Saturday, July 09, 2011

Launchers and Impossible Lessons from history.

NASA and DOD have spent a lot of effort on creating costing models for aerospace development. By the standard costing model the Falcon 9/Dragon should have cost 4B+ to develop. Using the most optimistic costing model it should have cost 1.6B. The documented actual cost is 390M. All of the traditional aerospace companies would have told you that it was impossible. Accepted "Facts" can be wrong. Spacex now has a backlog of > 3B. It looks like Spacex will be a business success with investment returns in excess of 10x. Elon has proven to be a brilliant business man in multiple fields.

Spacex has assembled a group of really talented people, many of them with a history in the traditional aerospace environment. If you read the bios on the Spacex web site here. You will discover that many of the senior engineering people came from large aerospace organizations. They used their experience to build the best rocket they could. They had a clean sheet of paper and enough resources to do the job.They fixed most of the business problems in the traditional aerospace model, they embraced vertical integration and the rejected the traditional aerospace supply chain. Win, win, win.

Now I'm going to ask you to get into your time machine.
Go back to the late 1970's. Take a large budget and go hire the best and brightest computer engineers from IBM, DEC, Prime,HP etc... . Give them a clean sheet of paper and allow them to fix any problems they see in the traditional computer business. Turn them loose and volia you have a killer minicomputer. It outperforms the DEC VAX and cost 1/2 as much. Its better than the rest of the industry in every way...... Instead of costing 120,000 or so it only costs 60,000. The orders pile up and the traditional computer companies would be worried. Meanwhile a guy named Wozniak with no degree and no experience designing computers is building a computer to impress his friends at the homebrew computer club. The Apple I soon to be an Apple II, in every measurable technical way the Apple II was inferior to the minicomputers of the day, except one, price. If you had asked the engineers from DEC,IBM, Prime, HP etc... to design you a useful computer that could be sold for less than 3000 they would have laughed at you.

So the question I ask is it possible to be a Wozniak in the space access area?

If its possible, you aren't going to get their via SBIR, because the SBIR evaluators with their reality closely tied to traditional aerospace model will be laughing.

You won't be able to do it by selling parts to other aerospace companies, IE Mr Wozniak did not start by building low cost memory cards to sell to DEC, the whole concept of modular cards and back planes as implemented in big computer land cost more than the whole apple I. The whole concept of separate bolt together components going into a launcher will need to be changed. The size scope and scale of what you build will not be appropriate for traditional aerospace.

Traditional aerospace customers will laugh at you and ridicule you... until one of you ends up unemployed and looking like a fool.

Clearly the microcomputer revolution was enabled by Mores law and the physics of space flight will not have any such exponential favoring factor.

Wozniak did not build any custom silicon, he use commercial off the shelf parts in new ways. I believe that modern CNC, 3D manufacturing and automated composite construction can be leveraged in a similar way.

This concept of comparing aerospace, old minicomputers and the PC revolution is not an original idea of mine. Charles at has used this comparison for years.
Frankly I never really got it. I thought it was a bit too much of a stretch. For the last 4 years I've been reading, studying and brainstorming on really low cost launch concepts. I've also been a rabid Spacex fan following what they are doing and cheering their success. In my though experiments I keep coming up with differnt solution concepts than Spacex. The Spacex solutions keep looking like traditional aerospace, but with much better execution, why is that? Is the traditional super high tech method the only way to achieve space launch? Then it dawns on me that Elon/Spacex hired the wizards from DEC, IBM, HP etc.. to build a better computer. In that context it ALL makes perfect sense.


Wolfkeeper said...

Really Wozniak's main genius was to ride a tech wave; he correctly picked the microprocessor wave and was able to leverage it.

SpaceX hasn't really done that, they've just cost reduced existing launchers. The economic models suggest that, although they may score more business for their company, overall they will reduce total revenues for the launch industry (and presumably put people out of that work and hopefully into more useful related work). Whether those models are true or not remains to be seen, but it doesn't matter that much to SpaceX themselves.

So we need new tech, but it's wrong to think it will be much smaller just because microprocessors were smaller; that's not the lesson; the lesson is that new tech is what makes new markets.

New tech is stuff like laser launch, airbreathing vehicles, launch loops; stuff that possibly actually can reduce greatly costs. Perhaps Jordin Kare is the new Wozniak.

Whatever happens though, it won't be pure chemical rockets; we've played that game, it's overplayed now. The new tech won't look like anything that happened in computing or anything else. Everything is always a little bit different or a lot.

Gaetano Marano said...

the ISS was a three-beds-only space-hotel, while, now, after its upgrade, the ISS is a six-bed-only space-hotel (a total of 12 astronauts per year with crew rotation) but, since it's an INTERNATIONAL space-hotel, two of these astronauts (four per year) are european, two (four per year) russian and ONLY two (four per year) american
well, the FOUR european and russian astronauts (eight per year) always will use the Soyuz to fly to/from the ISS, since it's a ready available, cheap and very reliable spacecraft
of course, the TWO american astronauts (four per year) will fly on Soyuz capsules from late 2011 to 2017 and the MPCV-Orion from 2018 to 2020, when the ISS should be de-orbited and burned in the atmosphere
so, the total number of american astronauts that will fly to the ISS should be around 9*4+2=38 but ONLY if each astronaut will fly ONCE
clearly, it's not rational to train an astronaut to fly only ONCE, then, each astronaut should fly at least four times in 2012-2020, reducing the total number of NASA astronauts (with some backup astronauts) to ONLY 10-15 between 2011 and 2020
and, of course, since the ISS is an "hotel for six" and ONLY TWO of them can be americans, ALL these 10-15 astronaut will fly to the ISS with Soyuz and Orion
so, when one or more of the "american Soyuz capsules" called "Dragon" or "Blue Kliper" or "CST-pollo" and "Dream(only)chaser" (that "should" fly with crews around 2016-2018) will be available, should NOT have a MARKET, since the ISS does NOT have enough "beds & breakfast" ALSO for the "commercial astronauts"
Shuttle era: 30 years (1981-2011) 135 missions, 900+ astronauts, 2000 tons of (high value & resupply) cargo to LEO (+ the astronauts and cargo launched with Soyuz and Progress)
Soyuz+Orion era: only 9 years (2011-2020) about 20 crew missions, 38 american astronauts, about 100 tons of cargo-resupply-only carried with Progress, ATV, HTV, Dragon and Cygnus
"commercial spacecrafts" era: it may happen only after 2016 and only for cargo, while, the crew missions may never happen in this decade, since... 1. the Soyuz and Orion missions will be more than enough for the ISS and... 2. after 2020, the ISS should no longer exist, so, ZERO places to go = ZERO manned and cargo missions

Stevo Harrington said...

One major difference between computers and rockets is that faster computers are used for creating faster computers via improved EDA software.
As for costs, NASA cost predictions are not very accurate. A better cost comparison would be to the EELV program, total development cost $1.8B for two rockets.
(from GAO-05-301)
Chemical rockets can do the job, airlines fly on 2-3x fuel costs, If launch vehicles could be reused, then the cost for each lb to orbit could be 2x the cost of 50 lb of fuel to get there, or about $100.
The development costs for a reusable vehicle are too high to justify the expense at the current launch rate.
The most expensive part of a launch vehicle is the insurance, which means that new entrants are at a serious disadvantage.
Point to point suborbital looks like a good market, until you realize that Concorde never made money and no-one is putting up the money to develop a supersonic bizjet.
My recommendation is to keep flying reusable rockets with the coolest features, the highest performance and the lowest opex that you can afford. Woz was having fun making a toy, and you should too.

Charles Pooley said...

Paul. Thanks for the mention. As you know, I've settled in Las Vegas to better the chances of having Microlaunchers reach "critical mass" . It's starting to work. With a local Meetup group I've started here there are several seriously interested.

I still believe in the microcomputer analogy as much as ever, and it will start multiple groups, companies shortly after the 1st Microlaunchers get off the ground.

This is meant to grow up outside the traditional aerospace culture, just as the Altair and progeny grew outside the mainframe builders.

In the short term, I'm still planning on a Nano Satellite Challenge together with a Cubesat launch service. That same launcher then can start sending 100-200 gram spacecraft to NEOs.

Stay tuned.

QuantumG said...

Like Ford, Apple changed the game by creating a new market. I love Pooley's rants (I'd wanna, I hear 'em almost every week on The Space Show) but when is he going to fly something? I'll laugh if InterOrbital fly something first :)

So, taking the lesson of Ford and Apple and others: is there a mainstream market for the currently niche product of space launch?

It's not just a question of price point. What's the killer app?

QuantumG said...

Seeing as we have Charles here, I rant a little more.

I really enjoyed when you used to talk about amateurs doing NEO flybys. The idea that a college level group could do a masters project and return practical scientific data.. that's game changing.

Charles Pooley said...

Quantum: college level group... Yes. There are over 100 Cubesats waiting for launch. Many more woulod be built if it were known they could be launched, for $50 thousand or so. The first involved would be building, flying on Microlaunchers first. Then competitors would appear. The movement will be on solid ground after there are several.

Game changing? Yes. Many directly involved will do what the micromputer did.

Practical data? Yes. How many NEOs have been photographed? Why not 100 or 1000?

Then evolution (as Altair begat what I type this on) will enable rendezvous, landings. And Lunar surface. Landed upon 19 times. Why not 1900? 19,000?

The Solar System is big enough and there will be no collisions.

ian said...

I think I agree more with Wolfkeeper and S. Harrington. I can't just extrapolate that a "small launcher" will work as well as the microcomputer. Maybe it will, but it's not an obvious extrapolation at all if only because the market, costs and technology scaling are completely different.

Have the UAVs disrupted the airplane industry, or the cheap hobby RC planes? Not really, although they're a huge new market. I don't think C. Pooley's reasoning is correct.

I do think the small new startup vs. big old established company comparison applies, because that's a social phenomenon that you can see it happening repeatedly over different times and industries.

QuantumG said...

I don't see why you need to displace anything. The microcomputer wasn't a social revolution because it provided cheaper computers to existing computer customers.. exactly the opposite. Existing customers were not interested in them.

While I certainly don't see how Charles thinks he's going to get $50k to LEO/escape launches, I'm not going to say he can't do it. I'll remain skeptical until he does, but that's just common sense.

Paul Breed said...

The original personal computers did not displace the existing market.
The existing market saw them as toys.

It created a new market, if one could do 50K a launch what kind of new markets would you have?

Clearly the existing space companies would laugh at you.

Paul Breed said...

>although they're a huge new market.

Your own paragraph negates your own point???

Its a NEW market.
What would happen if one could offer a 100K NEO mission?

Charles Pooley said...

Bottom line:

This will not be done by those who do not agree with the premise. It will be done by those who do.

Possible unknown regulatory costs aside, the hardware of each "ML-1" should be under $10,000 each. Quite sure of this--have been designing it for a while.

A launch fee of $50 thousand then makes room for profit, ROI (if there are investors).

An irrigation tube based rocket I designed in 1993-5 had a materials cost under $3/pound, and the part that got built worked.

ian said...

Paul, I dont think Im contradicting myself. I think its a matter of degree. I do believe strongly that a nanosat launcher is a good product and would open huge new markets, but I dont think its disruptive or would "change the world as we know it" like the microcomputer. Not even close. There are just fundamental limits on how cheap you can build it and honestly im not sure theres the potentially infinite market the micros had.

Do you know what I think would be disruptive and "change the world as we know it"? The first true RLV. That would have potentially zero cost, or just fuel and mission support. Think the Blue Ball if you had fo rebuild it every time.

That said, I think a nanosat launcher would easily pay for the RnD of a nano rlv.

My 2 cents.


Anonymous said...

A completely fresh take on an established market is a tricky thing to shoot for. You hear a lot about the few successes but not much about the many failures. That being said, the people who change the world are the ones who try. If you have the freedom to go for it, then do it. The world needs people who dream big.

Here is a crazy idea you might look at:

heroineworshipper said...

Personally don't believe the SpaceX balance sheet any more than Bank of America's. Paypal produced a lot of money. Having been through some of the SpaceX hiring process & failing, they seemed like a traditional, large corporation that isn't really interested in bedroom hackers & garage tinkerers.

If only there was a shottky diode of rocket propulsion.

Anonymous said...

My beef with SpaceX is that they talk a lot and fly a little. Their mainfest shows four more Falcon 9 launches this year, (3 with Dragon capsules), which, based on their current flight rates, seems really unlikely.

I understand that they are still in development mode with the Dragon, but the Falcon 9 itself should be matured enough that SpaceX is already conducting regular satellite launches. With such a "low launch price", and such a "highly reliable vehicle", you would thing their manifest would be twice what it is, and you would be seeing a flurry of actual launch activity. But you don't, which begs the question... "Why not?"

Wolfkeeper said it exactly right... SpaceX just cost reduced existing launchers. In that regard, they may very well be a business success. But for humanity to get off of this rock, it's not enough. We need something beside recycled and modernized old technology.

Pat Bahn said...


What you are discussing is wether or not a disruptive innovation can occur within the space launch business.

That is something I believe to be true,

The trick is you must be low tech, low cost and allow a new group of underserved customers to have access to what they are locked out of.

To compete against the majors you need to invest on their level. Certainly a well heeled investor like Musk can do that. However, how much will that market grow?

If you can create a new market, you can own 100% of its growth.

If you look at VG, you can see they've grown their end of the market significantly, prior to their existence space tourism was zero. Now they've booked up significant orders.

There are lots of other niches out there.

Carl Tedesco said...

Charles, $10K in hardware for an orbital rocket? Is this a volunteer work force? Do you plan for a salary? Just asking...

Max said...

I don't think current space technology can be compared to IT like that. Wozniak (and Gates, and Lord Sinclair etc. etc.) would all have been sitting dead in the water if the wave of microelectronics wouldn't have carried them on its back; and while I'm admittedly nothing more than a space enthusiast, I get the feeling our current chemical rocket technology is approximately the equivalent of the vacuum tube at best - wonderful tech but with severely limited potential.

Electronics was very much a nascent technology back then, with vast empty room to expand into, and ongoing miniaturisation kept its momentum up until recently (quantity - multiple cores - rather than quality seems to be the way we keep pushing Moore's law these days). Rocket technology was also like that once, but that was a long time ago. I'm just not sure "game changing" is possible as long as pretty much the same existing, established technologies have to be used to do it.

Also, I don't think the two scale the same way - a microcomputer could do anything Cray could, just less of it and slower. In space, you either reach orbit/escape velocity or you don't, and I tend to think micro-payload launches would have a seriously limited market outside academic circles.

But the thing is I'm not the one building rockets. You are. If you think you have an idea with that kind of potential, by all means, go ahead - honestly, I'd very much like to watch you try and see you succeed.

QuantumG said...

Max, analogies are never complete in all details. That said, NOFBX is a completely non-utilized new development in chemical rocket technology.. presumably they'll find some customers soon.. that could be used to fit the analogy of the 6502. I don't know how much millage you'll get with it though.

gravityloss said...

Actually Ian, the UAV:s probably have disrupted or are disrupting military aircraft producers and the whole chain and infrastructure of having the people in the planes: the pilots, the training, the remote bases, air tanking, radar suppression, supply lines, etc etc...

What used to be a wing of super expensive well trained pilots that was kept up in a remote location flying fast jets needing lots of ground and air support can now be replaced by a few technicians at the remote location loading a little fuel and a few missiles every now and then and a few quickly trained guys flying the drones from home turf.

To exaggerate of course. But would anyone really need Raptors or JSF:s in Afghanistan?

The same will happen to spaceflight. There will be massively less hands on people involved per flight and that will make it cheaper and much more flexible and reliable.

Paul Breed said...

Most transportation systems have fuel as about 1/3 the operating expense. Presently the propellant costs for any launch are almost in the accounting noise.

There is significant room to improve without any new magic technology.

peterfirefly said...

Honda started with an engine design they licensed to Toyota.

Memotech started with expansion RAM for the ZX81.

Logitech started with a Modula-2 compiler (and a mouse).

Microsoft did the Basic for other people's computers.


Btw, the VAX wasn't all that fast and many of the implementations were quite wasteful. The VAX 9000, for example, or the vector instructions. Or the commercial instructions. Or the fact that block moves were kinda slow. Or that it was so darn easy to blow the inner-most code cache because it was indexed using virtual (pre-translation) addresses so every little change of the address mapping required a cache flush -- they never got around to implementing a virtually indexed physically tagged L1 code cache.

mike shupp said...

Thought provoking. However, you missed the real irony -- DEC had actually pioneered in the small computer world with its PDP8/S -- which appeared in the late 1960's, and had a system box about the same size as the IBM PC. Alas, DEC sold so many of these to industry at $5-10,000 that they never guessed how many they could have sold to hobbyists for $500.

Ian said...

"Most transportation systems have fuel as about 1/3 the operating expense. Presently the propellant costs for any launch are almost in the accounting noise. "

Well, that's my point. You're throwing away the whole transportation system every time. Imagine if you threw away an airliner after every flight.

peterfirefly said...

MIT OpenCourseWare has a class on the shuttle:

Lecture 14 is about the ground operations -- refurbishing and checking the shuttle before its next flight is where almost all the cost is. It is not at all clear that that process is cheaper than throwing away a simpler rocket after each flight.

Andrew S. Mooney said...

I think that you have an interesting idea in that the success of the microcomputer comes from low upfront cost, with the actual capability of the design secondary. The original computers were frankly rubbish but they were working computers: As small as possible to do the job.

Rather than aiming for a market, you offer a capability in a payload size, and see what the market then does in response to it.

A problem though, is that this a weird perspective to sell/describe, and so a useful idea would be a quick/easy NAME for this intellectual approach. Any design actually needs to be smaller than the industry actually thinks is possible and viable.

The microprocessor = micro-launcher paradigm is not wrong, just problematic as it is too "debatable," as this posting shows. There is another illustration of an engineering precedent for this mindset. It also is something that doesn't involve massive levels of factual evidence that people then argue with you about.

In the field of model railways, there is a very small scale called Z Scale: Wikipedia describes it, where the tracks are 6.5mm apart.

The reason why it is called "Z" scale is because of the Marklin corporation, who in 1972 created it and did not imagine that it was possible to construct a commercial model railway scale at a smaller size. They used electric motors the size of sugar cubes as it was not thought possible to find any industrial process that could then provide cheap electric motors for them to work around to create anything smaller. It was named "Z" because it was deemed the last modelling scale: The smallest there would ever be.

35 years later and they have been beaten, in the form of "T Scale" trains, that use electric motors from mobile telephone and pager buzzers, and are produced by K.K. Eishindo corporation of Japan.

"T Scale spacelift" is a good name for what you're aiming at. It is smaller than the industry imagines, but maybe Z scale is more marketable as a MEMORABLE NAME: something people immediately recognize. It is an easy brand name for the approach.

Either way it works in that both T/Z names advertise/describe a subset-market against larger competitors. It is not massive, but it is always there. As Charles Pooley points out, there are lots of small payloads the "full size" industry just isn't interested in.

I would argue that the lack of a simple name for this approach is why few people are thinking along these lines. "Z Scale launchers" *are vehicles that always aspire to be as small as can economically be made.* It needs a name to outline that core idea, otherwise the idea is always forced to explain itself.

The small size of the system then gives low initial demands for industrial activity, investor heartburn and the difficulty of market entry, where, as SpaceX and Sir Richard Branson both illustrate, you otherwise have to make your billion somewhere else and then enter the rocket business.

Paul Breed said...

My obvious name for this concept is unreasonable or u-scale.


seb said...

U-scale like "the smallest Usable or Useful". Go for it :)

Ian said...

@peterfirefly I think it is obvious. I wouldn't use the shuttle as an example. This was an experimental vehicle that ended up operational.

Charles Pooley said...

Earlier post: Carl Tedesco said...

Charles, $10K in hardware for an orbital rocket? Is this a volunteer work force? Do you plan for a salary? Just asking...

Hardware parts cost only. That is easier to quantify and part of the early design process ongoing.

Andrew Mooney's comment:

First 3 paragraphs: he gets it--the basic premise. A main outcome of the early computers is that through affordability a community sparang up, start-ups started up and rest is history.

He wonders about a better name. I do intend to incorporate as a seperate entity a Nano Satellite Challenge entry with a Cubesat launch service. A process analogous to the funding, morphing of the X Prize entry into Virgin Galactic (but smaller scale).

Stuff about Z scale trains. Better example is the community into sub-gram model planes:

a separate Youtube about the actuators:

I plan to use that type of actuator for spacecraft solar panels, and 2 phase versions as tiny motors.

Next to last paragraph: I think the problem is not the lack of a name, but that this is too far off the train of thought of Newspace advocates. Too heretical. Most showing serious interest are not of Newspace and have no investment in that belief system.

As for Mother Nature--the physics: Absolutely do-able.

3tabs said...

I just found your site and wanted to say that I have really enjoyed browsing your posts.In any case I'll be subscribing to your blog and I hope you post again soon

aqui said...

Hey, there is a great deal of effective info above!

Model Baju Jas Pria Modern 2015 said...

Good article and a nice summation of the problem. My only problem with the analysis is given that much of the population joined the chorus of deregulatory mythology, given vested interest is inclined toward perpetuation of the current system and given a lack of a popular cheerleader for your arguments, I'm not seeing much in the way of change.

acil said...

new things and new experiences visiting this blog see mya blog jaket kulit asli.. I really like the articles and posts written on this blog thanks admin

shubham sapkal said...

How to check cbse result 2015 online for delhi Mumbai pune chennai
check 10th 12th class cbse result online
Delhi 12th 10th class cbse result 2015

shubham sapkal said...

gujarat 10th class result 2015
GSEB SSC result 2015 check online
surat ssc result 2015


Hopefully you won't stop the flow of such magical material! ταχυδακτυλουργός για παιδικά πάρτυ
μάγος για πάρτυ