Saturday, July 23, 2022

A Reusable SLS? UPDATED: July 31, 2022.

 Copyright 2022 Robert Clark

 The SLS is now projected to cost $4 .1 billion per flight. Because of that severe cost it is projected to only fly once per year. This can not form the basis of a sustainable Moon colonization plan. But suppose we could make the SLS reusable? It’s already known the side boosters can be made reusable as with the shuttle program. The engines on the SLS core stage were derived from the shuttle engines which were intended to be reused up to 100 times. However, since the SLS was intended to be expendable the shuttle-derived engines on the core were designed cheaper to be expendable. However, any rocket engine even an expendable in reality is reusable at least 10 times or more. This is because they have to be certified for several firings for testing purposes. This is described by the well-regarded space expert Henry Spencer:


From: Henry Spencer <>


Subject: RLV engines (was Re: X-33 Concepts: Lockheed, Mac Dac, Rockwell)

Date: Wed, 19 Jun 1996 13:03:12 GMT

In article <4q6am4$> (Andy Haber) writes:

>I think this is an area where critics can speak the loudest.  Today's 

>existing engines all leave something be be desired as true, good SSTO engines.

>This is mostly due to history.  Most engines (other than SSME's) were 

>designed for ELV's, not SSTO's.

Actually, this does not have a lot of bearing on their suitability for

RLVs.  Most ELV engines are, despite their application, reusable, because

they have to be developed and tested.  The F-1 was specified for 20 starts

and 2250s of life, the J-2 for 30 and 3750s.  Six F-1s ran over 5000s each

as part of the service-life tests.  DC-X's RL10s looked "pristine" after

20 starts; the RL10 is nominally rated for 10 starts and 4000s of firing.

>...In terms of using SSME's, sure those can used,

>although doing something to reduce the required level on maintenance on

>the existing engines is quite desirable...

Unfortunately, it probably can't go far enough.  Rocketdyne's own estimate

was that, with a *lot* of work, you could probably get SSME maintenance

costs down to $750k/engine/flight, which is unsatisfactory if you're aiming

for really large cost reductions.


If we feared danger, mankind would never           |       Henry Spencer

go to space.                  --Ellison S. Onizuka |


 Then even reusing the vehicle 10 times could result in a factor of 10 reduction of launch cost, if the maintenance cost could be kept relatively low. That quote about $750, 000 maintenance cost after a lot of work may seem low but from memory I recall it being in the range of $1 million to $2 million per engine after several years into the shuttle program.


 But how to land the SLS core? Starting the SSME’s is a complex process. Modifying them to be air-startable would not be trivial. Instead, I suggest using the method proposed for making the Centaur a lunar lander, multiple pressure-fed side thrusters for a horizontal landing. 

Robust Lunar Exploration Using an Efficient Lunar Lander Derived from Existing Upper Stages. 

 Note then that for a stage reentering to Earth broad-side almost all the reentry velocity is burned off aerodynamically just by air drag so that the stage reaches terminal velocity at approx. 100 m/s. For a stage nearly empty of fuel, this low amount of velocity could be cancelled relatively easily by pressure-fed thrusters with the thrusters running on just the residual of propellant left in the tanks.

 About the landing, there would be additional development cost for the horizontal landing thrusters.  But pressure-fed thrusters are a relatively simple technology. Compare for example the time SpaceX spent developing the Draco thrusters on the Dragon to the time developing the Merlin engine. And from discussion of the thrusters on the Starship they seem more like an afterthought compared to the cost, time, and complexity put into the Raptor engines.

 How about giving the RS-25’s on the SLS core restart capability? Again I’ll refer to the redoubtable Henry Spencer:



From: (Henry Spencer)

Subject: Re: One part Oxygen, two parts Hydrogen and BOOM!

Date: Sat, 14 Oct 2000 03:37:23 GMT

In article <>,

Pete Zaitcev <> wrote:

>> The SSMEs use "torch" igniters, little oxygen/hydrogen burners firing into

>> the preburners and chambers.  The igniters themselves are ignited by,

>> essentially, high-tech spark plugs.


>I see... obviously there cannot be a spark in a vacuum.

Not entirely true, but irrelevant -- when the igniter fires up, there's an

oxygen/hydrogen gas mixture there for the spark to travel through.

>Is the plug the reason engines cannot be restarted in orbit or

>there is more to the story?

There's nothing *fundamental* in the SSME which makes an in-space restart

impossible -- no one-shot parts or anything like that -- but it's a

complicated engine which has to be set up exactly right for a successful

start, and ground equipment (and gravity!) helps out with that.  It would

not be difficult to develop a variant which could start itself in space,

but there has been no reason to do that.


Microsoft shouldn't be broken up.       |  Henry Spencer

It should be shut down.  -- Phil Agre   |      (aka


 So likely it could be done by Aerojet, but I have no confidence they could do it in an affordable manner. Or more precisely, I have no confidence they would do it at an affordable price charged to NASA. For instance the RS-25 engine used on the SLS is derived from the SSME. It was expected to be cheaper than the SSME as it it used a lower parts counts and was not required to have the 100 times reusability of the SSME. But instead Aerojet charged more for this engine than the SSME even when accounting for inflation:

NASA will pay a staggering $146 million for each SLS rocket engine. The rocket needs four engines, and it is expendable. ERIC BERGER - 5/1/2020, 6:55 PM

 About the payload lost on reusability, a stage that goes to LEO can remain in orbit for a few orbits to come back over the landing site so minimal propellant is burned to return to launch site. 

 If we do use a large upper stage, then the SLS would not go to orbit and as SpaceX showed you would need minimal fuel burned if landed down range, and so minimal payload lost, rather than returning to launch site. However, there is then the cost of the upper stage. If it were the Ariane 5/6, the cost of the Ariane 6 being as low as $77 million, it should be even lower than that without the Ariane side boosters or upper stage. 

ARIANE 6 VS. SPACEX: HOW THE ROCKETS STACK UP The European Space Agency is planning to use the Ariane 6 for a variety of missions. ESA MIKE BROWN 1.24.2022 2:00 PM In January 2021, Politico reported that the Ariane 6 could launch for as little as $77 million. That’s a steep discount from the $177 million price tag for the Ariane 5.

  About the landing thrusters, I wouldn’t give a contract for it to any of the usual aerospace companies under NASA’s cost-plus contracts. Instead I would prefer doing it “in house”, so to speak. I was quite impressed by a team at Johnson Space Center led by chief NASA engineer Stephen Altemus developing an unmanned lunar lander for only $14 million development cost:

The Morpheus lunar lander as a manned lander for the Moon.

  The approach the NASA team used on saving costs was likely analogous to that used by commercial space in cutting costs.  No doubt also the pressure-fed engines being used rather than complex turbo-pump engines contributed to the low development cost.


 This report estimates the launch market as approx. $48 Billion per year by 2030:

Global Space Launch Services Market is projected to reach at a market value of US$ 47.6 Billion by 2030: Visiongain Research Inc October 05, 2021 09:33 ET | Source: Visiongain Ltd

 At a going rate of approx. $10,000 per kilo to LEO that would amount to 4,800 tons to orbit.  For a SLS lofting nearly 100 tons to orbit even in SLS 1 form, that’s quite a lot of launches it could take part in per year IF it could do it at a competitive price. If  it could do 10 reuses, that could bring the price down to $400 million per flight, or $4,000 per kilo, about the price of the reusable F9 when new, or a bit more than $3,000 per kilo of the used F9. But IF it could do 20 reuses, within the capabilities of some expendable engines, it would be $2,000 per kilo which would beat even the F9 used reusable price.

 In point of fact though the first four SLS vehicles would all use original Space Shuttle engines. Then likely each has dozens of uses left in their operational lifetimes:

Apr 6, 2021
RS-25 Rocket Engines Return to Launch NASA’s Artemis Moon Missions.


   Robert Clark


Could the SpaceX SuperHeavy be a SSTO?

 Copyright 2022 Robert Clark

 Running some  numbers for the SuperHeavy+Starship launcher, I was surprised to get that an expendable SuperHeavy alone could be SSTO with quite high payload.  Wikipedia gives the propellant mass of the SuperHeavy as 3,400 tons, but does not give the dry mass. We can do an estimate of that based on information Elon provided in a tweet:

 This is for a stripped down Starship, no reusability systems, no passenger quarters, and reduced number of engines.  But this could not lift-off from ground because of the reduced thrust with only 3 engines plus being vacuum optimized these could not operate at sea level. So up the number of engines to 9 using sea level Raptors. According to wiki the Raptors have a mass of 1,500 kg. So adding 6 more brings the dry mass to 49 tons, call it 50 tons, for a mass ratio of 25 to 1.

By the way, there have been many estimates of the capabilities of the Starship for a use other than that with the many passengers, say 50 to 100 , to LEO or as colonists to Mars, for instance, such as the tanker use or only as the lander vehicle transporting a capsule for astronauts for lunar missions.  But surprisingly they all use the ca. 100 ton dry mass of the passenger Starship. But without this large passenger compartment it should be a much smaller dry mass used in the calculations. For instance, the Dragon 2 crew capsule dry mass without the trunk is in the range of 7 to 8 tons for up to 7 astronauts. So imagine a scaled up passenger compartment for 50 passengers or more. That passenger compartment itself could well mass over 60 tons.

 So the dry mass estimate of a stripped down, expendable, reduced engine Starship of 40 tons offered by Elon does make sense. 

Based on this, an expendable Starship with sufficient engines for ground launch could be SSTO:

the ISP of the Raptors for both sea level and vacuum-optimized versions have been given various numbers. I’ll use 358 s as the vacuum ISP of the sea level Raptor. For calculating payload using the rocket equation, the vacuum Isp is commonly used even for the ground stage, since the diminution in Isp at sea level can be regarded as a loss just like air drag and gravity loss for which you compensate by adding additional amount to required delta-v to orbit just like the other losses.

 Then 3580ln(1 +1200/(50 + 50)) = 9,180 m/s sufficient for LEO.      

 But as of now, SpaceX has no plans of making the Starship a ground-launched vehicle. So we’ll look instead at the SuperHeavy. For an expendable version with no reusability systems, we’ll estimate the dry mass using a mass ratio of 25 to 1, same as for a ground-launched expendable Starship. Actually, likely the Superheavy mass ratio will be even better than this since it is known scaling a rocket up improves the mass ratio. So this gives a dry mass of 136 tons. Then the expendable SuperHeavy could get 150 tons to LEO as an expendable SSTO: 3580ln(1 + 3,400/(136 + 150)) = 9,150 m/s, sufficient for LEO.

 But what about a reusable version? Reusability systems added to a stage should add less than 10% to the dry mass:


From: (Henry Spencer)


Subject: Re: The cost (in weight) for Reusable SSTO

Date: Sun, 28 Mar 1999 22:37:10 GMT

In article <kemJ2.876$>,

Larry Gales <> wrote:

>An SSTO with a useful payload using Kero/LOX is easy to do -- provided that

>it is *expendable*.  All of the difficulty lies in making it reusable...

There are people who are sufficiently anti-SSTO that they will dispute the

feasibility of even expendable SSTOs (apparently not having read the specs

for the Titan II first stage carefully).

>   (1) De-orbit fuel: I understand that it takes about 100 m/s to de-orbit.

That's roughly right.  Of course, in favorable circumstances you could play

tricks like using a tether to simultaneously boost a payload higher and

de-orbit your vehicle.  (As NASA's Ivan Bekey pointed out, this is one case

where the extra dry mass of a reusable vehicle is an *advantage*, because

the heavier the vehicle, the greater the boost given to the payload.)

>   (2) TPS (heat shield): the figures I hear for this are around 15% of the

>orbital mass

Could be... but one should be very suspicious of this sort of parametric

estimate.  It's often possible to beat such numbers, often by quite a large

margin, by being clever and exploiting favorable conditions.  Any single

number for TPS in particular has a *lot* of assumptions in it.

>   (4) Landing gear: about 3%

Gary Hudson pointed out a couple of years ago that, while 3% is common

wisdom, the B-58 landing gear was 1.5%... and that was a very tall and

mechanically complex gear designed in the 1950s.  See comment above

about cleverness.

I would be very suspicious of any parametric number for landing gear which

doesn't at least distinguish between vertical and horizontal landing.

>   (5) Additional structure to meet loads from differnet directions (e.g.,


>        takeoff, semi-horizontal re-enttry, horizontal landing).  This is


>        guesswork on my part, but I assume about 8%

Of course, here the assumptions are up front:  you're assuming a flight

profile that many of us would say is simply inferior -- overly complex,

difficult to test incrementally, and hard on the structure.

>I would appreciate it if anyone could supply more accurate figures.

More accurate figures either have to be for a specific vehicle design,

or are so hedged about with assumptions that they are nearly meaningless.


The good old days                   |  Henry Spencer

weren't.                            |      (aka

 The 15% mentioned for thermal protecton(TPS) is for Apollo-era heat shields. But the PICA-X developed by SpaceX is 50% lighter so call it 7.5% for TPS.  And for the landing gear ca. 3%, but with carbon composites say half of that at 1.5%.

  But this would put the reusable payload at ca. 136 tons which is in the range of 100 to 150 tons of the full two stage reusable vehicle!

 How is that possible? A reusable multistage vehicle has a severe disadvantage. The fuel that needs to be kept on reserve for the first stage to slow down and boost back to the launch site subtracts greatly from the payload possible.  But for a reusable SSTO it can remain in orbit until the Earth rotates below until the landing site is once again below the vehicle.

   Robert Clark

Could asteroidal impacts be the cause of the coronal heating problem?

 Copyright 2024 Robert Clark   A puzzle in solar science that has existed for 150 years is the corona heating problem: Why is the sun’s coro...