Testing a Prototype 1U CubeSat on a Stratospheric Balloon Flight
Akaash Srikanth, Bharat Chandra, Binukumar G Nair, Nirmal K, Margarita Safonova, Shanti Prabha, Rekhesh Mohan, Jayant Murthy, Rajini G.K
TTesting a Prototype 1U CubeSat on a Stratospheric BalloonFlight
S. Akaash ∗ , Bharat Chandra , Binukumar G. Nair , Nirmal K. , Margarita Safonova ,Shanti Prabha , Rekhesh Mohan , Jayant Murthy , and Rajini G.K. Vellore Institute of Technology, Vellore Indian Institute of Astrophysics, Bangalore 560034
Abstract
High-altitude balloon experiments are becoming very popular among universities and researchinstitutes as they can be used for testing instruments eventually intended for space, and for simpleastronomical observations of Solar System objects like the Moon, comets, and asteroids, difficultto observe from the ground due to atmosphere. Further, they are one of the best platforms foratmospheric studies. In this experiment, we build a simple 1U CubeSat and, by flying it on ahigh-altitude balloon to an altitude of about 30 km, where the total payload weighted 4.9 kg andexamine how some parameters, such as magnetic field, humidity, temperature or pressure, vary asa function of altitude. We also calibrate the magnetometer to remove the hard iron and soft ironerrors. Such experiments and studies through a stratospheric balloon flights can also be used tostudy the performance of easily available commercial sensors in extreme conditions as well. Wepresent the results of the first flight, which helped us study the functionality of the various sensorsand electronics at low temperatures reaching about ± − ◦ . Further the motion of the payload hasbeen tracked throughout this flight. This experiment took place on 8 March 2020 from the CRESTcampus of the Indian Institute of Astrophysics, Bangalore. Using the results from this flight, weidentify and rectify the errors to obtain better results from the subsequent flights. Keywords : high-altitude balloons, attitude, stratosphere, CubeSat, payload motion, magnetometer.
High-altitude balloon experiments are becoming very popular as they can be used for scientific researchat fractions of the cost of space projects. Because of the low cost, they are perfect for both academic andeducational institutions. Besides that, at high altitudes we escape most of the atmosphere, thus avoidingthe need to resort to adaptive optics as in the case of ground-based telescopes (Hibbits et al ., 2013).Since balloons are launched to the altitudes of about 25–50 kilometers, the instruments are exposed toa harsh environment, where the temperature reaches about − ◦ C and pressure about 1 millibar, thusballoons serve as a good platform to test the functionality of the components before launching into space.Space payloads for UV imaging and observations can be tested for their performance on high-altitudeballoons (Mathew et al ., 2016; Sreejith et al ., 2016,b). Pointing systems are necessary for the study ofastronomical sources, and such systems can also be studied and tested using balloons. The design of theballoon-borne pointing system, which can be used to point and track objects with an accuracy of ± . ◦ and ± . ◦ , respectively, is described in (Nirmal et al ., 2017).Among the various instruments that are being used for space research, CubeSats are becoming verypopular among academic and research institutions. Their beginning was marked by a collaborationbetween California Polytechnique State University and Stanford University in 1999. The purpose ofsuch miniature satellites was to make research in space be accessible to the universities. Nowadays,even high schools have started sending CubeSats to space. These satellites are mainly used for climate ∗ E-mail:[email protected] time of the work, S. Akaash was a project student at the Indian Institute of Astrophysics, Bangalore. a r X i v : . [ a s t r o - ph . I M ] F e b onitoring and changes, biological science, the study of near-Earth objects, planetary science, space-based astronomy, stellar imaging and heliophysics (NASA, 2017).Like other instruments discussed above, CubeSats also need to be tested before they are deployed intospace. A study shows that though the success rate of the CubeSats slowly shows an increasing trend, itstill has a failure rate of about 22%. This can be improved considering that more and more of them arebeing launched over the years (Villela et al ., 2017). Thus, it might be better to test them and analyzetheir performance on a high-altitude balloon before actually launching them in space. Testing CubeSatscan reveal some of the problems in the subsystems. For example, the research article by NorwegianUniversity of Science and Technology (NTNU) describes such an experiment testing the functioning ofUHV and VHF radio on a meteoballoon, which revealed some of the thermal issues and problems withutilizing HAM radios in a CubeSat (Tømmer et al ., 2015). In Kimm et al . (2015), it was shown thatthere was a loss of contact with the balloon at very high altitudes. The study suggested that high-speedwinds, with a speed of about 250 miles per hour, caused this causing the swinging of the antenna. Astudy from another flight describes the overall preliminary performance test of the communication systemon the student CubeSat NAVIS (Mortensen et al ., 2010). Through the data obtained from a balloonflight, it was inferred that a 1U CubeSat is enough to receive signals from the Automatic IdentificationSystem around Greenland. Testing CubeSats on balloons can also prove to be a cheap and an excellentopportunity for students to get hands-on experience for building space instruments which are to belaunched in space (Davis et al ., 2019).In 2011, the High-Altitude Ballooning (HAB) program was initiated at the Indian Institute of Astro-physics (IIA) with the primary purpose of developing and flying low-cost scientific payloads on balloon-borne platforms and, also develop instruments that can operate on a range of near-space platforms,including CubeSats, minisatellites, or even in space missions. The results of initial tethered flights,performed at the IIA are described in Nayak et al . (2013). Safonova et al . (2017) provides a detailedexplanation of stratospheric balloon experiments, and describes the results from 9 such scientific experi-ments performed by the IIA balloon research group. Such experiments can be used to study the planetsin our Solar System, the Moon, comets, and even atmospheric emission lines (Sreejith et al ., 2016).In this work, we designed a prototype 1U CubeSat to verify the functionality of various sensors whichare used to measure such parameters as pressure, temperature, magnetic field, humidity. We describehere the ground-based tests, tethered flight tests, and finally, the results of the stratospheric balloonflight performed on March 8, 2020. The payload is a 1U CubeSat which has the dimensions of about 10 × ×
10 cm. The aluminium frameused was purchased from
Interorbital Systems . For this experiment, the CubeSat was designed suchthat the frame contains three layers of PCB connected with each other for proper functioning. The firstlayer is the microcontroller section, which acts like the brain of the CubeSat. The microcontroller is usedto give commands from the various sensors and receive the information from them. The second layeris the experimental layer, which contains the various sensors and electronic modules essential for theexperiment. The third layer is the power supply layer, which is used to supply power to the CubeSat.A flowchart indicating the connectivity of the various components has been shown in Figure 1. Perforated boards were used to build the prototype. The microcontroller we have used in the design isATmega328 with Arduino firmware. The microcontroller sends commands and acquires data from thevarious sensors on the CubeSat. There are 14 input-output pins, and the microcontroller board producesa 5V output which is the voltage required by most sensors. However, we also have a buck-boost converterso that we can supply power to devices that run at other voltages. This buck-boost converter is connectedto the battery which converts the unregulated voltage to a regulated 5V supply. This steady supply isgiven to the microcontroller. We also have a micro SD card module which stores the recorded valuesfrom all the sensors in the experimental section. Sometimes it so happens that when the SD cards aresuddenly removed without turning off the power supply in the CubeSat, the data gets corrupted. To We describe here the sensors and electronic components that measure parameters of interest, such aspressure, magnetic field, temperature, humidity and location. It is especially very important to calibratethe magnetometer before using it.
The magnetometer used here was the HMC-5883L triple-axis magnetometer. Change in resistance of theNi-Fe material in the magnetometer is detected through the bridge circuit which is then used to estimatethe value of the magnetic field. The resistance of the material changes because the Earth’s magnetic fieldaffects the flow of charges in the conductor. This sensor works based on the I2C communication protocoland the X , Y and Z values of the obtained magnetic field has to be obtained from the data stored in the8-bit registers. However the obtained value is not the true magnetic field as there are instrumentationerrors and hard and soft iron errors which have to be corrected. Instrumentation errors can be causeddue to some errors in fabrication of the device which can introduce scale factor errors, which are errorscaused in the proportionality constants relating the input and output or offsets, which introduces somebias in the magnetic field. On the other hand, hard iron effect introduces a bias in the reading due to thepresence of permanent magnetic materials, and soft iron effects are caused by external magnetic fields(Kok et al ., 2012). An example of the effects of these errors are shown in Fig. 3.Next, we perform the linearity test. This means that, as we change the input, we obtain a linearlyproportional output. In this case, the test was performed by slowly changing the azimuth by rotatingthe magnetometer with time. The azimuth is given by: φ = arctan( B y , B x ) . et al ., 2017) developedat the Indian Institute of Astrophysics, and rotate it with the help of a DC motor which is in turncontrolled by an independent Arduino UNO microcontroller board Fig. 4. The data obtained from themagnetometer was sent remotely to a computer. To do this, we had used a Raspberry Pi 3A+ to establisha remote connection with the magnetometer. We can also verify the linearity through the repeatabilitytest by looking at the input (time) vs output (azimuth) response.We tried several methods to calibrate the magnetometer data. However, none of them gave a properly4igure 4: Images show the setup which can be used to measure the repeatability and linearity. Theplatform rotates with the help of a motor which is controlled with the help of Arduino. The magnetometeris fixed on this system which remotely sends the data to a Raspberry Pi.Figure 5: Left plot shows the linearity test performed by a uniformly rotating the magnetometer byhand. Though it is not exact, it does give something close to a linear result. Right plot displays theresults of the repeatability test, showing that the magnetometer output is quite repeatable and that itis functioning properly.calibrated output. After some trial and error, we found that the ellipsoid-fitting algorithm was the bestto calibrate the magnetometer . The output after calibration is shown in Fig. 6. After some rigorouscalculations, we arrive at the following equation: h m = Ah + b , (2)where h m denotes the obtained magnetic field and h denotes the true magnetic field. The constants A and b are determined by the ellipsoid-fitting algorithm, which was performed in MATLAB. To determinethe constants, we need to obtain some raw data from the magnetometer by ensuring that the axis ispointed at various angles. The constants A and b were found to be: A = .
290 0 . . − . . . − . . . , b = − . − . − . . Detailed explanation: https://teslabs.com/articles/magnetometer-calibration/
Left : This image shows the uncalibrated magnetic field. It was obtained by keeping themagnetometer at various possible orientations. The different colours in the graph represent the rotationsperformed with different orientations of the magnetic field. We can see that there are several offsetsin it.Mag X, Mag Y and Mag Z axes represent the total magnetic field with respect to the directionsmentioned on the HMC5883L magnetometer and is measured in micro Tesla.
Right : This image showsthe results of the calibration which required using the ellipsoid-fitting algorithm in MATLAB.
This section also contains the Ublox Neo-6M GPS which needs to be supplied with a constant voltageof 3.3V. The GPS module needs to lock on to at least 4 satellites to get the latitude and longitude,and an additional satellite to get the altitude. GPS data contains 3 different types of information. Apseudo-random code tells us to which satellites the GPS is locked on. Ephemeris data tells the positionof that particular satellite, its health, date, and time. Almanac data shows orbital data of every satellite.Using the GPS data, we find out the latitude, longitude, and altitude. However, there are a lot of otherdata like flight velocity and satellite details which can be obtained. Care should be taken to ensurethat the baud rate of the GPS is perfectly set so that communication with the microcontroller is notaffected. The microcontroller must be programmed such that the data is requested from the GPS duringthe wait time of the GPS. If the data is requested during the processing time – the time when the GPSis acquiring the data from the satellites, it might lead to invalid or corrupted data.The DHT-11 temperature-humidity sensor was used for measuring both temperature and humidity.The humidity is measured by the level of water vapour in the atmosphere. There are two electrodesinside which form a capacitive system. When the water molecules stick to the surface, ions are liberatedthus changing the conductivity, and this can be used to measure the water level and thus humidity.Temperature is obtained from the temperature-sensitive voltage and current characteristics of a diode.When two identical transistors are operated at a constant ratio of collector current densities, the differencein base-emitter voltages is directly proportional to the absolute temperature. The temperature sensor inthe DHT-11 works only in the small range: 0 ◦ C to 50 ◦ C while the humidity sensor works between 10%or 20% to 90%.The pressure sensor used was BMP 180 pressure sensor and uses the I C communication. A pressuresensor uses a piezo-resistive sensor to measure the pressure. This piezo-resistive sensor is connected in awheat stone bridge. Now when pressure is applied to it, its resistance changes and because of this, theoutput in the bridge changes, thus indicating the pressure. BMP180 is also useful as it has a very goodtemperature sensor which works from − ◦ C to 80 ◦ C. Using the pressure reading, we can also estimatethe altitude from Eqn. (3) : = 44330 (cid:18) − pp o (cid:19) . , (3)where A is the estimated altitude above the launch point, p is the pressure obtained and p o is the pressureat the point of launch. Before the actual flight, a tethered flight, was performed on May 5th, 2019 from the Indian Institute ofAstrophysics (IIA) campus, Bangalore to check the functioning of the CubeSat (Fig. 7). CubeSat wasplaced in the Styrofoam box and wrapped in the bubble wrap for safety. The balloon was filled with H gas. The payload was launched to an altitude of about 200 meters above the roof of the institutebuilding. As expected, it took a while for the GPS to lock on to the satellites and, hence, we receivedvery little data on latitude, longitude and altitude. However, all the other sensors were working properly.Figure 7: Tethered balloon flight on May 5th, 2019.The magnetometer data was used to check if the correct result was obtained. After applying Eq. (2)to the X, Y, Z coordinates of the magnetic field of the Earth, we get the calibrated magnetic fieldas (17 . , − . , . µ T. The actual value of the total magnetic field in Bangalore as is 41.39 µ T, whichshows that our magnetometer was calibrated properly. The payload was finally launched to stratosphere from the CREST campus (13.133 ◦ N, 77.815 ◦ E)of the IIA, located in Hosakote, Bangalore. The entire balloon-payload system was launched at 3:09am on March 8th, 2020. The total weight of the payload, including other components, such as a Geigercounter (for another experiment which details will be communicated separately), radio-module and otherelectronics, was about 4.9 kg. We used three latex balloons to lift the payload: two balloons of 1.2 kg-type and one 2-kg-type balloons were filled with hydrogen gas. The set up used for the free floating flightis shown in Fig. 8.
During the flight, the data from the CubeSat were obtained only for about 1 hr. This translates to analtitude of about 11 km. The most probable reason for the short duration was because the battery wasnot insulated properly. We plan to rectify this in the subsequent flights. Here, we discuss the results of Verified with magnetic-declination.com/India/Bangalore/1132482.html A = 66420 (cid:18) − pp o (cid:19) . , (4)This modified equation (Eq. 4) is more accurate for low altitudes, however, its accuracy still needs to befurther tested for higher altitudes. Also, the RMS of such a fit is just about 7.369, much better than Eq. 3.The altitude estimated from the pressure reading is the relative altitude from the CREST campusaltitude. From these data, the average ascent speed of the flight was calculated to be about 3.68 m/s.Figure 10: Left : The absolute altitude data estimated from the GPS.
Right : The relative altitudeestimated from the pressure reading obtained with the BMP180 pressure sensor. The third image showsthe comparison of the altitude readings from the Pressure sensor and GPS. In the final image, the greendashed line represents the corrected altitude formula of Equation 3.Next, the temperature was measured using the temperature sensor in the BMP180 pressure sensorand the temperature sensor in the DHT11 sensor Fig. 11. The DHT11 temperature sensor did not workproperly below 0 ◦ C as its operating range is 0 ◦ C to 50 ◦ C. Moreover, its precision and accuracy are lesser,and than the one on the BMP180 sensor. The temperature sensor in BMP180 can work up to − ◦ Cand can give results in decimal points. Hence, we prefer to use its data.Next, the humidity data measured in % was plotted in Fig. 12. We see the sudden spike near thepoint where the temperature hits 0 ◦ C. We might have to investigate if this feature has to do anythingwith the freezing point of water.The pressure seems to show a smooth drop as we increase in altitude.The X , Y and Z values of the magnetic fields are also plotted in Figure 13. The raw magnetic fieldobtained from the tethered flight after calibration seemed to give the correct magnetic field (Section 3).However, while plotting the data obtained from the free-floating launch, we had found that the total10igure 11: Left : Temperature measured using BMP180 sensor.
Right : The temperature measured usingDHT11 humidity sensor. We can see that it doesn’t work properly below 0 ◦ C.Figure 12:
Left : Temperature and humidity data from DHT11. We can see a sudden increase in humidityat about 0 ◦ C. Right : Pressure (measured by BMP180) as a function of altitude.magnetic field keeps fluctuating with a mean of about 39 µ T to 40 µ T, between the 25 µ T and 50 µ T. Totest it, we had plotted the uncalibrated magnetic field. To our surprise, the raw magnetic field obtainedwas varying very badly between 40 µ T and 90 µ T. If the raw values oscillate, then we cannot expect theactual magnetic field to be stable even if the calibration was performed correctly. Hence more studiesneed to be done on this. It is possible that the electronic components like the radio module might haveinduced some kind of magnetic field which is being sensed by the magnetometer. Further, there areseveral current-carrying wires, which in addition to producing its own magnetic field, also keeps rotatingwith the payload. This would further induce an additional magnetic field in the system. Because of thesefactors, we might have experienced an oscillating magnetic field. Thus it is essential that we performmore rigorous tests on the ground, before we fly the entire system. The total magnetic field may beshowing some trend after 9000 meters altitude. However, further studies are required because of theinsufficient data.
Because we did not provide sufficient insulation, as soon as the temperature reached about − ◦ C (whichcorresponds to an altitude of about 10 kilometers), the battery stopped functioning. The following canbe concluded after studying the data from this flight: • Additional insulation has to be provided to the batteries so that they function even beyond 1011igure 13:
Top : Plots of X , Y , and Z components of the magnetic field. Bottom : Total magnitude ofthe uncalibrated magnetic field (
Left ) and calibrated magnetic field (
Right ). X Mag, Y Mag, Z Mag andTot Mag represents the X , Y , Z , and the total value of the magnetic field, respectively. Raw values ofthe magnetic field were changing rapidly with the altitude, and the total magnetic field may be showingsome trend after 9000 meters altitude; however, further studies are required because of the insufficientdata. kilometers of altitude. If that still does not prove to be sufficient, we might have to developa heating element with a feedback system. This might ensure that the payload is always at aconstant temperature. • It can be concluded that the magnetometer calibration has to be done in the presence of the entirepayload system. In other words, the calibration has to be done in the presence of instruments,such as radio module and Geiger counter, which are going to be working throughout the durationof the flight. The fact that this was not done during the initial design, might have caused thetotal magnetic field value to fluctuate during the flight even though it did not happen during thetethered flight.We plan to rectify these errors before the our next flight to further test the performance of CubeSat builtfrom the off-the-shelf components.
Part of this research has been supported by the Department of Science and Technology (Government ofIndia) under Grant IR/S2/PU-006/2012.
References
Hibbitts, C. A., Young, E., Kremic, T. and Landis, R. Science measurements and instruments for aplanetary science stratospheric balloon platform. Aerospace Conference, IEEE, Big Sky, MT, pp. 1-9(2013). 12ayak, A., Sreejith, A. G., Safonova, M. and Murthy, J. High-altitude ballooning programme at theIndian Institute of Astrophysics. Current Science, 104, 708, (2013).Sreejith A. G., Joice Mathew, Mayuresh Sarpotdar, Nirmal K., Ambily S., Ajin Prakash, MargaritaSafonova, and Jayant Murthy. Balloon UV experiments for astronomical and atmospheric observations.Proc. SPIE 9908, Ground-based and Airborne Instrumentation for Astronomy VI, 99084E (9 August2016).Safonova, M., Nayak, A., Sreejith, A. G., Mathew, J., Sarpotdar, M., Ambily, S., Nirmal, K., Talnikar,S., Hadigal, S, Prakash. A. and Murthy, J. An Overview of High-Altitude Balloon Experiments at theIndian Institute of Astrophysics. Astron. & Astroph. Trans. (AApTr), 29(3): 397-426, (2016).Mathew, J., Prakash, A., Sarpotdar, M., Sreejith, A. G., Safonova, M. and Murthy, J. Ultraviolet cosmicimager to study bright UV sources. Paper 9905-146, Space Telescopes and Instrumentation 2016:Ultraviolet to Gamma Ray, part of SPIE Astronomical Telescopes + Instrumentation, (2016).Sreejith, A. G., Mathew, J., Sarpotdar, M., Nirmal, K., Suresh, A., Prakash, A., Safonova, M.,and Murthy, J. Measurement of limb radiance and Trace Gases in UV over Tropical region byBalloon-Borne Instruments – Flight Validation and Initial Results. Atmos. Meas. Tech. Discuss.,https://doi.org/10.5194/amt-2016-98, (2016).NASA, Cubesat 101: Basic Concepts and Processes for First-Time CubeSat Developers. NASA andCalifornia Polytechnic State University, (2017).T. Villela, C. A. Costa, A. M. Brando, F. T. Bueno, and R. Leonardi. Towards the thousandth CubeSat:A statistical overview. Int. J. of Aerospace Engineering, vol. 5063145, no. 1, pp. 1–13, (Jan. 2017).P. Techavijit, S. Chivapreecha, P. Sukchalerm and A. Plodpai. CubeSat image transmission in JPEGcompression: An experiment on high altitude platform. 2016 8th International Conference on Knowl-edge and Smart Technology (KST), Chiangmai, pp). 164-168, (2016).M. Tømmer, et al . Testing of radio communication subsystems for the NUTS CubeSat on a meteorologicalballoon flight from Andoya in 2014. In Proceedings of the 22nd ESA Symposium on European Rocketand Balloon Programmes and Related Research, ESA Special Publication. ESA, (2015).Kimm, H., Kang, J. S., Bruinga, B., and Ham, H. S. Real Time Data Communication Using High AltitudeBalloon Based on Cubesat Payload. Journal of Advances in Computer Networks, 3(3), 186-190, (2015).Mortensen, Hans Peter, et al . NAVIS: Performance Evaluation of the AAUSAT3 Cubesat Using Strato-spheric Balloon Flight. Proceedings of the 61th International Astronautical Congress: 3742–3749,(2010).J. Davis, et alet al