In short: BFS can, even if not capable of SSTO with a payload get 40 tons or so to orbit with suborbital refuelling.

While we are waiting for the big reveal - I will flesh out an idea I've mentioned before.

I had intended for this post to have results of simulations and optimisations - this has not happened.

In order to simplify the numbers, the following assumptions are made:

1) BFS is capable of SSTO to LEO with 0 tons of payload and safely returning to the surface of earth.

_{This implies a number of things about the BFS that were explored more fully in that thread. Some changes in these directions seem to have been recently announced.}2) BFS is capable of landing safely and accurately +-1m.

3) BFS can transfer propellant at 100t/min to another similar vehicle, as is reasonable given the likely pressures in the tanks,

and flow in the pipework even without pumps. this is around the normal groundside fill rate of F9, proportionately. (see above

thread).

4) BFS can reenter fine at considerably lower velocity than nominal.

5) Enough ground support to launch two BFS is available, and one droneship.

In short, two BFS take off, on a slightly more lofted than normal trajectory, and at around LEO-2700m/s, with 100 tons of

propellant left each, match courses closely, and then cut engines.

They then proceed to transfer propellant so that one has a hundred tons extra, and the other only enough to land.

This naively gives the first BFS around 50 tons of extra propellant at LEO, or around 40 tons payload to LEO.

With effective gravity being around 3.5m/s^2 (inc RCS), although a somewhat more lofted trajectory allows you time - that time is not

infinite, and is directly traded with making the impact of the tanker on the atmosphere harder.

At the nominal rate of 100 tons/minute, it will take around a minute to transfer the propellant. Add another minute for slow

approach and 10s for separation, and you’re at 450m/s delta-v required on the tanker, if you need to reduce its velocity all the

way to the nominal vertical interface velocity.

The orbital vehicle is not affected by this, as it can miss the atmosphere with minimal cosine losses with a slight change in

thrust vector.

If you need to reduce the tankers interface vertical velocity all the way to normal, that takes the effective payload of the orbital vehicle down to around 30 tons.

‘Impossible’ ?

Routine testing must have already developed the ability to hit a nonmoving surface accurately and on target, at a terminal deceleration of several G, and a final acceleration of 1G, in the face of winds. It cannot have an error of more than several m/s.

F9S2 ignites the engine for boostback burns within 10 seconds.

The BFS RCS can do landing in 60mph winds, and RCS clusters have been mentioned as ‘10 ton’, ‘milligee’ several times,

which is consistent with a 0.5m/s^2 capability at 180 tons mass.

In more detail:

* ‘Park’ the BFS as you would for landing, 20m offset, pointing the opposite direction, at 1m/s together, and 5m apart in Z.

* At 0.5m/s^2 bring the vehicle axis into line.

* Mate two vehicles, slowing as you close.

* Apply ullage thrust.

* Dump pressurant partially out of orbital BFS, and use pressure to transfer 100 tons in a minute, around the same rate as it

is done on the ground. (the pipe diameter and length is consistent with this being easily possible with no pumps and 2

bar head).

* Undock and translate to clear the engines.

* One BFS ascends to orbit.

One lands on a drone ship.

The performance numbers for using this are similar to

the SSTO BFS thread.

If we assume launches cost $5M for two(**), then the numbers break down something like:

`Orbit Delta-v Capacity Launch Cost to SpaceX`

LEO 0 40 1 5

ISS 400 30 1 5

gto(*)1390 27 2 10

GTO 2500 17 3 15

TLI 3200 15 4 20

Mars(*) ? 150 30 150

*)Typical SpaceX reusable ‘GTO’ launches with reuse, nominal BFS mission to Mars.

**) Less fuel, no need for stacking, is designed to operate from much closer to a flat pad than BFR, simplified operations.

The count goes up lots for higher orbits, and it might even be cheaper to use a disposable BFR than to do this for Mars.

Is two years delay waiting on BFR worth $150M?

You’re spending $150M anyway on leaving the BFS on Mars, and more on the payload.

Numbers in this post are approximate, and if the arbitrary LEO-2700m/s is not optimal, there would be some margin.

I have not detailed the landing fuel required, as that is included in the 'can get to orbit and return' and affects only the 'dry' mass.

I have been assuming the total mass is of the order of 85 tons - 85 vs 100 tons does not change the numbers much.

If you believe it won't work, please supply specific reasons that are ideally backed by numbers, and not simply assuming BFS can't work anyway, or that because this scheme is silly it can't work. I am quite aware that this is a silly idea, and unlikely to occur in practice.