US Korea Institute

Monday March 27th 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 Largest Engine Test Yet

21 September 2016


On September 20, KCNA reported that Kim Jong Un had overseen the testing of a large new rocket engine at the Sohae Satellite Launching Station. They claimed this new engine had a thrust of 80 tons and would be use for a new space launch vehicle. Rocket engine thrust isn’t normally measured in “tons,” and there is some ambiguity in how to interpret that figure. It is also possible that the North Koreans are exaggerating; we can’t directly verify thrust from the images they released. Regardless, we can tell that the engine is substantially larger and more powerful than anything North Korea has tested before, even than the new ICBM engine tested in April, and the thrust may well be in the range of 160,000 pounds or 80,000 kilograms force.

And, while this particular engine may not be slated for an ICBM, we have now seen that North Korea can build large rockets using both solid and high-energy liquid propellants, to their own requirements. Whatever missiles North Korea may roll out in coming years, we can no longer expect to be limited to what can be cobbled together from old Russian cold-war leftovers.

The Engine

The plume coloration indicates that this engine probably uses the high-energy UDMH (unsymmetrical dimethylhydrazine) and NTO (nitrogen tetroxide) propellants used by the Musudan intermediate-range ballistic missile and the April 9 engine test—both derived from the old Russian R-27 submarine-launched ballistic missile whose technology North Korea illicitly acquired in the 1990s. However, a side view clearly shows the exhaust of a gas generator turbo-pump, rather than the more advanced staged-combustion cycle of the R-27. North Korea has previously used the gas-generator design on its (mostly) home-grown Nodong engine. The new engine does not have either vernier nozzles or jet vanes, suggesting the entire engine may be gimballed for steering purposes. This does not appear to be a copy of any Russian engine, but represents a mix of technologies well suited to a North Korean large engine development program.


(Photo: Rodong Sinmun)

(Photo: Rodong Sinmun)

(Photo: KCTV screengrab)

(Photo: KCTV screengrab)


It is reasonable to wonder whether North Korea might have had help in taking this unprecedented step, and two potential collaborators come to mind. China’s space launch vehicles, and some of their older ICBMs, use an engine called the YF-20 that is of the same size, performance and general technology as the new North Korean engine. However, there are visible differences between the new engine and the YF-20. We shouldn’t assume that China is helping Pyongyang. However, as we know, North Korea can sometimes connive with Chinese merchants to buy goods that are pretty obviously on the military side of the dual-use spectrum. For instance, while they probably couldn’t procure complete rocket engines, they could maybe get blueprints or critical components like turbo-pumps. Iran is another potential collaborator. There have long been ambiguous and difficult to pin down reports that Iranian engineers have been working with their North Korean counterparts on an “80-ton rocket booster.” The two nations have long collaborated in rocketry, with similar goals and technologies, and while their missile programs have diverged in recent years, their space programs are still in some respects a joint effort. Still besides the similarity in size, there is no additional evidence to prove this assertion.

The bottom line is that the North Koreans have the engine, it seems to work, and they can probably build more of them for whatever purpose they see fit.  The question is, what are they going to do with it? Given the size of the engine, they probably aren’t going to build an intercontinental ballistic missile (ICBM) with it. The engine is too big and powerful for the KN-08 and KN-14 road-mobile ICBMs under development in North Korea.

Space launch vehicles are another matter. North Korea has upgraded its Sohae launch site over the past few years to accommodate new, larger space launch vehicles that we have not yet seen. We may have just seen its engine. Or, more likely, one of its engines—liquid-fuel space launch vehicles typically use more than one engine on the first stage, and North Korean practice so far has been to use four engines. Iran, too, seems to favor a four-engine design. And a launch vehicle using four of these new engines would be about the right size for the upgraded launch pad and gantry tower at Sohae.

North Korea recently announced it plans to launch rockets to the Moon within the next ten years, along with launching increasingly capable satellites into Earth orbit. This engine would be suitable for launching modest unmanned lunar probes, along with geostationary communications satellites and various sorts of reconnaissance satellites in lower orbits. They still have a way to go on the necessary satellite technology, of course, but they are now one step closer to demonstrating a basic operational capability in outer space. We should start thinking about how we might live with a North Korea that has such a capability.

And, while this particular engine may not be slated for an ICBM, we have now seen that North Korea can build large rockets using both solid and high-energy liquid propellants, to their own requirements. Whatever missiles North Korea may roll out in coming years, we can no longer expect to be limited to what can be cobbled together from old Russian cold-war leftovers.

Reader Feedback

12 Responses to “North Korea’s Largest Engine Test Yet”

  1. Clay: Thanks, I got it! But I do not think that these nozzles are from the SA-2 booster. Also, for me it looks as if the photos in John’s article both indeed show the same engine running, and not an early nozzle test. KCNA’s video implies this:

    John: I agree on the textbooks. And please don’t get me wrong: I did not suggest that this is indeed a Russian engine – I just stated that we cannot rule it out yet, just as they may have received support from other countries as well. Or none at all. Still, all options on the table.

  2. Clay Ramsey says:

    Markus: Oops, I wrote it wrong, you actually have to search on ‘armscontrolwonk kn-08’.

  3. John Schilling says:

    Markus: Testing of jet vanes is generally done early due to the necessarily tight coupling with the engine design/operation. The same is likely true of combined-cycle verniers as seen in the Unha, but we have fewer examples of that configuration. What we do have is prior experience with North Korean testing. The first ground test images from their large solid rocket program, not only showed the jet vanes but showed data traces from the jet vanes being actuated. The first ground test images of the dual 4D10 system showed us the verniers, even though the 4D10 has a separate pump and feed system for the verniers and they can be tested separately.

    These may not have been the first tests of those engines, but they were the first tests whose images North Korea chose to release. So, to date, the North Koreans seem to test their steering hardware early, and they don’t go public until they are ready to show off the full package. I think it is a reasonable inference that the reason we don’t see e.g. jet vanes on the “80-ton” engine is that they don’t exist.

    Point taken on early German engineers and their protégés using “ton” as a thrust measurement back in the day. The North Koreans are a bit late to be directly benefiting from Operation Paperclip, but this may be a hint as to which textbooks they are using. But suggesting that this is a copy of a Russian engine design for no better reason than that the Russians made lots of engines and we don’t have details on all of them is I think rather irresponsible. If you’ve got any evidence beyond that, by all means share.

  4. Clay Ramsey says:

    Markus: Ah, I see the link itself has now been scrubbed. No fear though, just do a google image search on ‘armscontrolwonk’ and it will be on the third or fourth row. The one with a crowd of people and the KN-08 ‘missile’ (with nozzle partially visible) on the very left edge of the shot. You’ll see it’s just a pair of SA-02 booster nozzles (which are about 23 inches across).

    I don’t follow this site religiously, mostly only rockets and missiles and sometimes nuclear, but of all the things I’ve seen on this site, all have been misinformation except the one regarding the nozzle test a while back. That one seemed to be exactly as it was shown in the images and stated in the description. Also, it came with a official blurb from KCNA and a statement by Kin Jung Un himself. It stated 80 tons of thrust and was a test of valves and the combustion chamber configuration for a future satellite launcher. I did a sanity check on the nozzle thrust and quickly found it seemed to be what it said it was. I thought, probably cooled with water instead of aniline but who cares, at least it’s real. Two days later an image shows up of a complete engine with what appears to be the same size nozzle. That makes no sense. How can you go from the KCNA statement to a complete running engine in two days. What would be the need for nozzle and valve testing if the engine is already in a running configuration. Don’t believe it, the complete ‘engine’ is very suspicious with regard to the turbine exhaust using a pump discharge collector, obscured piping, turbine exhaust that is way too small for the amount of gas being discharged, and no video showing a credible run time. Not to mention that the combustion chamber seems to be glowing red along with parts in the area of the ‘gas generator’ (actually, it really is a gas generator!). I think I understand the mentality though, I’ve been through this enough myself. The thinking goes like this; ‘if we fool them, that means we’re smarter than they are’. Only the smart people fool the dumb ones, right? So when it’s all over everyone can lean back and say, ha ha, we were fooling them the whole time. Problem is, the ones being fooled don’t matter at all.

  5. Sorry, it may be too late, but I hope you do not mind if I add my two cents to this discussion!


    A nice and thoughtful piece of work! as you know, there are some general issues that I do not agree with, but I would like to discuss some other points here that caught my eye in this analysis. I look forward to your thoughts on that.

    1. Rocket engine thrust isn’t normally measured in “tons”.
    That depends where you are from, and who your mentors were. Wernher von Braun and his team referred to the thrust in tons, from what I heard even during their Huntsville days. This habit remained common in Germany for a long time, and the old generation of German rocket engineers still refers to rocket engine thrust in “tons” today. Soviet engineers also used “tons” to cite an engine’s thrust – you can still find this term in many recent books about the glorious past of various Soviet rocket institutions. I would not be surprised if they adapted the use of “tons” from their German mentors after the war.

    2. “The new engine does not have either vernier nozzles or jet vanes, suggesting the entire engine may be gimballed for steering purposes.”
    I am not aware of early static test runs of new engines that already involve the complete steering pack. If you ignite a new engine for the first time, you want to get it up and running, and you expect it to blow up. You do not add levels at complexity like steering elements at early tests. Therefore, would not agree that the lack of these elements suggests something.

    3. “This does not appear to be a copy of any Russian engine”
    You may be sure, but I think that we cannot know yet. There are indeed Soviet engines with the turbopump at the side. And Isaev alone developed much more than 130 engines, with 100 going into serial production. Not many people actually know all Soviet engines ever developed or produced.

    4. “North Korean practice so far has been to use four engines”
    The number of engines used depends on more factors, one major being the access to engines. Think of the Soviet N1 moon rocket – they had to opt for 30 engines because they had no other choice. Even though I agree that four engines seem sensible for the expected rocket, we should just wait and see.

    George, Clay,

    I agree with Clay that you can estimate the engine thrust with simple measures. All engines work along the same physical laws, which also means that they have to adhere to the same engineering principles, as Clay already pointed out. Clay, my estimate of the nozzle diameter is a little smaller, meaning something around 90 cm. I also think the chamber pressure is a little lower than you expect. Therefore, my estimate comes up at only 60 to 65 tons of thrust. It is hard to estimate the actual pressure, but it is very important to figure out if the real nozzle diameter. As you noticed, your link to the nozzle pic is not working. Could you post another link, or send me the pic?

  6. Clay Ramsey says:

    Curious, the link below *is* good, just not from this website…

  7. Clay Ramsey says:

    Oh no! Link broken. It’s the image of a bunch of guys standing around the back of a KN-08. I’m sure you can still find that image somewhere.

  8. Clay Ramsey says:

    George; in a way there is. Note the image in the link. The paragraph that follows it is from a comment posted earlier regarding the twin nozzle test several months back.×680.png

    “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.”

    I was speculating that those nozzles were SA-2 nozzles so that ‘verified’ the diameter to near 23 inches. From the twin nozzle test, I compared the exhaust jet diameter to two large sections of the fence railing in the foreground then compared that to the exhaust jet of the new single nozzle test. I got 37.72 inches and wondered if it was a 1 meter nozzle (but used the 37.72 number anyway).

  9. George William Herbert says:

    Clay, what are you measuring the nozzle exit against? It’s not like they’re holding a meter stick up against it…

  10. Clay Ramsey says:

    George; yes, you can come very close just by inspection. The only real unknown is the throat diameter, and that measurement is way less than an ‘order of magnitude’ uncertain. I’m sticking with the 80 ton (160,000 ish) value for the nozzle test image, and even more for the full engine image (assuming they are the same nozzle, I only did a rough measurement on it) since that one *is* operating at full design combustion chamber pressure and is not over expanded.

    As for being unsubstantiated, all my observations were made using the images provided by *this* website. They can be duplicated by anyone who wishes to check out what I was saying. The turbopump exit thing is on the same level as the ‘disco ball’, I just don’t think it would be made that way if anyone were to actually try and make one for real. They would look at the turbine wheel and say ‘that just doesn’t look right, let’s put an elbow there instead’. Does that mean it’s impossible? No, it probably would work like that, just seems odd to me. Also my same argument from the past still applies here, If pictures are being released to show us that this engine exists in flyable form, why not release videos or images that make it hard (impossible) to deny by showing a clear full duration run.

  11. George William Herbert says:

    Clay, with all due respect, there is more guesswork involved in trying to tell how big the engine is than supports more than at best an order of magnitude and one significant digit on thrust.

    All that’s really unambiguous (assuming as prelim analysis showed little photoshopping and no deceptive changes) is the general scale and that the exhaust behavior comes close to sea level expansion optimized.

    Everything else would rely on extrapolating a ridiculous amount and camera foreshortening assumptions unprovable from raw data.

    It’s not INconsistent with 80 tons thrust, and both NK statements and US State Dpt notices disclose such a suspected project. But you’ve got a lot of unsupported specificity there.

  12. Clay Ramsey says:

    Well, the nozzle image and the engine image are not of the same thing, and the description given for the nozzle image a few days ago was stated to be testing of the combustion chamber configuration and valves, consistent with the image. Don’t be too fast to so completely connect the two. If there is a statement regarding this complete engine, I missed it.

    You can make pretty accurate thrust calculations just by measuring the nozzle exit and making a wild ass guess at the expansion ratio. I got 160,120 lbs of thrust and subtracted a bit for the over expansion and then added the same amount because it seemed to be a little higher pressure than typical (compared to XLR87-AJ7, which I used for comparison). So yeah, 80 ‘tons’ is totally believable.

    Here’s a better image of the complete engine:

    Wow, is that thing real or not? Impressive if it is, and even impressive in a way if it isn’t. Would like to see the plumbing on the other side. Turbopump exhaust layout is kind of odd looking, and small too. Looks more like a pump discharge collector. Never seen anything like that before but who knows, nothing says it wouldn’t work, just kind of restrictive and not in the natural flow of the turbine gas. The whole thing doesn’t seem to be very hot, but the parts that are glowing red are the combustion chamber and bits of the turbopump (neither of which you would expect). But it is super clean, nothing is smoking from the heat or any vent ports or anything. I really don’t want to bad mouth it because it’s really a pretty cool looking engine and I can’t ‘prove’ it isn’t real, but it may be a very short run for demonstration rather than something ready for flight. A video of this one running would be *very* interesting.

Leave a Reply

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