Jeffrey LewisWhy ICBMs are Hard to Build

UCS’s David Wright has written a nice post at All Things Nuclear that replots a graph originally produced by IISS to explain that ICBMs are hard to build because the principal trick to increasing range is adding fuel:

The reason for this trend is that the primary way to increase the range of a missile is to make it larger and add propellant. (Changing to a more energetic propellant or reducing the weight of the missile structure will increase the range somewhat, but a large increase in range will require additional propellant.) The additional propellant and missile structure increase the missile’s mass. As a result, the engines need to accelerate all this additional mass as well as the original mass of the missile and payload, so the increase in missile size gives you much less increase in range than you might expect. What this plot shows is that this penalty becomes very significant as you go to longer ranges.

It’s the kind of thing you know in the back of your head, but it’s pretty stunning when you see in graph form how little other variables, like using a more energetic propellant or reducing the mass of the airframe, matter in actual practice.


  1. John Schilling (History)

    This appears to be the graph for a single-stage ICBM, and going back to the source confirms that yes, that is the case. It also appears to use gross weight as a metric for difficulty, without justification. Most of the gross weight of a large rocket is fuel and writing a purchase order for a hundred tons of rocket fuel is not much harder than for ten.

    Building a single-stage ICBM would indeed be Very Very Hard. Not because the rocket would have the impossible mass of One Hundred Tons! , but because truly herculean levels of engineering would be required to achieve the necessary performance. Which is why nobody ever actually does that.

    Building a two- or three-stage ICBM is about an order of magnitude easier. You have to get staging right, and nobody gets staging right on their first try, but the total mass is reduced by a factor of three or so (not the 10-20% Wright suggests) and the cost per mass is similarly reduced because you can now use e.g. sheet metal and stock engines rather than bleeding-edge lightweight high-performance technology.

    It is not a trivial undertaking. We can get a feel for it by noting that Elon Musk spent six years and on the order of a hundred million dollars privately developing a space launch vehicle with roughly ICBM-like performance, which crashed and burned on its first three flights (in part due to staging failures). That was starting with nothing but an open checkbook. Orbital Sciences Corporation required three years and $75 million to fly the first Pegasus, which would have made a nice air-launched ICBM, starting with a line of sounding rockets and a rented B-52.

    This is not a trivial undertaking. But it is one that requires realistic assessment, and graphing the performance of hypothetical single-stage-to-Moscow vehicles is simply irrelevant. ICBMs are what staging was invented for.

  2. George William Herbert (History)

    Taking off from what John said above (which I generally support), more space-launch like performance regimes for rocket stages would allow practical single-stage to MRBM and IRBM though.

    The rockets we generally talk about in proliferation circles – Scud and its many derivatives, its larger and stacked stage cousins and uncles, solid missiles with similar performance – have abyssmal performance by space launch standards. Some of that is because a lot of these are designed for long distance road transport, whereas rockets going to space are generally stacked gently on the pad (or sometimes, in a building and rolled to the pad on a large mobile firing platform, or rolled to a pad on a giant railway TEL which supports it rather firmly as in the case of Soyuz / Proton launchers).

    But – really – these rockets are terrible rockets. Heavy. Lousy propellants (some of which is a tradeoff for portability and storability). Hazardous propellants.

  3. Gregory Matteson (History)

    For a good read on how difficult the first ICBM was, Sputnik and The Soviet Space Challenge, by Asif A. Siddiqi is an excellent read. Further things to look at would be the failed British ICBM project, and the stalled Brazilian launcher efforts.

    Every successful ICBM program has shown that there is a lot more to a working ICBM than just a big rocket. In particular, staging and guided re-entry were major challenges, as is the length, expense, and very public nature of the testing program. People seem to forget that, contrary to the efforts of US and Soviet propagandists, the early ICBMs were stupendously inaccurate and it took literally hundreds of launches over many years to overcome that.

    A survey of the more or less parallel histories of the Bomb and the missile have led me to the conclusion that ICBMs and Space Launch vehicles are substantially more difficult and expensive than the H-bomb, and maybe more difficult than a basic Bomb program.

    • John Schilling (History)

      Be careful about comparing the development of ICBMs in the 1950s, and the development of ICBMs in the 21st century. An awful lot has changed – in particular, there are now many detailed and accurate textbooks on how to build large rockets and the various components thereof, mostly aimed at space launch vehicle designers but highly applicable to entry-level ICBMs. An awful lot of dual-use hardware on the market as well, most with only a fig leaf of MTCR control.

      By comparison, there are rather few good textbooks on nuclear weapons design, more public disinformation on the subject, and the dual-use nuclear technologies are more thoroughly regulated. So while the initial development of the ICBM may arguably have been more difficult than the initial development of the hydrogen bomb, this is unlikely to be the case for any Nth country trying to domestically recreate the technologies today.

      The one aspect of modern ICBM design that is still relatively obscure and without civilian applications, is low-drag reentry. That’s necessary if you want to defeat terminal-phase missile defenses and probably necessary if you want the accuracy needed to engage hard targets. But if a proliferator is primarily interested in deterrence, those aren’t major concerns. The prospect of an old Avco Mark 2 reentry vehicle with an antiquated W-7 fission warhead, falling on downtown Chicago plus or minus five miles, would still be plenty deterring to the United States. I suspect anyone who managed to actually build a working ICBM would do somewhat better in the warhead+guidance department, but even that minimum would probably suffice for the deterrent mission.

  4. Gregory Matteson (History)

    “By comparison” India, which has apparently had to go it alone to a greater extent than any other country, but has first rate domestic resources and technology, has had a great deal of difficulty and expense developing launchers. By my count theirs is the eighth separate attempt to develop an ICBM class vehicle, and no one in our press is proclaiming them as having a working ICBM in the absence of a visible test program. I think there’s no doubt an ICBM is within their immediate reach, but it has been much longer and more expensive to get there than it was for them to build the Bomb, also with notably lacking outside support.

    Of course we want to tamp down fears of Indian belligerence.

    This is in notable contrast to hysterical claims that a couple of failed DPRK launches, of questionable designs, constitute imminent possession of working ICBMs. Or to claims that a couple of Iranian small satellites constitute evidence of an imminent Iranian ICBM threat.

    I suppose that the main problem with what I’m saying is that everybody I read thinks Iran and the DPRK are getting outside help.

    • George William Herbert (History)

      Why would India make an ICBM? Their threat partners are all within IRBM range…

      They have a solid-fueled space launch vehicle, and several liquid-fueled ones. They make plenty big enough and high tech enough solid rocket motors to do an ICBM if they wanted. But rumors to the effect that they are developing on have never been substantiated in any way. Because, really, they have no fears about the US or Europe. Not getting along with the US at a geopolitical level at times is in no way equivalent to them seeing us as a threat or having any delusion that we see them as a threat. It’s a remarkably healthy and diplomatic geopolitical disagreement with India.

      Their farthest target of serious concern is Beijing.

    • John Schilling (History)

      First, what is your definition of “ICBM class vehicle”? Using 500 kg payload to 5500 km range as a cutoff, and including small satellite launch vehicles, I count twenty-four separate and original attempts to develop ICBM-class vehicles, in twelve countries. “Original” meaning no overt foreign assistance and no heritage of ICBM-class vehicle development in the organization doing the work. Of the twenty-four, seventeen were successful, with an average of seven years from full funding to first successful flight.

      Second, the only reason it has taken India so long to get to ICBM-class vehicles is that, as George Herbert notes, they haven’t particularly wanted or needed them. In 1972, they decided they wanted a small satellite launcher, and in 1980 they successfully flew the SLV. In the early 1980s they decided they wanted to at least know how to build serious ballistic missiles, and the SLV-derived Agni TD flew successfully in 1989. Then they played nice for a few years, but in response to Pakistani missile-related activities decided in 1996 to actually deploy something. Agni II, 1999. 2001, they decided they might want to reach China as well, and the Agni III flew in 2007. Possibly next year there will be a test of an Agni V ICBM, as a result of a decision made in ~2006.

      Whatever sort of “ICBM class” rocket India wants, India gets about six years later. With one or two failures in a three- or four-flight test program. This is par for the course, and about what to expect when any moderately industrialized nation decides to build such rockets without extensive foreign technical assistance.

      Yes, you do need the testing. If someone hasn’t yet successfully tested an ICBM-class vehicle, they don’t have an operational ICBM capability. But, A: one or two of the early tests almost always fail, and B: successful tests usually follow a year or two later, and IOC almost immediately after that if it’s a rush program.

      “Ha ha! Their rocket blew up; we don’t have to worry about them having ICBMs for many years or until someone else sells them the things”, is wholly unjustified complacency. If all they are doing is testing single-stage short-range ballistic missiles, then yes, you’ve probably got the better part of a decade to contain the problem. One or two failed tests of an ICBM-class vehicle, that is usually the last warning sign you get before deployment of operational ICBMs, and it’s usually not that much warning.

      If in fact it took a little outside help for them to get there, or a lot of outside help, that hardly matters unless you can fire up your TARDIS to kidnap the technical advisers ten years ago. They are where they are, however they got there, and they probably don’t need much help to make the last step.

    • Mark Gubrud (History)

      John, you call India “moderately industrialized” when I think what you mean is it’s pretty big and has enough resources and well-educated people to pull these things off when it decides to. You might draw an inference about Brazil, but Iran would be pretty iffy and North Korea downright dubious.

      Also, no matter what the reasons why building ICBMs is hard–and yes, size does matter; you haven’t talked about the problem of scaling up rocket engines, for instance–I think we can all agree that building ICBMs really is substantially harder than building smaller and shorter-range missiles, particularly if you are assembling the latter from surplus imported parts.

  5. Anne Bazuin (History)

    R max = 2 R arcussinus (Vo/Ve)^2/1-(Vo/Ve)^2

    R max is the maximum range of the missile
    R is the earth radius, 6368 kilometers
    Vo is the starting velocity of the missile above the earth atmosphere
    Ve is the escape velocity from earth, about 12 kilometers per second

    • George William Herbert (History)

      That’s the vacuum purely impulsive ballistic trajectory, yes.

      Real world – rockets burn over finite time ( = gravity losses to final velocity ), air drag matters on both ends to some degree, etc.

      High G launches minimize gravity losses, at the expense of a really harsh ride up through the atmosphere with acoustic loads and increased air drag losses. Nuclear warheads and ICBM guidance can be tough enough to handle that without flinching, so modern ICBMs tend to take the faster launch approach. But the losses from both cases (and suboptimal trajectory shaping, and so forth) are non-trivial. 1,000 m/s or more…

    • Murray Anderson (History)

      Nice equation. It needs a couple more brackets, as
      R max = 2 R arcussinus (Vo/Ve)^2/(1-(Vo/Ve)^2), and
      the escape velocity is about 11 km/sec.
      “arcsin” is more common in N. America – possibly “asin” is even more common now, since Excel uses that form.

  6. John Schilling (History)

    I call India “moderately industrialized”, because India has a $3.5 trillion dollar economy of which $770 billion is derived from industrial activity – such as, e.g., the production of ~2.6 million automobiles per year. Well-educated people are a seperate metric, but since you bring it up, India graduates about 180,000 new college-educated scientists, engineers, and mathematicians each year.

    Iran, by comparison, has only a $0.9 trillion economy but that still includes $400 billion from industrial activity, including ~1.4 million automobiles. And cranks out 88,000 STEM graduates a year, or did last time statistics were available.

    North Korea I wouldn’t believe any statistics they would offer, but as Iran and North Korea seem to collaborate in rocketry it is the total that matters – and all the usual jokes aside, North Korea’s contribution is not negative. We are comparing three moderately industrialized nations in approximately the same economic league, and what India can do Iran and North Korea together can probably do. Which is to say, about what the United States and Russia could do circa 1960, except that the production totals will be smaller.

    Yes, building ICBMs is harder than assembling SRBMs from imported parts. As the people we are talking about seem to be doing rather more than that, it’s not a relevant comparison. And “Ha ha! They’re just a bunch of illiterate third-world primitives; they can’t build missiles unless someone sells them an E-Z Bake Missile Kit!”, is just another brand of complacency. Even if they are fifty years behind us in terms of industrial and technological development, well, do the math.

  7. Mark Gubrud (History)

    I’m guessing your $400 billion “industrial activity” in Iran includes the oil sector, which would be about half of it. And from what I read, Iran definitely is making its own rockets but North Korea may be using a lot of surplus parts. I’m certainly not sneering at “illiterate third-world primitives” but just pointing out that how much a nation can afford to spend on missile ambitions (and how likely they are to be successful) depends on the level of resources they have to begin with, and India clearly outclasses Iran, while NK is a basket case.

    • George William Herbert (History)

      A cautionary tale for how hard it is…

      The DC-X vertical take off vertical landing rocket was done in the early 1990s, for about $50 million project budget for the first iteration. It was fairly big – gross liftoff around 19 tons – but the size wasn’t the primary cost driver, developing the VTVL operations and control technology and the rapid reflight operations were the hard parts.

      If someone asked in 2000 what a new similarly capable (in terms of delta-V, control, etc) vehicle would cost, the answer generally came back “$50 million, minus a bit”, the “a bit” being due to better off the shelf guidance electronics (better GPS and IMU systems) and there not being as much fundamental doubt about whether it can be done in the first place.

      The actual *combined* cost of 3 Lunar Lander X-Prize teams developments over the years was less than 20% of the DC-X costs, and one of those teams flew surprisingly well having been done by essentially 2 people in their spare time over 2 years.

      It’s not at all clear that rocketry is necessarily as hard and expensive as it was thought to be, based on “the old way of doing it”.

  8. Gregory Matteson (History)

    It remains noteworthy that the UK, a “First World” industrial power, in the 1960’s had the only strait up, throw-in-the-towel failure to complete an ICBM they believed they really needed. Of course they had the fall back that if they failed, they still had us.

    Regarding the DCX, could you really scale & warp DCX technology into an ICBM? That’s not clear to me.

  9. George William Herbert (History)

    DC-X isn’t an ICBM – or ICBM precursor, or analog.

    It’s an aerospace project, involving complex and precise control algorithms, high performance, and so forth.

    One of Armadillo Aerospace’s rockets recently did a flight to over 1,000 meters above launch point and back down. The center of the rocket stayed within a handful of inches of a virtual vertical pole going straight up from launch point for the entire trip up and back down, to landing. It looked like something going up a rail and back down again, minus the rail.

    Again – these aren’t even vaguely ICBMs. But the teams doing them are TINY – Armadillo is 7 people, Masten was about or slightly under that, and Unreasonable Rocket was 2 people until the day before they tried their Lunar Lander Cup flights (they had a lot of volunteers show up on flight day and immediately before). They’re introducing new materials into rocketry (Masten were the first to fly 5059 aluminum tanks on rockets), developing new rocket engine combustion chamber fabrication and cooling design approaches, much lower fabrication costs, integration of GPS with IMU systems, etc.

    They are building things a lot more primitive than “real aerospace” would do – but then, so is a Scud. And umpteen thousands of them later…

  10. Anne Bazuin (History)

    Talking about the “vacuum impulsive ballistic trajectories”of my ICBM’s
    L max =2R arcsin Vo^2/(Ve^2-Vo^2)
    H max = R Vo^2/(Ve^2-Vo^2)
    Question to mr. G.W.Herbert: What is the flying time of L max and H max ?
    You need to know the starting moment and the flying time of my ICBM at maximum range (L max) to reach your shelter in time.

    Answer to mr. M.Anderson: You are right, the earth escape velocity is 11 kilometers per second.
    Ve={(2×6,668×10^-11×5,977×10^24/6.368.000)}^1/2 = 11.188 m/s
    With thanks to Isaac Newton 1642-1727

  11. Allen Thomson (History)

    As long as we’re talking about ICBMs in various places, what do y’all think about assertions that Shavit has demonstrated “ICBM-like” performance by putting 300 kg satellites (Ofeq 5-9) into highly retrograde orbits (141 deg inclination at ca. 500 km altitude?

    • Murray Anderson (History)

      This means Israel can deliver about 600 kilograms to ICBM range, which gives it minimal deterrence against powers that far away. This exists without demonstrating accuracy or re-entry capabilities, since no one is going to risk a city over anything less than life-or-death issues.
      Israel probably doesn’t need this capability, since it has no serious enemies that far away. However, the same reasoning applies to Iran, which if it demonstrates similar capabilities would have minimal deterrence against a U.S. nuclear attack.

  12. John Schilling (History)

    The Shavit, and the Jericho III, are both ICBM-class vehicles, albeit limited to ~500 kg payloads in that role. This is an unavoidable consequence of achieving ICBM-class range by adding a small third stage to a two-stage IRBM. But 500 kg is where e.g. the MTCR starts worrying that a bad actor might be able to shoehorn in a complete nuclear warhead, and the Israelis are presumably rather better at warhead design than most players in that game. Why they would want that capability is an open question, but it’s not clear that they have actually deployed it as an ICBM and it is the sort of thing an existentially paranoid nation would want to have on the shelf just on general principles.

    The other interesting undeployed ICBM-class launch vehicle is Japan’s J-I, good for ton-plus payloads between continents. Flew only once, on a suborbital
    trajectory with a maneuvering hypersonic reentry vehicle as a payload, then put on the shelf. Now it’s coming back as the Advanced Solid Rocket, er, Epsilon Launcher, with a vaguely-defined mission but increased emphasis on manufacturability and operability. I have a partial writeup on that one that I need to finish one of these days…

  13. Gregory Matteson (History)

    The Israeli Shavit & Jericho III really don’t count for the purposes under discussion. According to public information they are built around “commercial” castor 120 solid rocket modules obtained from us, with undeniable full-up support. By no means can that be counted as a separate, independent program.

    Japan, however, got turned down flat for help on orbital/ICBM class vehicles; and for help on development of guided re-entry. This is something they are not terribly happy about. The costs and difficulty experienced by the Japanese program, as with the Indian program, support the notion that this is remains one of the most difficult undertakings on the slate for the ambitious powers of the world.

    • Allen Thomson (History)

      As I understand it, the Castor 120 was proposed for an export version of Shavit that went under the name of Leolink and, actually, never went anywhere. Shavit and Jericho first and second stage motors are made, AFAIK, by Israel Military Industries’ Rocket Systems Division.

    • John Schilling (History)

      I’m calling for a source on the “public information” that says a Shavit 1 or Jericho III uses Castor 120 solid rocket motors. The proposed LeoLink-2 satellite launch vehicle would have used the Castor 120, to provide improved performance and improved made-in-the-USA content for bidding on NASA contracts. LeoLink-2 was never built. I can find no source that says the vehicles which actually were built used Castor 120s, or that any Castor 120 was ever exported to Israel, or that the Shavit 1 used anything but a home-grown Israeli rocket motor. Presumably incorporating lessons learned from some quiet pre-1965 collaboration with France, for what little difference that makes.

      And again, what “costs and difficulty” experienced by the Japanese or Indian programs? I didn’t raise the cost issue when I first discussed the Indian program because I didn’t have those numbers readily available, but apparently you do. How much did India spend developing the SLV, or the Agni-TD? What critical technical difficulties did they encounter?

      If you want to talk about “most difficult undertakings”, let’s talk stuff that’s really hard. In the arms control arena, for example, let’s talk about modern jet fighters or submarines. Or even less-than-modern; anything that isn’t a complete joke in the 21st century, will either have to be imported or will tax the research and industrial capabilities of all but the richest nations. ICBMs are relatively easy on that scale.

    • George William Herbert (History)

      Greg –

      Your comment about Japan is somewhat confusing and perplexing.

      The Japanese H-1 launch vehicle was (literally) a homebuilt Delta-2 launch vehicle. Full plans transfer. Same fairly large solid boosters, etc. The US outright sold the design to them and helped with the engineering. They eventually developed the LE-5 LH2 burning upper stage engine on their own, a slightly smaller RL-10 class motor, and did other upgrades. But it was a Delta design to start with. In an earlier time, I got to crawl all over the H-1 pad (and H-2 pad before they flew it). I still have a bit of solid exhaust slag off the flame trench somewhere.

      H-2 was all their own, both the large solid boosters (which ended up as the J-1 / now Epsilon vehicle first stage) and the core liquid hydrogen stage and upper stage.

      Japan also indigenously developed the Mu series solid booster space launchers, which culiminated in a MX sized M-V launch vehicle with about 1.8 tons to LEO capability. That would support an ICBM range warhead of 3 plus tons, plenty for even a very very primitive atomic warhead. Plenty for a MIRV system, if it came to that. Plenty, even, for many many tons of conventional warhead to Pyonyang or Beijing, if it came to that.

      Japan’s own internal defense analysis community pointed out potential military implications of the Mu series and particularly M-V. When I first did the analysis on the launchers as missiles, back in 1992 (before M-V flew the first time), nobody in Japan would openly talk about that. Now, they openly talk about it.

    • George William Herbert (History)

      On the matter of Shavit and Castor 120 –

      The Castor 120 motor when loaded weighs twice what the complete Shavit launch vehicle does, and the basic Shavit will nearly fit inside the Castor 120 empty casing if you unbolt the nozzle.

      You need to find higher quality “public sources of information”…

  14. Gregory Matteson (History)

    I stand corrected on the Shavit. My public information was that they publicized planned launch services under the Shavit name, apparently with no sales, stating the boosters were Castor 120.

    As for the difficulties of the Japanese program; according to their Lambda rocket program, with US help had 9 launches with 4 failures from 1966 to 1974 leading to the Osumi micro-satellite, which was too small to count for ICBM class, at 24 Kg.. From 1966 they developed the Mu rockets. The Mu4S first attempt failed, second attempt succeeded in launching Tansei 1, at over 500 Kg, certainly an ICBM class vehicle. So, five years and five developmental failures.

    Being an open Parliamentary system, the Japanese have had very public debates about the high costs of their space program from day one, as have the Indians.

    India’s SLV program, from 1979-1983 was a micro-satellite program with 1 fail, 1 partial and 2 successes. The follow on ASLV 1987-1994 had 2 failures and 2 successes, but was still only 150 Kg. 1994 began their PSLV program with a successful full sized satellite. 15 years and 7 launches before they had something clearly ICBM class.

    I was basing the claim that the Japanese lacked US cooperation on my memory of Japanese representatives complaining very long and loud at US Congressional hearings that they could not get the help they needed because of missile non-proliferation treaty restrictions. I was not aware that some of their rockets were essentially copies of US designs.

    • John Schilling (History)


      You seem to be counting test failures as an indication of technical difficulty. If your standard for calling something “difficult” is that it didn’t work flawlessly the first time, that’s a pretty low bar. I’d wager large sums of real money the code running this web site crashed or otherwise failed the first half-dozen times through the compiler; perhaps running a web site is among “most difficult undertakings on the slate for the ambitious powers of the world”

      Test failures, are the *solution* to technical difficulty. Building ICBMs is, well, rocket science, and even card-carrying rocket scientists like myself would find it very very hard to simply sit in front of a drawing board and apply our Mighty Brains until we have complete blueprints for a guaranteed-to-work ICBM or space launch vehicle.

      But as with the hypothetical single-stage ICBM that started this whole discussion, nobody ever does it that way. What is actually done, is to sit at the drawing board, CAD terminal, modeling terminal, and laboratory bench until you have a vehicle that will probably complete at least half of the scaled-down mission before crashing, add instrumentation and telemetry, then see what actually causes it to crash when you light it off.

      Lather, rinse, repeat, until you can reliably complete the mission. If the mission is ICBM-like, it will take about seven years and two test failures of an ICBM-class vehicle. It will not matter much whether you are doing this in the United States, India, or Iran.

      Whether or not you want to classify a problem requiring this level of effort as “difficult” is up to you. But it is the sort of “difficult” problem that is successfully solved by most people who try it, in seven years or so, with no more resources than an Iran or even North Korea can readily devote to the task. And the first failed test of an ICBM-class vehicle, is usually the sign that a reliable ICBM-class vehicle will enter service in a year or two.

      Doesn’t always work that way, but it is the way to bet.

      Oh, and yes, you can always skew the apparent statistics by counting some or all of the years that someone was embarked on non-ICBM-class rocketry as time “required” to develop an ICBM. I believe the Indians have been working with metal-cased solid-fuel rockets for military applications since sometime in the 1780s. But the time between saying, “we should actually build an ICBM-class rocket this time”, to initial operational capability, is historically about seven years, in India and elsewhere, and even for institutions which never built any rockets of any sort before deciding they wanted one of the big ones.

  15. Gregory Matteson (History)

    I think the purpose of these discussions is to learn and improve the public discourse. I approach this not as an insider/engineer, but a member of the public with a BA in History and a 2 year certificate in Computer Technology, with way too much time on my hands, reading what passes for public discourse in the media.

    We can’t all be insiders

    It seems to me that the information on Japan from Mr. Herbert greatly strengthens the argument that ICBMs remain extremely difficult to build.

    • John Schilling (History)

      The information on Japan from George Herbert is that the Japanese license-built a vehicle of substantially greater than ICBM performance to plans provided by the United States, improved it, then built their own wholly-indigenous vehicle whose performance was so much greater than ICBM class that its strap-on boosters serve as the basis for an ICBM-class vehicle. Meanwhile, a second indigenous program simply went and developed an ICBM-class vehicle.

      George did not explicitly mention it, but the second team (JAXA) operated independently and in parallel to the first (NASDA), and JAXA put their first payload in orbit before the first American engineer showed up to help NASDA build its much larger vehicles.

      This does not support the thesis that building ICBM-class vehicles is “extremely difficult”. It somewhat supports the thesis that building large space launch vehicles with high-energy orbit transfer stages might be extremely difficult. I think either George or myself would be glad to try and explain the difference, on or off line. As you say, this is supposed to be a learning experience.

  16. Gregory Matteson (History)

    John, I think I do understand the difference. May have gotten a little out of my depth in later discussions. The thing that motivated me to speak up in the first place is the change in thought I got from reading the Siddiqi book I referred to in my first posting. That book was sponsored by NASA to further understanding of the Russians, and of the whole history of the space program.

    The experience was pretty much “every thing you know is wrong”. An awful lot of what we were told in my educational years; Apollo and before; was wrong. A lot of others from my generation have learned that much. Some people have come out scoffing, saying “Rocket Science” is overrated. I don’t think designing on 1.2:1 margins can ever be easy.

    I do remember as a kid getting up very early in the morning here on the west coast to watch the latest attempt to launch a satellite. It usually went up in a fireball, as I recall.

    I see a lot of stuff in the mainstream media trying to worry us that what we see in North Korea and Iran means that any day now we will wake up under the threat of annihilation from men who seem not to have a real good grip on reality.

    I suppose I’m trying really hard to justify the belief that it’s not really that bad.

    • George William Herbert (History)

      Not all rockets are done on on safety factors of 1.2 . Proton, for example, is mostly 2.0 . They assemble the rocket horizontally, by lifting the stages up with a single crane lifting a chain attached at each end. They are ok with people walking around on the stage as it’s lifted and moved into position.

      Small ballistic missiles often are very robust as well, due to the road mobility / TEL issues.

      NASA encourages safety factors of higher than 1.2 on most parts and requires them on others. Solid booster casings are often an exception, where there’s just so much of it that bigger margins are extremely difficult to justify given the performance impact. But even there, 1.2 is way low – the Titan IV SRMU PQM-1 test failed at 1460-something PSI versus nominal operating pressure of I think 860 PSI.

      The new companies have in many cases flown vehicles again after minor to medium crashes or tipovers on landing, just by turning them upright and inspecting for damage and dusting them off. Some crashes have been worse (some have totalled vehicles, to be sure). But a rocket that can fall over on landing, be propped up, and fly again in the same day? That’s not fragile…

  17. dove (History)

    Escape velocity and the resulting weight penalty is needed only if you have to go to space and back. How hard is it to take a shorter range cruise missile and add fuel/range to is ? My guess is if you have reliable GPS access, it might be less difficult than building an ICBM that has to go all the way up and then do re-entry.

    I would also guess (without any proof) that post Agni-V, which is 5000kms with a 1 ton warhead, india might focus on long range cruise missiles rather than an ICBM. Brahmos supersonic cruise missile is a good platform to base such a plan and India has gone some steps forward with testing a hypersonic version as well as a basic scramjet engine that can scoop oxygen from atmosphere to reduce fuel weight.

    Iran has demonstrated ability to build subsonic cruise missiles – basically glorified auto-pilot aircraft – which could be extended in range to reach farther.

  18. Gregory Matteson (History)

    ICBMs are national status symbols that have been used to cow the world population for 53 years. Why should they stop now? Russia had numerous piloted and un-piloted projects, sub & supersonic at the same time they were developing missiles. Kruschev in particular promoted missiles as economical a solution to the Soviet problem of deterrence, not as a practical weapon.

    The status and terror character of the ICBM threat seems to go against Indian national character. Further, India does not appear to have a need for global deterrence. Hmmm, I think that point was already made by others.

    As for cruise missiles, speaking strictly from my historical perspective, they have been around for a long time, and have serious credibility problems as a deterrent. From buzz-bombs to tomahawks, the invulnerability of cruise missiles has appeared lacking compared to missiles.

    In case anyone was not paying attention during the Iraq wars; the much derided Soviet style air defense in Iraq was routinely shooting down tomahawks on camera for the CNN audience; Pentagon denials to the contrary not withstanding. Until something truly horrific is accomplished with cruise missiles, such as nuking a city, no one is really going to believe in their deterrence.

  19. Mark Gubrud (History)


    I’ve decided your rule of seven years and 2-3 test failures is a very interesting thesis, especially since you write as if you have a database to back it up. Have you written a paper on this? It would be a very useful contribution to the arms control literature.

    I am still very doubtful that you can compare North Korea to India or that even Iran is equal to India in terms of resources it can bring to bear – money, technical resources, as in highly skilled and creative people, available technologies/techniques/parts/tools, etc.

    But you have stated a strong thesis, and in the long term, it has to be true that it just keeps getting easier to do something the US and USSR could do 50 years ago.


    • John Schilling (History)

      I’ve never quite got around to writing all this up formally, though I am now motivated to do so. Specifically including a graph, or at least bar chart, of mean time to ICBM development vs. size of the national industrial economy at the time. Though that will involve deciding whose economic statistics to believe for the Soviet Union in the 1950s 🙂

  20. Anne Bazuin (History)

    Waiting for answer from mr. G.W. Herbert on my question of sept.17 2010

    • George William Herbert (History)

      Sorry for the delay, I missed that there was a question in there.

      I forget the closed form of the time equation for ballistic range; it’s in trajectory textbooks, and I’ve used it, but I don’t recall off the top of my head.

      As an approximation – the ICBM is travelling “around” orbital velocity, which is 7,200 +- epsilon m/s. It’s trajectory is much more up and down than a circular orbit, but as a first order answer that’s not too bad. It’s 90 minutes to do a full orbit at orbital velocity, 40,000 km of circumfrence. So as a rough first guess – divide the range by 40,000, multiply by 90 min.

      That seems to give the “30 minutes” which are commonly quoted for US / exUSSR missile flight times, so it’s probably a good guesstimate.

  21. Anne Bazuin (History)

    Yes mr.Herbert, a low satelite (100 km) on a circular orbit has an orbital period of 86 minutes and a speed of 7.850 m/s. These are useful data in quick determining the flying speed of long range ballistic missiles. They fly half the globe, 20.000 km, in 43 minutes. It gives you some time for counteractions.
    The flying time of a ballistic missile on a range of 1.500 km, west Iran – Israel, is about 10 minutes. Very little time for counteractions. What to do ?

    • George William Herbert (History)

      As a rule, if the missile is going to hit in 10 min, you can try to shoot it down but probably not scramble your own launch in that time.

      And, probably, shouldn’t try – as the missile (for an Iran – Israel shot) could well be conventional warhead, and flattening Iran with a thermonuclear strike in response to a single or few conventional IRBM shots would be seen as unacceptable behavior.

      This is why your missile force lives in hardened bunkers in the side of mountains. So that you don’t have to worry that the enemy will destroy all of it with a few lucky shots, nuclear or not, and can ride out attacks and decide what to do afterwards, once you know what it was.

  22. Anne Bazuin (History)

    Mr. Herbert, thank you for your realistic answer.
    I don’t rely on my ballistic missile defence, even not when it is autonomous and quick acting with SM-3 hit to kill technics. It may fail.
    I want to decide afterwards how proportional my counteract should be and how aceptable my behavior is. If I decide to act, I can launch my ICBM’s from my hardned bunkers or my SMLBM’s from my submarines on strategical positions.