Great Mambo Chicken!
Paul Dietz
Transhuman Condition (Science Slightly Over the Edge)", but I acquired
another copy today. There's quite a bit about Truax in here (the
Snake River steam rocket and all). This book is a scream; I'd
recommend it highly but it's already been remaindered, I think. :-)
There's a nice bit about Sea Dragon. I quote...
--------------------
The Sea Dragon was a launch vehicle of stupendous proportions that
Truax had designed back when he was director of advanced development
at Aerojet General. The best perk of that high office was the $1
million budget that he could spend any way he wanted to. Truax used
it to test his pet theory that the *cost* of a rocket had nothing to
do with how *big* the rocket was. You could make a given rocket just
as big as you pleased and it would cost about the same as one that was
about half the size, or smaller.
This went against conventional wisdom and common sense, but at Aerojet
Truax collected enough facts and figures to prove its truth beyond a
doubt. Indeed, he'd been assembling the necessary data from the time
he'd been in the navy, where he'd had access to all sorts of cost
information.
Take Agena versus Thor, for example. These two rockets were
identical in every way: each had one engine, one set of propellant
tanks, and so forth; the only significant difference between them
was size. The Thor was far bigger than the Agena, but the surprise
was that the *bigger* rocket had cost *less* to develop.
"I was shocked to discover the Agena cost more than the Thor," Truax
said later. "The Thor was between five and ten times as big! I said
to myself, We've been tilting at windmills all this time! If all
rockets cost the same to make, why try to improve the
payload-to-weight ratio? If you want more payload, make the rocket
bigger."
The same anomaly cropped up again in the case of the two-stage Titan I
launch vehicle: the upper stage was *smaller*, a miniature version of
the lower stage, yet the smaller stage cost *more* to make.
It seemed irrational, but all of it made sense once you went through
the costs item by item. Engineering costs, for example, were the same
no matter what the size of the rocket. "You do the same engineering
for the two vehicles, only for the bigger rocket you put ten to the
sixth after a given quantity rather than ten to the third or
whatever," Truax said.
The same was true for lab tests. "The cost of lab tests is a function
of the size of your testing machine and the size of the sample you run
tests on, not the size of the product."
Ditto for documentation, spec sheets, manuals, and so forth. The cost
here was a function of the *number* of parts and not the *size* of the
parts. "There are absolutely no more documents associated with a big
thing than a small thing, as long as you're talking about the same
article."
By this time Truax had accounted for a healthy chunk of the total cost
of a given launch vehicle. About the only thing that *did* vary
directly with a rocket's size was the cost of the raw materials that
went into making it, but raw materials constituted only *2 percent* of
the total cost of a rocket. "Two percent is almost insignificant!"
he said. "And even with raw materials, if you buy a ton of it you get
it at a lower unit price than if you buy a pound. And this is
especially true of rocket propellants."
So if all this was true, if engineering, lab tests, documentation and
so forth didn't determine a launch vehicle's price tag, *what did*?
Essentially, three things: parts count, design margins, and
innovation. Other things being equal, the more parts a machine had,
the more it was going to cost. The more you wanted it to approach
perfection, the more expensive it would end up being. And finally,
the newer and more pioneering the design, the more you'd end up paying
for it.
"We came up with a set of ground rules for designing a launch
vehicle," Truax said. "Make it big, make it simple, make it reusable.
Don't push the state of the art, and don't make it any more reliable
that it has to be. And *never* mix people and cargo, because the
reliability requirements are worlds apart. For people you can have a
very small vehicle on which you lavish all your attention; everything
else is cargo, and for this all you need is a Big Dumb Booster."
--------------------
Paul F. Dietz
di...@cs.rochester.edu
"If I'd been in my grave, I'd have rolled over."
R. Truax on the decision to build the Space Shuttle
Chris Jones
quotes Robert Truax from Ed Regis's book _Great Mambo Chicken and the
Transhuman Condition (Science Slightly Over the Edge)_:
>Take Agena versus Thor, for example. These two rockets were
>identical in every way: each had one engine, one set of propellant
>tanks, and so forth; the only significant difference between them
>was size. The Thor was far bigger than the Agena, but the surprise
>was that the *bigger* rocket had cost *less* to develop.
The point Truax was making (fixed costs dominate rocket development, so make
'em big) could well be true. However, I think it oversimplifies quite a bit to
say that Thor and Agena were identical in every way save size: Agena's
restartable engine had to have contributed to its design complexity.
--
Chris Jones c...@ksr.com
Pat
Smaller scale is cheaper to build then large scale, within
the same orders of magnitude.
It's easy to build an Estes rocket or a small experimental
rocket.
It's easy to build a small kit car or kkit plane or boat.
it's WAY harder to build a 747, Mack Truck or ULCC.
I'd bet building SV's is a lot more work then building Titans.
I am sure between the Titan-Delta-Atlas-Thor, it's the same
scale, so just build titans.
But when you change Magnitude, look for an order of magnitude
cost rise.
The tooling is the killer.
pat
--
The greatest mistake you can make in life is to be continually
fearing that you will make one -- Ellen Hubbard.
Del Cotter
p...@access.digex.net (Pat) writes:
>It's easy to build an Estes rocket or a small experimental
>rocket.
How much does it cost to build an Estes rocket which is capable of putting
a payload -however small- into orbit?
>It's easy to build a small kit car or kkit plane or boat.
>
>it's WAY harder to build a 747, Mack Truck or ULCC.
How much does it cost to build a kit car which can carry cargo -any cargo-
through the air (747), across a continent (Mack) or across an ocean (ULCC)?
In other words, compare like to like.
>I'd bet building SV's is a lot more work then building Titans.
I wonder... I'd bet it costs as much to build ten -and only ten- Titans
as it does to build ten Saturns. Are Titans cheaper because more have
been built?
Truax may be exaggerating for effect, but I think he has a point.
--
',' ' ',',' | | ',' ' ',','
', ,',' | Del Cotter mt9...@brunel.ac.uk | ', ,','
',' | | ','
Pat
>In article <26icb8$6...@access.digex.net>
>p...@access.digex.net (Pat) writes:
>In other words, compare like to like.
>
If, you'd bothered reading my post, you would
have seen this.
>>I'd bet building SV's is a lot more work then building Titans.
>
>I wonder... I'd bet it costs as much to build ten -and only ten- Titans
>as it does to build ten Saturns. Are Titans cheaper because more have
>been built?
>
I would bet you are wrong.
When the choice is between say a Delta, Atlas or Titan,
The costs deltas are so marginal, that it's cheaper to standardize on the
titan.
But when they wanted to build the saturn V they
had to do things like builf the VAB (Largest industrial structure
in the world) The Crawler-transporter ( Largest machine of it's
time), Build pads an order of magnitude larger then
the others.
Now once all that was built, it might be less for saturn production,
but the parts are so massive, they are difficult to handle.
A titan can be moved by rail-barge. THe saturn needed special
air-craft, Giant barges, and plants close to KSC.
>Truax may be exaggerating for effect, but I think he has a point.
>
Truax still has a point, in that build the largest thing
within the size constraints of the form factor, known,
because the Engineering is not real difficult, but trust me,
the manufacturing is a bummer.
Steve Gunnell
In article <26icb8$6...@access.digex.net> p...@access.digex.net writes:
> I am not exactly sure if Truax is right.
> Smaller scale is cheaper to build then large scale, within
> the same orders of magnitude.
> It's easy to build an Estes rocket or a small experimental
> rocket.
> pat
You are missing the point. Truax said the parts count is what controls
the cost. So if you have the same parts count (and of course he was
talking about comparable parts sets) build it big.
-----------------------------------+-------------------------------------
Stephen Gunnell -----------------------------> ste...@eldred.dialix.oz.au
-----------------------------------+-------------------------------------
Pat
Parts count is Important. That's a Canon of Manufacturing ( My area
in school) but it's also Parts Count within a Form Factor.
Evrytime you change scaling, believe me, you incur enormous costs.
I am building a garden shed, no pieces longer then 16 Foot, long.
It's the same cost per foot and difficulty as teh Garage I worked
on 2 years ago. But if i were buiilding a 5 story steel building
it would be a seriously different matter.
A man and a helper can balance a 2x6-16. No-one can do that with
an equivalent steel beam. you need a crane.
A 45 foot sailboat requires similiar form factors to say a
Tug or Powerboat. Not a lot of special changes to the line.
but when you switch from Tugs to Tankers, you need much larger
Dry-docks, Giant cranes, etc...
IF you can hold the parts count, go for the largest design within
the form factor, but at some point that shifts.
Notice the constraint on the S-1, that was the ceiling heights
at STennis. To make the S-1 any larger one needed new buildings.
Tim Harincar
>In article <26icb8$6...@access.digex.net>
>p...@access.digex.net (Pat) writes:
>
>>It's easy to build an Estes rocket or a small experimental
>>rocket.
>
>How much does it cost to build an Estes rocket which is capable of putting
>a payload -however small- into orbit?
Well, you can't. Model rockets today, in moter sizes larger than D,
often use the same fuel as everyone else: AP/aluminum. These 'High Power'
rockets (they're beyond the legal defination of a 'model', and not
'amature' rockets) cross over in size with professional rocketry at
about an Arcas. So, if you want to scale up a non-professional rocket
to put a payload in orbit, you basically end up with something like a
Scout, since the fuel is the same. Hobby rocket materials - plywood,
cardboard, fiberglass, etc don't hold up well past mach 1 and in high
temperature areas.
However, Jay Apt, a shuttle astronaut, DID put a model rocket into
orbit - He brought it with him on a shuttle flight. So, if you look
at it that way, it cost him $4 to build it, and somewhere between
$500 million and $1 billion (depending on who yells the loudest :)
to launch it.
--
tim harincar
soc...@vx.cis.umn.edu
Tom Weinstein
> Parts count is Important. That's a Canon of Manufacturing ( My area
> in school) but it's also Parts Count within a Form Factor.
> Evrytime you change scaling, believe me, you incur enormous costs.
> I am building a garden shed, no pieces longer then 16 Foot, long.
> It's the same cost per foot and difficulty as teh Garage I worked
> on 2 years ago. But if i were buiilding a 5 story steel building
> it would be a seriously different matter.
> A man and a helper can balance a 2x6-16. No-one can do that with
> an equivalent steel beam. you need a crane.
Yes, that's true when you're building a shed. Fortunately, you don't
have to get each nail inspected by three inspectors who each have to
file a five page report on the quality of the nail. You don't have to
stress test each nail to make sure it won't fail when in operation. You
don't have to inspect each nail after it's been hammered in, and you
don't have a 50-step checklist on how to do the hammering.
I think the point is that for disposable rockets, the cost of
construction, and the materials cost are minimal when compared to the
cost of quality assurance. And the cost of quality scales with part
count.
--
TV is for selling things. Anyone who trys to tell you | Tom Weinstein
different is probably trying to sell you a TV. -- MT | to...@orac.esd.sgi.com
Nick Szabo
>I think the point is that for disposable rockets, the cost of
>construction, and the materials cost are minimal when compared to the
>cost of quality assurance. And the cost of quality scales with part
>count.
Hi, Tom! The argument is at cross-purposes here. I think we can agree that
cost is a function of both size _and_ parts count. The
cost/size function in transport typically has quantum jumps -- the
form factor point Pat made is very important. That's why
intermodal trasport settles on standard sizes for container cargo
(the boxes you see that go on ships, trains, and trucks).
The Cuban missile crisis stands out as an example of how critical form
factor is to rocketry. The Soviet move to put IRBMs in Cuba
stemmed from the fact that in the early 1960's, Soviet IRBMs fit
into their existing infrastructure: they could be carted by train
to port, put on ships, uncarted in Cuba and set up relatively easily
on mobile launch trucks that fit on normal roads.
The gigantic SS-6 ICBM was good at launching Sputniks but horrible
as an ICBM. It couldn't fit in silos, instead requiring huge
Canaveral-like launch complexes, non-standard gauge track to
move to the pad, etc. The parts count of the IRBMs and ICBMs was
roughly the same; the big practical cost was integrating the rockets
with the existing infrasturucture.
In 1960-61 USAF started launching spy satellites (especially the Discoverers,
more sophisticated and vastly more useful than anything the Soviets
launched in the early 1960's despite their much-hyped alleged lead
in the "space race"), exploding the "missile gap" myth (Discover pics
showd Sovs had at most 10-14 ICBMs in 1961, and hand't even completed the
ICBM testing program Sputniks were part of). Once we had unmasked the
bluffs behind the Iron Curtain, only way they could maintain a credible
threat was to escalate by actually putting nuclear-tipped missiles
within range of the U.S. While the SS-6 might have been cheaper on
paper, in practice it just didn't fit well with the surrounding Soviet
infrastructure, delaying and escalating cost of deployment.
Kruschev found it necessary and technically practical to ship IRBMs to
Cuba. Unfortuneately for the Commies, they didn't do it before
the Discoverers had unmasked the Soviet ICBM bluff, so Kennedy
could blockade Cuba without risking nuclear missile attack.
Form factor has been an even greater problem for the more ambitious heavy
orbital launchers produced since then. Saturn, Energia, Shuttle, etc. have
all required their own customized launch complexes,
satellite integration equipment, and even customized satellite
sizes which are incredibly costly. Not coincidentally, two
out of three are on the rust heap while the third squanders over
$5 billion a year mostly for the entertainment of astronaut groupies.
Incidentally, the Agena, which combined upper stage with spacecraft
bus and could be used on either Thor (now Delta) or Atlas,
was a splendid example of saving money with
transport standards. We launch over one hundred (100)
Agenas at the dawn of the space age, for everything from Discoverer
and SAMOS and MIDAS to lunar and planetary exploration, and saved
$10's of billions in the process. Unfortuneately, saving money does
not benefit defense and NASA contractors, so the common-sense strategies
borrowed from commerce and battlefield military experience at the beginning
of the space age have largely fallen by the wayside. NASA and its
contractors can make more money by building one-of-a-kind spacecraft,
building craft that, upon failing quickly motivate spending money on
replacements (eg Challenger, MO), or, taking the trend to the extreme,
spending $billions on lots of designs that don't need to be built or
launched at all (space station). Alas, it's a long way back.
Compare >100 Agenas to one Obsever-class (MO), one Mariner Mark
II (Cassinni, two if you count Galileo), and other sad contemporary
attempts at hyper-sized, under-funded "standardization".
--
Nick Szabo sz...@netcom.com
Gary Coffman
>to...@orac.asd.sgi.com (Tom Weinstein) writes:
>
>>I think the point is that for disposable rockets, the cost of
>>construction, and the materials cost are minimal when compared to the
>>cost of quality assurance. And the cost of quality scales with part
>>count.
>
>Hi, Tom! The argument is at cross-purposes here. I think we can agree that
>cost is a function of both size _and_ parts count. The
>cost/size function in transport typically has quantum jumps -- the
>form factor point Pat made is very important. That's why
>intermodal trasport settles on standard sizes for container cargo
>(the boxes you see that go on ships, trains, and trucks).
The boxes are sized to the least flexible system. There was
a problem for the railroads when the roadway load limit was 60k lb.
A flatcar could carry more mass, but could fit but one 65 foot
trailer on board. Now with 80k lb load limits and double trailers,
the capacity of the two match. Special RORO ships had to be
constructed in either case. The Navy has just let contracts for
19 new RORO ships at $1.1 billion each to accomodate the containers.
The reason I go into all that is that it's not just as simple as
picking a container and going with it. Each of the systems that
carry it have to be sized for it, or a multiple. External constraints,
such as roadway load limits and railway load and turning radius limits
constrain the feasible sizes. And of course some cargo doesn't fit
well with containers. Thus we still have various sized tanks for
liquid cargos, various sized grain carriers, and general purpose
boxcars and flatbeds for oddsized cargo.
The container saves a lot of handling when the cargo is shipped
by multiple modes, but is of little value when shipping is solely
by one mode.
>The Cuban missile crisis stands out as an example of how critical form
>factor is to rocketry. The Soviet move to put IRBMs in Cuba
>stemmed from the fact that in the early 1960's, Soviet IRBMs fit
>into their existing infrastructure: they could be carted by train
>to port, put on ships, uncarted in Cuba and set up relatively easily
>on mobile launch trucks that fit on normal roads.
The reason the Soviets put missiles in Cuba was that they wanted a
lever to get us to remove our Jupiter missiles in Turkey. The 90 mile
flight time was destablizing in the extreme. The Soviets were convinced
we intended to launch a first strike, and they were terrified by the
missiles sitting on their borders.
>The gigantic SS-6 ICBM was good at launching Sputniks but horrible
>as an ICBM. It couldn't fit in silos, instead requiring huge
>Canaveral-like launch complexes, non-standard gauge track to
>move to the pad, etc. The parts count of the IRBMs and ICBMs was
>roughly the same; the big practical cost was integrating the rockets
>with the existing infrasturucture.
I don't buy that. Unlike the IRBM that has to move to near the target
to be fired, the ICBM doesn't need to be moved to be within range.
It's certainly true, however, that the SS-6 required a launch complex
to setup, fuel, and fire the missile. The IRBM also needed to be setup,
fueled, and fired, but the equipment to do that could be carried in a
fleet of trucks. The road and rail network was not a consideration
for the SS-6 since it didn't have to be moved to be targeted and fired,
but the transport network was critical to the IRBM since it had to
be moved to get within range of it's targets. Neither was a very good
reactionary weapon due to the long time needed to fuel the liquid
engines. They were primarily useful only as first strike weapons,
but their inaccuracy even made that role less than desirable. Later
solid fuel missiles could be used as deterrents to an enemy first
strike because they could be readied and fired during the interval
of ICBM flight. Nearby IRBMs destablized even that launch on warning
strategy because the warning time became too short.
>Form factor has been an even greater problem for the more ambitious heavy
>orbital launchers produced since then. Saturn, Energia, Shuttle, etc. have
>all required their own customized launch complexes,
>satellite integration equipment, and even customized satellite
>sizes which are incredibly costly. Not coincidentally, two
>out of three are on the rust heap while the third squanders over
>$5 billion a year mostly for the entertainment of astronaut groupies.
Ignoring the usual cheap shot at NASA, the contrary example is the
Ariane series with it's multiple payload fairing. The latest Ariane
launch includes 7 separate payloads. Shuttle also works with multiple
payloads. The problem with Saturn and Energia is that no such multiple
payload capability has ever been developed for them, and the market for
heavy lift single payloads is limited.
>Incidentally, the Agena, which combined upper stage with spacecraft
>bus and could be used on either Thor (now Delta) or Atlas,
>was a splendid example of saving money with
>transport standards. We launch over one hundred (100)
>Agenas at the dawn of the space age, for everything from Discoverer
>and SAMOS and MIDAS to lunar and planetary exploration, and saved
>$10's of billions in the process. Unfortuneately, saving money does
>not benefit defense and NASA contractors, so the common-sense strategies
>borrowed from commerce and battlefield military experience at the beginning
>of the space age have largely fallen by the wayside.
Except that this is nonsense. What Agena did was shift the costs of
integration from launcher stacks to payload configuration. Each of
the different missions had to be configured to work in the same
Agena shell whether that was the best layout for the particular
mission or not. From the early film bucket droppers to Big Bird,
each payload had to be contorted into the same shape while maintaining
the correct mass balance. According to _Deep Black_ this ran the costs
of the NRO up to $4.5 billion a year during the period. That was so
high that the Air Force brass was squawking loudly that the black
programs were sapping the money needed to upgrade conventional
intelligence gathering, and indeed even offensive and defensive
weapon systems.
>NASA and its
>contractors can make more money by building one-of-a-kind spacecraft,
>building craft that, upon failing quickly motivate spending money on
>replacements (eg Challenger, MO), or, taking the trend to the extreme,
>spending $billions on lots of designs that don't need to be built or
>launched at all (space station). Alas, it's a long way back.
>Compare >100 Agenas to one Obsever-class (MO), one Mariner Mark
>II (Cassinni, two if you count Galileo), and other sad contemporary
>attempts at hyper-sized, under-funded "standardization".
You must understand NASA's roots as a R&D organization with a strong
emphasis on the D. NASA is not culturally very interested in science
data return. That's more the province of the NSF. NASA is primarily
interested in developing and operating hardware. This is a holdover
from the NACA days where developing new wing shapes, new propeller
shapes, etc was their reason for existance. Most of NASA is staffed
by engineers, not scientists. And most NASA management has the same
bent. To an organization chartered to develop new hardware, premature
standards are just a drag on the main activity. Each new mission start
is seen as an opportunity to do *engineering development*. The science
return, if any, is just a bone thrown to the science types at the end
of the project. This has nothing to do with making money for NASA, they
don't get to keep any money at the end of the fiscal year. Rather this
has to do with the types of work the people who make up NASA are trained
and culturally conditioned to do.
This is one reason that science data purchase isn't a good fit with
NASA. If we want to try science data purchase, NSF should be the
contracting agency. They are better equipped to determine what types
of science data are most valuable and to assign dollar amounts to
various purchase offers. NASA could then *bid* to fill those data
requests along with any commercial provider who cares to take the
development risks. But NASA's primary mission is to be a pathfinder
for new technologies. That seems to have been lost of late as NASA
new program starts have declined and primary emphasis has gone over
to operating existing hardware. Perhaps the operating side of NASA
should be spun off into another agency, perhaps under DOT.
Gary
--
Gary Coffman KE4ZV |"If 10% is good enough | gatech!wa4mei!ke4zv!gary
Destructive Testing Systems | for Jesus, it's good | uunet!rsiatl!ke4zv!gary
534 Shannon Way | enough for Uncle Sam."| emory!kd4nc!ke4zv!gary
Lawrenceville, GA 30244 | -Ray Stevens |