UK High Altitude Society

User Tools

Site Tools


PicoAtlas VI

How to track

  1. Radio setup
    • You need a radio that can do 70cm (430-440Mhz) and SSB (LSB and USB) - not all radios can do this.
    • Depending on how far away you are either a whip (for close) or a directional antenna (e.g. a yagi). This flight is low altitude so the range will be less then people are use to.
    • Tune to 434.000Mhz (may need to tune around as there could be some drift.
  2. Computer setup
    • Hellschreiber can be decoded by dl-fldigi (or fldigi). As it requires you to read the data it won't upload it automatically.
    • Select which mode you want to decode:
    • Chose either Feld Hell or Slow Hell.
    • This will select the mode and change the colour of the decode window to white.
    • Find the signal in the waterfall and click on it, the telemetery will start decoding.
    • Once you've got a telemetry string decoded go to Upload page. Enter in a callsign or nickname, your latitude and longitude are optional and the telemetry string.
    • The page will parse the data, check the string is valid and upload it to the server. It will be displayed on

Slow Hell

  • Slow Hell is a very slow mode - takes over a minute to transmit the telemetry. It will be better for long range decoding but you'll need to be careful to keep the red decoding bars over the signal.

Feld Hell

  • Much quicker mode - will be better close by (though you'll have to wait for the Slow Hell to finish!)

Telemetry String

  • No decimals just to warn however the Upload page will sort all that out.

Test Data

  • Here is a wav file (zipped) of both Slow Hell and Feld Hell telemetry. It 'should' be possible to play this back through dl-fldigi however I found that on Mac OS X it was slight too fast so the decoding didn't work - playing it through a loopback worked fine.


  • Arduino Mini Pro 3.3V/8Mhz
  • FSA03 Ublox 6 GPS
  • RM22 Radio Module
  • NCP1400 3.3V step up converter
  • 2xAAA lithium energizer

Flight 24/12/11

  • Very easy launch from Suffolk, filled balloon until ascent rate was approximately 0.5m/s. The radio signal was easily received by listeners including the Isle of Sheppy from an altitude of 358m. There were some problems with the php webpage for inputting the data which had to be fixed on the fly. The general impression was that it is possible to use a manually inputted setup, that Slow Hell worked much better then Feld Hell both for low signal situation but also it gave you enough time to input the data.

  • The altitude graph shows the very unusual flight path. The initial ascent rate was 0.5m/s and reached the predicted peak of 3.4km altitude, it then begun descending but then settled down at an altitude of 2.55km, floated for a short period of time and then begun to ascend again. This peaked just over 3km and then descended again. The peak is slightly sharper which might suggest that it sprung a small leak. It became difficult to track as it descended for the second time but looking at the recovered data is that it might have floated again around 1600m but then continued its descent, the last data at 880m.


Listener No. of strings
  • We also got reception reports from F5APQ, F6AGV and PD3EM especially near to the end of the flight.


  • It is possible to fly a payload on a single balloon
  • Hellschreiber payloads work, they require a lot more effort however it does mean that you can extract data in much poorer conditions. Slow Hell is probably better in these situations rather than Feld Hell.
  • It seems that these foil balloons do float, its difficult to work out exactly what has occured during this flight, theories include:
    • Natural float oscillations (larger then what has been seen in super-pressured latex flights)
    • Rain causing descent
    • Self-terminating leak.


  • Arduino Mini Pro 3.3V/8Mhz
  • FSA03 Ublox 6 GPS
  • RM22 Radio Module
  • LiPower step up regulator
  • 1xAAA lithium energizer
  • BMP085 pressure sensor


  • Try with a lighter payload to get a higher altitude
  • Explore floating again especially the unsual flight path
  • Annoy more people with making them manually input the data.


  • Turning the GPS on and off using UBX commands didn't really work - if it didn't get lock quick enough.
  • Using the power saving modes of the GPS does save some current.
    • Cyclic mode with 1s update + arduino + tx'er = 50mA
    • If it draws 50mA and we have 50% efficency with the step up circuit in theory it'll last for 10 hours ((1000mAh / 2) / 50)


Front and back of PicoAtlas6 MK2 without GPS


The launch of MKII didn't go well, the first launch resulted in the payload ending caught in a tree which took about an hour to get down. Once recovered batteries were changed and launch was attempted again. This time it was succesful however radio signal faded very quickly - nothing like the ranges of MK1. The payload was heard again at long range but very faintly suggesting that it was still working through the flight. The problem has been narrowed down to incorrect setting on the Tx/Rx antenna switch, it meant that at close range a good signal can be heard however the range rapidly drops off partially as the radio module isn't using the intended antenna.


  • Need to launch in good launch conditions - winds need to be less then 10mph
  • Range testing of payloads are useful - the error has been found on MK2 and tested on a new module



  • Arduino Mini Pro 3.3v 8Mhz
  • RFM22b
  • Pico Ublox 6 GPS
  • Lipower Step Up
  • 2xAAA


  • 2xAAA 15.6g
  • payload 10.9
  • total - antenna - insulation = 26.7
  • total - insulation = 28.4
  • total = 33.3g


  • The flight code can be found on github


Easy single balloon launch, filled at home to correct lift then taken in the car to the launch site. Once in the air it became clear that the signal was very weak - with a range of about 50km and so tracking became a challenge especially at the low altitude. Thanks to F5APQ and F6AGV it was possible to track for a couple of hours. The balloon managed to reach a stable float at 4500m for at least 2 hours before drifting out of range. It was heard again at 1830 suggesting that it was still floating but no data was collected.



  • It is possible to get a stable float with a single balloon system.
  • Launch was relatively easy by preparing everything before going out to the launch site.
  • The flight code worked well (though we weren't able to hear the transition into night mode)
  • The Pico Ublox GPS board worked well.


  • The radio transmitter had a poor output which made tracking very challenging. If it had been at the normal output we should have been able to keep track for much longer period of time. Currently being investigated
  • Once no one is able to hear the signal it is a challenge to find it again.
  • Occasionally the GPS repeated strings - needs better error detection.


Similar design to MKIII, based upon the arduino mini, pico ublox GPS and an RFM22b. Have been testing the antenna before flight to ensure that the payload is trackable as well as revising the code to avoid any problems.


  • Strengthened GP antenna with straws
  • Tested antenna over 1.2km (poor line of sight)
  • Revised / cleaned up code
  • Added measure of internal temp of arduino (not accurate)
  • Changed daytime loop to 0700 to 2000 as approaching summer time
  • Currently the GPS only uses ECO mode rather than power saving mode as its more reliable


  • Arduino Mini Pro 3.3v 8Mhz
  • RFM22b Radio
  • Ublox 6 GPS on a Pico board with chip antenna
  • Lipower Step Up board
  • 2xAAA in parallel soldered
  • Single layer of bubble wrap + tape
  • 1x 92cm foil balloon
  • Helium


Picoatlas 6 MK 4 had been ready for a number of months to fly, on a spur of the moment decision in an evening when the weather was perfect for launch (very little ground wind) it was decided to fly. The launch was from the North Kent coast with winds blowing directly north. Completed the payload weighed 29grams with an additional 2grams of lift aiming for a 1m/s ascent rate roughly (once inflated the ascent rate was measured and the helium volume altered). Simple launch the flight was picked up relatively quickly and travelled north over East Anglia. As planned at 2300 UTC it switched from continuous RTTY to slow-hell however due to a bug in the code it didn't update its coordinates. It continued to fly overnight floating at 4500m and when it switched back to RTTY at 0700 UTC was near Newcastle. It continued to be tracked during the day all the way to just north of Dundee when the batteries ran out and the signal was lost. The float was at that time at 5500m and was relatively stable - it seemed like it would have carried on floating if only we could have continued to track it.


  • Excellent flight, Pico payloads really can do long duration flights, the key points to achieving this are:
    • Single balloon (therefore need ultralight payloads)
    • Launch on calm days, winds need to be < 5mph to allow you to safely get the balloon in the air
    • Aim for an ascent rate of 1m/s to achieve float.
  • Payload worked very well, there were a few bugs (the time for RTTY switching weren't perfect, code now changed to switch back to RTTY slightly earlier, also the bug where slow-hell didn't give lat/lon/alt has been fixed).
  • Slow-hell doesn't really work, its just too tedious to transcribe the data manually though its an awesome protocol.
  • Measuring arduino internal temp is a waste of time as it generates its own heat


  • Switch to using the temperature sensor onboard the rfm22b
  • Stop using slow-hell and instead use periodic RTTY at night
  • Continue to improve power saving
  • Add lightweight solar panels to help power during the day
  • Use a complete PCB instead of wiring together as it will save additional mass
projects/picoatlas/picoatlasvi.txt · Last modified: 2012/07/06 06:46 by jcoxon

Donate Powered by PHP Valid HTML5 Valid CSS Driven by DokuWiki