Altitude Optimised Model Rocket

Model rocketeers often strive for their creations to perform well in a certain category, be it egg lifting (where you must design a model rocket to lift and recover an egg to the highest altitude), duration flights (where your rocket must stay in the air as long as possible), or best altitude. In need of a personal project, and with a passion for 3d printing, CAD design and physics, I set about designing the most altitude optimised model rocket I could, give the constraint of a D-Class model rocket motor. The aim of this project was to evaluate the practical usage of 3d printing in performance model rocketry, in comparison with existing manufacturing techniques. This instructable will detail my design process. I have conducted my tests in line with the scientific method, and attempted to explain my train of though as clearly as possible. I attend Magdalen College School in Oxford, and thanks must be given to this school for all the help it has given me in the development of this project, both through the loaning of a force plate, and the helpful advice of one of my physics teachers.

This investigation aims to evaluate the impact of additive manufacturing technology on enhancing performance of model rocket designs. The performance of existing rockets will be reviewed and both research and testing will be performed to scientifically iterate improvements to a 3D printed design to match or beat performance (against defined success criteria) of existing models. The data of these models will be lifted from the UK rocketry associations record tables.

Success Criteria

This investigation quantitatively considered the success of each design against these success factors:

1.     Altitude

2.     Velocity

3.     Acceleration

4.     Affordability

To be able to assess points 1,2 and 3, sensory components required addition to the design to gather data. As the rocket was designed and built during this project, a bespoke solution to coordinate sensory components was developed.

Test Controls

1.     The final rocket’s performance in terms of acceleration, speed, and height, collected from in-flight was compared against a computer simulation, to validate the results and assess the data accuracy.

2.      All designs tested had the same mass of altimeter.

3.     The engines used for the test were thrust-tested, to ensure the exact specifications were known. The same engine variant was used for all models.

4.     Only the body shape, materials, and construction methods were iteratively changed to ensure fair results comparison.  

5.     The same success criteria were used to assess commercially available rocket kits using standard components. 

A test on the motor was run to extract its thrust curve, useful for simulations and results analysis. The motor was fired onto a force-plate, attached securely to an aluminium extrusion test stand and held in place with a 3D printed bracket as shown in the photograph. A maximum force of 21.97 Newtons was measured. 

This must record data from two sensors; an accelerometer which measures acceleration in three axes and an altimeter. It must:

1.     Interface with these sensors over a communications protocol (I2C Protocol).

2.     Store the data on an SD card, which uses a different protocol (SPI Protocol)

3.     Have on-board power management to ensure individual boards are each receiving the correct voltage.

The Arduino nano circuit board was selected as it met these requirements (including plentiful coding support documentation) and low price. 

The sensors

1.     A pressure altimeter was selected (BMP180), with a refresh rate of 30Hz and accuracy +- 0.5m.

2.     An MPU6050 3-Axis Accelerometer was selected, due to low cost, reasonable accuracy (+-2G). This also came with an onboard gyroscope (to detect the rocket’s attitude), allowing further performance analysis. 

The Battery

Requirements:

1.  lightweight,

2.  compact,

3.  impact resistant.

4.  Must supply a minimum of 7v to the Arduino’s voltage in pin to power it.

A 2 cell, Small LiPo battery (typically used for remote-controlled cars) was selected and delivered a voltage of 7.4v in addition to meeting the other requirements 

The Program

This was then tested for component function and interfaced with a program, written in Python which did the following:

1.  Activated and tested the sensors

2.  Indicated a successful test by lighting a green LED

3.  Waited 40 seconds (to enable the launch technician to achieve a safe distance from the launch site)

4.  Read a baseline temperature and pressure

5.  Recorded acceleration, pressure and temperature and attitude every 10^th of a second

6.  Saved data to the SD card

Altitude is calculated from this information, which was done post-recovery, this saved processing power in-flight, enabling a high frequency of data collection. 

Improvements

After a flight trial on the version 1 rocket, it was found this design was too bulky and heavy for use in a performance model. To combat this, the accelerometer and SD card module were removed, and altitude data was written to the 500 bytes of internal EEPROM memory inside the arduino's processor. This meant 5 altitude readings per second were recorded over a period of 50 seconds, with each altitude reading consisting of two bytes. The program and board were updated. The removal of the SD card module in particular enabled the overall design to be far slimmer. Acceleration was now derived from the change in rate of the altitude readings.

With all the prerequisite controls and hardware laid out, the design of the V1 rocket could begin. This rocket was deliberately big and unrefined so I could practise construction techniques as I was still unfamiliar with the use of 3D printers and CAD. Firstly, our limitations had to be understood, starting with;

The Engine

Motors come as a pre-built unit, with a set impulse and thrust. The most powerful, easily accessible engine in the UK is an Estes D-12 motor. A D-12 motor has an impulse of 24Ns and an average thrust of 12N, meaning it produces thrust for roughly 2 seconds. This motor has a recommended take-off weight of around 300 grams. This is the main reason for its selection, as it has a wide margin, so a heavy, unoptimized rocket is supported.

Body Tube

The body tube was made from a cardboard tube as this was readily available, easy to integrate with 3D printed nose cone and fin sections, was the correct width for the data logger and crucial to one of the success criteria was also inexpensive and lightweight. 

With these two limitations in mind, a preliminary design, with traditional construction was designed in OpenRocket - a free open source rocket simulation tool. Using the values for the chosen body tube, and generating some 3D printable component designs of fins and nose cones, a rough working model of the V1 rocket was generated. The tool simulates the max altitude and acceleration using parameters such as tube shape, engine specification, nosecone profile and fin shape, producing the design shown in the photograph. Having verified this design would work, I hopped into fusion 360 to draw up some designs for the nosecone, adapters and fin sections which would require 3d printing.

The CAD model was developed to the same dimensions produced from the Open Rocket simulation. To save mass, the entire lower stage was designed as one component rather than the numerous traditionally discrete fins, couplers and body tubes presentonother designs. I also designed the fins to extend from the outer skin all the way through to the motor mount, enabling the fins to provide the bulk of the structure and for the core of the component to be hollow. this too is shown in the photographs in a cutaway style. Fusion 360's built in slicer was used to generate the gcode for the 3d printer.

Some advantages of 3d printing as a construction method were found to be;

With similar dimensions to existing models, the manufacturing method was shown to withstand crashes better and remove any possibility of assembly error. Provide prototyping rapidly and produced fins thinner than possible with balsa equivalents. The flexibility of this construction method also proved useful, as the nosecone could be made to any shape or width, and have a built-in coupler. The datalogger readouts are shown below. These will later be used to compare between designs.

As can be seen from the graphs above, the data was very similar to the predicted results. I am therefore happy that the 3d printer was able to produce parts very similar to those simulated by open rocket, at least in mass terms, as the rocket still accelerated as predicted. In terms of the quality of parts, the surface and poor chamfer quality meant the rocket was far less aerodynamically optimised than simulated, and the final altitude was lower. This meant more time needed to be put into the prints, as by reducing the layer height, increasing the time of the print, better surface finishes can be achieved.

Lessons Learnt

§ Being a pressure altimeter, when the ejection charge fired, the altimeter sensed a far higher pressure, dropping the recorded altitude suddenly. This can be seen in Figure 2.7. A better bulkhead will be installed to prevent this in future.

 The accelerometer and gyroscope data were not useful in the comparison and so are redundant. Acceleration can be deduced from time and velocity data. Removal will reduce rocket body diameter.

 The fins required chamfering post-print. This was laborious and added inaccuracy.

 The short length of the body and heavy rear meant nose weight had to be added to maintain a stable flight, compromising efficiency.

 The total material cost for the rocket was £6.89 with £2.50 required to 3D print and construct the body, and £4.39 for miscellaneous hardware.

 A heavy base necessitated large fins to lower the centre of pressure; increasing drag.

To enable focus on the right design factors to improve rocket performance, this relationship has been formulated for this study:

Ec = Ep + Ek + El

Ec: Chemical Energy of the engine before flight

Ep: Gravitational Potential Energy

Ke: Kinetic Energy

El: Energy losses

This equation demonstrates how the overall expended chemical energy of the motor is equal to the gravitational energy, kinetic energy, and energy lost to inefficiencies, in the form of aerodynamic drag.

At the point of apogee (the point of peak altitude, which is what we are optimising for), the rocket is moving neither up nor down therefore at this moment, there is no KE as there is no movement and so KE in the equation can be disregarded.

To maximise GPE, therefore increasing altitude, EI must be minimised, therefore reducing drag from the launch system and aerodynamic drag will be significant features of improving performance. The equation for GPE is:

Ep = m.g.h

Ep = Gravitational Potential Energy

m = Mass of rocket

g = Acceleration dueto gravity

h = Altitude

It can now be concluded that to optimise the rocket’s design for altitude, it must be made as light as possible, and the inefficiencies made as small as possible. The component design will be individually optimised to ensure the model satisfies these aims.

For maximum height gain, drag should be minimal. However, drag caused by the fin shape and location on the rocket body is necessary to lower the centre of pressure below the centre of mass. This is a critical idea within stability, as if these centres are not in the correct place, the rocket will be unstable. Ideally, this fin drag should only be caused when the rocket has an angle of attack greater than zero, meaning that when in flight with zero Angle of Attack (AoA), drag is minimal. The reason for this, is that ift the rocket has no AoA, it does not need stabilisation, and thus the fins should produced minimal drag. However, as soon as the rocket has AoA, it requires the restorative drag force from the fins to stabilise it.

Fin Shape

The pressure drag is derived from the length and thickness, bluntness of the tip and tail, and surface roughness. According to a report by Tim Van Milligan; by reducing the length of the fins, making them thinner, creating a symmetrical aerofoil plan and selecting tapering fins, the drag can be reduced hugely. (Milligan, 2017). The report detailed tests on several fin shapes with the same span and thickness, placed in a virtual wind tunnel with 100m/s of wind speed, with results shown in the attached photo.

As already established, low drag at 0 degrees AoA and the greatest difference between 0 and 5 degrees AoA gives the best design. While the elliptical fin best fulfilled these requirements, it proved too complicated to 3D print; as the support material meant an uneven finish. A clipped delta shape was therefore selected. 

The fin's cross section also matters. To determine the best shape, a test was devised;

Fin Cross Section Test

Hypothesis: The rearward bulge and symmetrical shape will deliver the smallest high-pressure region ahead of the fin and low pressure behind, as the airflow will be divided more gradually.

Methodology: A drag analysis tool ‘Fusion CFD’ was used to test these 3 fin cross-sections. The fins had their points of maximum thickness in various places, as shown in the image above, showing the 2D model of the fin's cross section:

Three designs were tested:

1.     Symmetrical, central bulge

2.     Symmetrical, rear bulge

3.     Symmetrical, front bulge

Control Variables: The velocities of air, and pressures were set to the same as the apogee fin shape test. A frontal airspeed of 100m/s was applied to 3 fins: each with different cross-sections but the same span, thickness, and root chord.

Results: As shown in the photos, the results indicated that a foil with a central bulge produced the smallest low-pressure zone behind the fin, and the smallest high-pressure zone ahead of the fin. The bulges were placed 75% of the way along each fin.

1.     Symmetrical, central bulge: 3.83 Newtons drag force

2.     Symmetrical, rear bulge: 4.51 Newtons drag force

3.     Symmetrical, front bulge: 4.91 Newton drag force

Analysis: The best shape for the fins would be a symmetrical, centrally bulging Shape 1. This is because:

  It has the lowest drag force

  A smaller orange section ahead of the leading edge of the fin (high pressure) is shown on this cross-section meaning that it is causing less compression ahead of the fin, reducing opposition to movement

  A smaller blue section (low pressure) behind the trailing edge of the fin creates less vacuum, reducing the suction on the trailing edge of the fin and less distortion to the airflow

Conclusion: The hypothesis was disproved. While the fin did indeed produce the least high-pressure section on the leading edge, the very fast re-converging of the airflow at the trailing edge produced a very large low-pressure region, sucking the fin back, and removing the advantages of the long leading edge. Shape 1 will be used on Rocket V2. 

Some other points of interest during fin design:

Interference drag between the body tube and model is also large, without a fillet; air flows in the harsh corner channel and becomes turbulent. Approximations have been made that a fillet of diameter of roughly 4% of the root chord of the fin provides the greatest reduction in fin-body drag without beginning to increase fin drag.

The final mass was reduced as much as possible, by making the fins as hollow as possible. A 15% infill was used.

The fins were made as thin as possible, to reduce their frontal area. This minimised the fin drag at 0 AoA.

The greater the surface area that is exposed to the flow of air, the greater the skin friction. The purpose of this component is to contain the parachute and motor as compactly and lightweight as possible. By reducing the frontal cross-section, the drag coefficient for the entire rocket is dramatically reduced, as less air must be displaced. Therefore, To optimise this component, it must essentially be made as short and thin as possible. The component was thinned to a mere 25 mm diameter. This was only 1 mm more than the outer dimeter of the motor. It was therefore as thin as possible. The length was minimised, but could have been shorter. The reason for this preservation of length is to bring the centre of mass up the rocket. Because all of the weighty components are stored at the top of the body tube, the decision was made to lengthen it a bit. This brought the centre of mass up. This meant smaller fins were necessary to bring the centre of pressure down, as the rocket would still have the centres in the correct orientation if the centre of pressure was allowed to raise a bit and as a result, overall drag decreased. To reduce the mass of the rocket, the body tube, motor mount and fin section was now combined into one component, with the nosecone being the opening for the parachute, rather than the rocket breaking open in the centre, as on the previous design. The electronics for altitude measurement had been reduced in size and were now stored inside the nosecone. A picture illustrating the ejection of the parachute has been attached.

The previous model rocket had used a launch lug attached to the rocket, which could slide up and down the launch rail. This provided the rocket with stability while it was getting up to speed. Since the rocket was now so thin, adding a lunch lug proved to upset the balance so much the rocket tipped over during flight, as the frontal area was so asymmetric. Therefore, the lug was removed and replaced with a launch tube system. the rocket slid up the centre, as illustrated in the photo. This tube was also designed in fusion 360.

The shape of the nosecone was arrived at through online research. It was understood that for subsonic flight, rounded ogive and paraboloid/elliptical shapes have the least drag, through examination of other performance models. To determine which should be used, Fusion CFD was again used, and the tended toward parabolic. This was then used in the final OpenRocket file, and the parameters of the shape adjusted. It was found that a longer nosecone lowered the drag coefficient. I assume this was because the air was divided more gradually. The length was increased until the mass started to increase enough to lower the predicted altitude. The final dimensions produced a 10cm long nosecone. The width of the part's walls were made a mere 0.4mm; only 1 layer thick on the 3D printer. This meant the entire component was a mere 8 grams.

These results are very similar. It is believed that the poorer altitude performance is due to some remaining surface roughness on the nose that was not accurately simulated. The nosecone proved tricky to sand, as to make it light enough to compete with its commercially available injection moulded counterparts, the component was only 1 layer thick (0.4mm). This resulted in a flexible construction, but with a tendency to break along the layer lines. This caused difficulties in sanding. The superior maximum speed may be caused by a defect in the engine, leading to longer burn time, as all other factors were the same as the simulation.

Lessons Learnt:

The lightening of the overall design, improvements in aerodynamic efficiency and removal of any launch rail drag meant more energy was transferred to the Ep store.

The light airframe resulted in immensely fast acceleration. This would not be ideal for a long flight, as greater speeds mean greater drag, but with short-burn motors such as these, this is ideal.

The very small amount of plastic necessary to create this design meant it cost very little to produce, and was a very fast print; the rocket was constructed from start to finish in 6 hours, with only 30 minutes of times from the constructor in sanding and support material removal.

The need for a 1-layer thick nosecone to reduce mass, meant that in the sanding and sealing process, it broke frequently. This required sealing with superglue. The component did survive flight, but then broke upon the fast landing.

While the current altitude record are still roughly 40 metres greater than my own (UK Rocketry Altitude Records | UKRA - United Kingdom Rocketry Association) for the D12 class of motor, the mass of these competition rockets is considerably less. I believe the relatively high mass (25 grams) of my own homebuilt datalogger is making the rocket heavier than necessary. Indeed the pico alt LO3 used by these record holders weighs in at only 10 grams. Overall, therefore, I conclude from this investigation that 3d printing technology, when applied in a performance manner to model rocketry is more than capable of holding it's own, and provides numerous construction advantages.

There are still a few areas where the commercially available product outperformed the final design however:

  These commercial products fewer specialised tools to manufacture and require little knowledge of aerodynamics and optimisation techniques to construct. The home rocketeer would undoubtedly find it easier to purchase a kit.

  The V2 rocket’s recovery system had reliability issues.

  Traditional balsa fins allow for more adventurous fin shapes, while retaining surface smoothness, as 3D printing requires support material.

In conclusion, 3D printing enables the production of parts able to compete with commercial products, and allows for easier optimisation of components, while removing margins for error during construction. The flexibility of manufacture offered by 3D printing, and the integration of many discrete components into one – such as the homogenisation of the fins, centring rings and lower stage, results in simpler components, lighter assembly and easier construction.