See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/366278128

Arduino Rocket Flight Computer

Conference Paper · December 2022

DOI: 10.1109/PACET56979.2022.9976324

CITATIONS READS

0 2,132

2 authors, including:

Petros S. Bithas
National and Kapodistrian University of Athens
137 PUBLICATIONS 2,765 CITATIONS

SEE PROFILE

All content following this page was uploaded by Petros S. Bithas on 27 October 2023.

The user has requested enhancement of the downloaded file.

 Arduino Rocket Flight Computer
Philippos Moschidis, Petros S. Bithas,
National and Kapodistrian University of Athens, National and Kapodistrian University of Athens,
Aircraft Engineering Department, Digital Industry Technologies Department,
Psachna, Greece, Psachna, Greece,
[email protected] [email protected]
Abstract- This paper describes the construction of a mechanism, or process that allows the rocket to return safely
flight computer, using a small model rocket, which is built to the ground [4]. To achieve that the mission profile is
with materials that can be easily obtained. In this context, a separated into a few steps.
comparison of the flight data with those of the simulation
has been performed. Based on this framework, an To start, the launch happens when the engines fire up,
introduction to the flight computer will be given and then its 2.1 seconds later there’s the main engine cutoff that’s when
construction together with all its sensors and its the engine stops burning. From that point, the rocket is
programming will be described. coasting for 5 seconds until it reaches apogee and the stage
separation begins. More specifically, the nose is separated
Moreover, a detailed analysis of the rocket used will be from the rest of the body of the rocket with the help of a
given. In particular, the design concept will be presented, small engine fire-up. A few seconds later, the parachute is
with the help of a computer program (OpenRocket), in deployed and after that, the rocket reduces speed until it
conjunction with the materials used, the procedure followed reaches the ground and performs a smooth landing.
for its construction, and finally, the test launches carried out
up to the final launch. Entering the last chapter, the flight In another paper [5] the goal was to compare some
data of the simulation will be presented, followed by the commercial and self-made solid rocket motors. In particular,
data of the launch that was carried out and analysed using a the paper highlights that the Klima D9 engine (the same
numerical computational and programming language. engine used in the proposed rocket) underperformed in
Finally, a comparison between the two data sets will be terms of total impulse and maximum thrust, differing on
made. average from the nominal values of 6.8% and 16.5%,
respectively, while they overperformed in terms of average
Keywords- Arduino, Flight Computer, Matlab, thrust by 4.1%, and showed good agreement on the average
Microcontroller, Project. burning time. Explaining some of the results that are present
later on. In similar investigations in the past, the scope was
I. INTRODUCTION to construct a small-sized model rocket and an introduction
Model rocketry is an activity that involves designing, for the OpenRocket computer program was given[6].
constructing, and launching self-made rockets. Model Furthermore, another paper on the construction and
rockets vary greatly in size, shape, weight, and construction description of flight computers is [7]. In another project, the
from detailed scale models of professional rockets to coding and analysis of the flight computer were given in [8].
lightweight and highly finished competition models. Finally, in [9] the design, analysis, and test of a High-
Powered Model Rocket were described.
Thousands of rockets ranging from 10 cm high miniatures to
large models reaching altitudes over 10 km are launched The methodology presented in this paper can be used to
annually. Model rocket motors with thrusts from a few familiarise with amateur rocketry. The remainder of this
Newtons up to several kilo-Newtons are readily available. article is as follows. In Section II, we present the flight
This activity is relatively popular and is often cited as a computer and its sensors. In Section III, a brief intro for the
source of inspiration for children to become engineers and rocket is given, while in Section IV we describe the flights
scientists [1]. that were performed. In Sections V to VII, we present the
simulation and the experimental data and how they were
Amateur rocketry has been started in the 1950s when
analysed. Finally, in Section VIII, we will compare the
hobbyists wanted to experiment with their skill of building
simulation data with the experimental data and after that in
rockets. Designing, building, and firing self-made motors
Section IX, a conclusion of the paper will be given.
was, however, extremely dangerous, and the American
Rocket Society has estimated that about one in seven
amateur rocketeers during the time were injured in their
hobby. In 1958, the first commercially-built model rocket
motors became available. Having industrially-made,
reasonably-priced, and safe motors available, removed the
most dangerous aspect of amateur rocketry. This along with
strict guidelines for the design and launching of model
rockets formed the foundation for a safe and widespread
hobby [2][3].
The subject in this paper is the construction of a flight
computer and a model rocket. With the help of the flight
computer, the data from the launch were collected from an
SD card after the rocket had returned from the launch and
then the data were analysed on Matlab. However, in order to
perform this analysis, the rocket had to safely return from
the launch without any damage. In model rocketry, safely
retrieving a rocket after it has been launched is usually a
primary objective, in which the recovery system is a device,
Fig. 1. Model rocket and mission profile

979-8-3503-9958-5/22/$31.00 ©2022 IEEE Page 1

II. FLIGHT COMPUTER Arduino's 3.3V pin. There are also three LEDs, which
The purpose of the flight computer is to calculate and indicate which is the current mode of operation of the
collect the data that will be available during the flight. To do computer. The red LED indicates that the computer is
this, a computer is needed which can give commands to the connected to the sensors, if the LED remains on it means
sensors to collect and calculate data: temperature, pressure, that there is a problem with one connection. If everything is
speed, altitude, acceleration, orientation, and displacement. okay, the yellow LED turns on.
In this case, because the duration of the flight is short and
the rocket is small, an Arduino microprocessor is used.
Specifically the Arduino Nano is almost as powerful as the
Arduino Uno microprocessor. Additionally, the Arduino
hardware is easy for beginners and the IDE is simple to
understand. But the main reason for the selection of the
Arduino nano is it’s size, weight and low cost as it is one of
the smallest yet very powerful. For the assemble of the
PCB I mostly used a soldering iron to create solder lines and
to connect cables from the microcontroller to the sensors.

Fig. 3. Flight computer diagram

This means that the computer is collecting data. Finally,

when the push-button is pressed, the blue LED turns on and
Fig. 2. Flight computer the computer changes from the ''data logging'' state to ''stop
data logging and safe disconnect’'. The software presented
in [16] is used by the flight computer for data logging and is
The first sensor is the BMP 180, it was ready to use out written in C++ programming language and loaded into the
of the box, there was no need for fine-tuning. The sensor is Arduino Nano with the help of the Arduino IDE.
designed to measure barometric pressure from 30000 to
110000 Pa (9000 m to -500 m above sea level) and III. THE MODEL ROCKET
temperatures from -40°C to 85°C with an accuracy of The general design of the rocket consists mainly of two
±1°C[10][11]. The unit comes with an integrated ML6206 parts, the main body of the rocket and its nose. The rocket is
3.3V regulator so that it can be used with a 5V divided into two bodies as after launch and while landing,
microcontroller such as the Arduino Nano[12]. the nose is separated from the rest of the body to open the
The second sensor used is the MPU 6050 which consists parachute to reduce the landing speed. In the launches that
of a three-axis accelerometer and a three-axis gyroscope. were carried out, more than one rocket was used and some
This helps us to measure acceleration, velocity, orientation, improvements were made after the unsuccessful launches.
displacement, and other parameters related to the rocket's The engines used are the D9-5 from Klima [17]. These
motion. To test the MPU 6050, Jeff Rowberg's example have a total thrust of 20Ns, a thrust of 9N, a total thrust
code was used, which includes a nice visual demonstration duration of 2.1s, and a delay of 5s. The 5-second delay
of the IMU's output. To measure acceleration, the MPU6050 means how many seconds after the engines have fired there
uses its on-chip accelerometer with four programmable full- will be an additional burn of the engines to separate the nose
scale ranges of ±2g, ±4g, ±8g, and ±16g. For the rotation of the rocket and deploy the parachute. For the process just
measurement the MPU6050 measures the angular rotation described to work properly, the rocket body, nose cone, and
of the rotation axis of the vehicle with programmable full- parachute must be connected. This was achieved by
scale ranges of ±250°/s, ±500°/s, ±1000°/s, and ±2000°/s attaching a rubber band to the inner wall of the rocket, while
[13][14]. at the other end of the rubber band, a ring was fitted which
The microprocessor powers and coordinates the two is connected to the parachute as well as to the rocket’s nose.
sensors mentioned. To store the data during the flight, since The first rocket (Aurora V1) used in the first launches
they will not be displayed on a screen at that time, a Micro without the flight computer had a D9-5 engine, the weight
SD adapter will be needed to collect the data so that it can of the rocket was 0.322kg. As shown in the figure below, the
be analysed offline. overall length of the rocket is 76 cm and its diameter is
0.68m. In the lower left of Figure 5 the total height (112m)
III. FLIGHT COMPUTER DIAGRAM
As seen in Figure 3, all the flight computer components
are connected to the Arduino, which powers them and also
determines the operation of each sensor, while the Arduino
itself is powered by three 3.3V batteries. The BMP 180 and
MPU 6050 sensors communicate with the Arduino using the
I2C protocol (it is a bus interface connection protocol
incorporated into devices for serial communication [15]), so Fig. 4. Aurora V1 (First model rocket) without the flight computer
the SCL and SDA pins of the two sensors are connected to
each other and from there a cable goes to the A5 and A4
pins respectively, while both sensors are powered by the

Page 2
Fig. 5. Aurora V1 with the flight computer
that the rocket would reach, the maximum velocity (38.5m/ The sixth launch, which happened on the 27th of
s) and the maximum acceleration (65.5m/s) are shown. January 2022, achieved a satisfying altitude but the
Four trapezoidal blades (0.09m root chord, 0.075m tip parachute did not open, fortunately, the rocket was not
chord, 0.06m height, and 45° arrow angle) were used at the seriously damaged so the repair was quick and simple. The
base of the rocket and the engine was stabilised with the seventh flight (10 February 2022) was successful, the
help of four rings. parachute deployed at the correct altitude resulting in an
extremely smooth landing. However, a problem with the SD
The installation of the flight computer inside the rocket card resulted in no data being recorded. The eighth launch
increased the total weight from 0.322kg to 0.421kg. This was the last and took place on the 11th of February 2022.
resulted in the maximum speed dropping to 26.1m/s and the There was no problem during the launch, but unfortunately,
maximum acceleration to 47.8m/s, while the maximum the parachute was late to deploy resulting in a high landing
altitude would be 65.9m meaning that the margin to deploy speed. This did not prevent successful data collection.
the parachute is minimal. Since it is not possible to reduce
the total weight of the rocket, the next solution is to use two V. LAUNCH DATA
engines in the rocket to achieve a higher altitude for the The flight computer can collect the data thanks to its two
parachute to open. sensors. The launch data are time, temperature pressure
which is understandable, but the other values, namely ax, ay,
and az (acceleration on each axis) and rx, ry, and rz (Euler
angles indicating the angular position of the sensor [17]) are
in raw form and it is not easy to draw a conclusion so they
will be analysed in the following chapter. Some of the data
collected during the flight are presented in the following
table.
Fig. 6. Aurora V2 (First model rocket) with the flight computer
VI. LAUNCH SIMULATION
Following this conclusion, the second rocket was built With OpenRocket (a fully featured model rocket
under the name Aurora V2. The only difference in the simulator that allows you to design and simulate your
design of the rocket is that a second engine was added. The rockets before you build and fly them) [18] at our disposal,
height and diameter remained the same. Looking at Figure it is possible to simulate the launch and create diagrams.
6, we can see that the total altitude was increased to 228m There is a wide range of diagrams to choose from, but since
(an ideal value for the parachute to deploy and reduce the the flight computer on the rocket calculates altitude,
landing speed as much as possible), the maximum speed is temperature, pressure, and velocity, we will limit ourselves
63.4m/s and the maximum acceleration is 99.2m/s2. to these. It is also possible to select weather variables and
IV. MODEL ROCKET LAUNCHES geographic coordinates so that the results obtained will be
more accurate.
In total there were eight launches, the first five with the
Aurora V1 Rocket. In the first three launches, the flight
computer was not inside the rocket as the aim was to check
that all the components that had been installed would work
properly.
The first flight, which took place on the 21st of October
2021, had some stabilising issues. 8 days later the second
launch took place. The launch started correctly but quite
early on, the rocket took a sideways inclination (due to
wind), resulting in insufficient time for the nose to
disengage and deploy the parachute. After the new rocket
nose was built, the first successful flight followed on the
25th of November 2021. The launch went exactly as
planned and during the landing, the nose was disconnected,
the parachute deployed and the rocket returned at a
relatively low speed.
Once everything worked as planned, the next step was to Fig. 7. Altitude, Temperature, Total velocity, Total acceleration and Air pressure
place the flight computer inside the rocket to record the
data. To achieve this, a hole was made in the rocket so that VII. LAUNCH DATA ANALYSIS
the computer could fit inside and was stabilised inside its
body with a fibrous insulating material (polypropylene). The flight computer can collect the data thanks to its two
sensors. Some of the data like time, temperature, and
At the fourth (31 December 2021) and fifth (6 January pressure are clear, but the other values, namely ax, ay, and
2022) launches it appeared that the problem of the rocket az (acceleration on each axis) and rx, ry, and rz (Euler
during the flight was the low altitude and the wrong weight angles) are in raw form and it is not easy to conclude so they
distribution, we decided to build a new rocket with two were analysed with the help of Matlab [19][20]. An
engines. This new design required more space inside the important data that the flight computer did not record is the
rings to accommodate two engines. To solve the problem of altitude at which it was launched. Using the data from the
the centre of gravity the flight computer was placed at a air pressure sensor on the BMP 180, we will determine the
lower point inside the rocket and three 3.3V clock type approximate altitude of the flight computer using the
batteries connected in series replaced the 9V battery used in relationship between air pressure and altitude (in Paschal's
the first flights (less weight). and meters respectively [21].

Page 3
 Looking at the altitude plot in diagram 8, we see that the However, afterward, there is a lot of turbulence due to the
rocket reached almost 160 meters at close to 50ts (note that fact that the rocket has finished burning the engines and
50 is not the seconds of flight but the fiftieth data record, a starts to return to the ground. As the parachute has not yet
timestamp is about 0.16 seconds). Then, it appears to be a opened, the rocket has several turbulences in all directions
sharp decrease in height, while near 120ts the reduction is up to the landing point.
smoothed out. This is because the parachute did not open
immediately so the rocket was free-falling.

Fig. 11. Axle movement diagram

Fig. 8. Altitude diagram The last graph (diagram 12) shows the speed. Up to 20ts,
we see that there is a continuous increase in speed as the
In diagram 9 the temperature is shown, at 60ts there is a engines are used up to that point. Then up to 40ts, there is a
sharp increase in temperature, which shows that at this point continuous decrease in speed because at this point the rocket
the ejection charge phenomenon is observed, i.e. the nose of reaches its maximum altitude. There is then a further
the rocket opens to deploy the parachute. While the pressure increase in speed (up to 90ts) as the rocket falls from the
plot (diagram 10) is exactly the opposite of the altitude plot, maximum altitude it has reached and begins its return to the
since as the pressure decreases, the altitude increases. ground. Finally, from 170ts until the rocket touches the
ground there is a sharp decrease in speed which occurs
because the parachute has opened.

Fig. 9. Temperature diagram

Utilising the raw values of the MPU6050, the rocket's Fig. 12. Speed diagram
velocity, as well as the motion in the x, y and z axes, can be
analysed. The way the flight computer is mounted inside the
VIII. DATA COMPARISON
rocket, the x-axis is the up-down motion, the y-axis is the
right-left motion and the z-axis is the in-out motion. Now that we have the plots from the simulated launch
Looking at the three motion diagrams (diagram 11), we can (right) and the actual launch (left), we can compare the
see that up to 60ts the values of all three axes are constant. values. Starting with the altitude (diagram 13), we note that
initially in both diagrams there is a sharp increase in
altitude, but in the simulation, the maximum altitude was
220 meters, whereas, in reality, it was 160 meters. During
the descent of the rocket in the simulation diagram after the
parachute has opened, there is a steady decrease until
landing, whereas in the actual launch diagram the parachute
deployed very late and therefore a sharp decrease is
observed.
The temperature graphs (diagram 14) differ considerably
from each other. Because the sensor is located inside the
rocket, the values are not very accurate due to the

Page 4

Fig. 10. Pressure diagram

 combustion of the engines. In the simulation, there is a Finally, in the velocity diagrams (diagram 16) we can
continuous decrease until it reaches the maximum altitude see that the simulation and the reality are very different.
and from there a steady increase until it returns to the initial Whereas at the beginning of both diagrams there is a
temperature. Below in the pressure (diagram 15) diagrams, constant increase in speed and a relative decrease after the
we see that the exact opposite of what happened at altitude end of engine combustion, in the simulation there is a
occurs. In the simulation when the parachute opened a constant decrease in speed until the parachute opens, but in
gradual increase in pressure is shown, whereas, in reality, reality, there is a new increase. Until the rocket lands, there
the pressure increase is steeper up to 120ts where the are noticeable fluctuations in the diagram with the actual
parachute has opened, resulting in a steady increase in launch which are imperceptible in the simulation. The
pressure. difference between the two diagrams is due to the fact that
the simulation shows only the lateral velocity.

Fig. 13. Altitude comparison diagram

Fig. 14. Temperatur comparison diagram

Fig. 15. Pressure comparison diagram

Page 5
Fig. 16. Speed comparison diagram

University of Rome. Italian Space Agency (ASI).

January 2022
IX. CONCLUSIONS
6. D. Rohini, C. Sasikumar, P. Samiyappan, B.
Building the rocket was easier than expected. However, Dakshinamurthy, Neeraja Koppula. Design & analysis
because of the use of different materials, it was difficult to of solid rocket using open rocket software. May 2022.
evenly distribute the weight. This could have been achieved International Conference on Advanced Materials for
by using a 3D Printer. The most difficult part of the rocket Innovation and Sustainability.
construction was the stabilisation of the flight computer 7. William Brown, Michael Wiesneth, Thomas Faust,
inside the rocket, which also had to be protected from the Nghia Huynh, Carlos Montalvo, Kent Lino and
engines. The solution was found by using a piece of acetate Andrew Tindell. Measured and simulated analysis of a
paper which had to be cut with great precision to allow the model rocket. February 5, 2018. Proceedings of the
nose to be separated from the rest of the body of the rocket Institution of Mechanical Engineers, Part G: Journal of
and a coating of aluminium foil which served as an Aerospace Engineering
insulating material. Another issue was the difference
between some diagrams showing the actual launch versus 8. https://www.instructables.com/Model-Rocket-Flight-
those showing the simulation. The flight computer data Computer-With-Arduino/
could be much more detailed if more accurate sensors were 9. Julia Bigwood, Caroline Kuhnle, Robert Connor, Alec
used, which could not be done due to high costs. Mitkov, Colby Gilbert, Nathaniel Rutkowski, Christian
M. Schrader, James Ternent. Design, Analysis, and
Thus we conclude that there is room for improving Test of a High-Powered Model Rocket-2. April 2021.
the measurements and the construction by using better Worcester Polytechnic Institute
sensors and more robust materials for the rocket. This
10. https://lastminuteengineers.com/bmp180-arduino-
conclusion and the fact that the difficulties encountered in
tutorial/
the course of the work were not discouraged but increased
the interest, are opening up the prospect of possible further 11. https://components101.com/sensors/bmp180-
work on this subject. atmospheric-pressure-sensor
12. https://store.arduino.cc/products/arduino-nano-every?
X. REFERENCES queryID=undefined
1. Model Rocket Workshop: a Problem-Based Learning 13. https://lastminuteengineers.com/mpu6050-accel-gyro-
Experience for Engineering Students J.I. Rojas, X. arduino-tutorial/
Prats, A. Montlaur and E. García-Berro Universitat 14. https://components101.com/sensors/mpu6050-module
Politècnica de Catalunya (UPC), Escola Politècnica
15. https://www.geeksforgeeks.org/i2c-communication-
Superior de Castelldefels (EPSC). December 2008.
protocol/
International Journal of Emerging Technologies in
Learning (iJET) 16. Code written by Matthew Minogue, Copyright 2020
2. Stine, H., Stine, B., Handbook of Model Rocketry, 7th 17. Rocket motors: https://www.raketenmodellbau-
edition, Wiley, 2004. klima.de
3. OpenRocket technical documentation For OpenRocket 18. Openrocket, https://openrocket.info
version 13.05 2013-05-10 by Sampo Niskanen 19. https://en.wikipedia.org/wiki/Euler_angles
4. T. Milligan, "The Different Rocket Recovery 20. https://www.mathworks.com/products/matlab.html
Techniques," Apogee Components, July 2017. 21. https://www.engineeringtoolbox.com/air-altitude-
5. Student firing tests and launches with commercial and pressure-d_462.html
self-made solid rocket motors: Mario Tindaro
Migliorino, Manuel Aiello, Matteo Berti, Marco
Rotondi, Simone D’Alessandro, Daniele Bianchi,
Munzer Jahjah, Marco Pizzarelli. Department of
Mechanical and Aerospace Engineering, Sapienza

Page 6

View publication stats