Reader's Favorites
Media Casualties Mount

Administration Split On Europe Invasion

Administration In Crisis Over Burgeoning Quagmire

Congress Concerned About Diversion From War On Japan

Pot, Kettle On Line Two...

Allies Seize Paris

The Natural

Gore Book Sales Tank, Supporters Claim Unfair Tactics

Satan Files Lack Of Defamation Suit

Why This Blog Bores People With Space Stuff

A New Beginning

My Hit Parade
Instapundit (Glenn Reynolds)
Tim Blair
James Lileks Bleats
Virginia Postrel
Kausfiles
Winds Of Change (Joe Katzman)
Little Green Footballs (Charles Johnson)
Samizdata
Eject Eject Eject (Bill Whittle)

Space
Alan Boyle (MSNBC)
Space Politics (Jeff Foust)
Space Transport News (Clark Lindsey)
NASA Watch
NASA Space Flight
Hobby Space
A Voyage To Arcturus (Jay Manifold)
Dispatches From The Final Frontier (Michael Belfiore)
Personal Spaceflight (Jeff Foust)
Mars Blog
The Flame Trench (Florida Today)
Space Cynic
Rocket Forge (Michael Mealing)
COTS Watch (Michael Mealing)
Curmudgeon's Corner (Mark Whittington)
Selenian Boondocks
Tales of the Heliosphere
Out Of The Cradle
Space For Commerce (Brian Dunbar)
True Anomaly
Kevin Parkin
The Speculist (Phil Bowermaster)
Spacecraft (Chris Hall)
Space Pragmatism (Dan Schrimpsher)
Eternal Golden Braid (Fred Kiesche)
Carried Away (Dan Schmelzer)
Laughing Wolf (C. Blake Powers)
Chair Force Engineer (Air Force Procurement)
Spacearium
Saturn Follies
JesusPhreaks (Scott Bell)

Biting Commentary about Infinity, and Beyond!

« Hail Lapham! | Main | Bill Clinton's Kryptonite »

The Value Of Documenting Trades

Jon Goff has an interesting post (if you're into rocket propulsion) on a technology that's been lying fallow for decades. But this post isn't about that concept per se, but a more general one:

I stumbled on this while trying to track down some old Aerojet papers about a sort of forced flow separation control technique that they researched back in the late 50s. I had noticed that most of the papers that cited the research talked about how Aerojet's had concluded that the approach didn't yield any net benefit, however the way they discussed it made me somewhat suspicious of their conclusion. You can sometimes get a sort of telephone effect with academic citations--where someone will read someone else's review of some obscure and hard to locate article, and instead of reading it themselves, they'll just summarize the summary, and before long who knows what the original article said. To make a long story short, I had good reason to be suspicious that there was something of that sort going on with this paper (especially since the two abstracts I was able to find online for their research seemed to directly contradict all the claims I've seen in citations of their work elsewhere).

Assuming that it was the case here, this happens more than you might think, and more than it should. This kind of thing, in fact, is the source of a lot of false mythology about space technology (e.g., highest vehicle performance is achieved with LOX/hydrogen, air breathers are the key to low launch cost, etc.). Many "rules of thumb" and conventional wisdom are based on a limited analysis, and used by people unfamiliar with their origin, or the underlying assumptions. I've written about an example of this before from my own early career:

Back when I worked at the Aerospace Corporation, a couple decades ago, I was fresh out of school, and sitting in a meeting with more senior people, discussing a conceptual design for a new military geostationary satellite. The subject was how to provide orientation. The two traditional choices were spin stabilization (many of the Hughes communications satellites used this technique) and active reaction control, which was more accurate, but limited the lifetime, due to depletion of propellant.

I (or someone, but I think it was me) suggested using gravity gradient stabilization (that is, taking advantage of the fact that a non-spherical satellite will naturally orient itself in the local vertical position, due to differential tidal forces between the line of the orbit and the small distances of the appendages from that line). The response of one of the supposedly experienced engineers was, "There's no gravity gradient at geosynchronous altitude."

I was a little surprised. "Oh, you mean there's not enough to do the job?" (I was thinking that perhaps he'd already considered it, and run the numbers.)

"No, there is no gravity gradient effect that high--it only applies in LEO."

Note that he wasn't making a quantitative argument, he was making a qualitative one. Low orbits had gravity gradient, high ones did not.

...What happened? Sometimes even engineers don't always apply good scientific principles. In this case, I suspect that he was an airplane guy who'd migrated into the space business (as often was the case in the beginning decades in the space industry), and had never really learned the fundamentals of orbital mechanics, or the underlying principles. Instead, he'd probably taken a space systems design course, and been given a lot of engineering rules of thumb, one of which was, no doubt, that gravity gradient can be used in LEO, but not in GEO.

And that's not a bad rule of thumb, as long as you understand where it comes from. Gravity gradient is indeed much less at twenty thousand miles altitude than at two hundred miles, and for most satellites could be considered, for practical purposes, to be non-existent. But we weren't talking about most satellites--we were looking at a new concept, much larger than anything previously deployed in GEO, with long booms and appendages that might, in fact be used for G-G stabilization. But because he didn't understand the physics, he mistook a rule of thumb for natural law, even though the law of gravitation says that the earth's gravity extends out to infinity, though it drops off as the square of the distance.

Often someone will perform an analysis, and people in a hurry will simply look at the bottom line, while ignoring the assumptions that went into it, which, if altered, might completely change the conclusion. Worse yet, sometimes the author hides the assumptions, making it even more pernicious (this, to me, is one of the primary reasons that we make so little progress in advancing a useful space policy--there are too many hidden assumptions on the parts of debaters, and everyone assumes that they're shared, when they're often not).

This is why it's important to properly document a trade analysis--so that when the assumptions change, it's easy to go back and determine whether or not the trade conclusion has, or whether or not it has to be redone. This is also why it's important to perform sensitivity analyses in the course of the trade--to make it easy later to determine, perhaps at a glance, whether an assumption change is critical or not.

I don't know whether or not the augmented thrust technology that Jon unearthed will find its way into future vehicles, but I'll bet that the original authors of the study didn't consider all of the potential applications for it when they published it--they were probably working on an engine for a specific vehicle concept. XCOR has been doing a lot of this kind of archaelogy of the early space age, and (at least it's my understanding) have found it a rich ore of untried but promising concepts. When one considers how much money was spent on the development of space technology in the early days (and how chaotic and largely undirected the vehicle development process has been over the last few decades), it would be surprising to me if there aren't a lot of old tricks in there that can find applicability in the twenty-first century. But one has to read carefully, and hope that the papers were documented properly. And when documenting our own results now, we should think of those who may be reading them in the future.

Posted by Rand Simberg at November 13, 2007 07:05 AM

Comments

Rand

Great overview. I found something similar to this in the early 1990's when I started working on microsatellites. It turns out that just about all of the tricks associated with making small sats productive that were heralded as breakthroughs then had been pioneered in the late 1950's when all we could launch were microsatellites.

After the blunder of Explorer 1's stabilization system (the Army guys thought that it would stabilize like a bullet flying through the air which was exactly the opposite of what happened) there was a period of an extensive study of the basic physics of stabilization and operation of spacecraft. Gravity gradient stabilization, spin vanes using differential sunlight, various shapes and sizes, were all studied during this period.

I only found out about this through a presentation that was made by some old timers at SCI (now Sanmina-SCI) (SCI was the first private commercial space firm in the world with SCI standing for Space Craft Incorporated). I was blown away with the concepts that they presented as I was going to the small sat conference every year hearing about all this wonderful new technology (that some of them were applying for a patent on) that I had seen in presentations about spacecraft design from 1958.

That being said, for every thing new today, there is the Pintile injector. It was brought from TRW to SpaceX as the bright future of injectors, when the old timers knew that it actually underperformed in relation to the "standard" design. This is why it is so important to work with the old guys that are still around, to glean from them their experience while fitting it into a modern understanding.

I got to see this in action at one of the X Prize flights. I introduced one of the old German rocket scientists to Elon Musk. Konrad Dannenberg knows just about everything there is to know about propulsion and after Elon explained the SpaceX system to him, he described the three principal weaknesses of the design in about 1 minute. At 92 years old that ain't bad.

Recently I was able to be with a friend who was interviewing Konrad about the early days of rocketry and I was blown away when Konrad said that the genesis of the design of rocket propulsion systems today is the diesel engine!! In thinking about it, it makes sense, but I have never heard even one person make the connection until that interview.

Posted by Dennis Ray Wingo at November 13, 2007 07:32 AM

And nowadays the analysis results get presented in N powerpoint slides, with N determined by the manager's schedule rather than the complexity of the issue.

I wrote up some of the LV analyses I did in detail a conference paper and later while working on my MS. But in everyday work documentation always falls off the bottom of the to do list.

Posted by Karl Gallagher at November 13, 2007 09:08 AM

...in everyday work documentation always falls off the bottom of the to do list.

I've noticed that in the last few years of consulting as well. A lot of engineers at large aerospace firms that shall remain nameless don't even seem to know how to write a trade analysis. As you say, it's documentation by Powerpoint and emails.

Posted by Rand Simberg at November 13, 2007 09:14 AM

This sort of thing happens in medicine all the time. Stabilization intervals, incisions and closures - things like that are often based on rules of thumb. No physician can remember all the argumenta ab initio, so we rely on the rules of thumb. However, there's always some ambitious hotshot who is convinced that ulcers are caused by bacteria or that glue is better than gut stitches for closing blepharoplasty incisions or something who pushes the rest of the world into taking notice. Sometimes those people are right (like the guy who proved that gastric ulcers were in fact usually caused by bacteria) and sometimes they're, you'll excuse the expression, dead wrong, as in that Brazilian guy who tried to treat cardiomyopathy by cutting out chunks of the ventricle.

But that's science. Just not rocket science I guess.

Posted by Jane Bernstein at November 13, 2007 09:27 AM

The science of human nature, perhaps ... in economics, it's about handling transaction costs, a fancy way of saying "one-stop shop." Mental one-stop shops range from rules of thumb to entire ideologies. The canonical example relevant to this forum is the first line of the CAIB report: "Building rockets is hard." Until the general public perceives rocketry to be no more problematic than aeronautics, it will remain dominated by the public sector and all the political/budgetary wrangling that goes with it.

Good to "see" you again, Jane.

Posted by Jay Manifold at November 13, 2007 10:44 AM

A lot of engineers ... don't even seem to know how to write a trade analysis.

At [big company] shortly after I hired in I handed a trade write-up to a manager. He handed it back and asked for a one paragraph verbal summary. So I stopped wasting my time on what they didn't want. Then I hit a point where I'd gone a year without writing anything but powerpoint bullets and my prose skills had atrophied. Took some work to get them back.

Hopefully blogging will produce engineers more practiced in writing things that will get looked up again later.

Posted by Karl Gallagher at November 13, 2007 11:09 AM

Dennis,
Interestingly enough your comments about pintle injectors happens to illustrate Rand's and my point very nicely. The old rocket scientists "knew that it actually underperformed in relation to the 'standard' design". The problem is, what are the assumptions, what are the details? We've found that for throttleable engines, pintles are amazing. We've been able to throttle very deeply without a variable geometry injector design, and without any noticeable lost in c* efficiency. Down to 30%. So yeah, once again, rules of thumb like "pintles underperform" are only true to an extent, and without more complete details on all the assumptions going in, you might miss a good corner of the trade space.

Not to pick on you, but it was a glaring example.

~Jon

Posted by Jonathan Goff at November 13, 2007 12:12 PM

Great stuff, Rand. Between your efforts and Jon's, I'd love to see more of this "meta" discussion percolate out and inform what often tends to be a shallow discourse of "which design is best?" -- or worse, "I know this design (which has yet to be flown, or maybe even demo'd) will prove the best."

Quantitative trade space and trade-offs... the profound difference between Apollo or Shuttle development (or the ESAS we seem to be getting) and a broad-front exploration of the trade space... these are so much the everyday fare of working engineers that you may not realize how small a part they play in the general, fuzzy vision of "technological progress" held by most citizens, legislators -- and to be honest, a lot of space fans.

Case in point: all the words and bytes devoted to "Apollo on steroids" -- essentially whether or not ELVs + capsules are "retro," and usually spinning off into sterile "the Shuttle was a lemon" vs. "surely with all we've learned since 1972 we could have a Real Spaceplane if we wanted..."

Those discussions are sterile not because they're based on wrong data or inadequate data, but because they're not even remotely how to think about the questions involved.

Which is why more discussions like these are all to the good. They won't turn everyone into engineers, but at least they help propagate a clearer idea, as Jon put it, of "what we don't yet know" and what it would take to find out.

Posted by Monte Davis at November 13, 2007 12:51 PM

Jon

Thanks for illustrating my point. I did not say anything about the best injector for a throttlable engine. When I made the statement that I made, it was related to the Spacx experience. They have made many major modifications to their design that has basically, as far as I know, rendered the Pintle obsolete in their designs (at least that is what they said in their court documents). In coming back and stating things in the way that you did, you did two things.

1. One uppmanship.

See I found a niche where what you said is wrong, see ain't I smart.

2. Not reading for comprehension.

When I made the statement that I made, it was in relationship to a standard design for a launch vehicle, not a deep throttling engine used for the lunar lander challenge or similar applications. Therefore your statement was an inappropriate branching off into a tangent that was not called for in the exposition. I have no doubt that a Pintle is good for your application, just not for the application that I was talking about.

I get a little snarky in this reply because I tire of young kids who think that because they have worked on one small part of a problem or technology that they then become the world's expert on all things space related.

The point, to keep this on topic, is that there is a great deal of technoarcheology that needs to be done to glean interesting technology that was researched at great cost decades ago that can be applied, whether generally, or in niches today. In fact we have a proposal into NASA headquarters right now for this exactly thing related to Apollo era Lunar Orbiter original imagery.

Posted by Dennis Wingo at November 13, 2007 01:36 PM

Relax, Dennis, Jon made a valuable point. I saw no oneupmanship in the post.

Posted by Lee Valentine at November 13, 2007 04:50 PM

"The point, to keep this on topic, is that there is a great deal of technoarcheology that needs to be done to glean interesting technology that was researched at great cost decades ago that can be applied, whether generally, or in niches today. In fact we have a proposal into NASA headquarters right now for this exactly thing related to Apollo era Lunar Orbiter original imagery."

An interesting and needed mindset. I am always spouting off at work that what makes our agency is not the physical plant, it is the institutional memory and that it should be defended vigorusly.

Nothing sadder than fighting a battle that was already fought several decades before at considerable expense and effort.

For all the criticisms, it seems that the armed services does a better job preserving and passing on institutional memeory than most other organizations. At least on the doctrinal level. I guess when the price paid was blood and the reward for failure is defeat and death, it becomes a thing far more cherished.

Again, the prisoners dilemma, how does taking time to archive information that may or may not become totally obselete and has an opportunity cost to properly do versues the abstract, hypothetical value to a future successor.

Posted by Mike Puckett at November 13, 2007 05:28 PM

sic, should be vigorously.

To add: Last winter I was looking over some data our predicessor agency collected in the late 1940's and wondered what the margin of error was on the analytical methods that were state of the art at that time and how much of the data was actually valid. Unfortunately, I could not easily ascertain what the lab standards were then. I have no idea what the minimum detection limits were with World War Two era analytical methods.

Now, current methodology is defined by the fed EPA thru the current federal regs.

Posted by Mike Puckett at November 13, 2007 05:36 PM

Sorry Denise,you are still just repeating folklore. Stating that a pintle injector is not good for a launch vehicle is just not true. Which feature makes it not suitable? TRW demonstrated LOx/Kerosene C*'s of over 98%. Isn't that good enough? And a pintle allows you to do a domed head end, which saves weight compared to a flat showerhead. Isn't low weight good? So good efficiency, low weight, hmmm, why isn't it good for an ascent engine? If heat transfer at the walls is your concern, do you know how that scales with pintle tip geometry? Or momentum ratio? Hmm?

The point is you are making a blanket statement, based on something that you have never investigated the detailed engineering trades on. If you had really done you homework on injector design you would not be so quick with a blanket condemnation.

Posted by Earl at November 13, 2007 07:03 PM

Earl

Who said anything about a blanket condemnation.

If the TRW data is so great, why did that injector not take over the world?

As another example, Beal found out the hard way that the so called benefits of H2O2 were not what the engineers claimed either.

Lets go for another one. Rotary Rocket claimed that with their high chamber pressures that they could do an SSTO. When discussing this with Konrand Dannenberg, his answer was, "one thing you must always understand, these systems leak". By focusing on one narrow area and neglecting the incredibly difficult task of integrating all those little combustion chambers with that rotary pump, you created a never ending cycle of complexity.

Launch vehicles are far more than just one little design part, they are an incredible trade between many variables and I have listened to this peanut gallery for 20 years exclaim how that if you just did x, y, or z, you could build inexpensive SSTO's that would fly for just over the marginal cost of fuel.

Still waiting. Even Elon has figured out that this Holy Grail is far harder and more expensive to reach than what the engineers spun to him. You want to compare their claims from 4 years ago to today?

Posted by Dennis Wingo at November 13, 2007 08:19 PM

Dennis,
these two statements:

'It was brought from TRW to SpaceX as the bright future of injectors, when the old timers knew that it actually underperformed in relation to the "standard" design.'

and

'I have no doubt that a Pintle is good for your application, just not for the application that I was talking about.'

are what I was referring to as a blanket condemnation of pintles for launch applications. did I miss something? If you can defend them with anything other than heresay ('old timers say', or 'why haven't they taken over the world') then I would be glad to see it. I have spent over $2.5M on pintle injector technology, and have at least a rudimentary understanding of the flow physics involved. I would not make any kind of statement about these or any other injector being only good for some applications, or bad for some applications. Unless you are deep inside someones trade space it is too hard to say anything.

Other than that, I agree with eveything else you said. Engineering optimization is an exacting science. To do it well, you need to throw out preconceived notions, heresay, folklore, blogosphere noise, even (gasp!) program manager input, and DO THE TRADES. Trust me, there is no "just do blank...". Elon thought he had some magic insight ("I'll build it like a computer!"), but actually claimed about other launchers (and I quote) "I don't understand why they cost so much." Not understanding why things are the way they are, and blindly assuming that eveyone else is stupider than you is a recipe for failure.

Rotary. Kelly. Pioneer. RpK. All said the same thing. 'We are really smart, trust us, and it will be different this time.' Well, people did, and they lost their shirts. Trust analysis. Question results. Dig for assumptions. Do the hard stuff. That is the only way to get anywhere.

Posted by Earl at November 13, 2007 08:58 PM

I would not make any kind of statement about these or any other injector being only good for some applications, or bad for some applications. Unless you are deep inside someones trade space it is too hard to say anything.

You know, you are right, I have not been deep into these studies nor have I spent $2.5M dollars on its development. I rely on the words of engineers who designed and built engines for decades and what they say about the different injector types. Their word is good enough for me on that one, so I can go on to solve problems that need to be solved.

I could have used ten other examples but this is the one that hits at the current fad of developing the next great American launch vehicles. I don't spend time on engine development because I can go out and buy engines that work from people who develop them. The big problems to solve today are out there not getting there, but that is another discussion.

That is an example of a macro cosmic trade. Your trade study told you to work on engines and mine on systems development. Both are absolutely necessary but at this time, if you want to make money, my trade said work on space systems. Only time will tell which will make it.

We need to all look at space every now and again in terms of first principles. What are we trying to do in space?

1. Make Money?
2. Work on Cool Space Stuff

Without 1., the opportunitites for number 2 are few. Without 1. which is a way to make money in space, what is the trade space and what is it that looks to be most profitable as that is where the space investor's money is going to go. That is where looking back in history at what was left on the scrap heap can score some big points. I do like what Planetspace wants to do in building an X24 variant, it would be very interesting to look at the economics of sending two people up to space, the number might change the way that the systems are done.

These are the kinds of trades that we need to make to see where we can push the ball forward fastest in the market and therefore get to have fun building stuff. That is the problem, money and money follows markets. Earl I could have used a dozen examples besides the Pintle injector. Fact is no one is using them (large engine manufacturers world wide). They may be wonderful and if they are then I look forward to the day that you take over the engine world.

Posted by Dennis Wingo at November 13, 2007 11:00 PM

Rand:
It's often called a "Paper-Virus" (which I found out after I read an SF-short story in Analog of the same name :o) meaning a certian conclusion or theory suddenly morphs to "fact" and becomes difficult to over-turn because it is now something that 'everyone knows' and becomes accepted without question.

Like the computer kind these virus' can quickly become the driving factors to inadvertantly (or purposly) supressing techniquies and areas of study by becoming the 'cold-fusion' of the day.

(My own example is an interview on the Space Show with a supposed 'expert' on Space Suit design who stated with no support that the Space Activity Suit {Skin-Suit/Mechanical Counter-Pressure suit) concept was un-workable and space suits would always be full-pressure designs. When questioned on the descrepency between what he was saying and the official reports findings he stated that the reports were all lies and that only "macho" astronauts had been used in the test so of course THEY didn't complain..... :o)

Randy

Posted by Randy Campbell at November 16, 2007 07:16 AM

These are the kinds of trades that we need to make to see where we can push the ball forward fastest in the market and therefore get to have fun building stuff. That is the problem, money and money follows markets. Earl I could have used a dozen examples besides the Pintle injector. Fact is no one is using them (large engine manufacturers world wide). They may be wonderful and if they are then I look forward to the day that you take over the engine world.

Posted by Firmalar at November 27, 2007 12:14 AM


Post a comment

Name:


Email Address:


URL:


Comments:


Remember info?