Saddlespike Nozzle



Altitude compensating nozzles have almost no flight history in spite of offering the possibility of higher performance results with lower performance equipment. In theory, a compensating nozzle allows a rocket engine to perform at optimum at any altitude and airspeed. Payload should experience improvements on almost any launch vehicle ever built. So why are they missing from the list of things tested.




I hope this other view makes a bit more sense. You are looking through the side open to the atmosphere with the side opposite bounded by metal through the low expansion ramp area. The aerospike section is the bottom line between endplates. The heavy black lines represent the endplates for both engine types, both keeping the expansion plume focused on the metal expansion ramp you can see. In the linear aerospike, plumes 2-9 are bounded from the sides by adjacent plumes. 1 and 10 are bounded by plumes on one side and metal endplates on the other. The saddlespike uses metal on both sides all the way down.




Money is the first answer almost every time. People with limited funds cannot affort to take serious chances on an unproven technology. Just getting something working is far more important than improving something they don’t have yet. This is probably the first answer for the serious newspace companies. Armadillo, Masten, XCOR and company would not have had any benefit from a compensating nozzle to date. As of October 2008, for any of them it would have been a waste of resources needed for current flight vehicles. The people with the vision don’t have the money, and the people with the money don’t apparently have the vision.

Uncertainty is very high on the list of reasons to leave this tech alone. An aerospike or other compensating nozzle may boost the payload for SpaceXs’ Falcon 1, it would definately have cost schedule time and reliability uncertainty that a young company just cannot afford. The RS 2200 that was intended for the X 33 never did a full test stand burn as far as I am aware. That Dirkson plus program didn’t answer even the basic aerospike questions, much less supply relevant flight history.

Technical issues are third on the list, or maybe tenth for not putting a compensating nozzle in the specs. There is a lot of throat area to be cooled on most versions of aerospike nozzle or expander deflector. There are either many combustion chambers or an annular chamber resembling a bicycle tire. Either one has to be fun to work with compared to a straight forward single large chamber with an understood nozzle type. Talk to anyone that builds hardware about trying to work with 20 or so small chambers that all have to play well with others all the time.

Several conditions must be met for any serious player to consider compensating nozzles.

1. It must be simple to develop.

2. It must supply a financial benefit to pay for the development plus turn a profit.

3. It must be at least as reliable and certain as the rest of the company hardware.

4. It must not increase technical problems for the company.

Current proposed nozzles fail all four.

I propose a variation on the linear aerospike with 2 chambers total instead of 20 or so.


Shrink the 10 chambers per side to 1 and extend the nozzle skirting on both ends to almost enclose both sides. With the endplates turned back into nozzle walls, the gasses expand against metal on 270 degrees of the arc with the atmosphere interacting with the plume on the remaining 90 degrees faceing outward. The recirculation between the two sides operate the same way as the orriginal linear aerospike. Turn it upside down and it resembles a saddle.

1. Any company that is currently burning two chambers simultaneously, could bolt one of these to a test stand within a few weeks. If it works well, moving forward should not be too difficult.

2. While new space is not flying many high altitude vehicles at this time, several are flying throttled VTVL machines. Compensating nozzles are not altitude compensating, they are pc/pa compensating. There is a financial motivation for having a vehicle that operates efficiently at 300 psi and 75 psi near the ground. You don’t want to try precision landings with intermittent flow separation. Nor do you want serious under expansion for the majority of the flight.

3. If it works at all, it adds nothing to the vehicle that can go wrong. If your engines work right now, they will continue to do so. The exception is if you decide to use vane surfaces in the exhaust for TVC. This engine will lend itself to a rudder/elevator type TVC better than most others.

4. This, if it works, adds no technical issues except base flow differences. It may be an improvement or detraction, but it will be an unknown to be answered.

The people I know in this business don’t have time to go haring after wild ideas. There might be at least one group out there that could try something like this and let the industry know if it works. A few percent performance increase would be useful in the coming years for people with profit motive, if it can be made to work.

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.

17 Responses to Saddlespike Nozzle

  1. gravityloss says:

    Interesting idea.
    Could you have a better picture? An MSPaint crude perspective sketch?

  2. Tim says:

    I was going to ask if you could incorporate one into an RD-180, but I thought of a problem. You could use rudder/elevator TVC and gimballing the entire thrust chamber/nozzle assembly to control pitch and yaw, but how would you control roll? split rudder/elevator TVC?

  3. Marcus says:

    I like the idea (as far as I can follow it) and I’ve been a big fan of the aerospike concept since seeing the first renderings of the X-33.

    I have to second gravityloss: a better picture would be greatly appreciated to improve my understanding of what exactly you are proposing (I’m having difficulties to visualize that 270/90 degree concept).

  4. John,
    While it isn’t clear whether this specific approach would work, you’re right that altitude compensation benefits VTVL designs more than other vehicle types.


  5. Adam Greenwood says:

    –Compensating nozzles are not altitude compensating, they are pc/pa compensating.–

    Acronym help, please?


    This is an instance of why I think the suborbital market could make such a huge difference for orbital access. There are lots more untested, potentially good ideas out there then there are people trying them. Once the suborbital market gets going, people will try them. Some of them will pan out.

  6. Jonathan Goff Jonathan Goff says:

    Pc = chamber pressure
    Pa = ambient pressure (ie the pressure outside the nozzle)

    What John was saying is that “altitude compensating” is somewhat of a misnomer. What such systems do is adapt the jet flow so that the exit pressure in the nozzle is close to equal to the local ambient pressure, regardless of what the chamber pressure is doing. This reduces overexpansion/underexpansion losses, and compensates for shifting Pc/Pa ratios.

    Traditionally, for most orbital vehicles, you’re not changing the chamber pressure that much inside the dense part of the atmosphere, so most of the change in the pc/pa ratio comes from the pa dropping. In such systems, the compensating allows you to take better advantage of the shifting pc/pa ratio.

    However, for VTVL vehicles you have to be able to throttle your engines for any practical approach. Throttling almost always means dropping your pc by dropping the flow of propellants into the chamber. As you drop the pc, your pc/pa ratio changes in a bad direction. Basically, for a deep throttling normal rocket, as you throttle down your fuel efficiency gets rapidly worse past a certain point, because your pc/pa ratio causes the engine to be badly overexpanded. If you design an engine so that it is just barely not having unsteady flow separation at landing throttle, you’re going to have a very short expansion ratio on your nozzle, and if your engine pressure is low, you’re going to get really crappy mission-averaged Isp.

    If on the other hand, you have an “altitude compensating” nozzle, as you throttle down, the jet flow will automatically adjust so that it is close to ideal expansion ratio, so you a) lose a whole lot less Isp as you throttle down, and b) can have an engine that doesn’t have to sacrifice as much performance in space in order to be able to throttle down.

    Basically, while altitude compensating is “nice to have” for other rocket forms, for VTVL it’s a “really, really nice to have” bordering on necessity.

    That make any sense?


  7. Adam Greenwood says:

    –That make any sense?–

    The East German judge gives your explanation a 9.5.

  8. john hare says:

    Gravityloss and Marcus,
    I inserted another view that might help. Never having edited a blog post before, it didn’t exactly go in the right place. My bitmap sketches will be crude for the most part, my engineering classes used pencils and drafting boards twenty mumble years back. But that was after sketch to scale on grid paper.

    Right in one. By using vanes to control the recirculation area, it might be possible to control the vehicle without the moving surfaces actually contacting the fast moving expansion stream. If that can work, then split rudders become feasible.

    I’m not going to the mat for this specific approach, I’ve got notebooks of others. Sturgeons law applies to my ideas. I will consider it a good day if someone that builds hardware tells me how stupid my orriginal was and how they fixed it.

  9. Marcus says:

    Thanks, I can now see what you are talking about. Since your saddle-spike only has one engine per side, the expansion of the engine exhaust is limited by the saddle-spike on one side, by the end-walls on two sides and the remaining side would be the “virtual” bell of the air surrounding it. Sounds interesting (at least to a layman ;))

    But having seen this beautiful picture of a XRS-2200 test firing again, I have to say that I seems they didn’t bother with sidewalls at all for the linear aerospike:

  10. john hare says:


    I remember endwalls though much smaller than I drew. If you are right, then thrust/weight and cooling would be much easier. Redneck style, I think in terms of building large endwalls on a test article, and taking it to the test stand with a portable grinder. Cut the walls back between each test fire until you find the sweet spot.

    You also corrected me on firing of the RS2200. I didn’t think it had ever been fired all up ever.

  11. gravityloss says:

    So the pictures actually show two designs, and not many views of one design?
    So the new thing that eliminates the multiple thrust chambers of an aerospike is a very short linearish aerospike in a normalish (if wider) bell nozzle? And that would keep the flow still attached at low chamber pressures since it would start to thin in the middle rather than the sides?
    Reminds me a bit of an expansion-deflection nozzle.
    Do you mean something like this (no 3d curvature shown)?
    do links work?

  12. Guy says:

    I am not a rocket scientist, but can you comment on these two
    ideas? Thanks

    The resultant hot gases are expanded through the nozzle 425 . Because the nozzle itself is manufactured of the solid fuel material (such as acrylic), the size of the nozzle expands as the motor burns. That is, at least a portion of the nozzle itself is burned and consumed because the nozzle is made of the solid fuel material. The initial flare of the nozzle 425 can be varied based on the altitude at which the motor is initially fired.

    And Also:
    The annular hybrid motor is an improvement to hybrid rocket motor technology that can be used for a wide range of applications. The technology will lead to improvements in hybrid payload mass fraction, increased performance during throttling, and reduced costs associated with casting hybrid fuel grains. The technology is highly scalable and can be used for applications ranging from small thrusters on satellites, to launch vehicle booster applications.

    Thanks, looking forward to any comments.

  13. john hare says:


    The two designs shown are suposed to be a traditional linear aerospike and the new saddlespike. The saddlespike is the only new type engine shown.

    If the right, left, and bottom sides of the box shown are open to the atmosphere, then yes. It is like two bell nozzles with doors cut in one wall to the outside, and connecting channel between the two for the aerospike area. The drawing was not supposed to imply wider bell than normal. It should thin at the sides rather than the middle and increase in thickness as altitude increases.

    To me, it does not resemble an expander deflector at all. I do have a track record of not understanding the meaning of a question however. Sometimes I might agree if I fully understand your meaning.

  14. James says:

    Here’s a reference to an aerospike prototype that flew in 2003:

    It wasn’t a linear aerospike though.

  15. James says:

    Jon and John,
    I’ve actually been fantasizing about an aerospike for quite a while. It’s been fantasizing because I am not an aerospace engineer and don’t have the tools, training nor time to do any serious design work (nor of course the money to really build anything :-).

    What attracts me most is the rather shallow depth and small size of the combustion chambers. This means that one could think about having a VTVL vehicle with a large, round bottom, wrapped around a spherical tank, in which the combustion chambers are embedded that could double as the TPS for orbital return. A material that I’ve thought about is silcon carbide as a nanostructured ceramic forming polymer, such as Starfire Materials sells:
    These materials are good up to 3300F and can be formed into a variety of shapes.

    One could combine this with transpiration cooling for the thrust chamber. Transpiration cooling seems to result in some ISP improvement for pump fed engines. There’s a great paper from DLRF about transpiration cooled thrust chambers (C/C rather than SiC though):
    A properly engineered transpiration cooling system could provide some thermal mitigation on orbital return as well.

    What’s the point of all this? Basically weight reduction through using lighter materials and through having one assembly perform two functions: thrust on up and TPS on the way back.

    But who knows, maybe it won’t even work, like I said, I’m not an aerospace engineer. And, even if it did, like John pointed out in his blog post, none of the NewSpace startups can try something like this, it’s too risky. In a startup, you get to push the envelope in one direction (for example, with Armadillo, it’s throttable, steerable engines) but that’s it. Any more and you’re risking commercial viability.

  16. john hare says:


    I’m not an aerospace engineer either. My education in this field is self inflicted and it shows. Lack of funds is my reason for throwing these ideas out there in the hope that some of it proves useful to somebody in the real..

    The aerospike/heatshield you suggest seems to resemble Rotary Rockets’ concept from the late 1990s. Except that their engines rotated. The main problem as you noted is the expense and risk. Someday it will be different. I got one paragragh from your transpiration link. It seems that they are working the problem.

    Using available thrust chambers and salvaged turbine assemblies And tested landing systems seems to be the way forward to me. The saddlespike could be tested in a weekend by glassing a couple of short bell nozzles together. A pump system could be designed around a salvaged aircraft turbine disk, turbine nozzle, shaft, and bearings. Armadillo could design a verticle landing system for an X38 type vehicle to eliminate the parachutes and uncontrolled landings, really important for injured crew. And so on.

  17. john hare says:


    Your links seem to lead to a short description of what they are going to do without any idea of how they got there. I didn’t get anything to form an opinion from them.

Leave a Reply

Your email address will not be published.