Pegasus I

MISSION COMPLETE (15/10/05) Pegasus I reached 20295.200m (66585ft) and travelled 63 miles. Took over 600 photographs






  • 1 x Gumstix Waysmall Linux based embedded computer
  • Nokia 5110 (with small battery)
  • Gps Reciever (Sirf II based and rs232)
  • 2 x MinDIN8 connectors
  • 1 x Nokia Dau9p Cable
  • 2 x Digital Cameras (1 Jam Cam (640×480)and 1 Premier DS-3080S(3Mpixels))
  • 2 x AAA battery holders (1 x 3AAA and 1 x 4AAA holders)
  • 9 x AAA lithium batteries (3 for gumstix, 4 for gps and 2 for Premier DS-3080S)
  • 1 x 9V lithium battery (for Jam Cam)
  • Polystyrene for payload container
  • Duct Tape
  • 2 x 256MB MMC/SD Cards


Warning most of the information here is from memory so it may not be perfect - if in doubt google is your friend (If i can find the information so can you!) Though of course if their are any questions get in touch and I'll do my best to help you.

Tracking System

GPS —→ gpsd – Gumstix (logs) – gnokii —→Nokia 5110—→ send sms to second phone containing coordinates and altitude


The GPS and gumstix communicate by serial - the gumstix is passive and therefore only receives information. 2 wires are required RxD (from the gumstix's perspective) and GND. First these have to be identified - I cut off the serial port off the gps and then identified the correct wires by using the female serial connector, an led and battery and a diagram of pinout for a rs232. By completing the circuit between the specific pin and its wire it was possible to identify the right wire. These are then soldered to the appropriate pin of a MinDIN8 connector (salvaged off a older mac printer cable) and plugged into ttyS0 (one closest to the power plug.) The pinout for the gumstix female MinDIN8 is here

Most serial GPS units are designed for laptops and so use PS2 ports to draw their power - for Pegasus 1 we had to convert this to batteries. Edward found an old battery case that took 4 AAA batteries in series which we decided was sufficient to power the GPS. The GPS unit we used had 4 wires coming out of it (RxD and GND have already been identified) and so the other two are +ve and -ve. Using the system of trial and error with the severed PS2 connector and a pinout it is possible to identify which are the correct wires and link these up to the gps. If the light turns on then you've got it right - if not STOP immediately or risk shorting out the gps unit (I did this and so had to get a new one :( )

Now that we have the GPS unit on batteries and connected to the gumstix power by using the command:

stty -F /dev/ttyS0 speed 4800

cat /dev/ttyS0

You should get a stream of data (the gps unit when on continuously sends data) - use Ctrl C to exit the stream.

Now using GPSD it is possible to extract information. Either you can telnet in to port 2947 or use the small C program gpsdlog (I adapted this from someone else and will post the source and credit next time I boot into Linux) to take the long, lat and altitude and dump it into the file gps.log. See the script gpslogger on methods of executing gpsd.

Nokia 5110

Now that we have the gumstix able to record its position it now necessary to get it to send the infomation on. As I wasn't going to use HAM radio I instead settled for GSM and found an old Nokia 5110, a slimline battery from a broken 6210 (I think…) and a serial cable from Ebay.

Phones work at TTL level while the serial ports they are connected to are at rs232. Therefore inside the cable there is a little bit of circuitry that converts the signals. This requires a small current which it draws from the serial port. Now the MinDIN8 on the gumstix aren't full serial ports only having RxD, TxD, RTS, CTS and GND. The phone cable draws its power from DTR and so to provide this power I used one of the spare connections on the serial port that the GPS connected to. As it can act as a GPIO I (with a lot of guidance from Craig on #gumstix on freenode) set it high therefore powering the conversion circuitry. Check the phone script for more information.

Now that the phone circuitry is working gnokii will work and so will allow you to connect to the phone. Therefore by running the command:

tail -n 3 /gps.log | /root/gnokii –sendsms xxxxxxxxxx

the last 3 lines of the gps.log are taken and sent as a sms to the number xxxxxxxxxx.


Gumstix – GPIO pads —→ Camera shutter

The Waysmall Gumstix has 20 contact pads which can be turned on and off by using echo and /proc/gpio/GPIOxx. First remove the plastic shutter button on the camera and solder two wires to the push button so that when a current is passed through it simulates the button being pressed. The connect one of these to the GND pad (pad 16 - with 1 being the pad next to the power plug) and the other to one of the other pads (I think I used 4). Pad 4 can be turned on by the command:

echo “GPIO in set” >/proc/gpio/GPIO29

and off by:

echo “GPIO in clear” >/proc/gpio/GPIO29

I wrote a simple shell script that turned this pad on and off every 30 seconds and connected both cameras up to the same pad (with an led that was then embedded into the polystyrene casing so that I knew that the cameras were working.) here

Packaging it all up

I found two pairs of polystyrene packaging that fitted together and after a bit of brute force adaptation cut windows for the cameras and worked out a way to arrange all the components. Originally I planned to have the GPS unit inside the packaging though last minute tests showed that it struggled to get a position lock so as it looked well sealed and protected I put it on the outside on top of the payload. The plan was to use all 4 pieces of polystyrene however the top section broke so we launched with the two smaller main pieces and a single big section on the bottom. Also we made a wireframe (made out of coat hangers) to go round the outside of the payload - to keep it together and also to create a place to attach the balloon and parachutes however testing with the parachute by throwing the payload off the top of my stairs showed that the main payload would break through the bottom piece of polystyrene. Therefore in an eleventh hour hack I made a new wireframe (again out of coat hangers) to go round the smaller polystyrene box.

Pictures of the finished payload (before the last minute changes.)


Most of the technical stuff involving the parachute was sorted by my friend Miki who had experience in model rockets. Over one weekend we made a parachute out of bin liner to test how they worked and spent most of the time throwing the parachute with appropriate weights out of windows to see if it would work. As proud as we were with our parachute we decided to not take the risk and order a proper model rocket parachute. With a bit of research we found out that a 36” parachute was appropriate for our payloads weight.

The plan was to not have to deploy the parachute but instead have it ready to open at launch. The main cord that linked the payload to the balloon ran throught the hole in the centre of the parachute and at this point (though so that the chute's cords were loose) there was a knot and a plastic milk bottle top (with a drilled whole through the centre) resting on it to stop the parachute slipping down. On ascension the parachute would be kept closed but when the balloon burst there would be no upward force and so the parachute would open, the cords tighten and move up the string away from the bottle top.


All this refers to the laws of the UK!

You are allowed to launch a meterological balloon in the UK as long as you get permission from the CAA (Civil Aviation Authority). They are exceedingly helpful - I sent an email to them and was directed to the correct person to talk to. He sent me the forms and once I had filled them in and sent them back he then coordinated with all the relevant organisations. You don't need to specify a launch date as they are able to give you a launch window. Once everything is sorted you are sent an official document that gives permission for the launch. All that is then required is for you to contact the CAA at least 48 hrs before the launch and if the permission document asks for it some other organisations. Due to the close proximity of the launch to Wattisham Air Base I had to also get in contact with them. Finally before the launch a NOTAM is issued and then its no turning back!

In regards to transmitting from a mobile phone at high altitude - actually no data is sent for most of the flight - the mobile phone used doesn't get reception above 650m so is unable to transmit and only managed to get reception back once it had hit the ground.

Setting Up

The Day Before

Helium for the launch was from here . They use BOC to supply their helium and I got a V sized canister (which was definitely sufficient). There a few difficulties in delivery due to living in the country however finally the gas arrived at 3.00pm the day before the launch.

The balloon itself was got from ebay (its amazing what you can get there) as was a 200g Totex balloon. I actually got two just in case we messed one up while filling them up.

Everything was got together the night before (while I quickly hacked together a new wireframe) in an attempt to reduce the things to do in the morning. However nothing was yet attached together as the payload still had to be sealed and the wireframe put in place.

Launch Day

Got up at 7 giving me 3 hours to get everything ready. I prepared the payload while my father prepared an adapter to convert the small nozzle of the helium cylinder to a large balloon. Borrowing a design from ballooning groups in the US we found some plastic piping and threaded a hosepipe through this and sealed one end. (If you want more information get in touch and I'll take a picture as we still have it). He also drilled the whole in the bottle top to act as the support for the parachute. At this point Edward, Jane and Tina arrived and so we split up. I continued to prepare the payload while the others went outside and begun to inflate the balloon (this is around 9.30). To get the correct amount of helium I prepared a bag full of tins that weighed the weight of the payload plus another pound of lift. So in this case it was 3lb in total - when this bag was lifted off the ground we had got the correct amount of helium (Apparently they didn't realise that I had added the extra lb of lift so added more helium - this may have saved the balloon - see later)

I powered up all the components either by turning them on or adding batteries and taped the payload up (I however wasn't getting any messages so I thought something was wrong so had to open the payload up - as I did this the messages came through!). Next I added the wire frame and taped that in place (I'm a firm believer in duct tape) and added a sealed plastic envelope contain a message explaining that the payload was harmless and my contact details if found (and the chance of a reward).

I took this outside and with the help of Peter tied the parachute to the payload while the rest of the people using a tablecloth kept the balloon from escaping. We decided to attach the payload to the parachute and then attach the balloon to this junction - therefore if something went wrong the payload would still have a parachute. We set it up as I explained earlier and then duct taped the knots to make sure it would hold. The balloon was attached by a single cord through parachute. With the balloon nozzle after tying it off with first nylon and then a plastic grip. The free end was then looped up and tied again and it was the loop formed that was used to attach the main cord from the payload to the balloon.

Once everything was attached we first released the balloon but held on the the main cord - slowly allowing more string through - then the parachute and then finally I let go of the payload and watched it gracefully but rapidly disappear upwards.

Search and Rescue

Once the balloon was released I received two sms messages - one at about 350m and then another at 650m and then silence for 2 1/2 hours. After a hour of packing up and then sitting around Tim and I decided to start driving towards where we thought it would go (using the two sets of co-ordinates that we had received and drawing a straight line) After driving for a bit we decided to turn back and just as we were a message came through. After the initial excitement we read the message which said that the gps coordinates were… NULL. The gps hadn't got a log but we knew that the payload was in once piece. We went back home and attempt to use an internet based phone locator service however due to privacy worries the phone has to acknowledge thats its being traced. So we though perhaps by phoning up vodafone they could help - as I was on hold a message came through with some coordinates (obviously a satellite had popped above the horizon). Put these into Google Maps printed out the location - jumped in the car and drove 80 miles to go and retrieve it.

Once we got to the road where we thought the payload was and found them to be fields. First using a handheld gps lined up the latitude and then walking along lined up the longitude. However this spot was a cabbage field so Tim and I spread out and he spotted something that looked red in the adjacent field. We ran over to find the payload totally intact and took this back to the car where I quickly opened it up - turned everything off. We then set off and I logged into the various devices via my laptop and downloaded all the pictures and log files.

There you go - from launch to retrival took pretty much 6 hours. I hope that this infomation was helpful. It was mainly from memory so it might need a bit of adjusting, now on to Pegasus 2…

Time vs altitude

Thanks to:

  • Construction: Edward
  • Miki
  • Dave
  • Tom and Helen (good times looking for a reed relay?)
  • Craig at Gumstix
  • The Gumstix Mailing List and #gumstix on freenode
  • The Launch: My Parents (Anne and Tim)
  • Peter and Charlie
  • Edward
  • Jane
  • Tina

Any Questions contact

missions/pegasus1.txt · Last modified: 2010/05/14 07:24 by jamescoxon
Except where otherwise noted, content on this wiki is licensed under the following license:CC Attribution-Noncommercial-Share Alike 3.0 Unported
Recent changes RSS feed Donate Powered by PHP Valid XHTML 1.0 Valid CSS Driven by DokuWiki