US Korea Institute

Saturday March 25th 2017
Subscribe to Our RSS feed@38NorthNK on Twitter
38 North offers informed analysis of events in and around the DPRK.

Subscribe for latest



North Korea’s Large Rocket Engine Test: A Significant Step Forward for Pyongyang’s ICBM Program

11 April 2016


North Korea’s April 9 test of a large liquid-fuel engine is a disturbing development that not only highlights the growing threat posed by Pyongyang but should also put to rest, once and for all, all claims that the North’s WMD programs are a hoax. In fact, the test demonstrated that North Korea has an even greater capability at a more advanced state of development than previously anticipated.

Specifically, three important conclusions can be reached based on this test:

  1. The North Korean test involved a tightly-coupled pair of propulsion units from an old Soviet submarine-launched ballistic missile (SLBM), known as the R-27 or the SS-N-6 “Serb.” North Korea has long been believed to possess this technology, but it was never confirmed before now. This engine uses high-energy propellants that would give a missile greater range than Pyongyang’s traditional mix of kerosene and nitric acid.
  2. Using this technology, North Korea’s road-mobile intercontinental ballistic missile (ICBM), the KN-08 or the KN-14 modification, could deliver a nuclear warhead to targets at a distance of 10,000 to 13,000 km. That range, greater than had previously been expected, could allow Pyongyang to reach targets on the US east coast, including New York or Washington, DC.
  3. If the current ground test program continues and is successful, flight tests of a North Korean ICBM could begin in as little as a year. Moreover, Pyongyang may be able to deploy this delivery system in a limited operational capability by 2020.


For more than a decade, experts have had reason to believe North Korea had obtained from Russia the technology for an old Soviet submarine-launched ballistic missile—the R-27 “Zyb,” or to old-school Cold Warriors, the SS-N-6 “Serb.” But until recently, the evidence was circumstantial: Russian engineers traveling to Pyongyang, parade models that look like the old Soviet missile, a discrepancy in the accounting of ex-Soviet hardware after the Cold War, but never any clear proof such as a North Korean test of Soviet hardware. We had begun to wonder if maybe some enterprising Russian schemers had sold the North Koreans a warehouse full of rusty hardware that could never fly again.

And from the day the KN-08 road-mobile ICBM first paraded through the streets of Pyongyang, there was speculation whether it might use this technology. The KN-08 is a much larger missile than the R-27, and while elements of the R-27 might be a good match for the upper stages, the R-27’s Isayev 4D10 engine was simply too small to lift the missile. It was possible that the North Koreans might be able to scale up the 4D10 engine, or cluster two or three of them together, but that would have proved challenging for their engineers. The simplest hypothesis was that they would use their old Scud or Nodong missile engines for the first stage—the same approach that had taken with the Unha satellite launch vehicle (SLV).

Figure 1. The April 9 liquid-fuel engine test.

(Photo: Rodong Sinmun.)

(Photo: Rodong Sinmun.)

For months, the North Koreans had been modifying their vertical rocket engine test stand at the Sohae Satellite Launching Station for testing larger rocket engines. They completed that effort last month and there were some signs that a test might be in the works. Conveniently, the North reported this test on the day I was attending a conference of American rocket experts, some of whom had worked with America’s old liquid-fueled ICBMs. The immediate consensus was that this was no Nodong or any other engine reflecting North Korea’s traditional missile designs. But it was a very good match for a pair of 4D10 engines clustered together into a propulsion unit for an ICBM. This ICBM would be more capable than previous thinking on what the North could build, and also possibly closer to deployment than had been anticipated.

The Engine

Looking at a close-up of the test, what appears from a distance as a single engine shows two exhaust plumes almost adjacent to one another, emerging from a hemispherical dome at the base of a stubby cylinder. The engines themselves are not visible. As the R-27 missile was designed for use in the tight confines of a ballistic-missile submarine, critical space is saved by putting the 4D10 engine inside the missile’s fuel tank with only the end of the nozzle projecting below. The Soviet Union went on to use this feature in more advanced submarine-launched missiles, some of which remain in service to this day.

North Korea appears to be retaining the submerged-engine configuration, but with two engines instead of one. As the 4D10 engine normally includes a complex assortment of pumps, valves and other plumbing wrapped around the nozzle and thrust chamber, this isn’t a simple matter of putting two stock engines side by side, but rather indicates a substantial modification of the original design. Possibly some components are now shared between the two tightly-coupled engines. It would still be better to use a single, larger engine—one small bit of good news in this test is that it seems to indicate that North Korea still lacks the ability to design (or buy) engines any larger than the 4D10.

Figure 2. Close up of the engine test.

(Photo: Rodong Sinmun.)

(Photo: Rodong Sinmun.)

Also visible are the exhaust plumes of smaller rockets on each side of the main nozzles, another feature of the 4D10 engine. The most efficient way to steer a missile is to mount the main engine on gimbals allowing it to swivel a few degrees in any direction, but that’s not really practical if the engine is inside the fuel tank. Instead, Soviet designers gave the 4D10 an independent pair of vernier engines on gimbaled mounts outside the tank. Those smaller engines appear here in about the same place as they would on stock 4D10 engines.

Perhaps the most spectacular difference between this test and anything previous ones conducted by Pyongyang is the clean, translucent orange exhaust. Scud and Nodong missile engines burn kerosene fuel, which almost inevitably produces soot and thus an intense, solid orange flame. The plume from these engines retains the orange hue, indicating carbon in the fuel but without the soot. Some of the internal detail of the plume as it leaves the engine and interacts with the surrounding atmosphere is visible. In the longer views, there’s no sign of the black, smoky exhaust we would expect from a kerosene-burning engine.

There are several fuels that could produce this sort of flame, but the only one we’ve had any hint of from North Korea is something called unsymmetrical dimethyl hydrazine (UDMH), which in combination with nitrogen tetroxide oxidizer is believed to power the small third stage of North Korea’s Unha SLV. These propellants are more efficient than the crude kerosene and nitric acid of the Scud and Nodong missiles, which is why the Soviets adopted them for the R-27. Until now, there was no sign that the North Koreans were using them in any large rocket engine.

In short, the North Koreans have demonstrated something we only suspected them of working on and have done it on a larger scale than anticipated. Moreover, these test images are sufficiently precise, accurate and detailed that Photoshopping or other fakery can be ruled out. The exact scale is difficult to discern, but the relative dimensions and geometry at least are consistent with a pair of closely-coupled 4D10 engines packaged in the base of a KN-08 or KN-14 missile fuel tank. And while still photos don’t tell us how long the test ran, the engines appear to have achieved a steady-state operating condition.

The Missiles

The recent test helps solve a minor puzzle that was raised when North Korea teased us with a carefully-cropped photo of the base of a KN-08 missile last month. That photo suggested two closely-coupled engines, but with nozzles a bit over 50 cm in diameter—too large for a Scud, and too short for a Nodong. True, a Nodong engine could have had its nozzle cut down to fit the KN-08 engine bay, but that seemed unlikely and inelegant. The nozzle of the 4D10 engine is just about right for what we saw. So are the positions of the vernier engines. And while we’ve never seen the base of the KN-14 missile (the Pentagon’s new name for what we had called the KN-08 Mod 2), the external configuration of the first-stage engine bay is identical so it probably uses the same twin-pack 4D10 engine.

Since the KN-08 and KN-14 are almost three times the weight of the old Soviet R-27, would two 4D10 engines be sufficient to power it? Three would certainly be better, but might be an impossibly tight fit. With only two 4D10 engines the missile would climb away from the launcher with an acceleration more suitable for an airliner than an ICBM. But unlike the airliner, the ICBM’s performance would increase dramatically as the fuel load burned away. The slow initial acceleration will cost the North Koreans a bit of performance, but unless they wind up launching with American missiles literally seconds from hitting the launch site, it won’t matter in the end.

In the end, the 4D10 engine is about 15 percent more efficient than North Korea’s kerosene-burning Scud and Nodong engines. And the submerged installation allows for bigger tanks carrying more propellant. Even with the lumbering initial climbout that will give the North substantially longer range. With a 500 kg warhead, the KN-08 is now estimated at having a range of over 11,500 kilometers, compared to 9,500 km using Scud-technology engines in the first stage. That’s enough to reach New York City or Washington, DC from North Korea. Even the two-stage KN-14 can reach over 10,000 kilometers with a light warhead, enough to cover all of the US west coast.

Future Prospects

The only good news from the recent test is that it didn’t include a complete first stage. Analysis of commercial satellite imagery suggests that the Sohae test stand was recently modified to use new propellants, so this may have been the first full test of the dual-engine ICBM power plant. But Kim Jong Un himself was present, and given the likely penalty for embarrassing Kim with a failed test, it is reasonable to suspect the engineers had high confidence from earlier testing of single 4D10 engines.

The most recent 38 North assessment of the status of North Korea’s ICBM program assumed an additional two to three years of ground testing would be required before North Korea would be ready to conduct the first flight of a new ICBM. However, based on last week’s engine test, and recent video of Kim Jong Un observing reentry vehicle ground testing, North Korea might be far enough along to conduct flight tests in as little as a year. It won’t happen tomorrow—they still have to actually build the first stage, and will almost certainly want to test the complete stage on the ground before it is launched.

If the North Koreans can launch an experimental ICBM early next year, chances are it will not work as planned. (North Korean missiles almost never work the first time.) But the timeline has moved up, and an initial operational capability for the KN-08 or KN-14 of about 2021 rather than 2023 is entirely possible. There is a small chance that the missile could enter limited service by the end of this decade. Large-scale deployment would require the ability to manufacture complete 4D10-class engines from scratch when they may still be using ex-Soviet hardware for key components like turbopumps and injector plates. But while this requirement may result in some delay, it is an obstacle that the North Koreans can overcome in time.

Whether the increased pace and visibility of North Korea’s WMD activities across the board will continue remains unclear. Certainly, Pyongyang has pulled back the veil normally surrounding research and development activities to give the international community a glimpse of its efforts. That should put to rest the reoccurring argument by some experts that the North’s nuclear and missile programs are a hoax or are fake, as Pyongyang clearly intended these actions to add credibility to their threats. This behavior may continue, may be a reaction to the current joint US-ROK military exercises or could be part of the run-up to the coming Party Congress. Whatever the reason, North Korea is clearly moving full-steam ahead with its threatening programs.

Reader Feedback

15 Responses to “North Korea’s Large Rocket Engine Test: A Significant Step Forward for Pyongyang’s ICBM Program”

  1. Rick Spurway says:

    Am I correct in thinking this rocket engine test stand is located at 39°39’9.87″N 124°42’50.37″E with the observation building located above at 39°39’17.34″N 124°42’35.87″E. The best Google Earth image is in the historical imagery dated 10/3/2014. You can see the sea in the background of a couple of the photos here…

  2. Max says:

    I’ve read Clay Ramsey’s comments with great interest. “[W]ho fools who?” Does DPRK’s capability in the nuclear and missile technology have “bubbles”? And if so, cui bono? It seems three different groups have vetted interest here: two DPRK supporter groups and one adversary group. The adversaries may have political and financial gains from the bubbles. For example, when NKorea recently fired missiles, Israel was quick to point out that the missiles used Iranian parts. I cannot confirm or disconfirm the Israeli claim, but it is clear Israel is a party to the interest group. Also, the recent US-SKorea talk about THAAD deployment has, to some degrees, been framed on them. The supporter groups have sympathizers as well as insiders. The bubbles certainly help “the friends of DPRK” by making the propaganda machine look more realistic. I am not surprised if some of the 38North contributors may be among them.

    To me, however, the most interesting but vastly overlooked group is DPRK’s military elites and their involvement in the bubbles. A clear example was the “stealth naval vessel” displayed by DPRK last spring. The vessel was nothing more than a disguise and certainly was not what they claimed it to be. It carried towering commercial-grade radars among other things (too bad I cannot upload the photo with this comment). Are the North’s military elites dumb? I think not. Why, then, did they publicly show it as a state-of-the-art stealth vessel? Who did they want to fool?

    Part of the answer, I think, may be found in the DPRK leadership structure and nature. In DPRK everything must be controlled and approved by only one guy. If that guy is happy and satisfied, everyone else can have a good night’s sleep. But that guy keeps popping up everywhere and “instructs and guides” his men on technological details and strategic developments; he gives them orders with a specific time frame and details. How does DPRK meet the kind of professionalism required in today’s world? Is the little Bozo the clown Mr. Know-All?

    The fate of Gen. Hyon Young Chol tells a lot. He was the North’s defense secretary with 50 years of experience in the military. But he was recently executed brutally with anti-aircraft guns because he tried to “advise” Lil Kim on military matters and “complained” about lack of his strategic wisdom. No one wants to speak up for Lil Kim anymore. Why bother to talk about the reality? The miserable failure of the Musudan missiles in recent weeks might be another case. About 20 were dispatched to combat units months before Kim Jong Il fell ill with a life-threatening stroke. But the DPRK military had never performed single test fires until now, and it seems NKorea has no idea what has caused the repeated failure. How are we to understand their mentality and audacity? DPRK is good at one thing: photoshopped evidence. So, who knows how big the bubbles really are.

  3. Clay Ramsey says:

    Regarding the SLBM test on the 23rd: That’s a big motor (actually, to my eye, it’s a cluster of 4 motors). I can’t identify them this time and have no idea where they may have come from. I can’t see anything in these images that definitively identify them as fakes, but considering the range it flew, I’m just projecting the trend outward. This is by far their best one yet though, and it had to have cost a bundle. My only real question is what kind of motors did they use for this latest ‘test’ and where did they get them. The radar data will show if that is the actual kind of missile that was tracked on the 23rd by measuring the acceleration and deceleration profile. If even the 18 mile missile was faked, we may not know about it, but everyone else will. I guess I’m being cynical, but all those earlier photos of various fabricated missile related things makes me slow to believe it. A cluster of short burn, high thrust motors would look just like that. A regular ballistic missile fired on a day with similar sky conditions out to 18 miles would be an obvious way to attempt a deception (of course, it ultimately would be detected because of the distinct sparkle radar signature finned rocks have, as well is the far too high terminal velocity). It’s odd to me, the pictures they released (other than these, which are pretty good) are outrageous. There is no way the US, Japan or RoK could be fooled by them. Am I to believe that any of them actually believes that DPRK has developed an engine that is in any way similar to a 4D10, or that even the tiniest shred of the KN-08 missile is real, or that the earlier SLBM ‘test’ was done using anything other than just an SA-2 booster? But for some reason, all the DPRK adversaries always portray these programs as potential threats, knowing full well they are not. Maybe pointing out the deceptions is the last thing they want. I’m beginning to wonder, who’s fooling who?

  4. Clay Ramsey says:

    Geez, I wonder if I’m getting into their head. This is like the 3rd time they reacted to something I posted! But this last one would have had to be some really frantic reacting, they would have had to pull it off in a matter of hours. Maybe this latest SLBM one was pure coincidence, but man, what a coincidence! Anyway (if they are listening) sorry, but I don’t believe any of it anymore. I can think of several ways that one could have been faked as well and considering the rash of really revealing (in a negative way) photos out there, there is no reason to suspect anything other than a fake out once again. Imagine a single engine version of the one in the ‘submerged’ test of the topic being discussed here. Pressure fed using an SA2 nozzle cooled with water. A small load of propellant to keep the weight flyable and burn time short enough not to burn through. Maybe fin stabilized (we’ll see, if they ever release believable photos). I guess I will remain skeptical until I see any evidence to suggest otherwise. So far, there is none, it all points to an elaborate multi discipline deception. There are plenty of engineers out there, and they don’t fake so easily.

  5. Clay Ramsey says:

    Was poking around and found yet another example of a DPRK missile program being faked using SA-2 parts.

    Musudan SLBM


  6. Clay Ramsey says:

    Ok, I figured it out. If you look at the nozzles of the supposed ‘KN-08′ in the link from my earlier comment, you will see that they are too thin to be regeneratively cooled nozzles, they are solids. If you assume the fourth guy standing from the left is 5’6”, then the nozzles work out to be about 23 inches across. From the video of the test firing, if you assume the red shed to the left has a 6′ door in it on the right side, you will find the nozzles of the test engine are also about 23 inches across. Now look at an SA-2 booster and see the stunning similarity. I think I’m ready to flatly state that DPRK is faking an entire ICBM program using old scrap SA-2 parts. The engine test submerged engine probably really is submerged, but submerged in water. They slapped an injector on the end of a pair of old expended SA-2’s and fired them for about 10 seconds.

  7. Clay Ramsey says:

    Dr. Schilling: The fuel being supplied separately to the verniers is very suspicious to me. If the verniers had their own turbopumps, they would not need to be separately supplied from the main propellant tanks in the ‘submerged’ system, and if they did not have turbopumps, they could be supplied from the main engines themselves. Besides, I found an image of a crashed 4D10, and the verniers don’t appear to have their own pumps, they are traditional style verniers. The over expanded vernier exhaust jet is also what you would see in an SAM sustainer (and they would have plenty of those laying around). Besides, if those really are 4D10 style engines (operating in a submerged role) that setup makes even less sense. The trouble making it all work with such a small tank (having to be resupplied from somewhere) would be needlessly complex. Just imagine keeping ullage pressure stable in a setup like that. Also, there are partial images of the supposed KN-08 engine bay showing what appear to be a refined installation (not submerged and looking not at all like a 4D10),×680.png nothing like the laughably crude engine on display in that video. If the actual running test engine is so crude, why does the installed KN-08 look so impressive. Simple, it’s *all* a well orchestrated (but ultimately kind of clumsy) fake.

  8. John Schilling says:

    James, there are a couple of images out there that are quite clear on this being two nozzles, see e.g. at 1:48, I think during startup (off mixture and low pressure). And two verniers on each side, which matches what you’d get from a pair of 4D10s. I’d prefer full video, but even with the stills I don’t think this is just photoshop.

    And you’re right, you don’t just bolt two 4D10s together and make this work. Not sure about building it entirely from the ground up – quite possible they are using legacy turbopumps, injectors, even chambers, but it is a major remanufacturing job at the very least.

  9. J_kies says:

    Dr Schilling; twinning a pair of Soviet era Isayev 4D10 engines is a level of post manufacture adaptation that is harder than just building new engines to the design drawings. From the images provided; I cannot discern a twin engine set especially with some of the image artifacts present that may indicate the usual DPRK photo-shopping. If the ‘clear’ gap is supposed to be indicating separate flow areas, downstream morphology does not clearly show a paired engine set.

  10. John Schilling says:

    Clay, the 4D10 vernier engines are in fact supplied with propellant separately from the main engine, with their own small turbopump system. This avoids the need to run high-pressure lines through the tank wall for the verniers, and allows the verniers to operate as a sort of post-boost system after the main engine shuts down (not important in a first-stage application, but very helpful for the original R-27). And the verniers here appear to be maybe half the size of an SA-2 sustainer.

    Also, you do need verniers (or control vanes) if you are using submerged engines, even two of them, because you can’t gimbal the main engines once you put them inside the tanks. It’s not worth the bother to do that on an ICBM, of course, unless it turns out that the only high-performance rocket engines you could buy happen to have been designed that way. This propulsion package has several features that don’t make sense for anyone designing an ICBM first stage from scratch but do match what we know of the 4D10.

  11. Clay Ramsey says:

    I understand the images are not very good, but I’m going to add an analysis that is a little different. Every description of DPRK liquid engines (wherever you look) almost always states Hydrazine, or UDMH but if you look at the pictures they all just scream Aniline. There is no way this engine is Hydrazine, UDMH or MMH unless something is seriously wrong (end even then, I’d say no way). It’s Aniline, something they can easily buy and isn’t even considered a rocket fuel by most (but it should be, since it’s hypergolic and really great for that very thing). As for the engine in the video, even though the ‘engine’ is real enough, to me it does not represent a propulsion system that would ever be reasonably adapted to a rocket or missile. In my opinion, it is a total fake. Note the following (reference the video). 1) the ‘verniers’ appear to be old SA-2 engines (note the over-expanded exhaust jets) operating under their normal as designed cycle. You can see they appear to be supplied propellant separately from the main engines. 2) There is no gas generator exhaust visible from the main engines. The main engines appear to be pressure fed. Beyond the fact that a closed cycle engine would be very difficult for DPRK to develop (or even copy), the fact that the verniers are supplied with separate propellant pretty much rules out that the main engines use any kind of trubopump system. This is opposite from logical practice, where the verniers are parasites to the main engine propellant system. In this case, they can’t be, the SA-2 style engines would not be able to ingest the high pressure prop from pressure fed main engines. 3) There are no control vanes visible in the main propulsion system, and no obvious mechanism to gimbal the verniers. Besides, it makes no sense to use verniers at all since all axis of rotation can be controlled using a twin nozzle arrangement (although there is president for doing it that odd way).

  12. Steven Hayden says:

    In a clear challenge to the United States, the KCNA proclaimed: “The great success made in the test provided a firm guarantee for mounting ANOTHER form of nuclear attack upon the U.S. imperialists and other hostile forces. Those words of “ANOTHER form of nuclear attack ” were carefully chosen. The conventional ICBM even with improved fuel [solid fuel or new liquid fuel] is ANOTHER form of attack to US mainland and was not the basis of earlier claims of ability to wipe out US Mainland. The DPRK supports “outside the box” approaches and refers to these secret methods as “korean style”. The leadership pursues and perfects creative problem solving methods and unconventional unique solutions to complex problems.The ICBM engine and 300 mm MLRS show that diligence and innovation are skills useful on conventional path. The primary means of “korean style” nuclear deliver remains obscure for DPRK national security reasons.

  13. Steven Hayden says:

    Necessity is the mother of invention.They show creativity and dedication. A group picture in Rodong after test shows and joy between testing engineers and the leader of DPRK.

  14. John Schilling says:

    The Unha/TD-2 second stage is only 1.5 meters in diameter, perhaps not coincidentally the same as the original R-27 missile. A single 4D10 engine would be just about right for such a thing. This cluster, tight as it appears to be, probably couldn’t be shoehorned into such a space. And if it did, it would have a burnout acceleration of well over 20g in a two-stage configuration.

    Knowing that the North Koreans have working 4D10 engines, I’d now expect that if they did redesign the Unha for explicitly military applications they might replace the Scud engine they are probably now using in the second stage with a single 4D10. But it does look like they have something in mind for a pair of 4D10s as well, in a vehicle with a base diameter of just under two meters and a launch weight of just under 40 tons. That means either the KN-08/KN-14 family, or something completely secret that’s about the same size and performance as the KN-08/KN-14.

    And yes, limited to road travel with a convoy of fuel tankers, etc. We’ve always known that. Really, I expect an operational system would spend most of its time in bunkers, only occasionally shuffled around to try and confuse the satellites as to which bunkers.

  15. J_kies says:

    Interesting analysis; what evidence do we have that this is not a new second stage propulsion system for the Unha / TD-2 vehicle? The innate problems with trucking fueled hypergolic liquid propellant missiles still arise independently of engine work.

Leave a Reply

Credit for photo of young North Korean girl: T.M. All rights reserved, used with permission.