (130 G) in the source to about 0.001 T (10 G) in the exhaust, an ion beam is detected for each propellant over a very similar operating pressure range (~0.023 Pa (0.17 mTorr) to ~0.267 Pa (2 mTorr)), as a result of spontaneous electric double layer formation near the exit of the plasma source. The characteristics of the ion beam versus operating pressure closely follow those previously obtained in argon, xenon and hydrogen. The ion beam exhaust velocity in space is found to be in the 17–19 km s−1 range in N2, 21–27 km s−1 range in CH4 and NH3 and 14–16 km s−1 range in N2O.

    Maybe H2O was not tested because oxygen ions are highly corrosive? Probably best to just ask them.

  • Alex Tolley May 26, 2015, 13:20

    @Eniac – thank you for the reference to the HDLT. Although the paper doesn’t reference water, in a news article the researchers joke that you could run it on “pee”. So I assume water should work too. Brian and I are going t0 look into this as a higher Isp engine solution.

  • Eniac May 26, 2015, 22:55

    I did not mean to poo-poo slings, really. They are some of my favorite means of propulsion. Good for throwing payloads directly, say from the moon to the Earth, with no propellant at all. Or, to propel an asteroid by throwing rocks off it, the cheapest possible propellant for the purpose. A maximum Isp of 2-4 km/s is not much, but if you are willing to throw away almost all of an asteroid, the remainder can be made quite fast.

    None of this will propel a space coach directly, but in the not too distant future both methods could play a critical role in the deep space water trade.

  • Michael May 28, 2015, 15:24

    Instead of launching water they could send up a more concentrated hydrogen peroxide version of which some will be decomposed to form breathable air. Hydrogen peroxide is much denser than water, two birds with one stone me thinks.

  • Alex Tolley May 28, 2015, 23:12

    @Michael. Interesting idea. At LEO the H2O2 is mostly decomposed with a simple catalist, the steam is cooled and put into the hull, whilst the O2 could be used to recharge the consumable O2 supply or even be used as part of the inflation gases. Reduces the need to electrolyze water for the O2 recharge, possible H2 purge and loss of a fraction of the consumable water and propellant. It could also be used to clean the hull interior to get rid of bacterial films. Given the huge excess of water, the peroxide could be quite dilute and safe to handle..

  • Michael May 30, 2015, 4:39

    Hydrogen peroxide can also be used in a fuel cell with a fairly high energy density, see article.