On-orbit refueling of the SpaceX BFR upper stage is necessary for trips to the Moon and Mars. In the presentation at the "International Astronautical Congress (IAC) in Adelaide, Australia, SpaceX CEO and Lead Designer Elon Musk (provided) an update to his 2016 presentation regarding the long-term technical challenges that need to be solved to support the creation of a permanent, self-sustaining human presence on Mars."

In the SpaceX video Making Life Multiplanetary after 23:40 the fuel transfer process between similar-looking upper stages is illustrated, with the tanker on the left and the ship on the right. Musk explains:

"They would use the same mating interface that they use to connect to the booster on liftoff… and reuse the propellant fill lines that are used when… the ship is on the booster. […] And then to transfer propellant it becomes very simple, use control thrusters to accelerate in the direction that you want to empty… transfer propellant very easily… from the tanker to the ship."

This is analogous to "gravity feed" replacing gravitational acceleration with propulsive acceleration, but I'm guessing that the magnitude of the acceleration would be an order of magnitude or more lower.

  1. Have I understood this concept correctly?
  2. Using some estimate of a control thruster's acceleration, how long would this pump-less, pressure-driven-less milli-gravity feed take? An hour? A day? Filling rocket tanks before liftoff takes perhaps tens of minutes with all the benefits of a refueling infrastructure on the ground. With only milli-gravity and no pressure assist, wouldn't this take a lot longer?

from https://youtu.be/S5V7R_se1Xc?t=1420

  • 1
    $\begingroup$ I don't think this concept was thought through very well. $\endgroup$ Commented Oct 6, 2017 at 16:49
  • 4
    $\begingroup$ @OrganicMarble if it was anybody else, I could dismiss it like that easily. But this guy usually has thought things fairly well through by the time he gets up on stage in front of a bunch of experts and video cameras. He may gloss over some rough edges, but usually doesn't say "wrong stuff". $\endgroup$
    – uhoh
    Commented Oct 6, 2017 at 17:05
  • $\begingroup$ I may be biased by my generally low opinion of the feasibility of all this BFR stuff. $\endgroup$ Commented Oct 6, 2017 at 17:23
  • 1
    $\begingroup$ @JCRM "We're big enough to take a few insults." - Lt. Cdr. Scott $\endgroup$ Commented Mar 16, 2018 at 20:12
  • 1
    $\begingroup$ I was referring to the typical SpaceX goal evolution, not yourself @OrganicMarble $\endgroup$
    – user20636
    Commented Mar 16, 2018 at 22:06

1 Answer 1


The reason for the acceleration is not so much to cause the transfer of the liquids, but to get the liquid to be at one end of the tank where the pump inlet is. Remember in zero-g, the liquid is not going to stay at one part of tank but will slosh all over and mix with "air" or whatever gas is in the tank along with the liquid fuel. Without a little bit of "G", even just a few milli-g's, you'll never get all the liquid to reach the suction pump without employing a bladder or collapsible tank. In that context, milli-g is clever, but SpaceX is not the first to think of it.

Milli-g transfer with no pumps, assuming equal sized source and receiving tanks, can only pressurize the receiving tank to 1/2 the pressure of the source tank plus whatever the "milli-g" force is. If you want more pressure, you must use a pump of some sort.

So milli-g is just one piece of the puzzle. The clever-er part is re-using some of the plumbing and mating hardware from the thrust phase for the refuel transfer.

Note that military jets sometimes use a "buddy tanker" system, aka one F/A-18 can carry fuel tanks for another F/A-18 to top off in flight. The second jet continues on the mission, the "buddy tanker" goes home. But the buddy tanker needs to have special tanks strapped under the wings, in lieu of other weapons. SpaceX is proposing to reuse the flight hardware for tanker duty instead of creating a special tanker. This continues the theme of re-use throughout the SpaceX ecosystem to reduce cost, risk and design effort.

  • $\begingroup$ Excellent explanation, thank you! OK this makes much more sense. In a similar way to what is being described in this answer, the thrust discussed is really for ullage. And also as is likely there, it may or not necessarily be above 0.001 g to fall under the "milli-g" term. $\endgroup$
    – uhoh
    Commented Mar 16, 2018 at 5:26
  • $\begingroup$ fyi I've just asked What do the terms milli-g and microgravity mean? How are they used? $\endgroup$
    – uhoh
    Commented Mar 16, 2018 at 5:31
  • 2
    $\begingroup$ I think they are proposing a specialised "tanker" version of the BFS eventually. Similar basic structure, but with unneeded components left out and consequently can carry more fuel. forum.nasaspaceflight.com/index.php?topic=43986.0 $\endgroup$ Commented Mar 16, 2018 at 11:47
  • $\begingroup$ Since the fuel and oxidiser are liquids the argument against milli-g transfer doesn't really hold up. If you waited long enough the "downhill" tank would contain all the liquid and the remainder of both tanks would be full of gas at more-or-less equal pressure, but almost all the mass would be the liquid. They might well need a pump to hurry things along though. $\endgroup$ Commented Mar 16, 2018 at 11:49
  • 1
    $\begingroup$ @OrganicMarble The more I think about it, vapor management will have to be a whole complex system, and I doubt any of it goes overboard (if avoidable). The RCS will be fueled by the vapors and you wouldn't want your RCS needs to be interfering with your ullage pressure goals so there has to be additional storage for gaseous propellant and an ability to pump it back and forth. Given that, it seems simple to maintain appropriate ullage pressure to facilitate the transfer. $\endgroup$
    – Saiboogu
    Commented Mar 18, 2018 at 2:29

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.