George

George was built to develop my experience with model rocketry after not having built or flown a rocket in over a decade.

Description

George is an Estes U.S Army Patriot M-104 kit designed to fly on inexpensive B and C motors. The Estes kit is based on the MIM-104 “Patriot” PAC-2 GEM surface-to-air missile designed and built by Raytheon. This particular kit was chosen mainly because it looked cool, but also because it seemed easy to build and was based on a well-tested, known-stable prototype rocket.

The kit was purchased from Amazon for $14.54, tax included. There were no significant additional costs associated with assembly, aside from the expected purchase of glue and sandpaper.

The name George is the result of a hasty decision made on launch day when I discovered that I would have to choose a name for each of the two rockets I would be launching. I was rushed and couldn’t think of anything better on the spot, so I went with George and Frank.

Details

All measurements are from the finished rocket.

• Height: 54 cm
• Body Diameter: 4.1 cm
• 4 fins
• 1 stage
• Width with Fins: 11.2 cm
• Paint: 1 coat white primer
• 1/8" launch lug
• Hexagonal parachute, 36.8 cm on the long axis, 22 cm on the short (side length 18.6 cm), with a 5.7 cm diameter hole in the middle. Total surface area 797 cm².

Construction

George was built during August of 2022. Construction was straightforward, and no significant modifications were made to the kit.

The fins were first sanded smooth on their large, lateral surface, then the leading and trailing edges were rounded to a semicircular profile. This profile was chosen solely because it seemed more aerodynamic; I did not spend a lot of time researching the shape.

The fins were then positioned on the body tube according to the fin template provided by the kit. Small pencil marks were made on either side of the cutout, then extended out to the length of the tube by using a doorframe as a straightedge.

All parts were test fit before being glued in place.

The fins were attached to the body tube using superglue (cyanoacrylate) along the entire proximal edge of the fin. An Estes Fin Alignment Guide (part 2231) was used to try to maintain alignment, but there was quite a bit of wobble using this method. Instead, I printed Justblair’s parametric fin jig with 0.5 mm added to all measured widths and diameters to account for the quality of my admittedly low-end 3D printer. No filets were made between the fins and body.

The engine mount assembly had significant play between the centering rings and body tube, so, following in Apollo 13’s tradition of tearing off parts of the flight plan, 5 mm wide strips were cut off the assembly instructions and glued with wood glue around the circumference of the centering rings until the fit was snug. The motor mount was assembled and attached to the body tube with cyanoacrylate.

Flashing on the nose cone was removed first with a hobby knife, then the nose cone was sanded with 400, 800, and 1200 grit sandpaper until it was smooth. The nose cone fit snugly inside of the body tube without any modification.

The parachute, shock cord, and shock cord mount used were all stock.

The finished rocket was placed on a 3D printed stand for painting. Time constraints prior to initial flight allowed only one coat of white primer to be applied, which was sanded smooth after it had been given time to dry. The yellow color of the upper (forward) body tube did show through somewhat, but the coloration is not noticeable at a distance. Despite the intervening time, painting has not been completed to date.

Simulation

A model of the rocket was built in OpenRocket 23.09 after construction of the actual rocket had been completed. In retrospect, waiting until all parts were glued together to weigh everything was not the right way to do this, because this led to mass estimations being used for several components. It seems I was a little too eager to get everything assembled and in the air.

A model of the craft, shown above, was built in OpenRocket. The mass component discussed below is visible towards the aft of the rocket.

After weighing, a 12.4 g discrepancy was noted between the measured rocket (heavier) and the simulated craft in OpenRocket (lighter). Initially, I tried to correct this by adding a 12.4 g mass component to the model with the same length and diameter as the rocket itself. This, however, incorrectly assumed a uniform distribution of the excess mass, which was not the case. The excess is hypothesized to have come from adhesives (and the paper strips added to the engine centering rings), which are concentrated around the motor mount and fins, and paint, which is distributed more or less evenly across the craft. The mass component was sized and moved until mass measurements and the position of the center of gravity predicted by the model agreed with those measured from the completed rocket. The final mass component was 10 cm tall and placed with its aft end in line with the aft end of the craft. This placement is consistent with adhesive applied to the motor mount and fins, and indicates that the mass added by the single layer of primer borders on negligible.

Stability, measured in calibers, is listed below:

After the model had been tuned until it closely matched the actual rocket, it became possible to make flight predictions:

Variants

George-L

In order to fly at night, some kind of exterior lighting is required. Discount Rocketry had a trailer on site and was offering small, battery-powered LED clusters for sale at the November 2023 ROCStock. I bought a pair of two for somewhere around $4 (I don’t remember the exact price), applied a little superglue to the underside of each, and slapped them onto opposing sides of George, close to the CG. I did absolutely no simulation or modeling of the new aerodynamics, loaded it with a C6-5, and launched.

It flew perfectly on the first try, and thus George-L (George with Lights) was flight-certified.

The light clusters weigh a bit more than I thought they would, and represent an 11% increase in mass. Due to their

Simulation

As I said, I did no simulation on site. The flight appeared similar enough, but out of curiosity, I updated my craft file using two of OpenRocket’s rail button presets for the lights.

The change in values relative to George is listed as a percentage.

The decrease in apogee is quite substantial, but this makes sense given the blunt face of the lights. In the future, it would be beneficial to find some low-profile or recessed lights for night flights. Still George is a blast to fly, and exceedingly reliable.

Flights

Lessons Learned

Design and Simulation

• Once it had been tuned, the OpenRocket simulation proved to be quite accurate and helpful. In the future, weighing components as they are assembled ought to lead to a much more accurate model, and therefore a much more accurate simulation. Mass inconsistencies can be handled as they arise, and would not need to be guessed at.

Construction

• The Estes fin jig is unreliable, and there is significant potential for fin cant. This is a problem with the design and material of the jig itself, and seems intractable: the clips which hold the fins to the jig are made of plastic, and clamp the fin to the jig. This forces the fin slightly to one side. If the clamping force were to be increased, would dig into the fin. If the clamping force were decreased, the fins would wobble even more and would not be true to the alignment lines along the body. Estes struck the best balance they could between too high and too low clamping force, and the result is adequate for low power rocketry, but I would not be willing to try such a jig with a medium- or high-power rocket.
• Using a doorjamb to draw alignment lines is terribly inaccurate. In the future, it would be best to use a metal angle extrusion to draw these lines.

Flight

• In order to track the distance traveled downrange by the rocket, the locations of both the launch pad and landing site need to be marked with GPS. This should be relatively trivial; it’s just a matter of remembering to do it.
• Similarly, a ruler or tape measure should be brought along: you never know what you may need to measure.
  • Checklists for tools, supplies, and pre-launch checks need to be developed, tested, and used. It may be beneficial to have these checklists on paper.
• It is going to quickly become possible to lose track of rockets as they ascend. I have seen colored powder ejected with parachutes as a means of easily seeing both where and when a chute is ejected. This would be very helpful, and deserves research.
  • Similarly, colored and/or reflective streamers added to the parachute or shock cord would aid in tracking the craft during descent.
• Tracking marks, such as alternating bands of black and white (or some other pair of highly contrasting colors) should be painted on the rocket to aid in investigating orientation and rotation.
• It is difficult to both record a launch and to watch it at the same time, and the desire to simply stand back and watch is always paramount. Instead of doing half a good job at either one, it would be good to plan research launches, during which all aspects of the flight are tracked and recorded, and recreational launches, where only minimal notes and no video need to be taken.

Changelog

1.2.0 - 2024-02-25
Added
- Variants and George-L sections
- Writeups for launches 3 and 4

1.1.4 - 2024-02-20
Fixed
- Fixed a math error in ejection charge timing 

1.1.3 - 2024-02-12
Added
- Clarifying information to some captions.
- Added version numbers to all elements using simulated data.
Removed
- Took out the first page's hero placeholder image.

1.1.0 - 2024-02-02
Changed
- Split into multiple pages for easier reading and linking

1.0.1 - 2023-12-11
Added
- Changelog
Fixed
- A few typos

1.0.0 - 2023-12-02
Added
- Initial release