Reverse Rocket

This is a post about an idea by Doug Plata. His idea is to put multiple propellant tanks on top with just enough structure to keep them intact and drop them in pairs as they drain. Under the tanks is the payload. Under the payload is a plug nozzle/heatshield that multiple engines expand against for altitude compensation. The configuration is chosen to be fault tolerant of both engines and tanks. This allows the use of somewhat questionable engines and tanks. The concept is specifically for the purpose of launching very large payloads on vehicles with relatively low development costs.

From the top down, this concept starts with a composite aerodynamic fairing that is also a fuel tank for kerosene. Next is a cluster of 19 tanks each of which holds either fuel or oxidizer with no common bulkheads. Next comes the payload volume which is the full diameter of the 19 tanks and whatever length required for the particular mission, with cylindrical sections added or removed as required. Under the payload volume is a full diameter plug nozzle that is also a heatshield for reentry. Against the sides of the plug nozzle are multiple engines of relatively low expansion ratio with the plug nozzle making up the difference at all altitudes.


The cartoon is a rough representation of Dougs’ concept. The payload can be a habitat cylinder 15 meters in diameter and 30-45 meters long. By launching it dry, all of the permanent fittings can be installed and tested on the ground as well as some of the transient components that are used early on.

The habitat is conceived as useful for an all up station in one go without the fitting problems of an expandable structure that is volume constrained at launch. It is also possible to design it in such a way that lunar mass could be added for radiation shielding in Lunar orbit or one of the L points. The rigid structure is also suitable for the Lunar surface with the capability of handling a thick regolith covering

The aerodynamic fairing that is also a fuel tank is handles aerodynamic loads only with both the fuel and pressurant gas providing support through the atmospheric portion of the flight. It is expected that mass of the shroud/tank will be on the order of 2% of the mass of fuel it holds. The shroud tank is sized to empty as the vehicle reaches low dynamic pressure at altitude when it is jettisoned.

The 19 tanks under the shroud/tank are protected from  head on pressure while in the atmosphere. As the shroud tank is jettisoned, the empty oxygen tanks in the 1 ring are sent off as well leaving only full tanks to carry. As each pair of either oxygen or kerosene tanks are drained on opposite side of the vehicle, the empties are kicked off from the 1 ring, then the 2 ring. The pressurant gasses in each tank are used as a cold gas thruster to ensure clean separation. There are as many as 6 tank staging events as the vehicle climbs out. Each tank being 3 meters diameter by 50 or more meters long, available propellant volumes are at roughly 350 cubic meters per tank. Tank mass can be on the order of 1% of propellant mass. Along with the shroud/tank, propellant mass can be on the order of 7,000 tons.

At the plug nozzle/heat shield, there are as many engines as required by a given mission. Since the concept is for launching massive one off missions, engine mounting must be modular similar to the tank concepts. Shrapnel shields and other safeguards must be designed in as the concept is for very low flight number vehicles with inherent infant mortality. Extra engines are a requirement for a couple of reasons. One is that it allows a more efficient flight profile than the normal thrust limited take offs of most launch vehicles. The other, more important reason is that available engines will be used with variable reliability and  availability. Careful attention to this detail should make it possible to use the remaining AJ-26 inventory of orbital-ATK as well as used Merlins and anything else the contractor can get his hands on. Unreliable engines can be compensated by having fail safe ways of shutting them down and jettisoning them. This approach allows buying engines from motivated sellers.

An additional advantage of Dougs’ concept is that it allows the multiple engines to use a variety of propellants. A mixture of kerosene, methane and hydrogen engines is quite possible in this configuration.

As the vehicle sheds propellant and tank mass, acceleration will rise. As it reaches the maximum desired, pairs of engines will be dropped in a manner similar to the way the tanks are treated. Each engine can be fitted with a decelerator and parachute as long as the velocity is low enough that there is a reasonable expectation of recovery. At higher velocities, expended engines will be lost as they are dropped. The ones that make it to orbit can be packaged into the heat shield for return to the ground.

The plug nozzle/heat shield has a multiple use as a heat shield in addition to its’ nozzle duties. In case of abort, engines and tanks are jettisoned and the heat shield  is used to protect the payload for a return to sea level. The idea is that the nozzle and vehicle shell may be lost, but the interior equipment could be saved for use on a replacement mission. For returning the payload from orbit, or atmospheric entry to another planet, the heat shield works in the normal manner. On a nominal mission where the payload is not returning to Earth, the heat shield is used to return the remaining engines and any other valuable gear that needs to return to the ground.

Figuring the payload to orbit is interesting. It turns out that with so many small staging events, the dropped tanks and engines can be treated as propellant for calculation purposes. With ~7,000 tons of Kero/LOX propellant, it is possible to place a 350 ton space station in orbit with well over 6,000 cubic meters of living and work space.  This is moreover, work space that doesn’t have to be cramped into tiny cubicals and corridors.

Posted in Uncategorized | 8 Comments

Getting My Numbers

This is an explanation of how I get BOTE numbers for such things as the comparison between LH2 and CH4 propellants as in Jons’ last post.

The focus Jon had was how much H2 do you need to carry to Mars or Venus to get a return stage if you do a straight LH2/LO2 as opposed to using in situ resources to create CH4 with the hydrogen you carried from Earth. If it was just as effective to use the hydrogen directly as using the intermediate step of converting it to methane with local carbon, then why bother with the conversion and the equipment required to carry it out. Also, if using a hydrogen stage for Earth departure, commonality of equipment is a plus for using a hydrogen stage for Mars or Venus departure. In effect, it is possible that the use of local carbon to convert the on board carbon to methane might not be a worthwhile step.

I is not a rokit sceintest so I tend to look for simplifications that I can do with a scratch pad and TI 30 calculator. A real rocket engineer calculating real missions will go far more in depth than I am capable of, but he tends toward getting paid for his efforts. So this method is worth almost what you are paying for it.

I picked Mars as the launch site for this BOTE because it is easy. I figured 6,000 m/s total V for the rocket because it seemed like a good start for a vehicle that has to reach Mars orbit and do other things once it gets there. Other things range from rendezvous with Earth departure vehicles to a surface return for more flights. If there were a good reason to figure a different V, you can do it during lunch and still have time to eat.

6,000 m/s stage with hydrogen figured as 4,500 m/s exhaust velocity and Methane as 3,700 m/s. Rocket equation gives mass ratios.                                                                                 propellant type                                          hydrogen                                      methane                     mass ratio                                                     3.79                                                 5.06

Then I translate mass ratio to percentage of propellant at lift off.                                                percent propellant                                                74%                                                 80%                propellant density                                               0.31                                                  0.94                 tank volume per ton                                        3.26 meters                                    1.05 meters     tank mass per ton at 20 kg/cu m                   65.2 kg                                               21  kg             tank mass as percent of GLOW                         4.8%                                                1.68%              engine T/W est                                                   80                                                       110                engine mass at 6 m/s at GLOW                        0.75%                                                0.5%             percent of propellant, tank and engine           79.55%                                               82.18%          percent payload to GLOW                                20.45%                                              17.82%        Glow per ton of payload                                      4.89 tons                                          5.61 tons      propellant mass per ton of payload                    3.6 tons                                           4.5 tons       hydrogen percent of prop load                              14.28%                                            6.25%          hydrogen per ton of payload                                 0.514 tons                                      0.28 tons

This is all just a fast way of getting a BOTE for a concept. Anyone could change a variable and have an answer for a different scenario in a few minutes. To me, the question that Jon posed is an interesting one. Is it worth doing all the conversions in favor of somewhat more hydrogen tankage?    I had assumed it was. Now I don’t know. It will take a detailed trade study for a specific mission set up to see which is better, or neither. Comments on Jons’ post brought up CO for a first stage. Another mentioned perchclorate in the Martian regolith at ~1% concentrations.

I think it is fair to say that there are far more possibilities than we normally consider, and that good answers are not always the most obvious ones.




Posted in Uncategorized | 2 Comments

Random Thoughts: Which is a Better ISRU Propellant on Venus/Mars–LOX/LH2 or LOX/CH4?

I’m not sure if someone has already run the analysis, but I’m kind of curious about which ISRU-derivable propellant combination is better for locations like Venus or Mars where there is plenty of CO2 available in the atmosphere, but limited water.

Assume for a second that water isn’t available in useful quantities. I’m not sure yet if the concentration in the Venusian atmosphere is high enough to be useful, and it’s not yet clear that there’s easily accessible water at most places on Mars–there might be, but it’s far from clear1. Assume for now that you have to bring your hydrogen with you, from Earth.

A few quick observations:

  1. One immediately obvious point is that if you have to BYOH2 you’ll probably want to bring the hydrogen as LH2, not water. While everyone complains about how hard it is to store LH2 for long durations in space, water is still ~89% Oxygen, which is a material you can get almost anywhere, especially if you have CO2 available, so for every kg of hydrogen you bring tied up in water, you’re lugging around an unnecessary ~9kg of oxygen. You can definitely make a dewar with active cryogenic cooling that masses far less than 9x the mass of LH2 you want to bring with you–it may be “harder” to do it that way, but is far, far more efficient.
  2. A typical O/F ratio for LOX/CH4 is probably around 2.8-2.9:13, which means that about ~6.5% of your propellant mass is hydrogen for LOX/CH4. For LOX/LH2, you’re probably looking at an O/F ratio of around 5-6 typically, which would yield ~14-17% hydrogen. So for every kg of hydrogen you bring along, you could get4 ~15.4kg of LOX/CH4, or you could get 6-7kg of LOX/LH2.
  3. If you’re limited by hydrogen you can bring, rather than dry mass, or volume, or other things, it’s not yet clear which of those will result in more payload in orbit, since the two have significantly different bulk densities and Isp values. That’s the analysis that would be fun to run. My guess is a lot will depend on the required delta-V5, whether you’re looking at 1, 2, or 3 stages, if you assume on-orbit refueling before the earth-return, etc.
  4. One way to cheat a little with LOX/LH2 would be to use a LOX-rich Thrust Augmented Nozzle (TAN). Basically, you have a core running at the more traditional 5.5-6:1 O/F ratio, while initially running the afterburning portion of the engine at a much higher mixture-ratio, possibly even higher than stoichiometric! As the rocket accelerates, you could throttle down this element and then shut it off. This is probably more useful for Venus ascent than Mars, but would allow you to get not only a much higher engine T/W ratio than you could realistically get normally with LOX/LH2 engines, but also give you more propellant per kg of brought hydrogen, because you’re shifting your mission-averaged mixture ratio to a very, very lean range.

I honestly don’t know the answer, and don’t have time yet to run the numbers, but I’m genuinely curious. If you have a fixed supply of hydrogen, which ISRU propellant method (using a Sabatier reactor to convert H2 and CO2 to LOX/CH4, or using a solid electrolysis cell to crack O2 out of the CO2 to make LOX/LH2) actually yields the most mass delivered to orbit or to an Earth return trajectory from Mars or Venus? Has anyone already done this analysis? If not, I may try to find some time at some point to run the numbers.

[Update 9:58pm on 9/5/2016: in case you’re curious what brought this on, I was thinking about Venusian Rocket Floaties again, and was wondering whether a Venusian launcher first stage would want to be LOX/CH4 or LOX/LH2.]

Posted in ISRU, Mars, Venus | 14 Comments

Semi-Staged Combustion

It is interesting comparing the two best known first stages in the US that use kerosene and LOX. The Atlas 5 and the Falcon 9 use a similar fuel in their first stages and then diverge in the technical aspects. The Atlas 5 with the RD 180 engine has about 10% higher Isp at sea level while the Falcon 9 Merlin has nearly twice the thrust/weight ratio. The over all Falcon 9 first stage seems to have a much lower dry mass ratio which makes up the difference in engine performance and then some.

There are going to be new vehicles designed by the various companies eventually that would like to benefit from the competitive advantages of both vehicles. A high thrust/weight ratio engine with high Isp that also has low dry mass is a desirable target. The more these features can be designed in, the more mass is available for payload, reusability, or both.

One of the engine cycles that is discussed from time to time is the dual chamber concept. It is more or less a gas generator cycle with an exhaust pressure high enough to inject into a lower pressure thrust chamber to burn with fuel or oxidizer to get useful thrust. I suggest it might be possible to get very near RD 180 Isp with very near Merlin thrust/weight with a variation of the concept. A low stage dry mass being part of the goal, I add in a few features that may be unique.

Semi-Staged CombustionIn the cartoon I have two high pressure chambers on the outside with a lower pressure chamber in the middle with an altitude compensating nozzle.

The black boxes in the tanks are the electric inducer pumps from the previous post.  They are to keep the propellants at high enough pressure to the main pumps to suppress cavitation as well as keeping required tank pressurization to a minimum.

The small blue tanks in the inter tank area are for the liquid hydrogen that serves multiple purposes. First the hydrogen feed hits a heat exchanger in the LOX  tank to keep it cold enough to stay liquid and suppress cavitation even as tank pressure drops. Then it hits a heat exchanger in the RP tank for the same purpose. Then it is used to cool the turbine blades the same way that jet engines use air cooling. Finally it burns with the excess LOX from the gas generator to produce thrust.

With the pumps providing pressures to the main engine similar to that of the RD 180, the Isp of them should be similar. About 10% of the propellant goes to the gas generator driving the pumps with a residual pressure of 300 psi after the turbine. If the 300 psi engine was a normal kerosene engine, one would expect an Isp in the 250s from that portion of the thrust system. With the lean (LOX rich) gas generator driving a hydrogen cooled turbine at much higher than normal turbine inlet temperatures, the warm hydrogen mixes with the hot oxygen as it is used for film cooling of the blades and burns in the secondary chamber above the throat. The hydrogen/kerosene/LOX engine at 300 psi could approach the ISP of the main engines due to the higher performance of hydrogen. Hydrogen usage will be a fraction of a percent of the total propellant load.

The compensating nozzle of the low pressure engine in the center would allow reasonable Isp of that portion at sea level, especially with the hydrogen component. The higher expansion ratio made possible would allow much higher Isp at altitude, which, with the hydrogen component, could give vacuum Isp higher than the RD 180. I think the potential result is low hardware mass combined with high first stage performance.

Posted in Uncategorized | 10 Comments

Electric Inducer Pump

It takes pressure in the tank to suppress cavitation in the pumps for rocket engines. It is customary to use helium for most pressurization due to its’ low molecular weight. Unfortunately, it can take a lot of helium in some cases. Some propellants can self pressurize under the right conditions, though here is where the molecular weigh becomes important. O2 with a molecular weight of 32 is eight times the mass of helium at 4.

Heating up the pressurant gas helps considerably, though helium can be heated up as well. Sub-cooling the propellant helps suppress cavitation and allows a lower pressure in the tank to be effectively pumped. I am going to suggest an inducer pump in the tank instead.

If the required pressure in the tank can be reduced from 30 psi or more to 5 or so as the vehicle reaches altitude, the pressurant gas quantity can be reduced by a factor of 6.  An electric inducer pump in the tank might make this possible. A pump that is bypassed early in the flight is gradually brought online with increasing power as the tank level drops and with it the head pressure.

header pumpThe waste heat from the pump can be used on the pressurant gas to reduce the required mass. The pump power can be gradually increased to keep a constant 100 psi  or whatever the spec requires to the turbo pumps.

The objections I have had in the past to electric pumps have mostly to do with the mass penalties of the electric motors, and to a much greater extent the battery mass to reach engine pressure. A relatively low pressure pump used as an inducer gets around some of this. If pressurant gas is reduced along with the elimination of their tanks, that should compensate for the all of the motor weight and some of the battery weight. If the batteries are those of the satellite payload, then there might even be a mass savings. Many of the satellite payloads have a considerable amount of their mass in batteries which might as well help haul the freight during launch when they are probably bored anyway.

Posted in Uncategorized | 6 Comments

Notchbell Nozzle

Several years back I suggested a type of compensating nozzle that should be inexpensive to build and test. Unfortunately the ones I made for demonstration with compressed air were hit and miss as I didn’t have the theory quite right. Hit and miss is not good enough for serious companies, so I mostly dropped the idea for a while. I thought a few of my acquaintances in the business might do something with the idea for a while. Now I think the idea has  been mostly forgotten as unworkable.

A few years back I did finally find the missing part of the concept and did nothing about it as I thought at the time that others had picked it up and moved on. Since I don’t think that has happened, I am going to repost the concept.


On the left is the engine with the notch showing on the right side. The notch allows the atmosphere to enter the bell to compensate for over expansion at lower altitudes. At higher altitudes and in vacuum the exhaust gradually uses the whole bell with some losses through the notch. This will allow a nozzle to be optimum at sea level when most are over expanded. It will also be nearly as good as a full diameter high expansion nozzle in regimes with the exhaust under expanded.

The missing ingredient in the prior concept was appreciation for the momentum of the exhaust at the notch site. The momentum, especially with the rounded notches that I was advocating before, would prevent the atmosphere from entering the notch in a controlled manner. The addition of a sharp edge at the notch to assure a clean break and a slight reverse on the notch edge to direct the exhaust inwards controls the momentum of the exhaust in a manner that allows the atmosphere to interact and provide pc/pa compensation at a range of back pressures.

A compensating nozzle allows lower pressure engines to operate more efficiently in a launch vehicle. They should allow a payload increase of 1-5% depending on the vehicle and the assumptions going in. For a VTVL that wants to operate at very low pressures in the landing phase, a compensating nozzle would be a very important upgrade, though the successes of Blue and SpaceX take some of the edge off that argument.

This is a public domain concept as I described it here years ago. So anyone that wants to see what I am talking about can build a quick and dirty nozzle to use with shop air. The ones have done were an air chuck and fiberglass. About $10.00 in materials. I know it works at 135 psi. Then you can try a higher pressure gas if it might be useful to you or someone you know.

Posted in Uncategorized | 9 Comments

Old Engines

Some before, and many after the accident criticized the use of old engines for the Antares. Many references were made to the long storage time with some mention of them as antiques. The refurbishment of these engines didn’t seem to slow down the slings, arrows and insults much.

My question is, at what point do the old Shuttle engines for the SLS reach the same category as the  failed ones on Antares? After all, at the probable flight rate, the used SLS engines will be in the same general age range at some point as the failed ones at Wallops.

Eventually there might be new engines, but how far down the road?

Posted in Uncategorized | 3 Comments

Working full time on the Martian surface is within US Radiation Worker limits

The US Radiation Worker annual radiation limit is 5 rem, or 50mSv/year. Divided into the 2000 annual working hours, that’s 0.025mSv/hour.

The Mars Curiosity rover measured an average dosage on Mars of 0.67mSv/day at about -4km altitude. That’s 0.028mSv/hour.

If you worked somewhere lower altitude, like Hellas Basin (-7km or even possibly -8km) or a place like Valles Marineris at -5km and valley walls nearby, you should be able to get that down to the 0.025mSv/hour of US Radiation Worker limits.

Or work French hours. 😉

Or limit it to 1000 hours per year like commercial airline pilots, the rest indoors.

So really, at least during solar max (when GCR is at a nadir), the surface radiation levels of Mars don’t seem like any insurmountable barrier at all, provided you can adequately shield everywhere else and provided you’re okay with US radiation worker limits. For instance, 3m of polyethylene at -5km altitude (Valles Marineris) gives you 22mSv/year (although the model I use for that calculation is questionable to me… it seems there are too many low-energy neutrons). 1m of water/PE is 48mSv/year. Of course, you can also bury below a bunch of regolith or burrow to achieve arbitrarily low radiation levels.

Long-term, I suspect we’ll find drugs that are effective. Or we’ll find out if the Linear no-threshold (LNT) model is correct or not. In any case, when there are millions of people on Mars, it’ll be much easier to produce high-quality data about relative risks for low-dose radiation and also easier to develop enough statistical power to show whether or not drugs like Amifostine can protect against low-dose chronic radiation as well as the acute radiation we know it’s effective for.

Posted in Uncategorized | 26 Comments

Homo Cosmicus: Vestibular Implants

EDIT: David Birchler mentioned in the comments that an implant is not required. The technology is called galvanic vestibular stimulation and it can stimulate the sensation of pitch, roll, and yaw. Since surgery is not required, it sounds like this really IS doable as a countermeasure for dizziness on landing (perhaps combined with training) and the sensation of coriolis in a short arm centrifuge. In fact, for the former, it looks like this is already being tested as a training tool for astronauts:
and here, it shows that GVS training can allow quicker adaption to different vestibular environments:
Here’s a company which is developing the tech (along with Mayo Clinic) synced to Virtual Reality: VMocion

For a few years, now, I’ve been convinced that one of the best ways of making human spaceflight affordable and even competitive with robotic spaceflight in some areas is by engineering as close to the human as possible.

For instance, to solve radiation issues, you could shield the entire spacecraft. But this requires a lot of mass. Mass-wise, better is to shield just the habitable areas. Better still is to focus on areas the crew spends the most time in, like the sleeping quarters. Better still (though weird and awkward) might be a garment with radiation shielding. And yet we can do better still: radiation countermeasures in drug form, which protects at the cellular level. Drugs like Amifostine. (But we need better ones.) Or even closer, the nuclear (as in, cell nucleus) level: You could propose either genetically modifying people themselves to produce radioprotectants (or perhaps engineering our microbiome to produce radioprotectants without requiring engineering of actual humans). In each case, the closer to the human you engineer, the lower the mass and ultimately the cheaper it is (at scale).

Another example of this would be microgravity. There have been drugs that have been used to help maintain bone density, such as those used for osteoporosis, particularly Bisphosphonates, which have already been tried on ISS (although there are more powerful drugs available which haven’t been tried, such as Forteo). But exercise seems largely sufficient for the usual durations. And on a larger scale, you could try internal short-arm centrifuges. And on a larger scale, tethers for artificial gravity. But at each point, the mass overhead becomes greater. So I prefer the drug-based countermeasures if possible.

Another effect is a sense of dizziness after astronauts return from long stints in microgravity. The dizziness doesn’t last too long, but it’s feared to prevent rapid escape from the vehicle after landing (say, on Mars) if there’s a problem. I think this is a corner-case-of-a-corner-case, i.e. you have to have a survivable landing but still have to have a reason to immediately exit the vehicle AND be close enough to other help while also not being too injured too move AND you have to be so dizzy that you can’t exit the vehicle.

But let’s say that’s the only showstopper to microgravity. (and it’s not a showstopper, but let’s say it is) Another possibility is vestibular implants. Some people actually have damage to their vestibular system from disease or injury, and so they can be given an artificial vestibular implant, using external MEMS gyros attached to their head, to restore a sense of balance:

First Successful Installations of Vestibular Implants in Humans

But because the signal is now synthesized, you can now modify it. You can impose the feeling of gravity on an astronaut in microgravity, prepping the astronaut for landing. Or perhaps smoothing out the strong Coriolis effect from short-arm centrifuges. You should be able to reduce the dizziness an astronaut feels after returning to gravity.

Also, it’d allow for some crazily-immersive VR.

Now, I personally think that the brain is already sophisticated enough that we can actually train ourselves to tolerate the Coriolis forces (and this is borne out of a study from MIT), and probably also learn how to combat the dizziness that is felt upon landing (perhaps by regularly spinning around just in the air while in microgravity). So I don’t think this is purely necessary. But I do think that long-term, we need to start thinking in this direction in order to make mass human spaceflight more feasible. Human spaceflight seems intrinsically expensive because of all the overhead required for humans in space versus robots. But we can engineer ourselves, much like we developed clothing to enable living in colder climates.

Posted in Uncategorized | 19 Comments

The Slings and Arrows of Outrageous Lunar Transportation Schemes: Part 4–Propellantless Horizontal Soft-Landing Methods

As mentioned previously in Part 2 of this series, one of the key elements in establishing a lunar beachhead is developing ways to safely and affordably land payloads and people on the lunar surface. In Part 3, we discussed a way of hard-landing bulk raw materials on the surface, but in this post, I’ll focus on two related methods for horizontally soft-landing equipment and people on the Moon, while significantly reducing the amount of propellant required compared to traditional rocket-based soft-landing.

While the idea of horizontally landing on a planet with negligible atmosphere may sound odd at first, there is a method to my madness.

Horizontal Soft Lithobraking
I’ve mentioned this concept in the introduction a previous blog post1, but to recap for those who haven’t read the previous post, back around the time of the Apollo Program, Krafft Ehricke2 invented a concept for horizontally landing payloads and people on the lunar surface.

Lithobraking Slide Landers

Lithobraking Slide Landers

Ehricke’s concept involved having a lander with ski-like skids land on a very long (10s of km long) bulldozed regolith track. The lander skids would drag along the surface exchanging momentum with the lunar regolith. In this way the rocket delta-V associated with landing could be significantly diminished. Ehricke even invented a new field of engineering called “harenodynamics” to study the fluid-dynamics-like properties of regolith particles in that situation.

While probably doable, and while the concept would likely significantly increase the payload deliverable from Earth, it’s not without its share of drawbacks and challenges. Obviously, prepping such a landing strip would be no small feat. You would have to clear all boulders, level the ground as smoothly as possible, probably brake up and rake the regolith, and then you’d probably have to redo the strip after each landing, as the landers spray regolith in every direction. You’d probably also want a really, really accurate gravimetric map of the approach trajectory, and some good navigation aids, to enable the vehicle to hit the track, and hit it with the right velocity vector.  This concept would likely result in a lot of wear and tear to both the track and the lander. Not to mention getting dust on pretty much everything, and being on the “that scares me, and I’m fearless” side of sketchy. But there’s also been a decent amount of research put into it, so it might be feasible.

That’s all I’ll say about his concept, but you can read more about it here on page 27. For the rest of the post, I’d like to focus on a newer variation on the theme that I think has real potential.

MFA™ Hover-Brake Horizontal Landing
This second concept is a combination of a concept I heard from a friend at XCOR about a decade ago3 combined with some newer technology from a cool company in the Bay Area called Arx Pax. Arx Pax is probably most famous for their “Hendo Hoverboard”, which can levitate while carrying a full-weight human about an inch over a copper sheet, without using superconductors.

The Hendo Hoverboard

The Hendo Hoverboard

Their Magnetic Field Architecture (MFA™) technology works by using four “Starms” or “Hover Engines” which are rotating discs with magnets arranged on them in a way that projects most of the magnetic field down from the disc into the conductive medium you’re trying to hover over.

Arx Pax Hover Engine 3.0

Arx Pax Hover Engine 3.0

As the Hover Engines rotors rotate, they induce a circular eddy current in the conductive medium, which creates a magnetic field that pushes back on the Hover Engines. Using two pairs of counter-rotating Hover Engines at the four corners of their hoverboard, they can push against the conductive medium with enough force to lift both the hoverboard and a human rider. They can also induce lateral forces by rotating the Hover Engines relative to the surface of the conductive medium. They think there may also be ways to modify their Hover Engines to induce push, shear, and pull forces on a conductive target.4 For those of you interested in learning more, Arx Pax just announced this last week that they’re now selling both a pair of the Hover Engine 3.0 modules shown above (capable of levitating 60kg at a height of 6mm over a 12.5mm thick sheet of 6101-T6 aluminum, with a system weight of <7kg each, as per this spec sheet)5 for $9,999 for the pair, as well as a MFA™ bundle kit with smaller hover engines and all the pieces you need to make and control your own hovering device for $1,589, both of which you can now order online here.

For this lunar landing application, you would need three elements: a long conductive track, multiple Hover Engines on the horizontal lander in order to levitate the lander over the track, and a magnet array on the bottom of the vehicle to decelerate the vehicle using eddy current braking with the long conductive track. The Hover Engines can help both keep the vehicle from touching the track, or from bouncing away from it, while also keeping it centered on the track, absorbing a lot of the shocks associated with a lunar landing, and providing additional deceleration force as the vehicle slows6. As the Hover Engines approach the conductive track, the repulsive force should increase, and as the Hover Engines get too far above the track, the repulsive force should drop off, thus providing at least some natural feedback The amount of power needed to keep the Hover Engines rotating is probably modest enough to be powered by an APU like ULA is proposing for their Integrated Vehicle Fluids system7. While there still are rather demanding requirements for hitting the landing track within the right cone of velocity vectors, the non-contact nature of this landing makes it somewhat less scary. And depending on the design of the eddy current braking magnets, the braking drag may be able to be applied gradually after the vehicle has established a clean hover on the track, if so desired.

One important question is still how much up-front infrastructure this may require. You’ll still need the bulldozer to remove boulders, and level the road. But you also no need a conductive layer of some thickness. The thickness is going to be driven by the float height, gravity, number and size of Hover Engines, and the conductivity of the conductive track. For earth hovering, they used a copper track of decent thickness8. If we needed that thick of a track, that could add up relatively quickly, even if we go with lightweight aluminum instead of copper9, to get the thickness down to much more reasonable levels.

Residual Resistance Ratio
Most people are familiar with superconductors, but did you know that if you chill high purity metal conductors to cryogenic temperatures, that their resistivity also drops off dramatically? In most cases not all the way to zero, but enough to make a real difference. You need a very pure alloy (typically 99.99% or more of the base metal), fully annealed, with minimal impurities. But if you can get the metal in that condition, the conductivity can go up by 50-100x or more compared to their room temperature values. For instance, the figure below (from page 1135 of this paper) shows the conductivity vs temperature curve for high-purity aluminum. As you can see, the resistivity at room temperature is ~2.7×10^-8 Ohm*meters, but at LN2 temperatures it’s 10x lower, and LH2 temps it is ~3400x lower.

High Purity Aluminum Resistivity vs. Temperature

High Purity Aluminum Resistivity vs. Temperature

If you could chill pure Aluminum cold enough to reach ~125x the room temperature conductivity of 6101-T6 sheet (ie resistivity = .025 Ohm*m), you would need ~100µm of aluminum to conduct electricity as well as the 12.5mm thick sheet of aluminum at room temperature used for the Hover Engine 3.0 modules. From the chart that’s somewhere around 45-50K. 100µm is a lot more workable than 12.5mm. The 100µm thick layer could be thermally deposited onto a microwave fused regolith track to provide a strong mechanical backing, while keeping the initially imported aluminum mass quite low10.

Running Some Initial Numbers
If you say had a 7.5m wide track with 100µm thick aluminum, the aluminum would mass approximately 2.03mT/km of track11. So if you could hard-land a 70mT sample of such high-purity aluminum 12, and could recover approximately 2/3 of the aluminum, that would let you lay about 23km of conductive track from a single hard-landing flight. If you were using that 23km track to decelerate something from lunar orbital velocity (~1800m/s), and assumed that you were landing on ~80% of the track length13, you could get the landing Gs down to ~9Gs (over approximately 20s). That’s only a little bit more intense than a Gemini launch on a Titan II missile and a bit gentler than a ballistic reentry in a Soyuz capsule, but probably within what a healthy human could handle safely14. With two hard landings worth of aluminum under the above assumptions, you’d be able to get the G rate down to a totally survivable 4.5G deceleration (with a 46km long track).

But how much would all the Hover Engines mass to do this? Not as much as you’d think. So, if you assume that you’ve chilled the aluminum enough that 100µm of pure aluminum can conduct as much as 12.5mm of 6101-T6 aluminum can at room temperature (40-50K as mentioned above), so it behaves similarly to the nominal track thickness, that means that for landing a 20mT payload attached to a 6mT ACES/Xeus/HoverEngine lander, you’d need ~64 of the Hover Engine 3.0 modules (say 8 pods of 8 Hover Engines each), which without further weight optimization would mass around ~450kg. You could probably cut that down substantially with clever design15, but that’s already in a similar order of magnitude to the mass of the landing kit for a rocket-powered Xeus lander.

As an aside, can you see why I was so interested in having a way to land large amounts of bulk raw materials on the lunar surface? Even though there’s plenty of aluminum in the regolith, and you could eventually setup facilities that could produce many tonnes of aluminum per year, being able to land that early-on dramatically lowers the cost of landing those facilities in the first place.

How to Chill
Getting back to the topic, can we really get aluminum cold enough to enable that level of conductivity? I’m honestly not sure. It may be possible to have the bulldozer that makes the path intentionally sink the track into the ground with berms around it in a way to keep the aluminum in shadow as much of the time as possible. If you combine that with some sort of “cryogenic selective thermal coating”16 that could help keep the temperature of the track passively cool for most of the day. Will it be enough to get the track down to the desired ~40-50K? I’m not sure. According to this chart from a site with data from the LRO Diviner mission, polar sites get down to around ~50K at night, so I don’t think it’s entirely crazy at least for polar locations. At more equatorial positions, even with a properly dug trench, you’ll likely need a thicker aluminum track to make things work.

Lunar Surface Temperatures over Time at Various Latitudes (from LRO Diviner Instrument)

Lunar Surface Temperatures over Time at Various Latitudes (from LRO Diviner Instrument)

One additional knob to turn is that apparently if you have a two layer track, with a highly conductive non-ferromagnetic upper layer, and a lower ferromagnetic layer, that the ferromagnetic sheet modifies the induced magnetic field in a way that actually increases the force pushing back on the Hover Engine. So theoretically, if you did a three layer design (first the microwave sintered regolith underneath, then a layer of melted NiFe material magnetically extracted from the lunar dust, and finally a layer of thermally deposited aluminum), you might be able to get a bit more bang for the imported aluminum buck.

Multi-Use Track
One other nice thing about a track like this is that it can provide two other uses other than just propellantless landing of people and payloads. First, the track is actually a pretty impressive conductor. A 100µm thick, 7.5m wide conductor has about the cross sectional area of a 1.25in (30mm) diameter rod. That’s a lot of conductor. And if chilled to 125x the room temperature conductivity of aluminum, you’re talking about a conductor equivalent to a 14in (350mm) diameter bar of aluminum at room temperature. A conductor that big could likely conduct megawatts of power over non-trivial distances. This means that you could have solar farms and installations located along the track, which all use the track as the high-power backbone to connect themselves to the main beachhead facility. If you had a polar facility with four tracks spaced 90 degrees from each other (say earthward, anti-earthward, and along and against the moon’s orbital velocity vector), you could locate solar farms along each track in a way that half of them are always in sunlight, so you wouldn’t need a lot of power storage capacity.

Second, the tracks also would serve as ideal highways/railroads for moving stuff between the main facility and other facilities down the line. Without air, and without rolling friction, these tracks could basically function as maglev railways linking installations up and down the track. The same Hover Engine kits used for landing could form the backbone of a hover vehicle that could rapidly move very heavy payloads up and down the track at very little energy cost.

And so long as you locate a track along a “great circle” route (ie a surface track that is in a plane that intersects with the center of the Moon), the tracks could be used for all three applications (landing, transportation, and power). While the first one or two could be built using Earth resources to accelerate their availability, once you have access to lunar aluminum, you can start putting tracks like this down wherever they are convenient, and as long and wide and thick as is convenient. I even have an idea 17 for how you could use tracks like this for propellantless launching of payloads, but that’s a post for another day.

It’s probably more complicated than this. These ideas are very conceptual, and could use a significant amount of further baking. In fact, if I get a chance to flesh this out further, I’ll probably do follow on posts as Part 4.1, 4.2, etc. when I get the time. But they at least suggest that there may be ways to eventually cut the cost of delivering payloads to the Moon in-half, by getting rid of most of the propulsive landing delta-V. And for setting up heavy infrastructure on the Moon, the sooner you can get that big of a landing-cost savings, the more your overall system cost goes down.

Next Up: Slings

[Update: I still need to dig into this more, but it may be necessary to put more thought/analysis into the eddy current damping part of this idea. All of the kinetic energy effectively has to be dissipated as thermal energy (by resistance losses by the eddy currents flowing through the metal). The problem is that as the metal heats up, it becomes less conductive. So there’s probably a mass limit that a given track thickness/length can realistically handle. The math is currently above my hand-calc levels, but I’ll do a follow-on Part 4.x post if I get a chance to dig into this more.

One other interesting idea I’d like to look at is using pure Beryllium as the conductor instead of pure aluminum. Beryllium has 2x the heat capacity of aluminum (~1800J/kg*K vs ~900J/kg*K), and has a much higher melting point (~1550K vs ~933K), so it is much better as a heat sink. It isn’t as good of an electrical conductor though, with 3/2 the room temperature resistivity of aluminum, but it is only 2/3 the density, so for a given mass/length, you get approximately the same resistance, just with 3/2 the cross-sectional area for the Beryllium. It’s RRR (ratio of resistivity near absolute zero compared to room temperature resistivity) of ~2000, which while not as high as the highest grade coppers or aluminums, may still be good enough to be quite interesting. Needs more analysis though.]

Posted in Lunar Commerce, Lunar Exploration and Development, Slings and Arrows of Outrageous Lunar Transportation Schemes, Space Development, Space Settlement, Space Transportation, Technology | 29 Comments