NASA's Space Launch System - an All-Liquid Alternative
by Ed Kyle, Updated 03/12/2011
An "All-Liquid" with a Droppable Booster Package
The recently passed "NASA Authorization Act of 2010"
directed the Agency to build a new Shuttle-Derived "Space Launch System"
(SLS). NASA has studied a number of "in-line" SLS concepts that would use
a pair of four or five-segment solid rocket boosters to lift a core powered by three to
five RS-25 engines derived from the existing Space Shuttle Main Engine (SSME).
The Act called for SLS to be developed in two phases.
Phase One would be able to lift 70-100 tons (63.5 to 90.7 metric tons or tonnes) to low
Earth orbit (LEO). Phase Two would include improvements designed to increase LEO
payload to at least 130 tons (118 tonnes).
An All-Liquid Alternative
Several all-liquid SLS alternatives are theoretically, if not
politically, possible. The Act stated that NASA should "to the extent
practicable" use Space Shuttle derived and Ares I components with existing U.S.
propulsion systems. That leaves only four liquid engine choices: SSME, J-2X, RL-10,
and the AJ-10 based engine used by the orbiter Orbital Maneuvering System.
A two-stage rocket with six RS-25s powering the first stage and a
single J-2X on the second stage could lift 63.5 tonnes to LEO or 23 tonnes to
geosynchronous transfer orbit (GTO). Both stages could be built using 8.4 meter
diameter Shuttle External Tank (ET) tooling. Or, alternatively, the in-development
5.5 meter diameter Ares I Upper Stage could be stacked atop an ET-diameter first
Comparison of Ares V, "5/5" SLS, and a "Core
Only" All-Liquid Alternative
Such a rocket would lift off on 1,136 tonnes (2.5 million pounds)
of thrust from the six SSME (ultimately RS-25E variants) throttled to 109%. The
rocket and payload would weigh no more than 946.7 tonnes (2.087 million pounds) at
liftoff. The 43 meter long first stage would weigh 668 tonnes loaded and 69
tonnes at burnout. The J-2X powered upper stage would weigh nearly 211 tonnes loaded
and about 21.8 tonnes at burnout. The rocket would stand 80 meters, including an
assumed 20 meter long payload fairing. If a 5.5 meter long Ares I Upper Stage were
used, the rocket's height would increase to about 88 meters.
At rollout, this two-stage launch vehicle would weigh less than
153 tonnes, including payload, about 8.6 times less than the rollout weight of a Shuttle
stack. It would nonetheless carry 2.7 times more payload to LEO than Shuttle.
For more deep-space payload, an already-developed Delta 4
five-meter diameter upper stage could serve as a third stage. In this application
the Delta 4 stage would then fire its highly efficient RL10B-2 engine multiple times to
accelerate to a higher energy orbit or to escape velocity. This configuration could
boost 32 tonnes to GTO, 18 tonnes directly to geosynchronous orbit (GEO), or more than 25
tonnes to escape velocity.
Use of Ares I and Delta 4 Upper Stages
Future growth to more than 66 tonnes LEO could be provided by
increasing RS-25E thrust to the 115% level, providing 2.64 million pounds of liftoff
Toward Phase 2
Meeting the 118 tonne Phase 2 LEO goal would require liftoff
thrust augmentation and stretched core stages. One approach would be to add existing
Aerojet Atlas 5 solid rocket motors, the most powerful U.S. monolithic (non-segmented)
solid motors, to the 8.4 meter diameter core. Eight such solids, boosting a 977
tonne first stage topped by a 322 tonne second stage, could lift 118 tonnes to LEO.
The core would use six RS-25E first stage engines and a single J-2X as before, but both
stages would have to be substantially stretched and strengthened..
Alternatively, strap on solid motors could be used to reduce the
number of costly RS-25E engines on the first stage. Six Atlas 5 solids could allow
use of only three RS-25E engines on the first stage, for example, while still getting 67
tonnes to LEO. Such a configuration, which would use a shorter first stage, might be
limited to cargo flights.
The use of existing solid rocket motors shared by EELV rockets
would eliminate the need to maintain a costly NASA-only solid rocket booster contractor
An all-liquid rocket designed to lift 118 tonnes to LEO would be
more difficult. A serial two-stage rocket would, for example, require perhaps 12
SSME type first stage engines and three J-2X second stage engines. The first stage
would have to weigh about 1,200 tonnes, requiring a an ET-diameter tank stretch to 77
meters, substantially longer than any previously contemplated stage.
Or, more practicaly, the first stage could be subdivided into
multiple identical "modules", with two acting as strap-on boosters and one as a
core. This design would reduce the number of SSME-type engines to about 10 (four on
each booster and two on the core), while also allowing use of a smaller upper stage
powered by only one J-2X. This design could lift 118 tonnes to LEO or, with no upper
stage, 78 tonnes.
Another alternative would use four Delta 4 Common Booster Cores,
each fitted with a pair of RS-25E engines, as strap on boosters attached to an ET-diameter
core powered by two RS-25E engines. This concept, which would lift 118 tonnes to LEO
or 73 tonnes with no upper stage, would require use of RS-25E engines rated for 115%
thrust, a capability that current SSMEs have for emergencies. If the CBCs were
powered by RS-68A Delta 4 engines, the core would have to be fitted with five or six
RS-25E engines to meet the 118 tonne LEO goal, and would only be able to lift 50-55 tonnes
to LEO with no upper stage..
A Droppable Booster Alternative
One interesting all-liquid alternative involves use of a
droppable "booster package" propulsion section in a "stage-and-a-half"
design. Convair's original Atlas used such a design, which ignited all engines on
the ground but shed the heaviest part of the propulsion system after enough propellant had
been burned to allow use of a smaller "sustainer" engine that was optimized for
operation in vacuum conditions. Atlas continued to fly with this configuration until
2004, even when Agena, Centaur, or solid upper stages were stackad atop the first stage.
The stage-and-a-half design was so mass-efficient that several Atlas rockets flew
themselves into orbit without use of a second stage.
U.S. "National Launch System" (NLS) designs of the late
1980s and early 1990s contemplated use of "stage-and-a-half". One design,
named "NLS-2", was powered by six Space Transportation Main Engines (STMEs) off
the launch pad. Four of the 295 tonne thrust LH2/LOX engines would have dropped
away, along with their associated plumbing and thrust structure, about 162 seconds into
the flight. The booster package would have carried away nearly 25 tonnes of dry mass when
jettisonned, or nearly 28% of the total core stage dry mass. The two remaining STMEs
would have powered the External Tank diameter core on toward orbit, or, if an upper stage
were used, toward a high staging velocity. NLS-2 would have been able to lift 27-29
tonnes to LEO without an upper stage or 47-52 tonnes with an upper stage.
A similar "Shuttle Derived" design would simply replace
the STMEs with RS-25s. Although RS-25 would not produce as much thrust as STME, it
would provide substantially higher specific impulse. An RS-25E-powered droppable
booster all-liquid design would end up carrying nearly as much payload as NLS-2 when flown
without an upper stage. It could lift more payload to LEO than NLS-2 if a tailored
upper stage were added. For example, a "6/2" design with six RS-25E
engines, four droppable, topped by a second stage powered by six RL10A-4-3 engines might
be able to lift as much as 68 tonnes to LEO or more than 29 tonnes to GTO. The
numbers would be 5-10 tonnes less if lofted trajectories were not allowed.
The first stage would gross 755 tonnes, and the scond 119 tonnes,
at liftoff. About 170 seconds after liftoff, the booster package would shut down and
drop, carrying away four RS-25 engines and about 25 tonnes of mass, representing about
one-third of the total first stage dry mass. The stage would weigh about 273 tonnes
after booster separation. Its twin RS-25 propulsion system would continue to burn
for another 220 seconds, until about 6.5 minutes after liftoff. The second stage
would weigh about 12 tonnes at burnout. Its engine cluster would produce nearly 66
tonnes of thrust for a total burn time of up to 743 seconds. The engines could be
restarted multiple times if required.
With no upper stage, the 1.5 stage core would be able to lift
29.5 tonnes to LEO by itself. This might prove useful for ISS missions.
Although not an optimum LEO design (it would cost more than Delta 4 Heavy), it would share
the same launch infrastructure as the 2.5 stage rocket, allowing for test options and for
The main benefit of this design is that it would allow for a
smaller upper stage, which could be powered by a cluster of existing, essentially
already-developed RL-10 engines. This design would provide future growth potential
through employment of higher thrust upper stage engines. The U.S. Air Force, for
example, has initiated planning for a higher thrust, lower-cost RL-10 replacement to be
used by the EELV fleet.
70 Tonnes, Big Enough?
Some space industry leaders think that a 70 tonne LEO rocket is
big enough for NASA's needs. For example, George Sowers, vice president for business
development at United Launch Alliance, recently said that a 70 to 80 tonne LEO heavy
lifter could support beyond LEO exploration missions, especially when combined with
on-orbit propellant depots. He called such a rocket "smaller and
lower-cost" than the proposed Shuttle Derived SLS options. NASA studies have
projected that a "5/5" 100 tonne LEO Shuttle Derived SLS could cost an average
of $3.6 billion per year, even at very low flight rates.
Eliminating the SRB contractor chain, and with it all of the
supporting infrastructure, should offer substantial savings. NASA might be able to
afford to fly a smaller rocket more often than a costlier 100 tonne LEO alternative.
A smaller rocket would be better suited to LEO work in support of the International
The trade for eliminating SRB would be a need to keep J-2X for
some designs, although other RL-10 based designs have been identified here. A
high-thrust second stage would be all but required for a traditional, serial-staged
"all-liquid" two-stage design. One alternative to J-2X might be an
air-start RS-25 engine. Staying "all-liquid" could thus provide an
impetus for developing J-2X - a program well into bending metal with plans for
engine testing in 2011.
SPACE LAUNCH REPORT
by Ed Kyle