NSL-56

On the afternoon of Sunday, June 12th, we flew NSL-56 to test out a different set of LoRa transmission gear. Paul L. set up a Raspberry Pi Zero with Dave Akerman's PiInTheSky system (http://www.daveakerman.com/?p=2101) and it worked well in ground testing. With a favorable jet stream in the area, he took the box from NSL-51 and loaded it with:

• GCM-320+ recording Geiger counter

• Mobius Mini Action Camera

• AP510 APRS tracker

• TK106 GSM backup tracker

• Outside Hung Compact Rain-Activated Pull-down

• Directed Arborial Recovery Node with Integrated Tether

• and the Pi Zero with camera, GPS, and LoRa modules

This was then connected to a 80cm chute and a 1000g cell will 2300g of neck lift from H₂. Tim W. and Jim C. and their families joined in to help with launch, chase, and recovery. Chris G. acted as remote Public Affairs Officer.

The flight projection for Sunday promised a lazy zigzag with a landing very near the launch point. Paul decided to launch from Lake Benson Park in Garner. That would put the landing area in a part of the county that had the least amount of forests.

Problems with the payload began at activation of the equipment. Paul hadn't performed integration tests with the various gear prior to flight and although the camera and Pi themselves were RF shielded from the tracking gear, there was a rather apparent interference issue. Once put together in the same box, none of the three trackers could maintain a GPS lock. The problem wasn't fully realized until the balloon had been inflated. So Tim was stuck having to hold the big balloon while Paul took time to troubleshoot. The team guessed that the local oscillator from the Pi GPS was causing interference for the other GPS receivers. Eventually Paul decided to simply cut the power line to the Pi GPS unit and continue with the countdown. This left the Pi without GPS info, but it would still continue to take photos and relay them via LoRa. After that, it was a matter of launching and keeping our fingers crossed.

The balloon rose nearly straight up and the crew was able to watch it for a very long time. LoRa data came down steadily and error free until the payload drifted horizontally out of range at an altitude of 11km. Sadly, we were also having issues with the LoRa gateway and only a handful of the packets received actually got sent on to the Habhub website. Due to the Pi no longer having a GPS, the system simply re-transmitted its launch coordinates. It also never scaled up its photo size as it was programmed to do at higher altitudes. So all Pi images were limited to 640x480.

During much of the ascent, the team stayed at the launch site and played with the LoRa receiving gear. Eventually they headed for the air conditioning of the chase cars and made way to the projected landing area.

The LoRa PiInTheSky system was set to speak in "Turbo mode". This is a faster, wider bandwidth configuration that is best suited for shorter range use. As the payload drifted overhead at 29km, we were still picking up some of the shorter telemetry LoRa packets, but there was considerable packet loss. The balloon was a an older Kaymont that was a year of out expiration, so it is not surprising that it only survived to 31.7km (104,008ft). Upon burst, the majority of the balloon attacked the parachute and completely tangled it.

This led to a descent three times faster than desired. The payload was spotted by the waiting chase crew as it screamed in towards suburbia. It punched down through a canopy of trees and the parachute/balloon mess snagged, bringing the payload box to a halt 2m above the ground.

=== Notes from Paul ===

This is what flight tests are all about. I have some RF tests to do to see what can be done about the GPS problem. What part needs the extra shielding? It may be easier to simply put the PiInTheSky in a small capsule hanging a few meters below the standard payload. The transmission mode should be changed from "Turbo" down to "SSDV" mode (Although I had problems making that work in the lab).

The TK106 had issues and only really worked once it was removed from the payload after landing. Is there an issue between the AP510 and the TK106 as well?

I will need to chat with Dave Akerman to see what I can do to improve the performance of the LoRa Gateway. We had decent Internet out in the field. There were others using SSDV that day so maybe there was contention on the server end.