Pumped Monoprop

Jon mentioned that I argue pumps over pressure fed. That is partially because the first thing I tried to do seriously in rocketry was build a pump for some friends for a project in Huntsville. I started with the base assumption that turbo-pumps were just too hard and something different was needed. I built a series of pistonless pumps that never did work properly. Steve Harrington with Flowmetrics uses the same cycle only he does it right with a totally different design. Mine was 8 dividers rotating in a cylinder. Jon and I went over it during an hour  layover in Salt Lake in 1999. That was back when you could get on a plane with large metal objects of uncertain purpose.

During the time I was trying to get these things to work, we were looking at modifying jet engines for use on an almost space plane. It was pointed out to me that the fans and turbines in a jet engine had a tip speed 5 times that required to pump the sorts of pressure we wanted. These 1950s vintage jet engines had fan and turbine tip speeds of 1,000 feet per second. A LOX pump with a tip speed of 200 feet per second can deliver on the order of 250 psi. So I started studying turbine and impeller technology.

This field is difficult, convoluted, expensive and covers several disciplines that I am not equipped to understand. So I did my thing and reread the references to figure out how to game the problem. Eventually I decided that the normal approach was wrong. The normal approach seems to be to get the very best rocket engineers, the very best pump engineers, the very best CFD crew, the very best etc and have them all do their specialty to some tight specs to the best of their ability. Over the years this method has turned out some good technical engines, just not all that affordable.

I think the people making the over all decisions need a general purpose expert to stand back and look at the whole picture. Then they need to do a whole system approach to the engine with an eye to making everyone’s job easier and more team oriented. I know that I am not qualified to do any of the real jobs involved here, but I do think I have identified an approach that experts could turn into an unbeatable combination, until the next unbeatable combination comes along. And competent newspace people could build affordable staged combustion engines with the performance, reliability, maintainability, and cost that would facilitate RLV development.

The sketch at the top is a pumped H2O2 monoprop idea. By putting all the components in a single pressure sphere, heavy housings become unnecessary. By stacking the components in order, pressure plumbing becomes unnecessary. Building a very large L* is less of a problem because it is all one symmetrical housing. Accessing the equipment for maintenance is relatively easy if it separates at the equator. All the parts can be removed through a very large opening.

At the top is a low pressure H2O2 inlet straight to the impeller. The 4″ diameter impeller at 12,000 rpms could create 250 psi minus efficiency losses. A residential water pump impeller is so overbuilt that you can use it if it is compatible with peroxide. The impeller discharges into a bowl volute which is much lighter and simpler than a normal one, at the cost of efficiency in converting the velocity head to pressure head. The catalyst pack top support and center bearing support are directly under the volute.  The catalyst pack surrounds the bearing housing down to the bottom bearing support and catalyst support.

The turbine nozzles connect the lower bearing and catalyst support to the sphere wall. The turbine is powered by the decomposed H2O2 to drive the pump. The sphere shape is to allow the use of a much larger diameter turbine than impeller. Impellers want to run at a few hundred feet per second tip speed, turbines really want over a thousand. The normal solution is a gearbox and weird plumbing. The sphere shape allows an impeller tip speed of 200 fps while the turbine has a 600 fps tip speed. Not efficient for a turbine, but the same 200 fps as the impeller would not have developed enough power, and a higher speed for efficiency would have mandated a gearbox. By using the whole H2O2 flow, even this medium turbine speed can power this system.

The turbine exhaust has more than sufficient time to get organized as rocket exhaust through the throat. There would also be plenty of time to afterburn some fuel in the sphere lower half if you solve the problem of pressurizing the fuel and cooling a lot of surface. The 250 psi theoretical pump pressure could be expected to yield about 100 psi below the turbine after the efficiency losses and pressure drops are figured.

I’m not really a fan of peroxide, but that is more a matter of cost and availability than anything else. Peroxide is a self pumping fluid if done right though. By using decent pump and turbine designs, the above speeds could be doubled which would quadruple the pump pressure and get you well on the way to a more efficient engine with good T/W. By using aerospace design practice, with engineers experienced at it, Final pressures of a few thousand psi could be realized in the lower chamber.

My bi prop ideas will come later. Hill and Peterson is my primary reference. Hutzel and Huang doesn’t get deep enough for good understanding of the issues. Sutton doesn’t help.

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.

25 Responses to Pumped Monoprop

  1. Habiat Hermit says:

    Interesting but the more I think about this the less I understand it. Am I right in thinking the impeller and turbine don’t run independently of each other, i.e. that they share the same axle and rpm and that the difference in tip speed only comes from the increased distance to the center?

    If I understand the setup correctly –which I probably don’t– the impeller and turbine will in total create as much pressure as the highest value either of them could create on their own (at the same rpm) –no multiplication involved as I see it.

    Let me try to describe my objection differently: the chamber curves out and then curves in so except for any difference between the inlet radius and the outlet radius it doesn’t have an impact. Inside the chamber there is essentially just one moving part: a combined impeller and turbine. This runs at the same rpm and is for all intents and purposes just like a single “fan”: the impeller can’t suck more fuel out of the fuel tank than the turbine can move further along: if the turbine is more efficient than the impeller at creating negative pressure (differential at the inlet, or positive pressure differential at the outlet) it doesn’t add (unless it allows the combination to run faster and it should but the throughput should be equal for the same total work being done so…) or multiply (at all) the work being done by the impeller or vice versa.

    My mind doesn’t bend around having the combustion take place in the lower part of the chamber but a regular combustion chamber can be added below the drawing can’t it?

    What am I missing?

  2. Have you looked at the Telsa turbine/pump combo? I’m not versed at all in its capabilities and the only examples of I’ve seen are low power applications, but has the aerospace community looked at them at all?

  3. john hare says:


    Jon and I discussed them about ten years ago. I don’t remember the others in the discussion. There is a problem with hot blades warping, weight, efficiency, and required development effort, IMO.

    I think it is more profitable to salvage a turbine engine with the right size turbine and have all the goodies developed by others. Jet and APU turbines already operate in a hot oxygen rich (lean) flow at high pressures. The ones that have 30-1 compression ratios are used at ~450 psi and well over 1,000 degrees turbine inlet temperatures with less than half the available oxygen burned.

    You need an A&P mechanic, a versatle engineer, and access to a boneyard instead of a large team of designers and builders. Liquid impellers are commercially available if you can’t salvage one. You need to build a housing for the different application and match the power to the liquid flow. Compared to that, developing a Tesla system seems a seriously daunting task.

    The J-85 and J-60 engines of 1950s vintage compressed about 70 pounds of air per second to about 120 psi. Pressurizing liquids is massively easier than compressing air. It should be possible to pump ten times as much liquid to 500 psi with an essentially unmodified turbine and shaft assembly. I remember picking up turbine disks or the driveshaft with one hand. The difficult part is going to be finding a turbine unit small enough to match current requirements.

    Now if you can find commercially available Tesla units, all the above is called into question.

  4. john hare says:

    Habitat Hermit,

    Yes the impeller and turbine run on the same shaft at the same rpm.
    All the pressure rise comes from the impeller. The peroxide decomposes, heats up, and expands through the turbine producing power to drive the pump. The impeller raises the velocity head to ~250 psi. The volute diffuses the velocity head to pressure head or just pressure. The high pressure peroxide is forced through the catalyst pack which decomposes it to steam and hot oxygen. The steam and GOX expand through the turbine nozzles and accelerate to drive the turbine. The 100 or so psi remaining is my estimate of remaining pressure after efficiency losses in the impeller, volute, and turbine. Plus the pressure drops through the cat pack and turbine.

    It is the same cycle as a turbojet engine except that liquid peroxide is pumped instead of compressing air, and the fluid is expanded by decomposing peroxide instead of burning jet fuel with the compressed air. In either case, a fluid is pressurized, then heated and expanded, and then used to drive a turbine with remaining pressure being the jet exhaust. Or in this case rocket exhaust. Think of it as three systems working together. First is the compressor or pump system that increases pressure. Second system is a combustion chamber or catalyst section which rapidly expands the gas/liquid. Third is the turbine which subtracts energy from the hot gas and turns it into work or shaft power. The turbine exhaust is at much higher pressure than the peroxide supply with something less than half of it being oxygen.

    The reason for the sphere shape it to velocity match rotating components in a simple and light manner and has nothing to do with the cycle itself. Afterburning in the lower half of the sphere is thermodynamically similar to burning in a seperate chamber below the sphere. It should be much lighter to burn in the sphere.

    I just remembered that I have seen impellers refered to as turbines. This could create a serious amount of confusion. If this is the case, then my use of turbine is very strictly as an engine that subtracts power to drive something else. I don’t always communicate well in print, so I would appreciate you bearing with me as I try to clarify my meanings.

  5. john hare says:


    Didn’t mean to misspell your name in the last. Sorry.

  6. Habitat Hermit says:

    Thank you for the reply. I have to apologize for my first comment, this was all my fault: I did make exactly the mistake you mention and thought the impeller and turbine were both working as impellers to increase the pressure rather than one driving the other. Worse: I should have known better but completely blindsided myself on the thought of having a new alternative to turbopumps –so I was just thinking pumps and nothing else and got everything wrong.

    And I didn’t even manage to spell my “name” right… Habitat Hermit is correct; you got it right the first time around despite my typo.

    P.S. I hope I haven’t made any reader cry and tear their hair in despair over my idiocy, once again I apologize. On the other hand I don’t mind if you all had a good laugh at my expense ^_^

  7. Paul Breed says:

    Why restrict yourself to peroxide?
    Exact same argument can be made for LOX only do an oxygen rich preburner in the space that presently holds the cat pack.
    Inject a volatile fuel like propane and put in some flame holders.
    One might have to inject this fuel as a hot fuel rich torch…
    With tight closed loop control over the mixture ratio one could have low temperature oxygen cool enough to even use Aluminum for all the rotating parts. Add an annular sonic choke below the turbine between the chamber and the one below this one.
    You would be injecting the “Real” fuel at this point. You would be injecting into a sonic gas stream. You just don’t get better mixing than that. The L* below this annular choke could be tiny. The space above the turbine would also be a source for high pressure gas that could be used for for things like attitude control thrusters etc…. Very interesting…..
    The big down side as I see it….
    If I built a turbo pump Steve H would never talk to me ever again ;-(


  8. Jonathan Goff Jonathan Goff says:

    Interesting idea. Might be possible. One challenge is that with LOX you really don’t want to have aluminum moving parts near aluminum. Great way to start an aluminum fire. If you look at XCOR’s valves for instance, you’ll notice that although the fuel valve is aluminum for the body and flanges, the LOX valve retains the original brass body, and only uses aluminum for the flanges. Almost all of the pump stuff I’ve seen for LOX uses moving parts and cases made out of some copper or nickel based alloy (ie Monel, Brass, Inconel, etc).


  9. john hare says:

    Paul and Jon,

    I had no intention of restricting it to peroxide. I was trying to present the basic idea at a conservative enough level to not run people off. The LOX version would combine the volute with the LOX injection manifold as one piece. Steve or XCOR could supply a fuel side pump into the upper fuel injection manifold. Below the turbine could either be HTPB for hybrid or a fuel injection manifold for bi-prop.

    On paper, salvaged and COTS parts for the turbo-machinery, plus $ome development work could have an economical staged combustion engine with performance comparable to the big boys. Afterburner pressures over 1,000 psi are not unreasonable and good combustion efficiency seems likely. T/W should be very good with the minimal plumbing and asymmetrical housings inside a single pressure sphere.

    There are some other tweaks I want to get into later after this one has been critiqued by any interested parties.

  10. The LOX preburner Paul mentions is what I’ve discussed with him over the last year or two, exploring his interest in collaboration, and previously in my site.
    A room temperature LOX preburner is very possible, if the desired chamber pressure is modest by Russian standards. There’s many metals that are compatible, and anodized aluminum just might work. In my calculations, a turbine tip speed of 200-250 m/sec is enough and pump speed 1/3 to 1/2 that is enough.

    I will do it someday.

  11. Jonathan Goff Jonathan Goff says:

    It might also be possible to electroless nickel plate the aluminum…but being careful is a good idea, if the oxide layer gets worn off, aluminum/LOX fires can be pretty intense from what I’ve heard.


  12. john hare says:


    One of the main points of the process is the thousands of used turbine assembies within walking distance of you, Paul, and Jon. Many of them are said to be for sale at a reasonable price. Mojave also has more Migs than Porshes, which implies plenty of A&P guys that might buy into an interesting project and bring their own turbines.

    The aircraft turbines already operate in hotter temperatures than you are suggesting with plenty of free oxygen. Some of the larger units have air cooled blades that allow them to run far hotter than any rocket turbine. Gaseous oxygen or fuel could be used to cool those blades instead of air so they could run with turbine inlet temps even higher than in normal aircraft service.

    Some of those turbines deliver more horsepower than any rocket ever flown. They are already developed and tested for reliability and long life. They are also at least as weight critical as rocket turbines. It is tough enough building a good rocket, why scratch build something else that is even harder than rockets.

  13. John:

    The surplus hardware you mention (turbocharger parts etc) are for high entropy, high temperature, low density gas, and are too large.

    A LOX stage combustion or expander cycle would use low temperature, dense O2 gas in the turbine. I think it would be easier to fabricate one of the desired characteristics than adapt one not designed for this use.

    As for risk trying aluminum. Stand back or use brass, monel, bronze, nickel…

  14. Paul Breed says:

    My apologies to Charles, the Low temp oxygen is most certainly his idea and my comments have their genesis in many talks with him.
    The new concept (to me) is having pump, turbine and chamber in the same pressure shell. How not to blow combustion products back into the LOX around the lox pump under all circumstances is an interesting problem.

    Also I don’t believe that aluminum will burn in O2 at less than 300 PSI or so. As this was discussed as a sub 250psi motor aluminum might work. It sure would be a lot easier to machine and a lot lighter than something like brass.

    One thought is that at small scales you have minimum gage tank issues and a 250 psi pump might not be worth much.
    A 12″ diameter 6061 tubular tank with 0.032 walls has a burst of 180 to 250 PSI all by itself.

    I also echo Charles comment that adapting something does not make sense.
    With modern CNC tools its almost easier to machine your own that try to adapt something that was optimized for a different design space.

  15. If you’d like to mix in some fuel for afterburning, and at the same time pressurize the fuel, and at the same time cool the turbine blades, how about hollow turbine blades carrying fuel? The fuel enters along the shaft at tank pressure, is raised to high pressure in the turbine blades while gaining heat, then ejects out nozzles in the backs of the turbine blades.

    * Rotating seal between low-pressure, low-temperature fuel and oxidizer. Could be worse, could be hot and high pressure.
    * If turbine is larger diameter than impeller, the fuel is delivered at higher pressure to a lower pressure portion of the cycle, leading to inefficiency. Fuel nozzles could be at a smaller radius than the actual turbine blade, which eliminates the advantage of blade cooling but slings the correctly-pressurized fuel radially through a axial oxidizer stream, which might be good for mixing.
    * Fuel in turbine blade must be prevented from getting close to changing state, because that would lead to large density changes which might lead to wacky pressure variation. If the pressure rise is fast enough, there will be no state change.

    Other tweaks:
    * If some fuel is injected at the back of the impeller, you might be able to burn LOX.
    * You could run the LOX down the middle of the shaft, and the fuel around it. LOX is always (?) denser than your fuel, so the tip speed required for given pressure is larger, so you want the fuel running around the outside.
    * If you have a small solid rocket motor eject into the space between the impeller and the turbine, it will start the turbine.

  16. Paul mentions minimum O2 pressure for aluminum burning. Actually, if not ignited it doesn’t burn at all. It is easy to anodize it with a strong oxide coating. Aluminum is worthy of consideration. Wasn’t the Agena engine made of aluminum?

    Minimum gage for low pressure tanks: the mass of the pressurizing means has no real minimum as the tank pressure is lowered. A heavier than minimum tank could be justified in terms of cost, durability. I’m thinking irrigation tubing…

    Also, a chamber pressure only a few times tank pressure will still be taking advantage of the several times chamber pressure for regenerative cooling, atomizing pressure etc. These things take substantial pressure drops and it’s better that be accomplished in some tiny whirling thing than heavier tank walls.

  17. john hare says:


    Some of the turboshaft engines operate at 30 atm and have turbine disk sizes that are in the range we are discussing. 450 psi turbine entry and 250 psi combustion chamber would be about the limit if they cannot be persuaded to work with denser cooler flows.

    If it would be easier to fabricate your own, there are some totally wicked possibilities for high Isp and T/W engines in small envelopes. Several of them involve cooling passages in the turbine blades for liquid regenerative cooling though. Designing and building it all from scratch could be a time and cash eater though.

  18. john hare says:


    The low pressure 250 psi or so might not be worth it to most people. It probably depends on how the helium bill affects their budget. There is also the issue that the pump adds to whatever tank pressure you are running. If you have 250 psi tank pressure and a 250 psi pump, then chamber pressure of 500 psi is attainable. Whether or not this is worthwhile is going to be a different decision to different groups.

    That I wish Armadillo good luck at the NGLLC does not mean I wish you ill. I want good luck for both of you, though only one can take the large purse.

  19. Tim says:

    Won’t spearating at the equator for maintenance put the seam/seals in the same place as the turbine rotor? I figure trying to get a decent seal in the same place you’re trying to get a small gap between the turbine blades and the chamber wall (the technical way to say that eludes me) will make things a little crowded. Maybe have a short cylindrical section (maybe an inch), so you’ve got room to put the seam and the turbine rotor in different places??

  20. Jonathan Goff Jonathan Goff says:

    I have no problem with using aluminum in rocket engines. I just get leary with using it in areas where you have parts moving very rapidly with tight clearances. I could be wrong, but all the literature I’ve seen (and all the cryo pump guys I’ve spoken with) have said that aluminum–even anodized aluminum–is a big no-no with things like LOX pumps. They could be wrong, but I would want to verify that before risking too much expensive hardware.


  21. Pete Zaitcev says:

    Tangentially, check this link (I hope translate.google.com can handle it):
    Initially it talks about the NK-33-1 with the extendible nozzle (and adding roll control), the costs of NK-33 derivatives vs. RD-191. Generally, the article is obsolete. However, the journalist also asked about the cross-over between aircraft engines and rocket engines at Klimov (re. John Hare’s point above). Notice the discussion of enamel covering (plaque) vs. nickel-based alloys. That damn oxigen!

  22. John Hare’s 2:28 am–Yes, easier to fabricate a turbine to own design than try to adapt something meant for something else. But cooling passages in turbine blades? Not needed for low (<500 K) gas. High T/W? I think, for 1 ton thrust or so, T/W could be over 200.

    John Goff, 8:15 am–Of course aluminum would be out for a delivery truck pump. But for a very small engine, maybe not to be ruled out.
    For quick prototypes, though, brass might be good. Machines and solders well.

  23. john hare says:

    Fuel through the turbine blades is exactly what I have in mind though it will be an extensive post. I believe I can address the issues you raise. Part of it is sacrificing efficiency for simplicity and ease of handling.
    I am thinking in terms of using the section out of the jet that seals the turbine tips and using it to mask the connection. An inch section or so would be a small price to pay if the concept works.
    I couldn’t figure out how to translate the page. It sounds like the sort of thing that interests me. Old and obsolete are often two very different things. My primary reference on air-turborockets was written in 1956 and is still valid for most of the basic points.
    I see your point. For your purposes, my salvaged parts method doesn’t work at all. For some other people, it might be worth looking into. That is the point I kept missing, that one concept doesn’t fit all, or even most, even if it works.

  24. Tony B says:

    This may be a stupid question as I don’t know much about pumps, I’ve been pretty much a structures guy my entire adult life.
    How does the pump start? I understand how once the peroxide hits the catalyst the pump keeps moving, I just don’t see how it starts. Is there a solenoid valve or something with the peroxide under some pressure to push it into the pump housing?
    Just curious,

  25. john hare says:

    The tank pressure must be high enough to move the liquid to the pump and not cavitate in the impeller due to low pressure vaporization. The 15 psi or so forces some peroxide into the impeller and through the cat pack. The perhaps 10 psi pressure differential across the turbine at start begins the spin up. Gradually the spin increases the pressure through the impeller and cat pack, which increases the power through the turbine, which increases the rpm, which increases the pressure developed by the impeller, and so on. It could take 1 to 5 seconds to reach full pressure depending on assumptions.

Leave a Reply

Your email address will not be published.