As of today (16th Feb 2012) I have moved the blog to my new personal website –

All future blog updates can be found over at


Data, Data And More Data

Following last Saturday’s successful flight of HABE 2 and the new addition of a “black box” style SD flight logger I have collected a wealth of data. Having spent some of yesterday writing a C command line program to parse through the log file, earlier today I managed to extract the information and pop it into an excel spreadsheet.

Having plotted some fancy graphs etc the results where fantastic. All of the graphs offer a unique story and contain lots of useful information – a great success. I’ve decided to distribute the excel spreadsheet containing all the raw data and graphs etc so people can have a look at all the information. It may prove useful to some. The excel workbook can be downloaded here (xlsx) and an older (xls) version here.

I’ve posted the graphs below for all to see:

A few brief notes on the graphs:

  • The temperature graphs have the ascent on the right (slightly warmer) and descent on the left (slightly cooler).
  • It can clearly been seen that the coldest temperatures faced reside in the 10-12km altitude bracket. As the balloon rises above this altitude it actually gets warmer contrary to popular belief! The science behind this: the upper layers (stratosphere) absorb UV radiation far more efficiently than the lower layers, therefore this acts as a heating effect – it also prevents extreme UV readings on the surface of Earth.
  • The solar panel readings are extremely interesting and probably contain the most information. On initial observation I didn’t see that much information, but on closer observation there is a lot said. Firstly – why the mass scatter of points everywhere?! Well, I believe this is due to the swinging of the payload (due to winds); as the solar panels were mounted on top of the payload and the sun was relatively low in the sky (Winter), this would have caused differing levels of light falling on the solar panels, indeed at some point they were probably in shade hence the many near 0v readings. Secondly you may ask, ok, why isn’t that the case for the lower altitudes (<10km)? The answer to this – the payload was in the clouds through this part of the flight. Clouds act to diffuse the light from the Sun and therefore are bright all around inside them – so whether the payload was on it’s side or pointing upright the light falling on the solar panels was much the same. On close inspection (you made need to download the full quality versions – excel workbook) the readings seem to fall into rectangles at lower altitudes – I believe this is the case because of the different cloud layers. From the photos taken by the onboard camera it’s clear that there were multiple cloud layers and each layer would’ve had unique properties – hence differing voltage/power constraints. Interesting!

I have annotated the first altitude vs SP1 voltage graph illustrating the different “rectangles” – take a look:


I’ve also done the customary 3D Google Earth plot of the flight path. You can download the KML file here and have a play about with it in Google Earth.


Last but not least, here are some interesting facts about the flight:

Max Internal Temperature: 19.6
Max External Temperature: 16.3
Min Internal Temperature: -26.8
Min External Temperature: -63.1
Max Altitude Recorded: 29958
Max Voltage (SP1): 7.67
Max Voltage (SP2): 8.96
Average Voltage (SP1): 1.88
Average Voltage (SP2): 2.51
Max Power/mW (SP1): 406.28
Max Power/mW (SP2): 419.23
Average Power/mW (SP1): 44.58
Average Power/mw (SP2): 58.92

HABE 2 – A Success!

It certainly has been a hectic few days – from the moment, Wednesday afternoon, I obtained permission for the weekend it’s been non stop! I am delighted though that HABE2 was a great success! I will try and write a full rundown of the flight in the next few days, but for now I’ll just give a few quick facts about the launch.

After waking up Saturday morning, I checked the prediction and nothing had really changed, so I gave the launch the go-ahead. The extremely cold conditions slowed down the launch setup due to very cold, numb fingers! In fact, for the last 10mins of setup/launch it started snowing. Despite the slower than usual setup everything was going fine until it came to the last part – the filling of the balloon. We had a faulty adapter! No helium whatsoever was flowing out! This was devastating – with the payload all ready to fly and transmitting it’s location perfectly it was looking like the launch would have to be called off. A quick chat to the other guys on the IRC channel, my fellow HABer/ist (not sure the correct phrasing!) in Worcester, Will Duckworth, came to my rescue after he drove his helium adapter to the launch site! A big big thanks to Will who saved the day! His HAB project is here.

So finally, about 1hr later than planned HABE2 took to the skies! The flight lasted about 2.5hrs and was recovered from a horse field near High Wycombe. At this point I’d like to mention how accurate the predictor is – yet again pinpointing the flight path to within a few km – incredible. The highest reported GPS position was 29, 958m (just 42m shy of the big 30!) – but, if the refresh rate of the GPS and the length of the cord connecting the payload to the chute and onto the balloon is considered I may well have just hit the 30km mark. However – this launch was never intended to go as high as my previous one; I needed a quick ascent and descent (which meant compromising on altitude) due to the predictions.

The camera worked perfectly resulting in some fantastic photos and a couple of short videos! The best can be seen here in my flickr set. See if you can spot the moon and aircraft vapour trails! I’m very pleased with how these turned out.

All the other main components worked great, the SD “black box” style logger has resulted in a significant amount of data being logged – I am now trawling through it all and I hope to have some nice graphics soon(ish). One interesting fact so far – the coldest temperature recorded during the flight was -63.1C!! That makes our -5C temp look strangely warm!

Keep tuned as in the next few days/weeks I’ll be providing a full analysis of the launch and hopefully producing some nice graphics.

For now – I’m very pleased with how it all went considering what could’ve gone wrong. If you have any questions please do drop me a comment below or email me or pop on the IRC #highaltitude

HABE2 Launch Tomorrow!

Well – what a frantic few days it has been… I’ll give a brief run through as the payload still need completing!

Wednesday afternoon: I was on the IRC and David Akerman (daveake) posted a link for the prediction for his launch this Saturday. It looked like a good prediction, so I decided to run one from my launch site; no intention at this stage to launch. It turned out to be a great prediction, actually landing in this country (the previous weeks had been landing in France or beyond!). There were two big issues at this point: 1) The payload wasn’t even completed and there were still some nasty bugs present in the code. 2) I didn’t have permission to launch.

After a quick call to the CAA I managed to obtain permission for this weekend (4th/5th Feb) – many thanks to David Miller at the CAA for getting this sorted in such short notice. As of yesterday evening, I now have permission to launch. The next big hurdle was getting the payload completed on time. Despite the huge number of hours I had put in the past 2 weeks, it still wasn’t stable – albeit a rather buggy beta version.

Over the last day and a bit I’ve fixed all the bugs I can see and built most of the payload stuff. Yesterday I ran a full systems check and everything worked apart from the camera – oh no, not this again… (you’ll probably remember that HABE1 was plagued by complete camera failure). This however, wasn’t linked to the HABE1 bug. Initially I thought it was a software issue as I’m running a modified version of CHDK that allows me to switch between photos and video (risky, I know..). However, it appears to be it was actually a failure of the alkaline batteries due to the cold temps outside. This meant I only got around 100 photos (only the 1st 30mins of flight) before the batteries conked out. For the flight I will be using lithiums which have a much better tolerance to cold temperatures thankfully… As I’m writing this the camera is outside with lithiums in running a final check. If it’s all positive then it gets the clearance to fly from me!

Lastly, I’ve just finished soldering all the solar panel logging equip up; I just need to pop the flight board in and solder up all the antennas (not an easy job at all). After – final checks, then time to pack the car for tomorrow.

The current prediction is ok – not great – but ok. I’ll be monitoring this throughout the day and tomorrow morning.

What’s new/updated on HABE2:

  • Software rewrite – this has resulted in a fair few bugs as I didn’t test as I was going along (mistake), though it’s resulted in a more efficient and readable software
  • Swapped old temperature sensors out for newer high precision ones
  • Uplink is present – simply to test range (let me add at this point I’m not expecting great range)
  • Camera now takes video as well as photos
  • “Black box” present on the flight – SD card logging system should log everything for review after the flight
  • Cutdown module – this can be activated via uplink and is automated by the flight computer should the payload stray outside a “safe” area
  • Solar panels – 2 high performance small solar panels present; I’ll be logging data using two different loads.

Launch details:

  • Launch site: Ombersley, Worcester.
  • Estimated time of launch: 11am GMT (setting up around 10am)
  • Flight time: approx 2.5hrs
  • Check here, twitter (@adamcudworth) or the IRC for the most up to date information regarding the launch. If you would like to come along to the launch please drop me a line either by email or the IRC (#highaltitude and my nickname: cuddykid).
  • Track the balloon live during the flight here
  • I would really appreciated any help – whether it’s turning up to help out with setup/launch, or helping track by listening in – thanks!

Hopefully the launch will be underway within 24hours! Over and out…

For the most up to date info during the launch, check the live tracker, my twitter or the IRC (#highaltitude). 

CAA Permission Requested

Just a brief update regarding HABE 2:

I have just sent off for permission to launch the balloon from the CAA. I have specified the last weekend in February and the first 2 weekends in March. I have two main doubts currently:

1) Will I complete the payload on time. There is still a LOT to do and none of the code that I have written has been fully tested yet. If a fairly major bug is present in the code then it will require a significant amount of time debugging.

2) Wind conditions – It’s a very windy time of the year and the high altitude winds tend to blow in the wrong direction (west to east) at this time of the year. Hopefully they’ll be fine for given dates *fingers crossed*.

A note on the side; if you haven’t already noticed there is now a brand new page called “Checklist” where you can track my progress. “Check” it out ;)…. (it’s in the toolbar at the top if you can’t find it!).

HABE 2 – Update

It’s been a long time since I last updated the blog so here’s a quick update as to what’s happening with HABE 2:


The code for the next flight is being mostly rewritten aiming to make it more efficient and easier to read/adjust where necessary. A few weeks back I spent a significant amount of time on the code and as a result it is near completion. There are still however a few more additions and tweaks needed.


Whilst the core tracking and telemetry system of HABE 2 will be the same as HABE 1, there are quite a few new additions that are going to be present on the next flight:

  • Attempt at an uplink – using some easyradios an uplink is going to be attempted. Note: I am not expecting great range, this is merely a test to see just how far an uplink can be maintained with a tiny 10mW transmitter
  • Solar panels – 2 high performance small panels will be flown purely to collect data to determine the plausibility of powering future flights using them
  • Tiny keychain video camera – an extremely cheap video camera will hopefully capture some video footage of the flight (although I’m not expecting anything great – at all!)
  • A570 video mode – using the CHDK software hack I have enabled easy switching between still and video mode. Plans on when video mode will be switched on are currently undecided
  • Cutdown module – whilst I have tested a couple of cutdown modules I’ve designed using nichrome wire I have yet to settle on a design and finalise plans here. A cutdown will be deployed for 2 reasons: 1) many flights with Hwoyee balloons (ones I’m using) have developed a mysterious floating capability resulting in the loss of payloads as the balloon floats off to another country, using a cutdown with a “geofence” will allow me to cut the payload and parachute away from the balloon if this occurs. 2) HABE Glider will require a cutdown module and therefore this flight will act to test out the chosen mechanism

Along with these additions the bugs from HABE 1 have been patched (hopefully!). Most of the systems are currently in prototyping stage (on a breadboard) and are yet to be soldered for the flight. There is the chance that some of the above won’t make it for the next flight and/or some other things may be added.

My aim is to launch HABE 2 before Summer. Current target is around Easter, perhaps before if I get a move on. Though, as always this depends on a range of uncontrollable factors: CAA permit, weather…

Over and out for now

HABE 1 – What Worked, What Didn’t

On the whole I was very pleased with how my 1st high altitude balloon flight went (despite the disastrous camera failure). Below is a summary of what worked and what didn’t.

Things that worked/I will not be changing in a hurry:

  • Telemetry system – the gps/arduino/ntx2 worked fantastically despite my doubts due to dodgy wiring and a botch job of the 1/4 wave ground plane antenna on board.
  • Backup tracker – worked exactly as described and I can see it proving critical to retrieval of future flights – a must have for every HABist.
  • Payload box – with an internal temperature sensor the temperature inside the payload was monitored constantly and the polystyrene worked very well at protecting the components from the approx -50C outside (got down to a minimum of about 0C inside).
  • Parachute – the 36″ chute worked very well with my 900g payload – although a initial very fast descent (to be expected due to lack of air resistance), the descent rate near the ground was around 8m/s.
  • Balloon – 1000G Hwoyee balloon went even higher than expected (seems to be a common theme with recent Hwoyee flights). Balloon was easy to fill and tie.
Things that didn’t go so well:
  • Camera – despite the extensive testing carried out before hand and the great shots it took on the ground, a last minute code change (changing a “6” to a “10”) resulted in the integer (1 000 000 rather than 600 000) being too large for the camera to accept. I had made the change to make the camera “sleep” (pause) for longer before starting to take photos as in testing the day before the flight it showed that the 4GB SD card may reach capacity very near apogee. However, I’ve subsequently found out that whoever wrote the code for the CHDK coded in an arbitrary limit of 999 999 for integers, therefore my value of 1 000 000 was 1 out from working. This actually translates to being 1 millisecond (1 thousandth of a second!) too long!! I have definitely learnt NOT to fly code that hasn’t been previously tested even if the most minor of changes is made. Hopefully on the next flight I will get some photos.
  • Buzzer – when attempting to locate the payload a buzzer is often very handy in case of landing in long grass etc. When the payload was found the buzzer was not on (which would’ve made locating it very difficult had the landing been in a worse location). After some debugging it appears as though one of the wires from the buzzer was removed from it’s slot during duct taping (during launch preparations). This will be corrected on future flights.
  • External temp sensor – an odd thing occurred with the outside temperature sensor. As the payload rose in altitude the outside temperature dropped, as expected, until it reached about -3.2C. Then it remained around -3C, fluctuating by a couple of 0.1Cs every string received. I know for a fact because of previous launches others have done that it gets a lot colder than -3.2C on the way up! As of yet, the reason behind this has not been exactly pinpointed, however, there are a couple of suggestions. The more likely suggestion is a bug in the code that prevented values being sent below -3.2C, this would seem likely as there is a lot of temperature conversion code – but, after a quick scan through the code a bug doesn’t seem to be present. The alternative suggestion is that I managed to fry the sensor in the testing process – it initially took a while to get the sensors working and I might have put the volts through the wrong pin – but, this also is strange as the internal temperature sensor experienced the same dodgy testing of mine and worked perfectly (though didn’t experience temps of -3.2C). Further debugging will continue…

HABE 1 – Partial Success!

Last Saturday (02/07/2011) afternoon after 2 years in the making, HABE 1 finally took to the skies. A quick debrief is given below:

After arriving at the launch site, a local farmer’s field, the essential equipment was unpacked and set up. I brought along a camping table, which in the end proved to be a great decision as it provided me with a clean, flat surface to place the laptop, radio and payload while getting prepped for launch. The day before I had decided on a launch routine – essentially 4 stages: turn everything on inside payload and fix the lid firmly on; attach string from the payload to the parachute; fill & carefully tie the balloon and launch!

A few weeks prior a guy called Will Duckworth who lives very near me got in touch and told me how he was doing a similar project and wondered if he could come along for the launch and tracking. I was amazed that there was another HABist in the area as I was firmly in the belief that I was very isolated with my Ballooning project being situated well away from the HAB hub of Cambridge! Will was a fantastic help throughout the day – from expertly tying knots (which I’m useless at) to helping track – so a big thanks to Will. Hopefully I’ll be around for his launch and in the future we can possibly turn Worcester into a HAB hub!

Right, back to the flight details. After the launch (which took around an hour with all the preparations), the initial telemetry was coming through with many errors in, however this was quickly identified to be the result of the yagi antenna pointing in the wrong direction (panic over)! Tracking went great throughout the flight with people managing to listen in from Northern Ireland & near Aberdeen – with just 10mW of output I was extremely impressed. So – telemetry was a success. The payload initially drifted over towards Stratford Upon Aven but in due course swung back to head right overhead (about 300m away!) the launch site and then on to near Tenbury where it again swung back around and landed near Great Witley. Map of flight below:

The balloon was filled to burst at 33km up however we watched in amazement as it just kept going higher and higher – eventually reaching 35,824m (the highest reported alt). This placed HABE 1 the 2nd highest payload of all time in the UK (according to UKHAS records) however, we were subsequently pushed down to 3rd place on Sunday when Steve Randall’s payload reached about 180m higher than HABE 1.

The parachute worked a treat as the payload had a reasonably controlled descent (falling about 8m/s near the ground). Once the payload started dropping below ~10km listeners dropped out as it disappeared over their horizon. As we were driving to the predicted burst site, there were periods were telemetry couldn’t be received so we had to make rough guesses where the balloon was until we could pull over safely and decode a few more strings. Once the payload was around 2km up – I sent an sms to the backup tracker (which I’d bought off eBay for about £50); not expecting it to work great, I was surprised when almost instantly I received a text back with it’s exact location! I continued texting it until the payload had landed in case cellular signal dropped out at lower altitudes (the area where we where had very poor signal).

With a location from the backup tracker I plugged it into an Excel spreadsheet I had made to spit out the co-ordinates in a google maps friendly format. On powering on tethering from the iPhone, it quickly became apparent that there was extremely poor signal. I had thought of this, so no problem – a Garmin Etrex was on hand. One problem though – I hadn’t used it for about 2 years and after about 10mins still couldn’t figure out how to use it! So, despite a clear signal (surprisingly) from the payload and a matching location from the backup tracker we had no idea where it was (apart from within a 2 mile or so radius around us!). After some driving around and trying to triangulate the position using the directional feature of the yagi we bumped into Will who’s computer had just died! However, he managed to remember a co-ordinate obtained just before it conked out (impressive with no pen & paper!) – on comparing co-ordinates they matched so we were fairly sure that we had the location; but we just didn’t know where it was! After heading in the vague direction where we were getting the best radio signal, cellular signal fortunately improved slightly – just enough so we were able to load up google maps and get a location. Screenshot (green arrow is location):

When we obtained the above location we were infact situated just below the “A” marker (on the road that the marker is on). We decided it was best to head south and then west and up a tiny little track which led us fairly close to the payload. After some hiking off road (up & down, up & down… sheep fields and over barbed fences etc) the payload was spotted – “that doesn’t look like a sheep to me” was the call when the white payload box was spotted!

Remarkably – it landed with no damage whatsoever to the payload box (possibly helped by the fact it landed in thick/high weeds). The parachute was nicely spread out and the balloon had shredded cleanly. So, very pleased, we headed back to the cars with the payload.

On arrival back at the cars, the payload box was quickly opened (after extensive removal of duct tape!) – the SD card from the camera was removed, inserted into the computer, then…. 3.94GB available. Oh dear.. 0 photos had been taken – what a disappointment after such a fantastic flight.


This article will be followed up with analysis of what worked & what didn’t.