Robert Clark wrote:
expendable or semi-salvageable while the upper stage (the orbiter ) is
reusable. As aesthetically pleasing as this configuration may appear
to some, from an engineering point of view this is precisely the
opposite of the correct way to design a partially reusable launch
system. Instead, the lower stages should be reusable and the upper
stage expendable. Why? Becasue the lower stages of a multi-staged
booster are far more massive than the upper stage: so if only one or
the other is to be reusable, you save much more money by reusing the
lower stage.

I don't say whether Zubrin's conclusion is correct or not, but the
logic in the above works only if "far more massive" always
translates to "far more expensive." I don't believe that's necessarily
the case.

Bob M.



Furthermore, it is much easier to make the lower stage
reusable, since it does not fly as high or as fast, and thus takes
much less of a beating during reentry. Finally the negative payload
impact of adding those systems required for reusability is much less
if they are put on the lower stage than the upper. In a typical two-
stage to orbit system for example every kilogram of extra dry mass
added to the lower stage reduces the payload delivered to orbit by
about 0.1 kilograms, whereas a kilogram of extra dry mass on the upper
stage causes a full kilogram of payload loss. The Shuttle is actually
a 100-tonne to orbit booster, but because the upper stage is a
reusable orbiter vehicle with a dry mass of 80 tonnes, only 20 tonnes
of payload is actually delivered to orbit. From the amount of smoke,
fire, and thrust the Shuttle produces on the launch pad, it should
deliver five times the payload to orbit of a Titan IV, but because it
must launch the orbiter to space as well as the payload, its net
delivery capability only equals that of the Titan. There is no need
for 60-odd tonnes of wings, landing gear and thermal protection
systems in Earth orbit, but the shuttle drags them up there (at a cost
of $10 million per tonne) anyway each time it flies. In short the
Space Shuttle is so inefficient because *it is built upside down*."
{emphasis in the original.}
_Entering Space_, p. 29.

Zubrin makes a key point about that dry weight of 80,000 kg of the
orbiter, which is essentially an upper stage, that needs to be carried
along to bring that approx. 20,000 kg of payload to orbit. That 4 to 1
ratio of the upper stage dry weight to payload weight struck me
because the upper stage for rockets is usually a quite lightweight
structure. Then the shuttle is quite poor on this measurement. I then
thought of the reconfigured kerosene version of the VentureStar I was
considering and realized that it was actually quite good on this
scale. It could carry ca. 125,000 payload to orbit with a vehicle dry
weight ca. 82,000 kg.
Actually the total shuttle system as a whole is even worse on this
scale. This site gives the specifications for some launch systems:

Space Launch Report Library.
http://www.spacelaunchreport.com/library.html

Here's the page for the shuttle system:

Space Launch Report: Space Shuttle.
http://www.spacelaunchreport.com/sts.html

You can calculate the total dry weight by subtracting off the
propellant weight from the gross weight for each component. I
calculate a total dry weight of 310,850 kg to a payload weight of
24,400 kg, a ratio of 12.7 to 1. In contrast the reconfigured
VentureStar has this ratio going in the other direction, that is, the
payload weight is larger than the vehicle dry weight.
This is important because the total dry weight is a key parameter for
the cost of a launch system. I looked at some of the vehicles listed
on the Spacelaunchreport.com page and all the ones I looked had the
total dry weight higher than the payload weight. For instance for the
Delta IV, it's a dry weight of 37,780 kg to a payload weight of 8,450
kg, for a ratio of 4.5 to 1.
For the Atlas V 401 it was 25,660 kg dry weight to a 12,500 kg payload
weight, for a ratio of 2 to 1. This was actually one of the better
ones. All the ones I looked at, all had a total dry weight
significantly higher than the payload weight, usually at least by a 3
to 4 to 1 ratio.
Then the reconfigured VentureStar would be important in that it could
reverse this trend (perhaps for the first time?) in making the dry
weight actually less than the payload that could be lofted to orbit.
Note that not even the original, planned VentureStar could accomplish
this, having a dry weight of about 100,000 kg to a payload capacity of
20,000, a ratio of 5 to 1.
The reconfigured kerosene-fueled VentureStar would have a greater
propellant mass using dense propellants, but the propellant costs are
a relatively small proportion of the launch costs. The more important
parameter of dry weight would actually be less.
Note also that the reconfigured kerosene VentureStar could accomplish
this feat of having a higher payload capacity than its dry weight,
while having a payload capacity that would be close to or exceed that
of all the former or planned U.S. launchers, and while being of
significantly smaller dimensions. See the attached image drawn to
scale showing some key U.S. launchers compared to the VentureStar.
Note that despite its small size it would be carrying more payload
than the shuttle, the Ares I, the Saturn V and nearly that of the Ares
V.
Another factor that I somehow missed when I first wrote on this was
the great reduction in launch costs. I somehow didn't make the
connection between the increase in payload capacity over the original
VentureStar configuration and that of the kerosene fueled one.
The development costs for the VentureStar or any launch vehicle are
figured into the launch costs. Then the estimated per kg launch costs
of ca. $1,000/kg for the original VentureStar are based on the late
90's estimated development costs for the VentureStar. However, a big
part of that development cost was due to the composite design which
was significantly more expensive then than now. Recall the point I
made before about the reduction in costs of composite materials
leading to auto makers including them more and more in passenger cars,
and this reduction in cost makes them now economically feasible for an
all composite SSTO.
Also, hydrogen engines and associated systems are generally more
expensive than kerosene ones. So the reconfigured VentureStar would
have a lower cost on that component as well. Then the total
development cost even including inflation for the reconfigured
VentureStar might be at or even below that of the 1990's estimates for
the original hydrogen-fueled VentureStar. This means the per launch
costs of the new version should be at or below that of the original
version.
But the reconfigured VentureStar can carry 6 times the payload of the
original VentureStar! This means the per kg launch costs would be
1/6th as much or only $166 per kg! This is such an *extreme* reduction
in launch costs over the current costs, that the calculation I made
for how much you could reduce the weight of the propellant tanks has
to be done in a more serious fashion.
Note that all the other systems for the VentureStar were progressing
well. It was only the relatively trivial problem of not using a strong
enough glue for the composite propellant tanks, that led to the
program being canceled. Then this is so trivial compared to the
complexity of the other systems and the importance of having a fully
reusable launch system is so clear, that a better course would have
been to open up a competition to find ways of getting the composite
tanks to work.
I gave a few different possibilities for lightweighting the propellant
tanks in section II of the first post in this thread. A few were
theoretical, not being tried before. However, the one involving
partitioned tanks is just basic engineering so I'll present a detailed
calculation for that in the next post.

Bob Clark

VentureStar, Shuttle, Ares I-V, and Saturn V size comparison.
http://i49.tinypic.com/2z3rup1.jpg