Loki boosted darts were developed by the US Army, beginning in 1946. They were supposed to be an anti-aircraft weapon, but they weren't accurate enough to hit anything, and therefore, five years and many tax dollars later, the project was canceled. However, the Loki dart lived on as a research rocket capable of throwing small payloads (like radar reflective chaff) to impressive altitudes, 100-150K feet.

Around 1970 the original 3" Loki motor was replaced by the Super Loki 4" motor for better payload capacity. Super Loki darts are still used today for collecting meteorological data. This information is all from Peter Always’s book "Rockets of the World," check it out for more background information and scale data on Loki rockets.

Basic Boosted Dart Theory:
A narrow un-powered dart rides on top of a larger diameter booster. At burnout, the difference in drag separates the dart from the booster. By keeping the booster light, most of the remaining mass, and therefore energy, is retained in the dart and it coasts to a much greater altitude than the booster.

Rocket Design:
I wasn't concerned with building an accurate scale model of the real Loki dart, I just wanted to try out the boosted dart principle. The smallest standard size tube that my BlackSky "AltAcc" accelerometer will fit into is 38 mm motor tube, so that set the diameter of the dart.

(Note: I had to remove the battery holder to get the AltAcc into a 38 mm tube. Also, I recently discovered that by grinding down the edges of the AltAcc's circuit board I can get it into a 29mm tube. I may try building a new 29mm dart when I have some more time.)

A 54 mm motor tube seemed the likely choice for the booster airframe. My home launch site has a 12,000’ waiver, an altitude easily obtained with a small K motor.

Booster Design Details:
The booster airframe is 54 mm phenolic tubing, with a single wrap of 6oz. fiberglass. I chose to use an Acme plastic fin can. The fin can is strong and did its job well, but there are two drawbacks.

First, the collar increased the diameter of the booster by about 1/4", to 2.5". Second, it's heavy, about 300g, and I’m sure that I could build a lighter fin can out of G10. However, the main reason that I stuck with the Acme can was that it was removable, and so with the fins removed, the booster would fit in my suitcase for the trip to LDRS 17. Removable fins also allowed different length motor tubes to be used.

The booster has an electronics bay and one parachute (the real Lokis do not recover the booster or the dart). These add weight to the booster, but are necessary as the booster is too large for a safe tumble recovery. Unlike the original Lokis, my booster is stable after dart separation.

For recovery, I started out with a 24" parachute, but later I reduced the size of the parachute to 18 inches, with a large spill hole (a PML drogue chute). With the 18" chute the decent rate was 50 fps, which is fast, but since the booster has a strong plastic fin can, it can survive a hard landing.

Dart Design Details:
The dart design is a modified "Cirrus Dart" from PML. The airframe is 38 mm phenolic tube with a single wrap of 6oz. fiberglass. The conical nose cone is solid urethane.

Because the dart has no motor, drag can be further reduced by using a full boattail. To make the boattail, I used a second urethane cone, making the dart pointy at both ends. The booster’s thrust is directed through the boattail, and I used several additional wraps of fiberglass around the airframe at the boattail junction for extra strength. The fins were cut from 0.063" G10 and are glassed to the airframe.

An AltAcc rides in a central section in the airframe for a standard "double-break" dual deployment setup. For maximum altitude, the optimal weight for the dart is varied depending on the motor used. Dart weight can be quite high for large motors, and so I needed a way to get a lot of ballast into a small space. Using loose lead shot or sinkers wastes a lot of space, so I melted some lead and poured it into a thin walled 1.5" aluminum tube.

Then I cut this into sections of different length with a hacksaw. Using this method you can get about 250g of ballast per inch of payload space. I used a payload section just behind the nose to hold the ballast, and this is section can be omitted for flights not requiring extra weight.

For recovery, a streamer and as much tracking chalk as possible is ejected from the tail section. After apogee ejection, with all that lead in the nose the dart falls, well, like a lump of lead.

Although the dart is capable of extreme altitudes, this really minimizes drift, but presented another problem; how to get a large enough chute and shock cord into the small airframe. I finally settled on a 30" hemispherical parachute, which gave a 17 fps decent rate when the dart was weighted to 1450g.

The parachute was connected with about 10 feet of ¼" tubular nylon. This has worked well so far, but it is difficult to pack it all into the small tube.

Dart to Booster Coupler Design Details:
The coupler between the booster and the dart is the most critical part of a boosted dart. The dart is long and heavy, with much of its weight forward. High thrust booster motors put lots of stress on the coupler. At one launch I witnessed, a full scale model of a Super Loki folded in half under boost due to a coupler failure, so I put extra effort into my coupler design and construction.

To avoid failure, the coupler must hold the dart in precise alignment with the thrust axis, and not allow any side to side wobble. At the same time it must allow a frictionless separation at booster burnout.

My final design has the coupler holding the dart using several close fitting plywood rings and a pilot pin that fits into a hole in the dart's boattail. The coupler was constructed so that in flight most of the thrust is transferred to the dart through the coupler’s aft bulkhead and the pilot pin.

The remaining bulkheads are designed primarily to provide lateral support. These bulkheads have differently sized holes to support the conical boattail at several intervals.

During construction, the boattail was wrapped with fine-grain sand paper and lapped into the holes so that they fit snugly and provided a maximum of contact area. At the aft end of the coupler, a pilot pin (¼" inch bolt with the head cut off) fits into a hole in end of the solid urethane boattail. To make the pilot pin hole, I cut off the boattail tip, and center drilled a 5/16" hole. I then pounded a section of brass tubing into the hole to provide extra strength, and to reduce friction with the pin.

This was the rocket’s maiden flight. Under J800 power, liftoff was quick (24.3 Gs), and at burnout the dart separated cleanly and then vanished. All of my flights have shown that the dart is hard to track visually, because it is moving fast and makes no tracking smoke. I never saw the dart’s tracking chalk, and while I was looking for the dart, I lost track of the booster. Luckily, I happened to be looking in the right direction when the dart's main chute deployed, and so I watched it touch down and happily headed off after it. When I arrived at the landing site, there was the booster about 10 yards away! The dart and booster were in excellent condition, not even a ding in the paint.

The AltAcc data shows a MECO velocity of 1050 fps, a little shy what RASP predicted, but pretty close. I had intended the motor ejection charge to be for back up of the AltAcc only, but the booster coasted for longer than expected, and the motor ejection fired slightly before apogee.

Flight #2
Date: 9/12/98
Location: Tripoli South Carolina’s Super Sod farm. Orangeburg, SC
Weather: Temp ~93° F, light surface winds ~5-10mph, clear skies
Launch Site Altitude: 200’
Launch Site Mach: 1153 fps
Motor: Aerotech J460-M
Booster Dry Weight: 0.9 Kg
Liftoff weight: 2.581 Kg
Dart weight: 0.846 Kg

Flight predictions:
RASP and Digitrack predicted about 8300’ for the dart with a MECO velocity of 977 fps. I prepped the dart without the ballast section and it weighed in at 846g, which was very close to Digitrack's estimated optimum weight for the dart.

Flight observations:
Dart Apogee: pAlt 7415' iAlt 7610'
Booster Apogee: pAlt 3789' iAlt @ejection 4291'

The boost was nice and vertical. At burnout, I saw the dart separate and then vanish. The medium delay I selected for the booster turned out to be way too short. I should have known this from the J800 launch, but I forgot. The booster ejected at high speed and there was a loud pop from the chute opening suddenly. At this point, it was too high to see clearly, and I was sure the chute had stripped. However, as the booster descended it became apparent that it was under canopy. When I picked it up I found that the chute had popped a couple of shroud lines.

The dart produced a large cloud of tracking chalk at apogee, and although I did not see it descend, I was again looking in the right direction to see the main chute open at 500 feet. The dart was recovered with no damage about 100 yards from the launch pad.

The best thing about this flight was that both AltAccs recorded good data. The moment of burnout is interesting, and the acceleration graph makes it very clear why the dart separates. After the dart separates, note how quickly the booster slows down. Also note the high velocity at the time of booster ejection. After this, the booster velocity in the graph is bogus. I have no idea what caused the second spike in the booster acceleration, maybe the chute was tangled for a while and then opened? The pressure data for the booster is very strange too.

One of the great things about a recording accelerometer is that you can use the coast phase acceleration data to measure drag. This means no more guessing at the Cd for flight simulations. Here’s a drag vs. velocity plot made from the flight data. It’s very noisy in the low velocity region, mostly due to the 8-bit resolution of the AltAcc data. However, this is less of a problem as the speed increases, and when the velocity is up over 0.5 Mach, the drag data is really nice. You’ll notice that there is no slow speed data for the booster in this plot. This is because the booster ejected early, at high speed, and scrambled the remaining flight data. From the plot you can see that the dart has a Cd of 0.35, and the booster’s drag is between 0.7 and 0.75 after the dart separates.

Flight Predictions:
This was an experimental launch and fellow North Carolinian, Mark Lloyd, was kind enough to offer me one of his experimental 38 mm J motors. Mark called his propellant "Diamondback Lite", as it was a modified version of Jim Mitchell’s (of Dynamic Propulsion Systems) famous "Diamondback" formula. I was pretty excited about the flight until one of Mark’s other motors over pressurized and turned one of his rockets into phenolic powder. After that I was nervous to say the least! Without knowing much about the motor’s thrust profile it wasn’t possible to run simulations, but by now the boosted dart had nine flights on it, and was characterized well enough that I knew I would not bust the waiver.

Flight Observations:
Dart Apogee: ????
Booster Apogee: pAlt 3683’ iAlt 3499’

I was all out of fingernails by the time Jim Conn pressed the go button, and then holy smoke did that puppy take off!! Only the high speed motor drive on my Nikon was able to catch the multiple shock diamonds chasing the rocket off the pad. What I saw was a flash of light, and an empty launch pad. Searching the sky, I didn’t see anything at all, but someone else was shouting that they had it in sight. It was quite windy that day so there was no question which direction to look in, and following the line pointed out to us, Mark and I set out after what we were told was the dart. After crossing more than a mile of sod, we caught up with what turned out to be the booster. The dart was never recovered.

The AltAcc data from the booster showed that the rocket had pulled 55Gs at liftoff and reached a MECO velocity of 752 fps. Pretty awesome performance from a six grain 38 mm motor! This flight really proved my coupler design too, if 55Gs didn’t fold that rocket in half, I think I’m pretty safe for most any other flight.

Using the MECO velocity and altitude from the AltAcc data, it is easy to calculate how high the dart went using basic physics. These types of calculations, including how to find the dart optimal mass were covered by John Boles in his article "Rocket flight performance mapping" (HPR November, 1997). Using constant air density I calculated dart apogee to be 7058’.

Another exciting thing you can do with a recording accelerometer is to calculate a motor thrust curve from the flight data. This is particularly valuable in a case like this flight, were the motor was an unknown quantity. The curve shows a total impulse of 658 N-sec (a fetal J) with an impressive thrust peak at over 1300 N. If the final tail of thrust is ignored, the average impulse would be over 800 N.

Flight #4
Date: 7/4/99
Location: Fourth of July Freedom Launch Orangeburg, SC
Weather: Temp ~100° F, surface wind 0-5 mph, clear skies
Launch Site Altitude: 200’
Launch Site Mach: 1159 fps
Motor: Aerotech J460
Booster Dry Weight: 1.3 Kg
Liftoff weight: 2.135 Kg
Dart weight: no dart

Flight Predictions:
After losing the dart, I decided to test the booster by itself. The idea was to weight the booster to match the total liftoff weight of the earlier J460 flight (2,581g). However, as I forgot to bring my lead weights to the launch and so I had to launch it about 450g too light. For this flight I replaced the coupler section with a small 54mm payload section and a nosecone. RASP predicted a Mach 1 flight reaching about 8,000 feet.

Flight Observations:
Booster Altitude: pAlt 6679’ iAlt 7076’

Looking back on this flight always gives me the chuckles. It was a hot day, and after I set the rocket up on the pad, I gave my flight card to the LCO and headed back to my sunshade. I find it much easier to track high velocity launches if there is more distance between me and the pads, and I wanted to listen for a "mach pop". When it came time to launch my rocket the LCO paused, but did not turn off the PA microphone he was holding. I (and everyone at the launch) could hear "Is this right? A J460 in that? That rocket isn’t three feet tall!" Well actually, it’s about 3 1/2 feet tall with the nose cone, but I don’t blame him for wondering about my sanity. After further discussion with the RSO, he finally went ahead and launched it, calling, "heads up!"

Zoom! A nice fast boost, and I was able to track it all the way to apogee. The rocket arced into the wind a bit. The AltAcc data showed that MECO velocity was 1149 fps or 99% of Mach. I find that this is pretty common, the rocket reaches the sound barrier, but doesn’t quite break through. The sound barrier is also apparent in the drag data, where the Cd increases as the rocket enters transonic velocity.

Conclusions and future plans:
I’ve had a lot of fun with this project so far, and I’m anxious to get back to flying boosted darts. I’m building two new darts, one will be 38 mm again, and the other will be 32 mm. The only significant change will be to reorganize the dart’s interior so that I can get my new Walston tracking transmitter in there.

The major thing that the flight data has shown me is that I’m not using large enough motors to really take advantage of the boosted dart principle. Heck, the control flight went almost as high as the dart did using a J460, so why am I bothering with the dart? Where is the added performance? Well, it comes down to the mass of the booster. Looking at flight #2, at burnout the motor has accelerated a total mass of 2,190g to 950 fps. However, the dart is only 38% of that mass, and all of the energy spent accelerating non-dart mass is wasted. For the J800 flight this ratio was slightly better, with the dart being 47% of the MECO mass, but still not too good.

There are two solutions to this problem. First, I can make the booster somewhat lighter by not using the Acme fin can. Possibly more weight can be shed by switching to a more exotic airframe material (like carbon fiber). I could also remove the AltAcc from the booster and just use motor ejection.

The second solution is to use a bigger motor (this is always my favorite solution). A bigger motor increases the boosted dart’s advantage over a conventional rocket by increasing the dart’s optimum mass, and therefore making the dart a larger fraction of the MECO mass. This means that a greater percentage of the motor’s energy will be contained in the dart, and so the performance bonus is greater. My flights show that when using a 38 mm dart, J motors can at best do only slightly better than a traditional 54 mm rocket. However, having crossed the break even point, I believe that there is much greater potential in the K and L motor impulse range. I have proven that my design can withstand large accelerations, and I think my current rocket has a good chance of staying together through a K1050 boost. My simulations indicate that I may be able to break 20,000’ with this motor. If that works, then who knows? Maybe a DPS L3900?

About the author:
Jeff Taylor has been flying HPR since 1997 when he certified at RATS V. He now makes his home in Raleigh, North Carolina and does the majority of his flying at the Whitakers, NC launch site. You can contact him by e-mail at This email address is being protected from spam bots, you need Javascript enabled to view it  or This email address is being protected from spam bots, you need Javascript enabled to view it .