The Launch Part 1 – Mistakes Were Made

[First off, I’d like to thank the guys at UKHAS and HABHUB for all their support during this project. Without them, this would have been more than a disaster.]

[Also, this first part will only contain details on the preparation before launch day itself. Part 2 will contain all the pretty pictures and the story of the flight, but read on if you want to know about all the many things that went wrong beforehand.]

On the 24th June 2015, we finally launched our high altitude balloon, and it was, quite frankly, an absolute miracle. Let me tell you why.

To start with, we were very close to not even having a launch site. To launch a balloon in the UK, the Civil Aviation Authority (CAA) has to give to permission for your launch window so that they can notify airplanes to tell them not to crash into your balloon. Obviously to launch you need a launch site, and we originally applied for five or six locations ranging from the rather ambitious (the roof of the EEE building at Imperial in London) to the more conservative (Farley Mount). All of these applications were rejected.

However all hope was not lost because if a launch site has been approved in the past, then the CAA only needs 72 hours notice before granting a launch window. Luckily for us, the kind people at Cambridge University Space Flight (CUSF) were willing to let us use their launch site at Churchill College (thank you!). Unfortunately they were only on campus from the Monday on the 22nd, and our project presentation was on the 25th, so that only gave us a 3 day launch window to get the launch in before our deadline.

Usually launch windows are around a week long to have wiggle room depending on the weather, and the dates made sure that we’d only have 3 days. On the 24th, the day before our presentation, we got lucky and had great weather – sunshine and a gentle easterly breeze. We were lucky, but that really is just the beginning of our good fortune.

Sorting out the helium was also a close call. To get the gas to fill our balloon with, we ordered three small canisters from BOC, and had them delivered to our department stores. What we didn’t receive was a regulator to make sure that the air flow happened at a reasonable speed. If you happened to be in our labs when we collected the canisters, in fact probably if you were within the same building, and heard what sounded like a gunshot going off, that was the sound of us opening the valve and frantically closing it as fast as possible!

This was the day before launch so needless to say, we panicked. Running around campus and asking the chem eng department for regulators brought no success, so we were in trouble. Thankfully, later in the day Adam Greig from CUSF told us that we could use their regulator to save our bacon yet again.

We were also literally 15 minutes away from missing the 5pm deadline for booking the minibus to get us and the helium to Cambridge in the first place! If you didn’t know, it’s illegal to bring canisters of helium on public transport; we unfortunately didn’t consider this until the day before. All this trouble was because of an argument about the morality of reimbursement, and because of it we almost had no way of getting to the launch site! On top of that, if we hadn’t been lucky enough to have had Jake in the team, there would have been no chance of getting a minibus in the first place because he’s the only one who could book one from our union. Did I mention we were lucky?

And then, at 11pm the night before launch, came the biggest disaster of them all. Our SD card was corrupted, and had to be reflashed to install a fresh OS because it suddenly lost the ability to mount the main partition. Fortunately we had all the code backed up in a git repository; unfortunately, upon putting it all back in there, none of the telemetry from our transmitter was working properly. This could have been due to the new OS, or some missed packages, but the code was compiling fine, just not transmitting complete sentences and we could not work out why.

So we only had one choice – start from scratch. Our labs shut at midnight, so the only choice was to bring the pi home and to try and revert all the way back to the original code in Dave Akerman’s github. By the time I got home it was 1am, and we had to be back in school by 7am to get to Cambridge in time for a 10am launch. That left around 5 hours to get the whole thing done from scratch. Yikes.

This is the part of the post where I’m going to give a huge shoutout to Dave Akerman, the man who designed the Pi in the Sky that we’re using for communications. Without the code he wrote, we really would have been dead in the water. Waiting for the fresh code to install was the most anxious I’ve ever been, but the code ran and was being received perfectly by dl-fldigi.

The rest of the night/morning was spent adding as many features to the new code as possible, such as configuring the interface to the the screen or taking pictures with the camera. The only problem was that I didn’t have the payload with me at home, so I couldn’t test the code past the C compiler telling me that it compiled. The whole thing could have crashed as soon as the external hardware was attached, so the worries were far from over.

We now finally move onto launch day. It’s 7am and we’re packed up and ready to head out. Mostly. I won’t name names, but one member of our group failed to wake up in time, and we had to leave him behind. I won’t lie; part of me wished that I could stay as well because of the fear that everything would fall apart at the last minute. It almost did, but missing his alarm ended up being entirely his loss, because the day of launch brought the biggest surprise of all…

Which will be covered in Part 2 of this blog post – the launch itself. Thanks again to CUSF for all their help!

[I’m not sure where else to put this, so I’m going to quickly list some other pitfalls we almost fell into because the theme of this post is mistakes:
-Don’t touch the balloon without latex gloves on! Greasy hands (or greasy noses) degrade the silicon in the balloon.
-We almost attached our GPS antenna the wrong way up.
-Our antennas weren’t made correctly until a few days before launch, because we didn’t follow the correct tutorial! (
-LCD quickly becomes CD at high altitude temperatures, so we should have used an OLED screen.
-Again,! If I could start over I’d read everything on there, then read it all again, then after a few weeks read it again because the stuff on there is critical because of everything that can go wrong.
-In terms of Bluemix, we should have made use of the HABHUB database api to funnel data to our outputs because it’s not feasible to track telemetry data all the way yourself.
-Relevant to point above: the (incredible) network of receivers that UKHAS set up to help balloonists track data was invaluable. It’s also all volunteer driven! However you shouldn’t rely on them entirely for tracking because any telemetry you upload helps them help you so don’t slack off!
-There’s probably a bunch of other things that we forgot about. If you have problems go to the IRC chat at #highaltitude (link). The guys there friendly and very experienced. We also enjoyed their wit, which you may see in Part 2 where there’ll be snippets of the chat for you to enjoy.]