An Overview of High-Altitude Balloon Experiments at the Indian

Institute of Astrophysics

Margarita Safonova∗, Akshata Nayak†, A. G. Sreejith, Joice Mathew, Mayuresh Sarpotdar,S. Ambily, K. Nirmal, Sameer Talnikar, Shripathy Hadigal, Ajin Prakash and Jayant Murthy

Abstract

The High-Altitude Ballooning programme began atIndian Institute of Astrophysics, Bangalore, in theyear 2011 with the primary purpose of developingand flying low-cost scientific payloads on a balloon-borne platform. Some of the science goals are studiesof the phenomena occurring in the upper atmosphere,of airglow and zodiacal light, and observations of ex-tended astronomical objects such as, for example,comets, from near space (20 to 30 km). A brief sum-mary and results of the tethered flights carried out atCREST campus are given in Ref. 1. Here we presenta complete overview of the 9 free-flying balloon ex-periments conducted from March 2013 to November2014. We describe the launch procedures, payloads,methods of tracking and recovery of the payloads.Since we fall in the light/medium balloon category,the weight of the payload is limited to less than 5 kg— we use a 3-D printer to fabricate lightweight boxesand structures for our experiments. We are also de-veloping in-house lightweight sensors and controllersto use in the balloon flights. The flight and scien-tific data obtained from the different launches, andfuture plans for the development of a fully-fledged3-axis pointing and stabilization system and a low-cost star camera-cum-sensor for forthcoming balloon

∗For correspondence (E-mail: rita@iiap.res.in)†Akshata Nayak is a Ph.D. student at the Jain Univer-

sity, Bangalore 562 112, India; M. Safonova, A. G. Sreejith,Joice Mathew, Mayuresh Sarpotdar, S. Ambily., K. Nirmal.,Ajin Prakash and Jayant Murthy are in the Indian Institute ofAstrophysics, Bangalore 560 034, India; Sameer Talnikar andShripathy Hadigal at time of writing were internship studentsat the Indian Institute of Astrophysics, Bangalore 560 034,India.

flights are briefly discussed.Keywords: high-altitude balloons, scientific pay-

loads, free-floating flights, upper atmosphere

1 Introduction

Approvals from Director General of Civil Aviation(DGCA), Ministry of Defence (MoD) and AirportsAuthority of India (AAI) are necessary to performfree-flying balloon experiments. These permissionshave to be requested at least 8 to 12 months be-fore the beginning of the program. We have appliedfor these permissions in 2012 and have conductedour first free-flying balloon experiment on March 3,2013. In 2013–2014, we have carried out a total of 9launches. There have been failures in recovery, work-ing of the electronic systems and even loose connec-tions in wires during flights. But we have learnedfrom our experiences and put together some success-ful flights. The following sections give a detailedoverview of the all these launches.

2 Essentials for the balloonflights

2.1 Permissions necessary to carryout the launches

Since Bangalore is a hub of AirForce bases along withairports and flying training schools, we had to takeprior permissions from DGCA and MoD, Delhi. Inaddition, we need local permissions from HindustanAeronautics Limited (HAL), Bangalore International

1

arX

iv:1

506.

0153

8v1

[as

tro-

ph.I

M]

4 J

un 2

015

Airport Limited (BIAL), Chennai airport as a south-ern region headquarters, Jakkur flying school, Yela-hanka AirForce Base, AirForce station at Chimneyhills, Bangalore. It is a regulation that we informthe abovementioned offices about the balloon launchtwo weeks in advance, specifying the payload details.Launches are only allowed after obtaining all the re-quired No-Objection Certificates (NOCs).

2.2 Balloons

Many types of balloons are used for high-altitudeballoon experiments. Latex balloons, also popu-larly known as sounding or weather balloons, aredesigned to reach 35–40 km and burst, after whicha parachute is deployed to safely carry the payloadback to Earth. There are also thin plastic balloons,called zero-pressure balloons, which can float at highaltitudes for days. We decided to use latex balloonsas they are readily available in India from Pawan Bal-loons, Pune.

2.3 Gas used for the balloon filling

Helium as an inert gas is considered the safest for fill-ing the balloons. However, because its is expensiveand due to scarcity of pure helium2, we eventuallyswitched over to a much cheaper commercial hydro-gen. The price of a 10 cu.m. helium cylinder (99.99%purity) is Rs. 17,000 as compared to Rs. 800 for a7 cu.m. commercial (99.99% purity) hydrogen cylin-der. We buy gas for each launch from Sri VinayakaGas Agency, Bangalore.

2.4 Payload box

First Model The first payload model (Fig. 1,Left) was a rectangular structure made out of Styro-foam with had dimensions of 20 cm × 20 cm × 20cm. The base thickness was 2 inch, and each sidewas 1 inch thick. It was wrapped in aluminum foilto provide the basic insulation as well as the radarreflection. We, however, found that this shape gaverise to considerable spinning motion of the box.

Figure 1: Left: First model; Right: Second model.

Second Model The second stage of develop-ment was to overcome the disadvantages of the firststage. The shape was selected to be cubic (15 cm ×15 cm × 15 cm) with wind vane structure. The idea ofhaving such a design is to align the payload always inone particular direction i.e. along the wind directionto limit the spinning of payload. This structure wasmade from lightweight plastic fibre and open in thebottom and top to allow movement of the air throughthe vanes (Fig. 1, Right). This payload structure wasmuch smaller and lighter than the first design. Thethickness of the base and the sides was retained asin the previous model. However, the spinning mo-tion was still substantial and there was insufficientinsulation for the instruments inside the box.

Third Model The third and final model (Fig. 1,Right) was the one that has been used for actualflight experiments. To provide the insulation we haveused Kapton tape, which withstands a temperaturedown to −269◦C.3 The wind-vane structure with thesame dimensions, box material and thickness as inthe second model was retained. However, the topand bottom openings in the second model were cov-ered with Styrofoam. A suitable ring structure wasconstructed to attach the payload to the parachute;this ring structure also helped in reducing the spin-ning of the payload. This payload was covered withbright coloured tapes to be easily noticeable duringits recovery. We also pasted team members’ mobilenumbers on the payload box in case of it being found

2

by the public.

2.5 Payload and the Flight computer

To capture the darkness of space, horizon and travelpath of the balloon, we decided to use a camera on-board. Initially, we used a Canon Ixus camera butfound it to be too heavy for the payload and wereplaced it with a tiny Raspberry Pi (RPi) cameraprogrammed for specific exposure time. For the ini-tial flight, we used a BASIC STAMP 2 micro con-troller (MC) module-based flight computer. Since itwas found insufficient to support the necessary sen-sors and computations, we later switched to a Sin-gle Board Computer (SBC). A flight computer usingRPi SBC was developed in-house for this purpose.4

It computes the attitude of the payload by combin-ing data from an accelerometer, a magnetometer anda gyroscope located on a single chip. The attitudeinformation in terms of Euler angles is combined ona RPi with the output of a global positioning system(GPS) chip to obtain the celestial coordinates of thepointing direction.

2.6 Flight Termination Unit (FTU)

Balloons drift off if winds are high during the flight.To restrict the drift to another state or territory, wehave designed two FTUs for the balloon system. Onesystem is based on a timer circuit and placed belowthe parachute (Fig. 2) and the other – Arduino-basedsystem with a GPS – is placed inside the main pay-load box. The FTU uses a thermal knife to cut theload line between the balloon and the parachute.

1. Timer-based FTUThis mechanism consists of a timer circuit whichis set for a duration of 2 to 3 hours dependingon the type of experiment. It is initialized justbefore we release the balloon. The timer systemis based on a tiny MC, which provides a trig-ger signal after the programmed time to heat upthe nichrome wire wound across the load line tomelt it and sever it for the deployment of theparachute.

2. Geo-fencing based FTUThe geo-fencing system uses GPS to obtain lati-tude and longitude information. It sends a signalto heat up the nichrome wire to melt the load lineif the payload crosses the predefined latitude andlongitude.

2.7 Parachutes

Parachutes used in high-altitude ballooning are usu-ally made out of ripstop nylon. We have obtainedparachutes from Rocketman Parachutes Inc, USA, asthey are of good quality and inexpensive, speciallydesigned for high-altitude balloon payload recovery.We use parachutes of two different sizes, 7- and 8-footdiameter that have lift capacity of 4.0–4.9 kg and 5.4–6.8 kg, respectively. We also have small parachutesof 1-kg lift capacity from Aerocon Systems Inc, USA.In some flights we connect these parachutes serially.

3 Free-flying Balloon launches

3.1 Launch 1, (March 3, 2013)

This flight was our inception to the free-flying balloonexperiments. To keep it simple, we used a small cu-bical shaped box to house our payload (Fig. 1, Left).This flight was basically a test flight for the electron-ics and a GSM-based GPS tracker.5 It was an earlymorning flight scheduled for 6:30 am, but was de-layed by one hour. We have used 3.4 cu.m. of heliumto fill the 1.2 kg balloon. The total weight of thepayload was about 1 kg, so we have used 1-kg liftcapacity parachute from Kohli Enterprises, Delhi (afree-sample parachute given for this launch).

Predictions for the flight To enable safe recoveryof the payload, we carry out predictions of the land-ing location using http://predict.habhub.org. Inthis flight, the predicted landing location was closeto Malur, Karnataka, near the actual location of re-covery. Fig. 3, plotted using Google Earth, shows thepredicted path as well as launch, burst and landinglocations.

3

Figure 2: Balloon–payload structure.

Payload A low-cost flight computer using BASICSTAMP 2 MC module of Parallax Inc (Fig. 4, Top)was developed in-house to measure atmospheric pa-rameters, such as pressure and temperature, alongwith latitude and longitude positions, using varioussensors. These parameters could be recorded into aUSB device (pendrive). The bread-board model ofthe MC was used for the flight. The latitude andlongitude positions could not be transmitted live be-cause a radio license is required to carry out the livetransmission over a radio band. To live-track the bal-loon, we obtained a small GPS unit called Satguidetracker (Fig. 4, Bottom left), which worked on GSMsignals (Tata Docomo network). To capture images

Figure 3: Prediction for March 3, 2013 flight.

during the flight, we placed a Canon Ixus 115HS cam-era on the payload box (Fig. 4, Bottom right).

Figure 4: Top: Basic Stamp 2 MC board, Bottomleft: Satguide tracker, Bottom right: Canon Ixus 115HS camera used onboard.

Recovery The FTU was programmed for 1 hourcut-off. We, however, lost the signal from the Sat-guide at 7.50 am and the search at the predicted land-

4

ing location failed. After two days, local residentscalled and informed us that they have found the pay-load. It was recovered in a bad condition and we lostthe parachute, camera and Satguide GSM tracker.

Flight Analysis and Summary We lost the GSMsignal due to poor network (Tata DOCOMO) inMalur-Hoskote area. In addition, GSM networks donot work above 3 km altitude. Though we had re-covered the MC, it did not record any data, proba-bly due to loose wires in the circuit board. We havedecided to use Printed Circuit Board (PCB)-basedelectronic circuits for next flights. It is also possiblethat the balloon was underfilled. A detailed study ofballoon filling procedures seemed essential before thenext flight.

3.2 Launch 2, (June 30, 2013)

After a failure in tracking using GSM trackersin the previous launch, we decided to use radiotracking for the second launch. We incorporatedthe radio tracker from Dhruva Space Pvt. Ltd.(http://www.dhruvaspace.org) to live-track theballoon. The total weight of the payload was 3.2 kg,therefore we have used two helium filled balloons: 3-kg and 1.2-kg, using 3-kg as a master balloon (Fig. 5).Two 1-kg parachutes were connected serially. We in-cluded an in-house developed RPi-based attitude sen-sor to serve as the flight computer.4

Figure 5: June 30, 2013 launch with two balloons.

Predictions for the flight Using the availablewind data, we found the predicted location of landingto be in Ramnagara district, Karnataka.

Preparations for the launch

Accuracy of the RPi sensor Testing of the in-house developed attitude sensor (Fig. 6) mounted onthe telescope was carried out for this flight (Fig. 7)prior to the launch. For details of development andtesting, see Ref. 4.

Figure 6: RPi based attitude sensor.

Payload

Payload boxes We had two payload boxes thistime.

• Our payload comprised the in-house flight com-puter programmed to store data onboard. Anattitude sensor with a webcam looking down wasinterfaced with the RPi. The attitude sensorcontinuously logged the RA and DEC of the pay-load pointing direction, as well as the latitude,longitude, UTC and altitude. The webcam wastaking low-resolution images every minute. Atimer-based FTU, set for 90 minutes, was pro-grammed to detach the payload from the bal-loon.

5

Figure 7: Attitude sensor mounted on the telescope.

• Dhruva Space payload box contained a telemetryunit and an RPi, interfaced with an infrared (IR)camera. The plan was to compare the visible andIR images of the ground to find the chlorophyllcontent around Bangalore region.

Telemetry Unit The telemetry unit, precon-figured with a call sign (Micro-Trak All In One),was necessary for this launch for live tracking. TheMicro-Trak All In One (AIO) (Fig. 8) is a complete,self-contained, water resistant, portable, 10-watt Au-tomatic Packet Reporting System (APRS) tracker,a TinyTrak3 controller chip, a Byonics GPS2OEMGPS receiver, and an Standard Male (SMA) antenna.It operates using 8 conventional AA batteries or a12V power supply. The MT-AIO has been confirmedto run for nearly 8 days when used with typical AAalkaline batteries, transmitting every 2 minutes.

Figure 8: Dhruva Space telemetry unit in a Pelicancase.

Recovery We lost the radio signal near Mugabalavillage, north of CREST, Hoskote (The last radioreading was at 10.2 km lateral distance). The payloadwas recovered after two weeks near a small village inRamanagara district following a phone call from lo-cal residents, near the predicted landing location. Werecovered all the electronic components, except radioantenna, from the landing site.

Figure 9: Azimuth-Elevation variations recorded onLaunch 2, (June 30, 2013).

6

Flight Analysis and Summary This flight wasto test our attitude sensor and the radio communi-cation link between the payload and ground station,provided by Dhruva Space. However, both telemetryand IR camera failed to transmit and record data.Our RPi also failed 40 mins after the launch. Thesharp temperature variation (around −30 to −40◦Cat an altitude of 5 km) caused the power supplybattery of the RPi to drain, probably because theelectronic module (EM) was not properly insulated.From the attitude sensor data output plot shown in(Fig. 9), it can be seen that the azimuth and eleva-tion of the pointing direction were varying randomly.We attribute these variations to the random natureof winds in the upper layer of the troposphere, at∼ 5–10 km altitude. Variation in the elevation wassmaller which indicates that the swinging motion ofthe payload box was less pronounced.

Figure 10: June 30, 2013 GPS plot of the flight path.

3.3 Launch 3, (September 3, 2013)

The September launch was a test flight in preparationto capture the images and spectra of the comet ISON.We used one helium filled 3-kg balloon and two 1-kgparachutes connected serially (Fig. 11). The totalweight of the payloads was 2.4 kg. The launch wasat 6.20 am and the payload landed at 7.40 am.

Predictions for the flight The predicted landinglocation was 25 km from CREST campus along thesouth direction and the landing location was close tothe predicted (Fig. 12).

Figure 11: September 3, 2013 launch with the rope-ring mechanism.

Figure 12: Prediction for September 3, 2013 launch.

Preparation for the launch

Rope-ring mechanism In the previouslaunches, the payload was subject to jerks duringthe release. To ensure smooth release of flight train,we decided to use a rope-ring mechanism. Thismechanism provided vertical suspension and stabil-ity of the payload at release. The rope connectedto two poles was passed through the ring. Thering was used to support the whole balloon–payloadstructure. Balloons and the parachute were attachedto this ring from opposite sides. This way, thepayload was suspended vertically and stabilizedbefore the launch. Finally, the rope was cut to allowthe balloon to have a smooth launch.

7

Payload The payload consisted of one box con-taining a Medium Personal 1 (MP1) GSM trackerfrom Ingo labs, Hyderabad, with Airtel SIM card,Arduino-based flight computer with temperature sen-sor, and an FTU, set for 45 mins. Dhruva’s pay-load had a Radio AIO unit (same as in the previouslaunch) and an Arduino-based GSM tracker with aVodafone SIM card.

Recovery We lost the GSM signal when the bal-loon rose above 3 km. The radio receiver wasworking all the time. The payload was recoveredon the same day at 9.20 am in Oblapura village,Karnataka, 26 km away from the CREST cam-pus. Both radio and GSM–GPS trackers trans-mitted the exact location of the payload. Thecomplete video of this launch can be found onhttps://www.youtube.com/watch?v=NwQBl4tfRjk.

Flight Analysis and Summary The predictedmaximum altitude was 12 km. But the altituderecorded by the radio tracker was 12.9 km (Fig. 13).The GPS positions, continuously transmitted by theradio tracker, were used to calculate the ascent anddescent rate of the payload. The average ascent ratewas found to be about 7 m/s, and the descent rateat touchdown was 5 m/s, as expected.

Figure 13: The graph of altitude vs. time of thepayload recorded on Launch 3, (September 3, 2013).

3.4 Launch 4, (October 13, 2013)

This launch was to test the gimbal (Model Aviation,Bangalore) — a 3-axis gyro-stabilization platform de-signed to point and stabilize the telescope onboardthe flight (Fig. 15, Top). We used two 3-kg balloonsfilled with helium. The total weight of the payloadswas about 5.5 kg. Therefore, we used an 8-ft Rock-etman parachute. The launch time was 6.15 am.

Predictions for the flight The predictions for theflight path were carried out and the expected max-imum altitude was 25 km. The balloon burst wasexpected to be close to Tiptur, Karnataka (Fig. 14).

Figure 14: Predicted path of the balloon for October13, 2013 launch.

Preparations for the launch Laboratory testswhere conducted to verify gimbal’s working temper-ature and stability of the pointing direction. Theground level tests were performed in the mechanicalworkshop by suspending the gimbal by a crane toprovide conditions similar to the balloon flight.

We had conducted insulation tests for the payloadbox prior to the launch by placing one open and onecompletely insulated Styrofoam boxes into a refrig-erator with an ambient temperature of −22 ◦C for 2hours. An Arduino-based temperature data loggerwas kept in both boxes for equal intervals of time.LM75 was used as the temperature sensor and thetemperature inside the box was recorded every 15seconds. The temperature plot is shown in Fig. 16.

As can be seen from the graphs, the inside temper-ature drops much faster in uninsulated box comparedto the insulated box. Since our usual time of flightis about 2 hours, it is advisable to properly insulate

8

Figure 15: October 13, 2013 launch. Top: Gimbalwith mounted telescope used in October 2013 flight.Bottom: UV Spectrograph (Mayapro 2000, OceanOptics, USA).

Figure 16: Graph showing the temperature test re-sults of the insulated (blue) and uninsulated (red)box.

the payload boxes to ensure working temperature forthe electronics.

Payload We had two boxes this time. Our payloadconsisted of gimbal and its battery, RPi with attitudeand temperature sensors, two GSM trackers (Airtel

and Idea SIM card), Arduino-based cut-off mecha-nism and a GoPro camera. Dhruva’s payload boxhad the radio AIO unit.

Recovery The payload was found two weeks laterby the fishermen near Udupi, Karnataka and re-turned to us. Most of the components, includinggimbal and electronics, were lost.

Flight analysis and Summary The initial ascentrate was 7.5 m/s. From recorded radio data we in-ferred that one balloon burst at the altitude of 20.5km, while the remaining balloon floated at about 19–20 km altitude. The timer-based FTU failed andthe balloon along with the payload eventually crossedinto the Arabic sea via Karwar coast, Karnataka. Wefollowed the balloon through radio till the last re-ported position of about 150 km into the Arabic sea(Fig. 17).

Figure 17: October 13, 2013 path of the balloon.

3.5 Launch 5, (November 24, 2013)

The objective of this launch was to test the spec-trograph (Fig. 15, Bottom) acquired to observe at-mospheric lines and extended Solar System objects,in particular, comet ISON at perihelion. We used 4helium filled balloons – one 3-kg, two 2-kg and oneunderfilled 1.2-kg balloon. The total weight was be-low 6 kg. We used one 8-ft Rocketman parachute(Fig. 19).

9

Predictions for the flight The wind data predic-tions gave the location of landing near Nandi Hills,Karnataka (Fig. 18).

Figure 18: Prediction path for November 24, 2013launch.

Preparations for the launch We have used thenew gimbal: a compact model, with just 2-axis sta-bilization from NIT, Surathkal, Karnataka, with atelescope mounted on it. The telescope body wasfabricated by Cosmic Labs (Electronic city, Banga-lore). The gimbal along with the telescope was testedin the lab for its mounting. The telescope body waspainted black to avoid scattering of the straylight.The light from from the telescope was fed to the thespectrograph though the optical fibre.

Payload The payload included three boxes. Onebox contained the spectrograph, telescope, gimbal,attitude and temperature sensors, two GSM track-ers (Airtel and Idea) and a timer-based FTU. Thesecond FTU was in a separate box attached directlyunder the parachute. The flight computer remainedthe same, with an RPi as the controller. Dhruva’spayload box had the radio AIO unit.

Recovery The payload was recovered near Nandihills, Bangalore, a little ahead of the predicted land-

ing location. Three balloons burst before the pro-grammed cut-off time and the remaining underfilledballoon served as a flag which helped the recov-ery (Fig. 19). The payload was successfully trackedthrough radio and GSM trackers.

Figure 19: November 24, 2013: Recovery photo withthe flag balloon and two parachutes.

Flight analysis and summary The launch wasscheduled for 2 am but got delayed by 35 min-utes. The flight data showed that an altitude of∼ 16 km was achieved. Although all the electronicsworked throughout the flight, the spectrograph didnot record any data due to broken USB connection,most probably at launch.

3.6 Launch 6, (February 16, 2014)

This launch was scheduled at 4.30 am to observe at-mospheric lines or airglow at twilight. The launchgot delayed by 30 minutes and was carried out at 5am. We used three 2-kg hydrogen-filled balloons toprovide the desired lift of 5.5 kg.

Predictions for the flight As the winds were pre-dicted to be high on the launch day, we expected theballoon to drift faster at high speeds. The predicted

10

landing location was found to be 115 km away fromCREST campus.

Preparations for the flight 2-kg balloons werefilled so as to lift a weight of 5 kg each: 2 kg to liftthe balloon itself, 2 kg to lift the payload and 1 kg –a free lift.6

Payload The payload was the spectrograph withthe optical fiber (serving as an aperture) pointing atthe horizon, the USB camera placed on top of thepayload box facing upwards to the balloons (Fig. 20,Bottom), and a horizontal RPi camera to look at thehorizon. The upward camera was programmed tocapture images of the balloons every 30 seconds. Wehad two FTUs: one timer-based and one geo-fencingbased, for redundancy. Since our scientific objectivewas observation of atmospheric lines, we did not re-quire a stabilization platform (gimbal) for this flight.

Recovery The FTUs failed to cut the load line dur-ing the flight and the payload drifted towards AndhraPradesh. The balloons burst and the payload landedaround 130 km from CREST campus (Fig. 20, Bot-tom). The payload was recovered successfully withall the electronics in working condition.

Figure 20: February 16, 2014 launch: photos fromonboard USB camera. Bottom: open parachute dur-ing descent.

Flight analysis and summary Timer-basedFTU failed to worked in this launch. The maxi-mum altitude obtained was 26.916 km. We also ob-served a premature balloon burst (before reaching theburst altitude) through photographs (from the up-ward USB camera). The photograph taken by RPicamera of the sunrise, Moon and Earth’s horizon areshown in Fig. 21.

Figure 21: February 16, 2014 launch: Photos takenby the RPi camera: Top left: the Moon, Top right:the sunrise, Bottom: Earth horizon at 25 km.

3.7 Launch 7, (May 4, 2014)

This launch took place at 9.37 am with the samepayload as in the previous flight using four hydrogen-filled balloons (Fig. 22, Right). Each balloon lift wascalculated to be 5 kg (Fig. 22, Left).

Predictions for the flight The predictionsshowed the balloon path towards Mangalore-Bangalore highway.

Preparations for the flight

• Spindle Fabrication A rope spindle was fabri-cated in IIA workshop from mild steel. It

11

Figure 22: May 4, 2014 launch. Left: Balloon show-ing a lift of 5 kg, Right: 4 balloons used for launch at9.15 am.

served better purpose for the rope-ring mecha-nism. The spindle can be seen in Fig. 22.

• Temperature tests carried on April 16, 2014

– Low-temperature chamber at MRDG, CES,IISc campus;

– Temperature of the chamber: −80 ◦C;

– Time duration: 1 hour 20 mins.

This experiment was carried out to checkthe temperature properties of the Pelicancase (trademark of Pelican Products, Inc.http://www.pelicancases.com/Micro-Cases-C1.aspx)and the working of the final timer cut-down circuitat very low temperatures. Two Pelican cases, onewith the cut-down circuit and the other with thetemperature sensor (MSR temperature data logger),were kept for 1 hour 20 mins in a low temperaturechamber (MRDG, CES, IISc, Bangalore) at anambient temperature of −80 ◦C to simulate theconditions at high altitudes. At the end of theexperiment, the cut-down circuit was still working.

Payload The payload remained the same as in theprevious flights. One balloon was underfilled to actas flag balloon (Sec. 3.5 and Fig. 19).

Figure 23: Temperature test result carried out onApril 16, 2014.

Recovery The payload landed in Kadehalli nearthe Mangalore–Bangalore highway. The local resi-dents gave a call to the team and we recovered thepayload next day with all the electronics and cameraintact.

Flight analysis and Summary From the USB-cam photographs we inferred the following:

• One balloon burst 15 mins after of the launch.

• The ropes from the burst balloon got entangledwith the parachute and prevented it from open-ing.

• The FTU worked but since the ropes got entan-gled, no separation occurred.

• Two remaining balloons had burst on time.

• The remaining balloon was diffusing the gas andacted as a parachute for the descend of the pay-load.

• Maximum altitude achieved was 25 km.

3.8 Launch 8, (June 15, 2014)

The experiment was launched at 8.43 am with thesame payload as on previous flight. Each of the twoballoons were filled with hydrogen to lift the totalweight of 4.5 kg to achieve the desired altitude.

12

Predictions for the flight This time the predic-tions showed the same direction as on May 4th, alongthe Mangalore–Bangalore highway, as the wind pat-terns are generally the same for this particular timeperiod (Fig. 24, Top).

Figure 24: June 15, 2014 launch. Top: Predictedpath of the balloon travel. Bottom: Actual path trav-elled by the balloon.

Preparations for the flight

Details of the temperature test carried onMay 28, 2014

• Low-temperature chamber at MRDG, CES, IISccampus;

• Temperature of the chamber: −80 ◦C;

• Time duration: 3 hours.

This test was carried out to check the survival ofFTU circuits kept at low temperatures for extendedtime. The FTU circuit was programmed to trigger at5 different times: 1, 1.5, 2, 2.5 and 3 hrs. The FTUcircuit worked every time, and the battery voltageswere recorded (Table 1).

This shows that the FTU battery lasts at least 5cut-downs and the circuit can keep operating inside aPelican case at down to −80 ◦C outside temperature.

Table 1: Cut-down time and battery voltage data.

Cut-down Time(Hrs) Battery Voltage (V)1 7.61

1.5 7.382 7.14

2.5 6.953 6.99

Payload The payload included the UV spectro-graph, two RPi-s, attitude sensor and temperaturesensors, USB camera facing up towards the balloons,RPi camera, 3 GSM trackers (Airtel, Idea and Voda-fone SIM cards), 2 cut-off mechanisms in separateboxes (one set for 2 hrs and another for 2.15 hrs) andDhruva’s radio tracker.

Recovery The payload landed in Hulipura near theMangalore–Bangalore highway and was collected bylocal residents before the recovery team reached thespot. The residents have returned the payload to us.

Figure 25: Inside temperature of the payload boxrecorded on Launch 8, (June 15, 2014).

Flight Analysis and Summary The launch wascarried out smoothly with all the systems ready ontime. The actual path was exactly as the predictedpath (Fig. 24, Bottom). One balloon burst prema-turely within 10 mins of the launch. Second balloonburst at 10.20 am. The first FTU worked at set timeof 10.32 am. The RPi-1, which included spectrographand PiCam, and the RPi-2, which included USB cam-

13

era, attitude sensor and temperature sensor, workedcontinuously throughout the flight. The temperaturegraph for the internal sensor showed the maximumtemperature of +53 ◦C (Fig. 25). The maximum al-titude reached was 28.48 km.

3.9 Launch 9, (October 12, 2014)

The launch was scheduled for 6.30 am. Three 2-kgballoons were used to lift the total payload weight of5 kg (Fig. 26).

Figure 26: October 12, 2014 launch at 6.15 am.

Predictions for the launch The predicted land-ing location was 85 km away from CREST campus,near Savandurga forest, Karnataka.

Preparations for the launch Prior to the launch,extensive vibration tests were performed in the lab onthe payload. In addition, the usual temperature testswere carried out and the GSM trackers were testedfor the performance.

Payload The payload remained the same as in theprevious launch: the spectrograph, RPi camera, USBcamera, attitude sensor, two temperature sensors,one inside and and one outside the payload box, twoGSM trackers (Airtel and Aircel SIM cards), androidphone (with IDEA SIM) and the FTUs in two sepa-rate boxes, one set for 2 hrs and another for 2.15 hrscut-off time.

Recovery The payload landed almost 110 km awayfrom CREST campus, at predicted location nearSavandurga forest, Karnataka. Though the radiotracker stopped working 10 mins after of the launch,the Airtel GSM tracker worked throughout the land-ing and transmitted the exact location of the payload,which was successfully recovered. The android phonealso provided the location but with low accuracy.

Flight Analysis and Summary One balloonburst prematurely within 20 mins of the launch. Theother two balloons burst 90 mins after the launch.Both FTUs failed to work because of low dischargecurrent from the batteries. The radio tracker failedduring this launch. GSM trackers were used to locateand recover the payload. All the scientific instru-ments on the payload worked throughout the flightand data was successfully recovered. The maximumaltitude achieved was 26 km as estimated from sim-ulations and temperature readings.

4 Future work

The future work goals include:

• Development of two-way communication link be-tween the payload and the ground.

• Design of the payload with stabilization andpointing platform for balloon launch.7

• Development work towards increasing the accu-racy of pointing mechanism.8

• Development of the image-intensified UV detec-tor on telescopic system for spectroscopic andimaging applications to fly onboard the balloonpayload.9

Acknowledgments:

The authors are thankful to Air Force stations(Mekhri Circle, Yelahanka, Chimney Hills, Banga-lore), HAL, Chennai Airport Authorities and JakkurAerodrome for providing the necessary NOC to carryout our free-flying balloon experiments. Authors also

14

thank Dr. Annapoorni Rangarajan and students ofher lab for providing low-temperature chamber facil-ities for temperature tests and a UV-lamp for spec-trograph calibration.

References

[1] Nayak, A., Sreejith, A. G., Safonova, M. andMurthy, J., High-altitude ballooning programmeat the Indian Institute of Astrophysics. CurrentScience, 2013, 104, 708-713.

[2] Nuttall, W. J., Clarke, R. H. and Glowacki, B.A., Resources: Stop squandering helium. Na-ture, 2012, 485 (7400): 573575.

[3] Navick, X.-F., Carty, M., Chapellier, M., etal., Fabrication of ultra-low radioactivity detec-tor holders for Edelweiss-II. Nuclear Instrumentsand Methods, 2004, A 520, 189-192.

[4] Sreejith, A. G., Mathew, J., Sarpotdar, M., Mo-han, R., Nayak, A., Safonova, M. and Murthy,J., A Raspberry Pi-Based Attitude Sensor. Jour-nal of Astronomical Instrumentation, 2014, 3,1440006, 1-10.

[5] SatGuide Tracker User Manual, SatNav Tech-nologies, Hyderabad, India (2012).

[6] Marshall, B. How Helium Balloons Work.01 April 2000. HowStuffWorks.com website,http://science.howstuffworks.com/helium.htm,Retrieved 01 June 2015.

[7] Nirmal, K., Sreejith, A. G., Mathew, J.,Mayuresh, S., Ambily, S., Safonova, M. andMurthy, Jayant., Pointing System for theBalloon-Borne Telescope. 2015, in preparation.

[8] Mayuresh, S., Murthy, Jayant and Agarwal,V. K., Development of a Miniature Star Sensor.Presented in XXXII Meeting of the Astronomi-cal Society of India, March 20-24, 2014, Mohali.

[9] Ambily, S., Mayuresh, S., Mathew, J., Sree-jith, A. G., Nirmal, K., Safonova, M. andMurthy, Jayant, Development of Detectors for

Balloon and Space Flights. Proc. Golden Jubilee(XXXIX) Conference of the Optical Society ofIndia, February 20-22, 2015, Calcutta.

15