The school was one of the better state schools in West London - the kids were 17 at the time. At the end of the school year they give them a couple of weeks to do a project of their own making. This particular project was the kids initiative. They got their first Muon detector working in these two week, and flew it to Germany and back on a cheap flight. The ballon was then an extra addition a few months later. The Muon detector design was based on one from MIT http://cosmicwatch.lns.mit.edu/detector but they redesigned the circuit boards themselves because they weren't confident about making surface mount boards, and they wanted to separate the scintillator board from the rest, to give them a chance to change things later. The kids did all the circuit board design themselves, and assembled everything themselves. I put them in touch with the right people to get circuit boards etched. They built two Muon detectors - one nicely cased with a display, and one lightweight one for the balloon flight. The head of physics at the school lent them a beta source to test with (his main instructions were "don't lick it!").

I helped with some of the comms software, but they did most of the other software themselves too. As for the balloon, again, they did that themselves - they rented the helium cylinder, researched the regulations, how to get a NOTAM issued, arranged the use of a disused airfied as a launch site, phoned air traffic control to get final permission to launch, etc. Beforehand, they did all the testing they could manage without being airborne, and got the hang of driving the sdr dongles they used to receive the two RTTY streams. They also discovered GPRS transmissions really interfered with the Muon amplifier, but that wasn't a big issue if it was only turned on for landing.

On the day, the main thing the two parents did was drive the two chase cars, as the kids hadn't got a driving license. Most things went right first time. They had two redundant GPS receivers and two RTTY transmitters, running completely separate software, one on a raspberry pi zero, and one on a teensy. The GPS chipsets need configuring into high altitude mode, or they stop working above 10,000 meters. But this is impossible to test in advance. The RTTY transmitter that was relaying the Muon results gave a much higher bit error rate than the second one, but the second one's GPS chipset turned out to not be in high altitude mode. The balloon sprang a leak at 16,000 m, and descended gently afterwards. We had been aiming for the expected landing site, and as it dropped back into the jetstream, it started to get away from us, despite us driving as fast as we dared with a large antenna on the car roof. Fortunately two other HAB guys were receiving too, and relaying position information back as we started to hit the limits of what we could receive from the working GPS. As the balloon descended the second GPS came back to life, and we were able to track again until it dropped too low. In the end we got to within a kilometer of where it landed, but the reason we recovered it quickly in the end was because the lovely lady whose pigsty it landed in phoned the number on the payload.