An Updated Propellant Depot Taxonomy Part VI: Roving Depots

In the last post in this series, I discussed the idea of large-scale depots for human spaceflight applications, which operate in fixed, low-orbits. While the final post in this series will investigate human spaceflight depots that operate in fixed higher orbits1, in this post I want to talk about situations where you want to refuel with large amounts of cryogenic propellants in an orbit where a permanent depot doesn’t make sense — in temporary high-elliptical orbits on your way into/out of a deep gravity well like Earth’s. The background on these depots involves getting into the weeds on some orbital dynamics, but I’ll try to keep it as understandable as possible for the layperson2.

Orbital Dynamics Background for Roving Depots

For an outbound interplanetary mission, there are two obvious places to do refueling — one at the first safe stopping point after leaving the planetary body3, and other is in an orbit that’s just shy of leaving the planetary body’s gravity well4. For a high-thrust/low-Isp departure, like you’d have with a chemical or nuclear thermal rocket, this would be a highly elliptical orbit, with the periapsis as low as possible5, and a apoapsis as high as practical6. There’s just one problem — in order to leave on an interplanetary trajectory from a highly elliptical orbit, that elliptical orbit’s periapsis has to be in the right place7, going in the right direction8, at the right time9. As you can probably guess, the odds of a specific highly elliptical orbit for one interplanetary departure trajectory lining up with another specific interplanetary departure trajectory at a specific point in the future is really, really low.

An illustration of hyperbolic departures, showing the locus of periapses/injection points on the opposite side of the earth from the desired departure asymptote (credit: Bate, Mueller, and White)

While I was working with Mike Loucks and John Carrico on our three-burn departure orbital dynamics papers (reviewed previously on Selenian Boondocks here, here, and here), I realized that there might be a way around this problem of elliptical orbits not lining up for future missions, especially if you had a reusable, mobile depot capable of both entering into and then returning from that highly elliptical orbit.

Illustration of the 3-Burn Departure method described in the two linked AAS papers. An interplanetary mission stack (1) starts in a low-orbit depot’s orbit after topping-up, (2) performs a burn to enter a highly-elliptical phasing orbit, (3) performs a plane change maneuver at apogee to align the departure plane, and (4) performs the final interplanetary departure burn when the spacecraft is back at perigee in the right place at the right time.

One or more such “roving depots” could work with reusable tankers10 and a low-orbit depot to preposition propellants into a highly-elliptical orbits for specific departure missions, with the mission stack for a given mission only having to travel out and rendezvous with the depot on the last time the departure highly-elliptical orbit plane lines up with the low-orbit depot’s plane. Once the mission stack has refueled, and left the roving depot (on its way to doing its final plane change and departure burn), the roving depot can return to the low-orbit depot the next time its plane lines up with the low-orbit depot, enabling it to be refueled and prepared for its next mission.

Conceptual Illustration of a reusable space tug. A roving depot would likely look similar, except possibly with more tankage and robotics capabilities. (Credit: Commercial Space Development Company)

Roving Depots

Application: Significantly enhancing the performance benefits of using a low-orbit depot, by providing one last chance to top off before heading out into interplanetary space. This can range from topping up a smallsat interplanetary stage to assembling and fueling a large interplanetary human mission or multi-ship convoy.

Location: As described earlier, roving depots are mobile depots that start a given mission at a low-orbit depot, maneuver into a highly elliptical orbit that is targeted for a specific interplanetary departure, and then returns to the low-orbit depot after the refueling operation is over.

Size: Depends strongly on what is being refueled. These could be as small as 100’s of kg for refueling a smallsat launcher interplanetary stage, up to 10’s to 100’s of mT for refueling larger human spaceflight missions.

Propellant Types: The propellant type for a roving depot will be driven by the propellant type used for the mission stack stage that is performing the interplanetary injection. For smallsat interplanetary stages, this is almost certainly storable, while for human spaceflight missions this would likely be LOX/LH2 or LOX/CH4.

Other Considerations

  • For roving depots and tankers operating around planets with an atmosphere, you’ll almost always want to use some form of aerobraking or aerocapture system11. Because you want your highly-elliptical orbit to have a relatively low periapsis for departure performance reasons, it only takes a little burn at apoapsis to lower the periapsis far enough for practical aerobraking/aerocapture.
  • Something you may have noticed is that the distinction between a roving depot and a tanker is sort of blurry. As I see it, they exist in sort of a continuum with expendable tankers on one end and reusable roving depots on the other extreme. The key differences are that roving depots would be more likely to have more significant propellant cooling (active and passive) capabilities, likely be designed to handle a wider range of client vehicles, and likely carry a lot more rendezvous, prox-ops, and grappling hardware.
    • I think the tankers that top off a roving depot in its mission orbit would likely be just minimalist upper stages with an aerocapture system, with as much of the smarts and complexity offloaded to the roving depot as possible. Minimizing the dry mass per unit propellant hauled from the low-orbit depot to the roving depot.
    • I think the roving depots, since they move less would strike a balance of complexity/robustness between the tanker and a fixed depot. You don’t want to go too crazy by throwing dry mass at problems like RPOD12 reliability, and mission robotic flexibility, but you want to be able to make the tankers as dumb and lean as possible, by offloading capabilities to the roving depot as much as possible.
  • This idea of roving depots can be combined with in-space assembly/manufacturing in the highly-elliptical orbit, sending up parts/materials/propellant for the mission every few months when the low-orbit depot lines up with the departure plane, assembling the overall stack, and then only sending the mission crew up on the last time the low-orbit depot’s plane aligns with the departure plane.
  • For chemical propulsion, travel to interplanetary destinations like Mars and Venus typically is only feasible once every 1.5-2yrs (depending on the synodic period). The use of roving depots can allow you to spread out the propellant launches for missions to a planet like that over a longer period of time. Instead of having to launch all the propellant and hardware for a given “launch season” all in the few months leading up to that season, you may be able to set up multiple roving depots, aligned for that departure opportunity, and then top them up over the course of a year or more. This allows a much smaller fixed low-orbit depot to support a lot more mission capacity than you would otherwise think, because the low-orbit depot wouldn’t need as much surge capacity, since you could likely plan things in advance.
    • This also suggests to me that for busy planetary systems like Earth, if roving depots take off as a concept, they’d likely significantly outnumber the low-orbit fixed depots.
  • One drawback to using roving depots and these highly elliptical parking orbits is that you end up putting a lot more van allen belt passes on your hardware than you would otherwise. While you typically won’t have your crew onboard until late in the process, your electronics, especially on your roving depot, will take a lot more radiation than it might otherwise.
  • On the plus side, you spend a lot less time in LEO where MMOD13 issues are worse, so your mission hardware doesn’t need as much MMOD shielding for how long it is in orbit. Additionally, since you spend most of your orbit far away from the planet, these highly-elliptical phasing orbits tend to be much easier for long-term cryogenic thermal storage.
  • One other consideration is that the longer you spend in the phasing orbit, the more orbit adjustment maneuvers you’ll need to perform. This may put some limitations on how long you can practically build things up in a phasing orbit for a given mission. If you’re just pre-staging propellant long in advance for a large convoy mission, you may be able to let your orbit drift a bit, and only trim things up shortly before departure, but we’d need to run the numbers on how practical that is. This is more of the case for orbits with very high apogees/long periods, especially in multi-body systems like the Earth/Moon system where, lunar perturbations become more of a problem with high apogee phasing orbits. In theory, it may be possible to craft an orbit that starts in a lower, shorter-period phasing orbit initially, but that then boosts up to a higher phasing orbit shortly before the final refueling of the mission stack. Long story short, there are lots of knobs to twist on optimization14.
  • One other problem I’m handwaving away as probably solvable is the complexity of rendezvousing with a roving depot in a highly-elliptical orbit. This will definitely be a lot different than the relative dynamics of rendezvous with objects in LEO. Though fortunately this should be pretty similar with trying to rendezvous with a facility in NRHO, so it’s something that probably has had a lot of recent thought put into it.
  • Roving depots around planets with lots of moons will definitely be more challenging from an orbital dynamics standpoint, but could be really enabling for missions to the gas giant planets. Especially if you want to do return trips, and don’t have something as cool as an Epstein Drive available yet. In some ways this reminds me of the idea of using base camps and/or storage depots when planning expeditions up mountains, or the early Antarctic expeditions. Once you have the right building blocks in place, there’s a lot you can do.

When you think about what you can do with the combination of low-orbit depots and roving depots on both ends of the mission, especially supported with ISRU and reusable launch, you can actually do very large and capable missions without needing super-heavy lift vehicles. It’s kind of amazing what you can do with a refuel early, refuel often, reusable space transportation architecture.

Next Up An Updated Propellant Depot Taxonomy Part VII: Human Spaceflight Fixed Depots (High-Orbit)

The following two tabs change content below.
Jonathan Goff

Jonathan Goff

President/CEO at Altius Space Machines
Jonathan Goff is a space technologist, inventor, and serial space entrepreneur who created the Selenian Boondocks blog. Jon was a co-founder of Masten Space Systems, and the founder and CEO of Altius Space Machines, a space robotics startup that he sold to Voyager Space in 2019. Jonathan is currently the Product Strategy Lead for the space station startup Gravitics. His family includes his wife, Tiffany, and five boys: Jarom (deceased), Jonathan, James, Peter, and Andrew. Jon has a BS in Manufacturing Engineering (1999) and an MS in Mechanical Engineering (2007) from Brigham Young University, and served an LDS proselytizing mission in Olongapo, Philippines from 2000-2002.
  1. I’ll be honest that I’m currently ambivalent about if higher-orbit depots make sense other than in the Earth/Moon system, but more on that in the next post.
  2. I’ve spoken with my partners at SEE about the idea of putting together some cool explainer videos with animations, to better illustrate some of these ideas, but I wanted to get the basic concept down in writing first
  3. as was discussed previously in Part V
  4. Basically the last place you can refuel on the way out of town
  5. To get the maximum benefit from the Oberth Effect when you do your final departure burn
  6. so you’re traveling just under escape velocity when you get to the periapsis
  7. i.e. somewhere on the locus of periapses/injection points
  8. i.e. with your plane aligned with your desired departure asymptote vector, and your orbit heading into the locus of periapses heading in the direction of the departure asymptote
  9. i.e. with your orbit hitting the periapsis at the desired time in the departure window
  10. Potentially simplified versions of the roving depot, or upper stage derivatives.
  11. This is one of those areas where Magnetoshell Aerocapture could potentially be really, really awesome if it pans out. Having a lightweight, rapidly throttleable aerocapture system that leverages and benefits from systems a roving depot would already need could be really interesting. But if it doesn’t pan out, there are plenty of other more boring and proven aerocapture techniques that could be used
  12. Rendezvous, Proximity Operations, and Docking
  13. Micro-Meteorites and Orbital Debris, ie natural and manmade debris that could collide with a spacecraft and damage it
  14. This would be a fun topic for a Master’s Thesis or PhD dissertation
Jonathan Goff

About Jonathan Goff

Jonathan Goff is a space technologist, inventor, and serial space entrepreneur who created the Selenian Boondocks blog. Jon was a co-founder of Masten Space Systems, and the founder and CEO of Altius Space Machines, a space robotics startup that he sold to Voyager Space in 2019. Jonathan is currently the Product Strategy Lead for the space station startup Gravitics. His family includes his wife, Tiffany, and five boys: Jarom (deceased), Jonathan, James, Peter, and Andrew. Jon has a BS in Manufacturing Engineering (1999) and an MS in Mechanical Engineering (2007) from Brigham Young University, and served an LDS proselytizing mission in Olongapo, Philippines from 2000-2002.
This entry was posted in Commercial Space, ISRU, Launch Vehicles, Lunar Exploration and Development, Mars, NASA, NEOs, Orbital Dynamics, Propellant Depots, Space Development, Space Exploration, Space Transportation, ULA, Venus. Bookmark the permalink.

10 Responses to An Updated Propellant Depot Taxonomy Part VI: Roving Depots

  1. This does indeed blur the line between tanker and depot. I’ve been using the following taxonomy to keep things straight:

    1) Lift tankers (LTs), which bring prop up from Earth’s surface.

    2) Accumulation tankers (ATs), which sit in VLEO and receive lift tankers’ prop until there’s enough for a particular mission.

    3) High-orbit tankers (HOTs), which move from VLEO to whatever energy orbit is needed to top-off a payload mission.

    These roles fit very nicely for a Starship tanker, where one vehicle could assume all three roles, but there’s nothing to prevent different vehicles from assuming fixed roles in the architecture. I do agree that HOTs would need to be able to aerobrake or reenter to be reasonably efficient.

    That said, I’m a bit skeptical about HEEO HOTs if they’re in ~LEO+2500m/s orbits, for a couple of reasons:

    a) The Van Allen Belt problem is a fairly big deal, and it’s a huge deal if you’re talking about human missions.

    b) Because of the VAB radiation problems, you’d like to make the process of refueling extremely efficient, so that there’s a very good chance of completing it in one orbit. But the amount of time for RPOD near apogee in an HEEO is fairly limited.

    Instead, it seems to me that it might make more sense to go with an L1 or L2 halo orbit (NRHO being a likely candidate) instead. You’re obviously trading a fair amount of propellant efficiency in terms of getting the HOT to the target orbit, but you’re gaining a lot of operational simplicity for the actual refueling operation. RPOD is quite simple. You can take as much time as you need for actual fueling and checkout. Crews can be stood off in something like the Gateway during fueling ops. Then, when everything’s good to go, it costs about 450m/s to do a fast lunar flyby into a TEI, which can then be used to hit a broad range of hyperbolic exit trajectories after perigee.

    Another possibility: an extremely high-energy HEEO. I’m thinking of something like C3=-2, which is pretty much what TLI does, but timed so that interaction with the Moon is unlikely for a couple of orbits. This would give the HOT and the payload mission several days for RPOD and refueling / checkout ops.

    One other thing to consider: accident debris. Fueling accidents have a decent chance of producing real-live explosions, which will be massive debris-generating events. That’s not a huge problem in VLEO, where debris will decay fairly quickly. It’s not a huge problem in an L2 halo orbit, because the debris will mostly clear the refueling target fairly quickly, and be flung either out of cis-Earth space or onto the lunar surface. But an HEEO is a disaster for an accident, because you’ll wind up with a lot of debris that has a higher perigee than the original orbit, putting it into orbits that are stable for centuries or millennia.

    Using the C3=-2 HEEO wouldn’t do as well as a halo orbit from a debris-mitigation perspective as a halo orbit, but you could count on everything eventually being flung away by the Moon. It might take a while, though.

  2. gbaikie says:

    If the SpaceX Starship is refueled at highly-elliptical orbit can it get to Mars with significantly shorter travel time?

  3. Jonathan Goff Jonathan Goff says:


    Possibly. When messing around with the porkchop plot, there are often shorter trips if you’re willing to deal with a higher departure and arrival Vinf. The problem is I don’t know how much higher of an arrival Vinf Starship can realistically handle, and I don’t remember off the top of my head if there are shorter trips where the arrival Vinf is still reasonable but using a higher departing Vinf. So the answers is probably “yes, to a point.”

    But another interesting point is that you might not need as big of a stage to send a 100-150mT crew cabin to Mars. If you’re using LOX/LH2, and refuel in LEO and from a roving depot, I think you could make that happen with a much smaller, lighter transfer stage.


  4. There’s nothing that says that you can’t split some of your extra delta-v between increasing your departure v∞ while also reducing your arrival v∞ with a braking burn.

    Figuring out where the optimal split is between the two isn’t really doable with publicly available porkchop plotters, unless one is willing to fight with GMAT (which I’m not). But I’d be a little surprised if you couldn’t hit 120 days if you had a completely full Starship, payload 100t, starting from a C3=-2 HEEO. It’s a huge number of lift tanker launches, laddered together in extremely painful ways, but if you really want to get there fast…

  5. gbaikie says:

    “..If you’re using LOX/LH2, and refuel in LEO and from a roving depot, I think you could make that happen with a much smaller, lighter transfer stage.”
    I think NASA should do 3 crew only to Mars in 3 months, and use chemical rockets to do non-hohmann planetary transfer- and not need high delta-v course correction when get to Mars distance.**

    “But another interesting point is that you might not need as big of a stage to send a 100-150mT crew cabin to Mars. ”
    With Starship I thinking of hohmann and arriving at Mars distance and requiring a lot delta-v. And/or requiring Starship’s aerobraking ability. So Hohmann with patched conic but better than 6 months, say 5 months or less??

    ** This would not need heat shield, but craft doesn’t go to low Mars orbit {or the surface} but instead gets to Mars high orbit and crew gets on another spacecraft which can get crew to the Mars surface. And this other craft gets to Mars, using hohmann.

  6. Jonathan Goff Jonathan Goff says:


    I’m skeptical it would be worth it, but agree that in theory it’s probably feasible. As you put it, throwing some delta-V at nulling out a high arrival Vinf is possible. But you’re also right it would be a lot of tanker flights.


  7. There’s also been some discussion of modules that are kind of a cross between a depot/tanker and an orbital booster stage. Not sure if this is orthogonal to your taxonomy or not; it’s a roving tanker, but it’s also a reusable kick stage. The name “pusher/tanker” had been discussed.

    The basic conops is that a payload vehicle does RPOD and tops itself off from the pusher/tanker. Then the pusher/tanker boosted the payload vehicle into the highest energy Earth orbit it could manage with its remaining prop. It would then detach and aerobrake back into LEO. Meanwhile, the payload vehicle would use the next perigee to do its interplanetary injection burn.

    The main advantages to this are twofold:

    1) You get the usual staging advantages.

    2) It’s a very safe conops, because there’s only one RPOD and refueling, which occurs only in LEO, with the sole remaining risk to the payload being undocking after the boost. That makes it extremely likely that you only incur one extra VAB transit, and you can manage boil-off quite precisely.

    The main disadvantage is that it’s a weird geometry. You can do nose-to-nose docking/fueling with an eyeballs-out boost, or you can to do nose-to-tail docking, but then the nose has to avoid the payload ship’s engines, which gives you an odd fairing problem to aerobrake back into LEO.

    It’s still a lot of prop, but it’s considerably less than if you had to ladder-fuel a free-flying tanker/depot into a high energy orbit.

  8. TRM,

    Yeah, I’d consider that a variant on what I was discussing here. It’s an interesting idea. Definitely less capable, but as you say, it’s also simpler in some ways. Definitely worth keeping in the tradespace.


  9. Greg Hullender says:

    Sorry if I’m a little confused here, but are you talking about fueling the depot in LEO and then boosting it into HEEO or are you talking about putting the depot into HEEO in the first place and fueling it there? The former seems to make the most sense to me, although it means the depot won’t be completely full when it reaches that orbit. (So just make the tanks bigger.)

    The latter seems to add all kinds of complications. Not only do tankers have to spend more fuel to match orbits, I think it greatly reduces the number of launch opportunities for tankers, since the orbital period is so long. Further, having matched orbits, the tankers would have to do aerobraking after a long (~10 days) orbit, meaning this would tie up a lot more tankers than simply fueling in LEO would.

    If you could boost a full depot using a different propulsion system (e.g. ion drive), that would be the biggest win–fueling the depot entirely in LEO but still putting it into HEEO for the payload mission to dock with. I don’t think we’ve got any ion engines up to that task, though.

  10. Jonathan Goff Jonathan Goff says:


    For the roving depot concept, yes you would top the depot up in LEO, but as you notice, you’ll burn about half that prop getting out to the HEEO. If that’s enough for you mission you can just leave it there. But if it’s not enough for your mission, you could fly subsequent tankers to rendezvous with it and fill it up. It’s true that you’re only going go have opportunities every say few weeks (depending on the apogee of your HEEO orbit), but that should be fine, most likely you’ll only need a few tanker flights to get there, and yes the tankers will have to aerobrake later, which means you’ll have to have a larger number of tankers in play. But at the end of the day, that’s the cost of getting a lot of prop pushed uphill into an HEEO.

    You could definitely move the depot and/or tankers out to HEEO using a high-power EP system. It’s a trade I haven’t really run the numbers on, but when I discussed this idea with one of the EP tug companies, that was exactly their concept. LEO depot, tug the roving depot uphill to a staging orbit, tug up tankers if needed, and then use either aerobraking or the tug to bring back the empty roving depot when you’re done. It’s all just a question of running the numbers on which approach is most affordable. Which is going to depend on things like the unit cost of the EP tugs, how fast they can do the mission, how many missions you can amortize their development cost over, how expensive chemical tankers would be, how cheap propellant in LEO gets (either based on RLVs driving down earth-to-LEO costs, Lunar ISRU, or better yet, both).


Leave a Reply

Your email address will not be published. Required fields are marked *