Buzz2 aka Buzznik was my second attempt at getting a payload up in amongst the top of the UK high altitude table for amateur weather balloon launches. Buzz1 burst earlier than planned, though was otherwise a successful flight recovered in a field in Cambridgeshire.

Gaining high altitudes is an inexact science because of variables such as the atmospheric conditions, launch conditions, balloon variability etc. However the basic plan is to use a relatively large balloon and a lightweight payload. Larger balloons burst higher, and a smaller payload means you need less gas for the same ascent rate. Also balloons from the Chinese manufacturer Hwoyee seem to burst higher than others, though not always (e.g. Buzz1). The final variable is the amount of gas, but this is where it gets tricky. Less gas means the balloon needs to be higher to burst (because it needs the same pressure differential between inside and outside), however it also means less lift so there’s a tendency for the balloon to run out of lift and to float. A floating balloon can go a long way (the last UK floater got as far as Poland and was lost).

For Buzz 2 I didn’t want the balloon to get that far (!), but I also wanted it to cross the English Channel. The forecast winds for the launch day meant that without a float it would land somewhere in northern France, though previous predictions said South West Belgium.

Too much gas would see it land quite close to the coast (bad – I didn’t want another sea landing!), but too little could mean us giving up chasing the balloon as it headed east. Also, with the balloon covering such a large distance, I wanted it to stay up at high altitudes for a while, where hopefully the wind speed would be a bit lower, to give us a chance to catch up in the chase car. So my plan was for a lowish ascent rate of about 4 – 4.5m/s (normal would be about 5) which had a small risk of a float but maybe, just maybe, enough time to get nearby before we lost signal as it came down.

Timing then was crucial, so I prebooked a Eurotunnel crossing for a time where there were several trains within 10-20 minutes of each other in case we were early or late. I then needed to launch at a specific time (and most launches tend to be quite a lot later than planned!), then get moving in the chase car sharpish.

For Buzz2 I used the electronics taken from the recovered Buzz1. I decided not to use the same payload housing, because that didn’t have enough insulation meaning that the radio transmissions drifted in frequency very quickly as the transmitter got colder (down to -53C!). I had a foam polystyrene ball left over from the stuff I bought for Buzz 1, so I put everything in that. I removed the tiny camera (very poor quality) and excess flashing LEDs, and replaced the set of 3 AAA batteries with 4 AA batteries for longer life (in case of a float). Finally, since Buzz payloads have to actually look like space vessels (Buzz1 was a flying saucer), I arranged the aerial and ground plane wires to make Buzz2 look like Sputnik. So here’s “Buzznik”…

The day before launch I put the payload together and checked that it was still running OK and showing online, and simplified the software a little (fewer lights to flash!). I also got the car PC set up ready for tracking the payload and uploading the car position to the live tracking map. Normally I use my car but that’s curently in bits at the garage awaiting a part that is apparently now unavailable, so instead I installed in Julie’s car. Into that went a car PC for handling the telemetry and uploading the car positio, a netbook for online chat, and Samsung Galaxy Tab to show the map. The roof was then adorned with aerials for UHF radio downlink from the payload, GPS to get the car position, and 3G aerial. The latter connected to a 3G dongle and wireless access point that provided all the computers in the car with an internet connection throughout the journey (except, unfortunately, whilst in the channel tunnel!). Using a single internet connection meant a single bill for roaming in France and Belgium, and I opted for a £5 deal with T-Mobile for the time abroad.

The morning of the launch was quite windy, and normally I’d probably have decided not to launch, but I had of course already paid for the chunnel tickets! Also, the wind did die down for periods so I knew that although the balloon would be a handful during gusts, there would be times where we could measure the balloon’s lift (very important) and a time to launch without the payload being carried horizontally (as happened to Cloud2!). So final preparations were made, the car electronics all powered up and the payload started and put online. With bags and helium packed in the car we set off to the launch site just down the road, put down the ground sheet and started through the process of inflating the balloon and connecting the parachute and payload.

There were, as expected, some alarming moments where the wind tried to pull the balloon out of my hands, but I just held it as safe as I could and waited for the calm period that would follow. Here’s Buzz2 just before launch. The eagle-eyed will see what I forgot to do ….

…. the aerial wire (poking out the bottom) is bent towards me and then down. It got that way during preparation for launch and was supposed to have been straightened out. Bent aerials don’t work too well!

As I said earlier, the amount of helium added is key to the flight achieving its targets. For mine I wanted enough to give a “neck lift” of around 800g, which meant that the balloon would just lift the filler (around 150g) plus a water bottle that I’d pre-filled to 650g. However after stopping the inflation at what I thought might be about enough helium, when I measured the lift it was too high. I did try getting some helium out but it’s at such a low pressure that takes forever to do. The idea of sucking it out, and the ensuing hilarity of talking in a squeaky voice, got old after a couple of lungful’s so I opted to tie the balloon off and launch anyway.

With a large balloon and small payload the initial ascent is quite slow as the balloon is basically under-inflated and forms a rather un-aerodynamic floppy shape. Here’s Buzz2 shortly after launch:

With Buzz2 whizzing eastwards, we had little time to waste. So after dropping the helium cylinder off at home we set off to the M4, then east to the M25 and south to the M26 then M20. The balloon made better progress than us reaching speeds of 100mph and higher, but eventually it reached slower winds as it gained altitude. In the car we tracked using the computers I mentioned earlier, some of which are here:

Tracking wasn’t helped by that dodgy aerial. We didn’t know the reason at the time, but telemetry reception in the chase car was poor, and everyone else who was receiving the telemetry for me complained that the signal level was much lower than normal. I suspected an aerial issue (there’s little else to go wrong), but the proof didn’t come till I checked the launch photos later.

We eventually caught up with Buzz as we got to the channel tunnel. By then Buzz was creeping up the UK altitude record table for amateur latex weather balloon flights. This was a very exciting time! The ascent rate was high meaning that although it moved up the table quite quickly, I was convinced all along that it was just about to burst. As well as watching the altitude on the map, and coming in directly from the balloon telemetry, I was chatting online to other enthusiasts including the Australian chap who held the existing UK record. The chat room is a friendly and very supportive place with some gentle ribbing as the UK record was brought back into British hands “for Queen Elizabeth and the realm” as one put it.

Here I am on the phone to a friend who called at that time. I think this photo was taken just after Buzz2 broke the previous highest record.

During this time we were held up for security checks at Eurotunnel. I had the Yagi aerial on the rear shelf, plus GPS, UHF and 3G aerials on the roof. After the usual checks the security guy comes over and asks, “Sir, why do you have so many aerials on your car?”. LOL! He seemed happy enough once I explained.

Shortly after that, the balloon burst having achieved an official altitude of 40,986 metres (just under 25.5 miles, or 134,468 feet). Unofficially it reached 41,008 metres, but although 2 people received that altitude nobody received a complete telemetry string containing it. So now I hold the record (but probably not for long), having beaten the previous record by 411 metres. It also stands at #2 in the world record table for latex balloons. Here’s the current UK record table:

And the altitude plot over time …

So now attention turned to where the payload might land. I didn’t include a “cutdown” device this time, so the descent speed was dependent on how much latex was still attached, and whether that tangled itself up around the parachute. Initially the descent was very rapid because of the thin air at that altitude, but it did slow afterwards. Unfortunately though the received signal level was poor (that aerial again, plus the payload was spinning) and the last reading was at 23km altitude (normally I would expect someone to have received a position down to about 4km). Here’s the predicted landing path:

Our chunnel train left about this time, so we didn’t get back online (WHY don’t Eurotunnel put wifi in their trains??) till we got to France. We pulled over at the terminal petrol station while I signed up for T-Mobile’s roaming deal, then I went straight to the highaltitude IRC channel to see if there was an udated landing prediction. Sadly not, so I had a choice of spending the afternoon in a possibly fruitless search for the payload, or give up and have a day out in Bruges (our backup plan). We opted for the beer and waffles! Initially we were going to try searching on the way back from Bruges, but by then we were both very tired so we just headed for the chunnel.

We did though have a great time in Bruges – a lovely little city that’s not very far from Calais. So here are some photos from our day out.

*** UPDATE ***

This afternoon (the day after the flight) the payload was recovered in Belgium by a fellow enthusiast. It sounds like a difficult recovery, with the payload in a muddy field a long way from the calculated landing point. This is the final flight path:

into this field:

where it was retrieved by Peter of the Belgian HOWest team:

This is Buzznik ready to be packed up and sent back to me:

Also, see The Register and Newbury Weekly News

On Sunday, with great help from some friends, I successfully launched and retrieved the balloon, recovering complete video footage of the entire flight.  Not everything worked (see below), but nevertheless I’m very pleased with the result.

The morning started with at around 9am, feeding the helpers with tea, coffee, bacon butties and egg muffins.  People need proper healthy nourishment to perform at their best!  With the first essential task of the day completed, we walked down the road to the launch site at 10am.

When we got there, the village green was occupied by a cricket tournament.  There were so many cars that they were also filling the “plan B” field next to it.  I knew there was a game on but thought it was a single 6-a-side football game, not a 6-a-side cricket tournament!  After talking though the options with someone from the tournament, we opted for an area at the far end of the field, near some tall trees giving us some wind cover.

We then set up, with some of the team inflating the balloon while I and a friend built the payload.  The first problem happened straightaway, when I noticed that a wire to the computer board had broken.  I’d intended to bring a portable soldering iron, but didn’t actually do it, so we had to walk back and fix the board at home, before returning to the field (armed this time with soldering iron, just in case!).

So, the payload was then built up with cameras set up and installed first, and the computer, GPS and batteries installed afterwards.  I then noticed that the signal level on my receiver was much lower than it should be.  Not good – there’s very little that could be fixed.  We checked the battery voltage, and that this was getting to the transmitter.  No problem.  Then we checked the antenna connection.  Again no problem.  So it seemed that the transmitter itself had partially failed in some way.  Nothing here can be fixed, so I wasn’t expecting a good tracking result.

In the panic I forgot to check that the backup tracker – a GPS/GSM module – was working.  It wasn’t.  For some reason it didn’t get a GPS lock at all though the GSM side worked well for a large part of the flight.  Maybe it ended up upside-down.

With the payload finished, albeit not working 100%, it was time to attach to the parachute.  This was all straightforward and didn’t take long.

The balloon seemed to have deflated a bit from how it was earlier, so I decided to over-inflate a little in case there was a leak in the balloon rather than in the filler.  I set the amount of helium by measuring the lift.  To do that I attached a drinks bottle filled to weigh 2kg – the amount of lift I wanted.  I then adjusted the amount of helium to just hold the bottle up, then added a bit more to be sure.  This compromised the final height a little but I wanted to make sure we didn’t have a slowly deflating balloon that never burst.

Final job before launch was to tie the balloon to the parachute, and release the filling hose.  I made knots the best I could, doubling up everything and taping with duct tape.  The balloon neck was then doubled back and tied closed with tie wrap and tape.  Next – the launch!

We fed the line out gently, fortunately with no noticeable breeze so it was all very easy.  Finally, with the complete stack ready, I held it down and did a countdown before releasing it to fly straight up.  I said “bye” at the time, deliberately because with the transmitter problem I thought I’d never see it again!

I then remembered that I’d not checked the backup tracker.  This is a separate device which takes a mobile phone SIM card and contains a GPS receiver.   You call the SIM card’s number from a mobile phone, it then answers and immediately drops the call.  It then sends a text back with the location.  So I called; it answered; it dropped the call; it sent me a text.  Zeroes.  It had no GPS.  I realised then that the backup tracking wasn’t going to help, and that the poorly performing primary tracking system being the only hope.  We picked everything up and walked back to the house.  Everyone else seemed very happy, but I was quiet and dejected.  I walked back with a heavy head.

Back at the house Julie put the kettle on for tea, some helpers put the TV on for the F1 race, and I wandered upstairs to my computer to post a status report on the UKHAS IRC channel.  There, High Altitude Balloon enthusiasts gather to talk about their projects.  The channel gets very busy during a launch where people are helping to track the balloon.  The way this works is that they set up a radio ham receiver to a PC, tune in to the low power transmission from the balloon, and decode the signals to reveal the balloons position (longitude, latitude and altitude) plus speed and direction.  All received data is sent to a central server which drives a map and can also send its data to Google Earth, so anyone with an internet connection can watch the balloon’s position live.  Pretty impressive!  However, I wasn’t expecting much because of the transmitter fault I’d seen earlier.

So, I logged in to the IRC channel, and loaded up the map.  To my surprise the map had the balloon’s location on it!  I then found out that the position had been decoded by someone in Yorkshire!  This meant that my transmitter was actually working properly – there’s no way it could be faulty and still working at that distance.  I was amazed and relieved!

After a while, other receivers showed up on the map too.  One was in Northern Ireland, about 320 miles from the balloon!  Another was in Holland.  Unbelievable!

The tracking was working great, with the balloon moving steadily upwards and steadily eastwards.  This continued until the balloon got to a certain position east, when it suddenly jumped from Berks to Essex!  Some of the UKHAS guys soon spotted the error, which was a bug in my GPS translation code.  Apparently I wasn’t the first to fall into that particular trap, and it’s quite a common problem.  Fixing it in the data was easy, so I had a way of translating the transmitted position into the actual position which I then displayed on Google Maps.

Bug aside, the tracking was working so well that I forgot about sending out a chase car to the predicted landing site.  So when I did remember it was too late to get to the area before the balloon payload landed.  We all went out, in two cars with one set up for tracking.  We took different routes to the same area.  I sat in the back of the tracking car, using a netbook to decode the received data, and a magmount antenna on the roof of the car picking up the signal.  I had a really good signal for a large part of the journey.

As we got closer though, the balloon burst and fell out of the sky.  I could see from the data that it was falling quite quickly.  Because of that we lost the signal with still some distance to go.  We got to the expected landing area, but found no signal at all.  So I decided that we should go to the last known position of the balloon.

On the way the receiver was just playing us static.  Not good.  Then, for just a brief moment, I heard the familiar warble that the balloon transmits.  We pulled over as soon as we could, and I got out of the car to connect up the more sensitive, and very directional, Yagi aerial.  That then picked up a clear but weak signal from a specific direction only.  So I knew from that what direction we should drive in.  The signal seemed too weak to decode, but anyway I connected the receiver to my netbook and was delighted when the decoder software displayed the data with very few errors.  We now had a better GPS position to aim for!  The data showed that the onboard computer had rebooted, presumably because of a heavy impact.  It also showed that the GPS had stopped updating (presumably because the payload was upside down or perhaps on its side).  There was a very good chance though that the GPS position was pretty close to where the payload actually was.

So, back in the car, and using the mag-mount antenna again.  We entered the GPS position into the car sat nav, and headed there with, for the first time since launch, the belief that we would probably be able to retrieve the payload.  As we got closer the signal came in and got stronger.  We found somewhere to park within a couple of hundred yards of the GPS location, then got out and walked down the hill towards where the GPS said the payload was.  I had the receiver connected to the Yagi antenna, waving it around to direction-find the payload’s position.

All around us were trees.  Very tall and quite dense trees.  It seemed very likely that the payload would have landed in a tree rather than on the ground.  I pointed the Yagi upwards and the signal got stronger!  Not good – there would be no way of retrieving a payload from up there.

It soon became clear that the payload was to the left of the road, and after a while the signal got to be very strong and apparently directly to our left.  It then got weaker and was behind us, so we turned round and went to the central point.  This was next to a house and fortunately the owner was in his garden.  I approached and gently explained why 7 geeks were very interested in the contents of his garden, and why we had an aerial and a warbling radio receiver!

We wandered around trying to get the location from the Yagi.  We had a general idea of the location, and then I thought I heard an extra high pitched tone as well as the usual warbling from the receiver.  So I switched the receiver off, and could still hear the whistle.  No, it wasn’t my tinnitus, this was real!  The others could hear it too, but the retired house owner couldn’t.  We walked around trying to decide where the whistle was coming from, then one of my helpers Rhona spotted the Hi Viz tape on the payload.  It was in next door’s garden!  We were then informed that we probably wouldn’t get a good response by going to see the gentleman’s neighbours, so an alternative approach was used

The payload had obviously hit the ground quite hard.  The insulation was battered but everything inside seemed to have survived unscathed.  First I checked the camera for photographs – only 33!  It had switched off before launch!!  Next I checked the video – yes, that had worked.  It had 14GB of video and played back fine.  We then chatted with the house owner for a while.

After showing him some video and went back to the chase cars to celebrate with lunch and some champagne. I removed Commander Buzz Lightyear from the payload so that he could join in!

I then put the camcorder SD card into my laptop so we could view the video.  We saw some quite stunning views, and I couldn’t wait to get back home to see it more clearly. The payload and the video coverage both clearly showed why the descent was too rapid.  The balloon did not shred fully and about 600g of the original 1000g was still attached to the line after the burst.  This line and the latex wrapped themselves around the parachute cords, closing the parachute more and more till eventually almost the entire cords were tightly twisted together.  Fortunately I chose a parachute that was larger than theoretically necessary, and this helped to prevent the payload from descending even faster.

So, although not everything worked, we did get complete video coverage and we recovered the payload.  I learnt a lot for next time.  There are many things I will do differently, and hopefully I will be in less of a panic if things go wrong!  Panic just made things worse.  I know that this experience will help a lot when I plan and execute my next HAB project!