Target Fixation

guest blogger john hare

Work has picked up for us in the last several weeks, so I have been focusing on trying to make money instead of posting anything. Second reason for not posting is that several of the people that responded to my concept posts managed to effectively challenge the ideas in ways that I hadn’t expected. There was a far higher quality of critique than I usually see without writing a check. Continuing on with further possibilities on a concept in doubt is not very attractive, unlike throwing something controversial up once or twice. Over half the ideas on my short list of ideas to post now have known possible flaws that I have to rethink before moving on. Thank you to all the people that managed to get through to me exact problems in my conceptual visions, and to the ones that added to the ideas.

I had something of a grand scheme for a systems approach that seems to be better than most concepts I am aware of, to me at least. By breaking things down into the smallest subsystem I could describe, many answers to questions I hadn’t thought of came forth. I believe that if I had simply tried to throw the whole thing out there at once, it would simply have been dismissed (properly) as another  flawed champagne idea on a beer budget. The accurate information to divert me to other channels wouldn’t have happened.

The short version of the total idea was an HTHL flyback stage with a second stage that hits a tether. Both stages torus tank based with the upper nested in the donut hole of the first.  Two tipjet propellers on the first stage to provide cruise thrust and wingtip vortex control to get the induced drag down. Cagejet tuborockets in the vertical tails of the first stage for drag compensation and thrust to mach 1.6 (lightweight intake limit), with very high pressure, altitude compensating rockets kicking in at 20,000 or so feet and mach 0.6. Second stage uses whole bottom of vehicle as very high ratio aerospike from mach 6 to tether at mach 20. Tether reboost with 750 second tetherrocket. And so on  

In fighter aircraft there is a term called target fixation with a pilot so focused on his target that nothing else matters. Sometimes that focus results in getting kills, the scoreboard of fighter warfare. A high level of focus is required to hit a high speed twisting turning target. It has been noted many times that successful pilots are hunters, not hunted, and a very high level of self confidence is a job requirement. That only half the participants in a successful dogfight fly away is simply not useful information. 

Sometimes that focus results in getting shot down himself by the people on the other side that he excludes from his attention. Without a wing man or warning system that he will listen to, fighter pilots have an even lower life expectancy than normal with unfriendly people doing their best to do it to him first. When your buddy is yelling break break, it’s time to slam the stick over, stomp some rudder, and dump chaff and flares.

I am seeing a lot of this behavior in the rocket business, with many organizations so fixated on one target that they simply cannot see the guns on their six, and won’t listen to the people calling the break. Target fixation is a useful concept in the rocket business to distinguish between concentration on getting a job done, and setting yourself up for failure. The cannons of the fighter have a parallel in the march of technology, and the capabilities of the competition. Your business can be shot down by another company that is faster, smarter, or more agile in delivering what the customers want. The technology is simply one of the tools they use to get on your six.

The Griffenshaft is the most visible example at the moment, with billions poured down a rat hole of a flawed concept. The dozens of groups and members of his own calling the break seem to be unable to get his attention. The target fixation prevents the consideration of the quality of the target, whether there is a better target, and even if he is going to run out of fuel/funding pursuing this one. It is even worse if the wing men know they will be grounded and lose flight status if they call the break and interrupt his concentration.  You wouldn’t mind him getting shot down so much if he wasn’t taking so much useful hardware down with him. 

Commercial space has a few target fixations of its’ own that just might need to be addressed. I share in several of these so if I am pointing a finger, there are three others pointing back at me. The question in each case is whether we need to call a legitimate break, or if we are distracting people trying to make the shot.

Hydrogen is high on the list. I have joined many others in saying that hydrogen is more trouble than it is worth in almost all cases. Jon has pointed out that there are examples of existing hardware that are  mass competitive with  dense fuels for upper stages.

Clustering modules is very popular. Develop one stage properly and cluster as many as you need to get the job done. While attractive in some ways, getting 20 or more stages to play well together all the time seems like it might be tougher than building a bigger vehicle. Some very smart people are on both sides of this one. 

The single massive, do all rocket is one that comes up all the time. The Shuttle replacement or the Saturn replacement or the you name it is a dangerous fixation if a really good use for it is not justified. I think Kistler bit it on this one.

Pressure fed for simplicity is a standard for most newspace companies. You know where I stand on that one. With the rocket equation being what it is, a pump seems to be the inexpensive option compared to the sheer size penalty for low pressure orbital rockets. I have also seen plenty of complexity in the so called simple pressure fed systems. I never heard of a battery powered pump for a rocket until Paul started working on one.

High tech gets high performance. Sometimes high tech is just gingerbread. If it is not really needed, don’t waste your money trying to get wall to wall data on a pig that is designed wrong to begin with. 

ELVs being cheaper to develop is getting busted now, though some still believe it is cheaper to analyze problems than to test them. I don’t really care to depend on a vehicle with no flight history after a test flight program of perhaps two flights by similar vehicles. 

Vertical takeoff being lighter could possibly be flawed. A standard ELV on a runway trolley on a slight incline could possibly improve performance and reduce vehicle mass slightly. Side benefit is less ground infrastructure and therefore more potential launch sites.

Simpler is cheaper an more reliable. Not always.

There are many assumptions out there that really need to be examined. Focus on a single objective can result in making that objective, or in getting shot down by the competition that you ignore. Having been shot down in business a few times myself, I hope my friends out there have good wing men watching their six both technically and financially that they will listen to. I also hope the jamming from us wannabes doesn’t distract them from good targets.

The following two tabs change content below.


I do construction for a living and aerospace as an occasional hobby. I am an inventor and a bit of an entrepreneur. I've been self employed since the 1980s and working in concrete since the 1970s. When I grow up, I want to work with rockets and spacecraft. I did a stupid rocket trick a few decades back and decided not to try another hot fire without adult supervision. Haven't located much of that as we are all big kids when working with our passions.

Latest posts by johnhare (see all)


About johnhare

I do construction for a living and aerospace as an occasional hobby. I am an inventor and a bit of an entrepreneur. I've been self employed since the 1980s and working in concrete since the 1970s. When I grow up, I want to work with rockets and spacecraft. I did a stupid rocket trick a few decades back and decided not to try another hot fire without adult supervision. Haven't located much of that as we are all big kids when working with our passions.
This entry was posted in Uncategorized. Bookmark the permalink.

40 Responses to Target Fixation

  1. Bill White says:

    Hydrogen? BOTH of the recent COTS winners use kerosene on the first stage, correct? And I read that Taurus II uses a solid fuel 2nd stage. Do either COTS winner use hydrogen, anywhere?

  2. Gary C Hudson says:

    Some excellent thoughts in this post, John.

    While I was long known as an SSTO VTOL guy, in fact, I don’t care about the technology at all. All I have been looking for for the past forty years is low cost access to space. If I could achieve that goal using pixie dust, I’d do it. I think you are dead on when it comes to the issue of not becoming overly enamored of a particular concept.

    The flip side of that coin is when you change (for your own good reasons) everybody jumps on you about “not staying the course” or being a rocket dilettante. Kistler was an organization that couldn’t see its way to making a change. About 2000 I tried to convince them to build a Falcon-9-type vehicle with their available NK-33s and avionics, abandoning the RLV for the near term, but failed. If they had taken it up, they’d be holding SpaceX’s NASA contracts today. But they wanted to “stay the course.”

    On the subject of modular vehicles, let me say I have both designed them and opposed them (as well as competed against them, i.e., AirLaunch’s QuickReach vs. Microcosm). Today I tend to come down against them, but could be persuaded that they could be made to work with some effort. Pressure-fed vs. pump-fed is an interesting problem. I have a metric I use when listening to someone talk about a design. If they dismiss pressurization as easy and pumps as hard, and spend their time talking about their engine, then I know they don’t know what they are doing. Whenever I have worked on pressure fed vehicles, I have found the pressurization system costs me $2 for every $1 I spend on the engine itself.

    Simpler is by no means always cheaper. Commercial aircraft are wickedly complex and fantastically reliable.

    FYI, Mitchell Clapp was the first to suggest a battery powered pump to me, back about 2002. For certain types of low energy stages, it is worth trading. As I recall, we looked at it for RASCAL.

    Bottom line is to use historical data and studies but always keep an open mind. Oh, and don’t ever build a rocket with a fineness ratio greater than ten!

  3. Gordan says:

    @Bill White: Yes, none use hydrogen, though Musk contemplated a high energy upper stage in the past. It would probably be worth it only for GTO or escape trajectory payloads. It’s tempting to contemplate such stages, but one wonders how the price/kg to GTO would compare with regular F9(H). I’m thinking KISS wins here and big time.

  4. Gary C Hudson says:

    Hydrogen has its place. Even though I was one of the early proponents of densified LOX and hydrocarbons for SSTO concepts in the late 1970s onward, I think hydrogen has several advantages that need to keep it in trades. First, it burns very cleanly in atmosphere and engines and obviously with high performance and great cooling, second, in crashes and leaks it is less dangerous that propellants that don’t vaporize; third, it makes sense for air-launched vehicles where GLOW is a major (if not the major) design factor; and fourth, it can be obtained, along with LOX by electrolysis or other similar techniques and doesn’t depend upon extraction industries, so long as one has electrical power available. That may be important someday.

    Finally, it will be most likely important if one is going beyond LEO and using ISRU.

  5. Jonathan Goff Jonathan Goff says:

    What Gary and Gordan said. Sure, I wouldn’t use Hydrogen for a terrestrial ground-launched first stage. But like Gary says, there are several places where it looks like the tradeoffs might really favor LH2. While Old.Space tends to be biased toward LH2 more often than it should be, I think New.Space tends to go too far in the opposite direction.


  6. John Carmack says:

    Gary — I have heard you make the comment about expensive pressurization before, but I would like to hear some elaboration about it. At Armadillo we are currently pursuing two routes to improve performance over blowdown, and neither of them are very expensive.

    Modern carbon bottles for high pressure helium are good and inexpensive, and we only need a high pressure ball valve under computer control to act as an extremely high flow regulator. You can make a pressure fed TSTO with mass numbers like that, although admittedly with a tiny payload. Using heat exchangers or Trydine with a catalyst to reduce dry mass will certainly add complexity and expense, but I have a hard time seeing how it is of the same order as the engine.

    We recently started flying a vehicle with self-pressurized lox-methane, which is extremely simple, and seems to offer a lot of systemic advantages. I understand that there are issues to be dealt with regarding propellant conditioning to get optimal performance, but still, I don’t see how that can add up to all that much. Did the Air Launch work turn up a lot of unexpected problems?

  7. Gary- I am an N-Prize contestant and I had some interest in Target Fixation. I have noticed that some of the other teams seem to be fixated that they can get a small conventional rocket into orbit on less than $2000 USD worth of fuel. I have not been able to find a design that could achieve orbit and still be say half the size of the Space-X rocket. It just wont add up. Can I get a realistic opinion on the smallest conventional rocket that can archive orbit?

  8. Roger Strong says:

    Munroe- Wouldn’t the cost of the launcher – and the launch itself – be so much more than the fuel costs that those fuel costs are not really an issue?

    I’ve seen this cited as another form of target fixation: Scram jets on a HTHL launcher to use less launch fuel, but which add much more complexity and cost.

  9. David Summers says:


    Oh, and don’t ever build a rocket with a fineness ratio greater than ten!

    Why is that? Don’t you like pointy things? (Seriously though, why do you say that – isn’t that lower than several existing designs?)

  10. Gary C Hudson says:

    That’s actually a quote from von Braun. An acquaintance of mine knew von Braun (as did I…shows my age) and reminded me of it in the context of Ares 1. Interestingly, this acquaintance is a former launch vehicle designer for NASA JSC…

    The reason for the comment is that while it is possible to go longer than ten to one, it moves into uncharted territory and there is virtually no reason to do it when starting a new design. If one is modifying an existing booster (the Titan series comes to mind) then just maybe the trade comes out in favor of higher fineness ratios. But it is not necessary with a clean sheet design. The problem with Ares 1 is that the silly idea to use “Shuttle-derived” SRBs sent NASA down the wrong path from the get-go. Of course, we now know the “shuttle-derived” SRB is nothing like what was promised. I like to joke the only thing common between the current SRB and the Ares 1 are the clevis pins that hold the cases together. Maybe a bit of an exaggeration, but close.

    My consulting rate is pretty high. 🙂 More than the cost of the propellant… But the short answer is that I think an orbital booster can be designed in the 10K to 20K GLOW range. It is possible to build an very small orbital launcher for $2K of propellant cost, but in the end, that would turn out to be perhaps 0.001-0.004% of the total development cost for such a vehicle.

    The response would be fairly long, so perhaps we’d best take it to email. I will say that we decided methane was a much better propellant for VaPak systems than propane for purely operational reasons, not related to Isp or density.

  11. “””””””Hydrogen is high on the list. I have joined many others in saying that hydrogen is more trouble than it is worth in almost all cases. Jon has pointed out that there are examples of existing hardware that are mass competitive with dense fuels for upper stages.””””””””

    I’m intrigued by what I would guess are near-term “HEDM” (High Energy Density Matter) propellants. I’m not talking about anything esoteric, but, rather, kerosene derivatives (I think “derivative” is a correct designation [?]) like quadricyclane and JP10. They are both more energetic and denser than RP1, and have been produced in quantity at — I believe — reasonable prices. I also understand that their is a proposed technique of pre-chilling these propellants to make them even denser in the launch vehicle’s fuel tank(s). I read in one post somewhere that quadricyclane burns a bit rough, and needs to be mixed with RP or JP.

  12. “””””””Bill,
    What Gary and Gordan said. Sure, I wouldn’t use Hydrogen for a terrestrial ground-launched first stage. But like Gary says, there are several places where it looks like the tradeoffs might really favor LH2.””””

    Hence the concept of a tri-prop engine for SSTO. Such engines have been built (as you know — the RD-370/371 series comes to mind), but are complex. The best solution — if you’re going to that route, in my opinion — is the Thrust Augmented Nozzle concept Aerojet came up with, where you couple this “TAN” with a LH2 engine, using RP or other hydrocarbon for the TAN “afterburner”. It’s simpler that a triprop engine, and, as serendipity would have it, it appears to solve the nozzle expansion ratio conundrum.

    Hasn’t been flight-tested, though.

    Oh, and for the record, even if such an engine could allow for a SSTO RLV, I would be inclined to put the boost-phase portion of the propellant in a drop (external) tank to minimize the size of the orbiter, and to minimize weight penalty, and make the system more robust. The tank can be parachuted into the ocean and recovered.

  13. Sage says:

    “Target fixation” is not only about “how” to get a job done, it’s about focusing on a specific in deference to the big picture. In the commercial space industry, one fixation is on “launch vehicles.” Most think about “heavy” launchers as a way to provide services to those who want to launch “big” payloads. And because of this, the industry has generally focused on developing large launch systems, which is clearly the wrong path towards “space access.”

    There is no “magical minimum” GLOW for a launcher, given it retains structural integrity and has the necessary performance and mass ratio. Considering the effects of increased drag loss and certain system dynamics due to scaling, a launcher still needs to be large enough to put its payload into orbit; the lighter (and smaller) the items such as relative structural mass, avionics, and payload become, the smaller the vehicle can also become.

    So long as the vehicle maintains integrity during flight, possesses the required mass ratios, and accounts for any reductions in performance dynamics due to scaling issues, the vehicle can be designed as small as one can reliably engineer it. Without getting into specific solutions to minimum gage issues, the practical size is mostly limited by the required avionics hardware, cooling considerations and engine efficiencies, structural integrity and acceptable drag loss, and the actual size of the payload fraction.

    Although it’s not any easier to design a tiny launch system than a giant one (and it may in fact be more difficult in certain regards), it is much cheaper and easier to actually fabricate, test, and launch it.

    In my view, hydrogen is a relatively bad choice for first stages. From a pure physics viewpoint it makes some sense for upper stages where the overall density is less important, and the increased Isp comes into its own.

    Pressure-fed systems are generally easier to engineer and implement, but there are enough trade-offs (and performance hits) to make pumps attractive for the first stages. Depending on the system, a pump-fed first stage may indeed make the whole launcher easier (and cheaper). I’m not really convinced that upper stages gain enough benefit to make it worth the extra hassle of using pumps, usually. And, pressure feeding generally provides a more expedient path to actually firing up an engine. So, much of this is really goal dependent.

    The idea of electric pumps is not new. They have actually been used in flight, for instance, on the secondary propulsion system of the Agena upper stage. I myself talked with NASA regarding the potential of electric propellant pumps more than a decade ago (prior to my work with JPL), and more recently through formal proposal submission regarding specialized rocket applications. Some companies (including Lockheed) claim certain IP on variants, most probably unenforceable as such has been both discussed openly for a long time and used long ago (I think even Goddard used electric pumps at some point during testing).

    Anyway, without getting into specifics, for certain applications electrically driven systems can rival a gas-generator cycle if implemented correctly (if employing cutting edge technology); this is not true across the board, but there are applications where electrical systems become competitive. However, composites can also get you respectable results. So again, it all depends on your goals…


  14. Habitat Hermit says:

    Monroe L King Jr. are you familiar with SS2S? Granted their aim seems to be “just” reaching space rather than orbiting but multiply the volume of that rocket by let’s say ten and it’s still a fairly small rocket.

    Gary C. Hudson wrote:
    “Oh, and don’t ever build a rocket with a fineness ratio greater than ten!”

    I thought I might be missing a joke but now I’m confused; is that ratio & statement based on measuring the width between the tips of the fins rather than the maximum width of the body?

  15. Gary- Thank you 10K is way too high as I suspected, the comment about the launcher doesn’t pertain to the N-Prize because the new rules, actually the old rules but there are some rule changes that reduce the goal to FUEL. It’s more complicated than that but it is the main concern. I just wanted some realistic number on the size of a small conventional orbital launch vehicle.


  16. Monroe says:

    I’m sure SS2S is going to get there (I have chatted with Richard Nakka) but that is to “Space” not orbit. “Go-fast” already proved a space shot. What we are looking at is cheap orbital insertion from that point exactly. Specifically with a Light Gas Gun. The acceleration is a huge problem however for a track able satellite. Thanks for any comments I know the N-Prize is Laughable to most real scientist, But that wasn’t the idea in the first place. I might be interested in some of that consulting if your really interested in an N-Prize attempt and you can contact at


  17. Jonathan Goff Jonathan Goff says:

    I definitely agree with your points three and four. Pressure fed makes sense in some cases, but there is a bit to be said for pumps. There was an interesting analysis in the latest AIAA Journal of Propulsion and Power about electric driven pumps vs. pressure-fed systems. As you say, there’s benefits and drawbacks to both approaches, and Your Mileage Will Vary depending on system details.

    I kind of like the idea though of using brushless electric motors as a sort of electric “transmission” for turbopumps though. By doing that you can avoid having to do interpropellant rotating seals, and adding in something like a TAN system with its own pump becomes a lot easier. For instance, imagine taking an RL-10, and splitting off the LOX pump entirely with it’s own electric motor. You then put an electric motor/generator on the LH2 side, and take the excess power that isn’t being used by the LH2 pump, and converting it to energy to run the LOX pump. Get rid of the big gear box, make startups a lot simpler (since you have a great high-torque motor to get things started), get rid of interpropellant seals and stuff, make adding TAN easier…And with good brushless DC motors being so high power efficiency, you really don’t lose much power compared to a more traditional gear-based system.


  18. Monroe says:

    “There is no “magical minimum” GLOW for a launcher”
    There is a minimum it may not be magical but it is there. My point is and always has been it’s to big to make an interesting attempt for a completion such as the N-Prize to costly for such a small Prize spending that much money just doesn’t make since when you could go for the X-Prize if you can achieve orbit with something that big.
    I don’t suggest it cant be done but John- did say “unless your a multi-millionaire find another hobby” and I have to agree any conventional attempt would be too expensive. Even an unconventional attempt without a quarter-million dollars to spend for permits and insurance in the US would still be hypothetical.


  19. Don’t very small launchers suffer from some inefficiency due to a basic solid geometry phenomenon? Larger launchers can hold proportionately more propellant in relation to their dry weight.

    This is not a knock on very small launchers, merely stating what I believe is a fact.

    I think that there is a significant role for very small launchers because of the potential market for both micro-and nanosats.

    It also seems to me that very small launchers would be ideal for a ground-based boost assist, like a catapult tower or gas gun. A radio-tower tall structure supporting a gas gun-driven catapult could impart at least high-subsonic speed to a small launcher, thus dramatically increasing its payload capacity.

    The other option is balloon launch (nothing new there!)

  20. David Summers says:

    Well, a really small solid would probably be pretty easy – as long as it didn’t have to do anything fancy. When I looked into the scaling papers, they had some pretty silly assumptions to make small solids look bad. To a certain extent, the problem you have with evaluating “the smallest launcher” is exactly the same “target fixation” that the main article describes – when you are doing the trade, you weigh your “special sauce” implementation of X against a standard implementation of Y.

    An example: in one paper on minimum launcher size they made the solid rocket lose. This was very hard, because according to their initial assumptions the solid worked the best. But they were presumably liquid guys, and “knew” that wasn’t true – so they added some assumptions for the solid, that it had a minimum diameter and could not be throttled. I’m sure that Estes would be very interested in this minimum diameter theory, and while solids are not throttled per se I don’t think a truly constant thrust curve has ever been achieved…

  21. David Summers says:

    Larger launchers can hold proportionately more propellant in relation to their dry weight.

    Oh, on this – in general, this is not true. As a tank’s volume goes down, the tank mass goes down at exactly the same rate – there is exactly 0 economy of scale in tank design. That does fail when you reach “minimum gauge”, though, where a sneeze will rip your rocket in half.

    (Metals can be made arbitrarily thin, but external forces subjected on your rocket have some parts that do not scale well.)

  22. Tim says:

    Are pressure-fed and pump-fed cycles mutually exclusive? I have to wonder if there’s a moderate performance rocket with simple low performance pressure feed and a simple low performance pump sitting in a trade space somewhere that noone can see because of the pressure-versus-pump debate.

  23. Monroe says:

    In the heat of the atmosphere a pressure system is really cheaper and reliable and makes a lot of sense especially for shorter duration. In the vacuum of space a pump seems to make more sense especially for longer duration more reliable operation. depending on the mass of the launch vehicle a pressurized first stage and pump second stage seems a likely combination considering the mass of the first stage is pretty large and under greater acceleration and operating in warmer conditions. a hybrid would have the good qualities of both but also the weight of both and the complexity of both. But what is the goal here? Cheaper or easier. John lost me at:
    “Using heat exchangers or Trydine with a catalyst to reduce dry mass will certainly add complexity and expense, but I have a hard time seeing how it is of the same order as the engine.”
    I could see using heat exchangers to the engine bell. Trydine can be explosive with Helium and sounds pretty risky, but I’m sure he knows what he’s doing. Helium has awesome heat transfer ability and it would cool the nozzle. You could have probably 2 exchangers on the bell and use one as a redundant in case of failure and still be lighter and less complex. But what do I know?


  24. Sage says:

    First, as a quick comment to your last post… Actually the opposite of what you state is true, and I mentioned this in my above post already. If you are going to use a pump on any stage, and you could only pick one, the first stage is where you would use it. One reason for this is due to the nozzle expansion ratio and the pressure at which you need to run the engine; the nozzle needs to expel the gases to the atmosphere which limits the expansion ratio on first stages, and thus reduces the engine performance – by designing the engine for higher pressures you can mitigate some of this loss in the nozzle design (and also increase the combustion efficiency in the TC). A pressure-fed system requires heavy tanks that could hold the higher pressures; but at lower pressures, pumps don’t realize the same relative advantages. Outside of the atmosphere you can employ very large expansion ratios and lower pressures – there, the complexities of higher pressures start to yield diminishing returns system wise. Because of the feasibility of lower pressures, the tank weight can become smaller until it is only slightly higher than what you might need anyway to feed a pump system. Thus, upper stage performance on a pressure-fed system can be quite respectable if implemented correctly. There are still applications for exoatmospheric pump-fed devices, but such is not necessarily optimal for a small launch vehicle.

    The minimum size launcher depends primarily on the factors mentioned in my earlier post, and for an N-Prize attempt that’s principally a function of how small you can make the avionics and rocket frame, while achieving the required mass ratios, engine performance, and structural integrity, while also accounting for any scaling effects (including increased drag loss); there is no “magical minimum.” A physical minimum may indeed exist (i.e., based on constraints applied to the task, technical approach, etc.), but such is not near the size of a reasonable N-Prize launcher within the N-Prize guidelines.

    Note that the N-Prize is about doing the nearly impossible; it’s an intellectual challenge. The rules are quite brilliant and the nature of the contest puts it square into the realm of intelligence, not cash. No offense to the X-Prize organizers, but I have never been that enamored with the X-Prize rule sets/contests. One of the reasons is because of the large “cash bias;” the N-Prize, which is more difficult than the original X-Prize, is fairly devoid of this.

    I would strongly disagree with anyone who claims that rocket technology is solely in the domain of the “super wealthy.” As with many things, money can be an important factor. But, as far as the N-Prize is concerned, even if you do have plenty of money, it’s not really going to help if you don’t know what you’re doing. It’s quite easy to throw away tens of millions with almost no results in that way. In other words, it is quite unlikely that all the money in the world would have helped a Roman soldier build a nuke; on the flip side, plenty of money didn’t help Graham Bell beat the Wright’s to heavier than air powered flight.

    If you’re trying to making “bigger” engines, then you’ll need “bigger” money for the infrastructure, materials, insurance, testing, etc., etc. You will need all these elements for small rockets too, just on a smaller scale. The smaller the rocket, the smaller the infrastructure costs, and the smaller the material costs, and the cheaper the fabrication costs, and the less the potential damages (and thus the less the insurance/liability costs). Everything becomes less, except for the complexity of the system (which stays about the same or may actually increase). And, such design complexity is an “intellectual” challenge more than a “cash” challenge.

    As you noted, certainly electric pumps have some advantages. I did notice that paper in the Journal of Propulsion and Power but haven’t analyzed it yet; I may submit my own studies in the future (depending on my available bandwidth and such), but if not (and if there’s enough interest) I might do a public white paper on the topic.

    As a quick note though, in modern closed-cycle systems you have virtually no pumping loss; a switch to a basic electric system will result in a fairly large hit. In open gas-gen cycles you have pumping losses and the switch to an electric system can be fairly competitive (in some applications); as such, you get most of the advantages you mentioned (and others) in certain applications. But personally, I would leave the (beautiful) RL-10 alone; I seriously doubt that you would get a better engine by moving to electric drive given the current state of technology and that class of device (but that’s just my opinion).

    Current systems work pretty well; the use for electric drives (in my view at least) is mostly application specific and targeted to where such makes sense. As you scale the engine down, raw turbine performance gets worse. Electric motors also get worse, but tend to retain their efficiency better; the scale constraints and dynamic is also different. One of my proposals for such pumps was directed towards small rocket applications as a compromise solution between performance and complexity. My analysis suggested that an advanced and properly implemented electrically driven system could theoretically outperform standard pressure-fed and open-cycle systems at smaller scales. While I’ve devised methods to make electrical systems more competitive with closed-cycle designs, these are generally more complex solutions and can approach (or even surpass) the complexity of traditional closed-cycle engines (so you’d need a really, really good reason to move that way).

    Anyway, I’d love to discuss this more, but I think I’ve already “over posted”…


  25. Sage says:

    The items small launchers suffer from are minimum gage issues, increased drag losses, certain efficiency and cooling concerns, and the fact that certain items, such as avionics packages can only be made so small with existent technology. Structural integrity can become a problem; it’s made a bit worse because the rockets need to have relatively smaller diameters to compensate for the drag losses. Engine performance may become somewhat affected at exceptionally small scales, depending on the engineering, but you typically won’t have extreme engine efficiency problems.

    I pretty much agree with most everything you said. And yes, some “studies” are just laughable. Most start out with their desired results in mind, and then carefully construct the study to “yield” those desired results.

    It’s not that they are mutually exclusive throughout a whole launch system (for instance, pump-fed first stage and pressure fed upper stage is reasonable). But, for a single particular engine it does make engineering sense to use one system. Think about how a pump system typically works – you use the propellants for running the turbines and cooling the engine. So, if you have already designed a pump system for one propellant (including engine pressures, tank size/weights, cooling, plumbing, etc) then you’ll need to have a really compelling reason to run the other propellant with a pressure feed noting that the design will be quite different — and if you manage this, you’ll likely be inherently limited to the pressure-fed limitations of the system. So if that’s the case, you mind as well just use the pressure feed since you’re then not going to realize significant added performance by splitting the feed (and you’ll also have more complexity). I should note that “low-performance” pump is almost (but not always (in certain applications)) a contradiction in terms, because the general idea of using a pump is for better overall performance. Also, note that propellant pumps can be driven from a single turbine.


  26. David Summers says:

    Tim, Sage:

    How about this for a useful (maybe) mash of pump and pressure fed: A LOX/LH rocket engine, where the LOX is pressure fed and the LH is pump fed. The pump is powered by the pressurized LOX, simplifying turbine design (cold gas/liquid instead of hot gas).

    The advantage is that LH tanks are heavy, and LOX tanks are light – so the light LOX tank holds high pressure without messing up margins, and the heavy LH tank is extremely low pressure to make it as light as possible. You don’t need a gas generator anymore (well, you do have to pressurize the LOX tank somehow – but maybe you can get by with a blowdown system).

    Um, surely this must have a use somewhere?

  27. Sage says:


    One more step and you’ve “almost” reinvented an expander cycle (and flipped it around for the LH2 case). Let me explain… You need a pressurizer for the LOX to run the turbine – but, why would you need this? Since you have already conquered turbine development and are going to need to address engine cooling anyway, you’d might first think to use the LOX as coolant and the high pressure regen output to drive the turbine (and by doing that, you’ve reduced the oxygen tank’s pressure handling requirement and the requirement of the pressurizer). But, if you’re doing that, you’ll probably (pretty much certainly) decide to flip it around and use LH2 for the coolant — and then, you’ll arrive at something like the RL10 (one of my favorites). Now, for a hydrocarbon fuel you could do a similar thing, but in that case you would probably use LOX as the coolant (rather than the fuel). Anyway, in the case of an expander, you might be thinking you could maybe get some higher power with external pressurant (even at the expense of weight), but a system performance improvement is doubtful, and we’re probably looking at upper stages with LH2 anyway, so you’ll want the higher efficiency of an expander.


  28. Tim says:

    I was thinking in terms of using both a pressurant gas and a pump to feed each propellant into the chamber to acheive ahigher chamber pressure than could be reached by each system alone, rather than using a different method for each propellant. I suppose I’d define a low performance pump as one that provides a relatively low chamber presure.
    I don’t necessarily think that using multiple pressurising methods leads to better performance, I’m just concerned that veiwing such systems as rivals obscures solutions that use both methods.

  29. Monroe says:

    Humm. Alright then. Anyway I found a nice paper on Hybrid’s that explains the Trydine system John mentioned and it was quite interesting
    It may be useful to someone.


  30. Monroe says:

    There are a lot of problems with hypothetical hardware. I haven’t seen anything novel about pumps in a Very long time. I mean physically all this is not new and has been pondered over for decades. A realistic launcher can be designed using what ill call common components of the trade and it’s not anything really new. It will cost tons of money and it wont be less than 7k lbs and that’s with flimsy everything. I don’t think that is an intelligent attempt for an N-Prize rocket is all im saying. It would be intelligent for an aerospace company that is going to launch nanno satellites. And forget the N-Prize winning that wouldn’t matter.


  31. john hare says:


    One of the points of pressurization and pump is the ability to bolt a new type of pumped engine to a thoroughly tested pressure fed vehicle. By making the engine compatible with other companies’ vehicles, your market expands. I think it is safe to speculate that Lynx, Pixel, and MA-02 will each have much more investment in getting the total vehicle right, than in the core engine. The ability to bolt on an engine that gets another 25 seconds Isp could be attractive to several companies when they are starting to reach the performance limit of a particular vehicle and want more.

    With good pressure into the pump, the suction characteristics are relaxed to the point that you can almost ignore cavitation, which makes the pump both easier to do, and less critical to the vehicle. The two methods can possibly be designed in as safety back ups for each other.

  32. David Summers says:

    John, another way to look at it is that most of the currently flying engines already are pressure fed / pumped hybrids – it’s not like the tank hold the propellant at ambient pressure! I’m pretty sure that most tanks are at 10 atmospheres or so (I’m not sure of the numbers – but non-trivial pressure).

    Turbopumps require a certain amount of pressure to suppress cavitation – and they also are a pressure multiplier, so higher inlet pressure = higher outlet pressure.

  33. Sage says:

    You’re incorrect in this case. A conventional launcher is an excellent and intelligent choice for an N-Prize attempt; your conjectures of hardware mass limitations and other items are simply not compatible with the physics.

    In the case of a “low performance” pump as you mention (i.e., low chamber pressure) your overall mass savings with a pump are going to be small (if any). As I mentioned, practically all pump fed systems still require a pressurized tank (to some extent). But, using both systems for increasing chamber pressure (much beyond the tank pressurization requirements currently used for pumps) suggests to me an increase in complexity (and probably weight), but there may be some uses (for instance, maybe as a possible multi-string/backup implementation and/or potentially reducing pump inlet requirements). Still though (and regardless of the merits/drawbacks of such a system), I do agree with your point that we should be careful not to exclude synergistic approaches (and that’s true in general).


  34. Monroe says:

    We shall see I hope! 🙂 keep plugin.

  35. One issue relevant to the notion of “target fixation” are the huge expenses and long development times that seem to have become an accepted part of the aerospace culture, at least with government-funded programs. This forces program managers and policy makers to lock into a concept which will not survive challenges from more agile competitors. Of course, it has been the culture for so long because agile competitors have been nonexistent.

  36. Alan Kruppa says:

    David, I saw your point about tanks holding the propellant at 10 atm or so. Isn’t that more for structural integrity against buckling of the tank skin? Or is it just a nice benefit born of the need to provide some inlet pressure to the turbine? Forgive me if that’s a stupid question.

  37. john hare says:

    My understanding is that pumped systems usually shoot for 2-4 atm in the tank, while pressure fed hit about 20 on average. For small vehicles though, minimum guage on the tanks tends to diminish the mass savings possible.

  38. David Summers says:

    Alan, I think it is a little of both. Most importantly, the turbo pump mass goes way up if the tank pressure is too low – so there is a relatively easy trade off there in favor of higher tank pressures. The structural integrity benefit seems to be mostly incidental – it is still somewhat rare for designers to rely on “balloon tanks”, as they are known (only early Atlas, as I recall). It is true that all liquid rockets I know of rely on it somewhat, though. (I think the Space Shuttle was supposed to be the exception – it was spec’d to not require tank pressure at any point, but they missed the spec by a bit during the worst case loading.)

  39. Alan Kruppa says:

    Thanks David. That makes sense. I’m a bit naive about turbo pumps but I think I’ve got this correct. Turbo pump mass vs. tank (i.e. inlet) pressure will approximately follow an inverse (negative slope) logarithmic curve, right? I understand how tank weight scales linearly with pressure, so I’m asking the question to better understand the trade off.

  40. Sage says:

    Tank pressure in a pumped system is to provide positive pump feed, and in many tank designs, also increase the structural rigidity. You need a positive inlet pressure, especially with centrifugals; note that the output pressure is related to the positive head, and increasing this reduces some of the inlet requirements and affects items such as cavitation onset as well; the physical characteristics of the pump design are thus affected by the inlet pressure. As you pump from the tank you need to maintain positive feed which is generally achieved through an adjunct pressurization system. As John noted earlier, about 3 – 5 bar is a reasonable target for pump fed systems but this depends on your design; 10 bar is not unreasonable.

Leave a Reply

Your email address will not be published.