Geoff FordenClustering Engines, Adding Up the Weights

In an unfortunately title article in Jane’s Intelligence Review (“Safir When Ready: Iran sets its sights on long-range capability”; sorry, its subscription only) I discussed why I interpreted pictures of the test-stand version of a Safir second stage as indicating Iran was developing a new advance in its missile program: a clustering of two engines that are gimbaled for thrust vector control.

In yesterday’s post, I discussed why eliminating the jet vanes could be used to lower the dead weight associated with storing the smaller amount of fuel required to produce the same change in velocity.

Today, I want to estimate how much dead weight is either saved or used up in the engines alone when you cluster them. (The last post in this series will try to put it all together and discuss how much Iran will have to improve its technology in order to get an ICBM capable of launching a nuclear weapon at the United States.)

This post might be excessively techno-wonkish but as an added bonus, for those of you who are willing to wade through the details, it should yield some interesting insights into the level of Al Samoud/SCUD/Nodong engine technology.

As can be imagined, a missile’s engines constitute a major portion of the “dead” weight contribution outside of the payload. Engine weights have, therefore, been a major area of research with much work going into reducing the weight per ton of thrust. A wonderful book, if you can get it, is Aerospace Vehicle Design: Spacecraft Design (vol. 2) by K. D. Wood. It basically covers every important aspect of designing missiles and spacecraft and, most importantly for this analysis, has a ton of empirically derived graphs. I’ve reproduced a couple of these below. Engine weight, including the turbopump and piping, as a function thrust is shown on the left while the turbopump weight alone is shown on the right.

I’ve added three points to the plot on the left: the Al Samoud II, the SCUD-B and the Nodong engines. The first two are based on actual measurements while I estimated the third, the Nodong, from an admittedly arbitrary curve that I drew using the Al Samoud and the SCUD points and the rough shape of the two curves surrounding it: the World War II range of rocket engines and the range identified as “Attainable Range, 1960-70.” Remember, this is a blog and not a refereed journal! (The book was published in 1964.) This “SCUD-family” of engines falls nicely between the two. Does that mean that the SCUD technology is worse than Western technology dating from the 1960s? Not at all! SCUDs were designed for rough handling and do not use the high energy fuels that the racehorses of the era, the space launch vehicles, did; namely cryogenic fuels. Both of these factors have the effect of increasing the engine weight per ton of thrust.

There have been a number estimates for the thrust of the Nodong/Shahab missile which can be used for estimating the engine/turbopump masses using the graphs above. I’ve used the thrust from a set of parameters for the Nodong originally published by Robert Schmucker in estimating the Nodong/Shahab/Safir 1st stage engine weight, including its turbopump and associated piping, as weighing 330 kg.

You can use the graphs above to do the same for your favorite missiles. Let me summarize a few of the weight combinations that I find most interesting:

*These are measured quantities. All others are estimates or derived from measured quantities.

So what does this tell us?

The Safir does not manage to save much weight (just 2 kg) by using a single SCUD turbopump as opposed to one turbopump for each Safir engine, assuming the turbopump is sized for the individual Safir engine. Iran might, however, save an addition 14 kg if they developed a new, single turbopump for the cluster of two engines. But would that be worth it, considering the time and resources they would have to devote to developing a new and unique turbopump? Obviously, Iran has decided it would not. I cannot help but agree with them. Perhaps when their space/missile industry has developed enough to afford the human resources to devote multiple teams to multiple turbopump design project they can. This points out one of the most important of scarce resources to any country just starting to develop new rockets: trained and skilled manpower!

We know now that Iraq was planning on using the airframe developed for the Al Samoud and stuffing a second Volga engine in it for a two engine cluster. (This possibility, which we considered very likely, was a crucial factor in why UNMOVIC proscribed the Al Samoud II.) Here too, the proliferator would have gained about 14 kg by developing a new, single turbopump for the cluster. Here too, the proliferator chose not to develop a new turbopump, again presumably because the resource costs for such development would have been too high. (Also, Iraq never did manage to develop its own turbopump!) On the other hand, a SCUD turbopump would be considerably overpowered for a two Volga engine cluster and would actually weigh more (142 kg vs. 120 kg) than having each engine retain its own turbopump, the course Iraq actually planned on using. Would Iraq have gained other benefits from using the Volga turbopump? Would it have been easier for them to balance the thrusts of the two engines by controlling the two turbopumps? Or would it be more difficult? Both, it seems to me, would have required considerable R&D but I would guess that a single turbopump would be easier. Of course, I could be wrong; thoughts, Wonk readers?

Finally, a cluster of four Nodong engines—such as has been reportedly used in the first stage of the Tae’podong II—saves well over 100 kg in weight if it uses a single turbopump developed and optimized for that configuration. Is that enough to justify a proliferator developing a new turbopump? Perhaps we will only know for certain when we start seeing images of that stage appear in public. We will put those weight savings in perspective when we compare the two strategies for developing ICBMs: stacking existing missiles one on top of the other vs. developing new airframes for existing engines. (I tend to think of these two paths as the North Korean path vs. the Iranian path, but perhaps I’m jumping to conclusions too much!)

I am postponing the fourth post in this series, which will consider the US estimates of Iran’s missile development program, until Monday to consider the discussion we have had on the Sejil.


  1. Jochen Schischka (History)

    Good work, Geoff!

    Just one thing to consider: the Turbopump of the Volga/Volchov/SA-2/Guideline has a mass-flow of 15.88 kg/sec, while the Scud has one of 57.83 kg/sec; The chamber pressure of both engines ranges at about 7 MPa; So it would theoretically (after some modifications) be possible to supply four Guideline-chambers with one Scud-turbopump.

    And until i see hard, photographic evidence of a Taep’oDong-II (shouldn’t this be Taep’oDong-B, since that would be a major modification or rather a totally new design?) with four (DF-3/DF-4-like?) clustered NoDong-engines, i refuse to believe in such rumors. I can’t exclude that possibility, but it sounds somewhat improbable to me (i’d rather expect something Safir-like…).