South Korea’s Analysis of North Korea’s Rocket Debris

January 25, 2013
David Wright
Former contributor

South Korea recently released a short summary (in Korean) of its analysis of the pieces it recovered of North Korea’s first stage from its Dec. 11 satellite launch. Parts of the analysis have been showing up in the South Korean press in the last few days.

Here is a translation that I did of that summary (with the help of Google and a student who wants to remain anonymous). Corrections from readers are welcomed.

Much of what is in the analysis is already known from previous reports. The first stage of North Korea’s Unha rocket uses Scud propellant—a fuel that is essentially kerosene with an oxidizer of red fuming nitric acid. It is powered by a cluster of four Nodong engines. The body is made out of an aluminum-magnesium alloy and shows relatively crude construction.

Two findings are particularly interesting. The first is that the first stage was steered not by jet vanes like the Scud and Nodong; jet vanes stick into the rocket exhaust and turn to divert some of the thrust. Instead the analysis states the Unha first stage uses four small “auxiliary engines,” sometimes called “vernier engines,” that can turn to add thrust in different directions and thus change the direction of the thrust enough to steer the rocket. Signals from the rocket’s guidance system  could be used either to rotate jet vanes or these small steering engines.

Figure 1. Nodong missile with one of the four jet vanes circled.

The existence of these small engines is not a surprise, since they are what we have assumed North Korea used to power the Unha third stage, but no one knew they were used in the first stage.

Adding these auxiliary engines to the cluster of four Nodong engines makes the propulsion system more complicated. But this steering method is preferable since sticking jet vanes into the rocket exhaust reduces the thrust by a couple percent. Instead, the thrust of these auxiliary engines adds to the overall thrust of the first stage.

The analysis states that each of these steering engines provides 3 tons of thrust in addition to the 27 tons of thrust from each of the Nodong engines, giving a total thrust of 120 tons for the first stage. The analysis does not indicate what the “3-ton” figure is based on, but it seems too high for several reasons.

These engines appear to be very similar to the steering engines of the Soviet SSN-6 missile, and one possibility people have discussed is that North Korea and Iran obtained or copied the SSN-6 vernier engines. However, the SSN-6 used two of these engines for steering and they had a combined thrust of 3 tons, or 1.5 tons each.

Moreover, the SSN-6 engines were fueled with more advanced propellant (UDMH + NTO). In the Unha these steering engines instead use the kerosene + RFNA propellant combination used in the main engines. That would seem to drop the thrust below 1.5 tons.

One interesting question is whether the North Korean steering engines are able to use UDMH + NTO as well as kerosene + RFNA, since UDMH + NTO is the propellant frequently assumed to be used in the Unha third stage and the Iranian Safir second stage. If the North Korean engines can’t, then we need to rethink the propellant and performance of those stages.

In addition, Markus Schiller and Robert Schmucker looked at the diagrams in the analysis and saw another problem with assuming these engines each provide 3 tons of thrust. As Markus wrote in an email:

“According to the flow chart, the fuel for the verniers is diverted after the regulator. The regulator itself looks exactly like the standard Nodong/Shahab 3 engine’s fuel flow regulator. With the claimed 3 tons of vernier thrust, this would mean a diversion of 11% of the fuel flow. This seems a little too much to work without adjustments. However, for a vernier with less thrust, this might work.

“Also, if you have your Nodong engines running with standard mass flow, plus an additional 10 or 11% flow for the vernier engines, you would need 10% more propellant in the first stage to get the generally accepted 120 s of burn time. This would be 6 extra tons of propellant mass, and with that, it gets difficult to fit the propellants into the given tank volumes.”

For these reasons, it seems likely that the thrust of the steering engines is well below 3 tons each, and likely less than 1.5 tons each. It’s true that the thrust increases from the sea-level value as the rocket goes to higher altitude, but if the sea-level thrust was 1.5 tons it would only increase by about 20% to 1.8 tons as the rocket reached higher altitudes (see details below).

So the engines would add perhaps 5% to the thrust of the Nodong engines, assuming their presence doesn’t reduce the performance of the main engines. That appears to be 5 to 6% lower than the South Korean team assumed.

The second interesting result of the assessment is that a number of parts were purchased from abroad, especially various sensors and some of the electronics, which it notes are commercially available products. These were apparently identified by markings, etc. The assessment then states that it presumes the other parts of the missile were made by North Korea.

The most interesting question is whether or not the Nodong engines were built by North Korea. It would be nice to know more about the South Korean panel’s thinking on this. The absence of clear evidence that the Nodong engines were made elsewhere does not convince me that they were made in North Korea, although that is certainly possible. There is ongoing debate about whether North Korea bought these engines from Russia, bought production equipment and built them in North Korea, or reverse engineered them and built them in North Korea. I don’t see this analysis as shedding much light on this question.

After looking at the photos in the summary, Markus wrote:

“The main engines were manufactured in the typical Russian way, with the well-known soldered corrugated metal sheet design. This is the design that Iraq never successfully managed to implement, even though they tried hard. It is quite a demanding manufacturing process, and this does not really add up with the poor production quality of the tank.”

This doesn’t prove anything, of course. North Korea may have had considerable technical assistance in learning to build Nodong engines and may have applied quality control in that production process at a level they didn’t bother to apply to constructing the propellant tanks, which are said to show uneven welds and crude workmanship.

So if the South Korean analysts have concluded that the Nodong engines, turbo pumps, etc. are in fact built in North Korea it would be interesting to hear more of their thinking.

Two other interesting things in the analysis: The first stage has two antennas, called “flight termination antennas” in the analysis. People have wondered whether North Korea’s could terminate a rocket that started going haywire after launch, and this suggests they have that ability—at least with the first stage.

The rocket also has “acceleration” and “braking” motors that are used for stage separation. Six acceleration motors on the interstage section between stages 1 and 2 are used to boost stage 2, while four braking motors on stage 1 slow it slightly. This causes the two stages to separate enough that they do not run into each other.

The South Korean analysis and especially the photos are very interesting. Hopefully more details will eventually become public.

Some details on the auxiliary engines:

The figure below compares the picture of the auxiliary engines from the South Korean report (left) to similar small engines displayed by Iran (right). If they are being used for an upper stage, like the Unha third stage or the Safir third stage, the nozzles of these engines are extended (as shown in the center) to work better at high altitude. The debris  clearly shows the nozzles are not extended, and they appear to have a nozzle diameter of about 20 cm.

Figure 2. The vernier engines from the Unha (left) and those displayed by Iran (right). The red line on the right shows how much the nozzle in the center picture has been extended. (Source for Iranian engine photos:

The increase in engine thrust in going from sea level to high altitude is roughly the nozzle area times the change in atmospheric pressure. In this case, this gives an increase of about 0.3 tons of thrust for each auxiliary engine, which is about a 20% increase if the sea level thrust is 1.5 tons.