NSL-75

The North Carolina coast was visited by Hurricane Dorian on Sept 5 & 6, 2019.

Paul L. had storm-coverage duty at his work overnight on these days, but he had the daytimes open. He hatched a plan to fly at least one flight over the storm. On Sept 5, Paul quickly assembled two complete flights with payloads, parachutes, and such using spare parts from recent flights:

Flight A: "Sunrise Flight" with AP510 tracker, a Mobius Mini video camera, and low-light camera; stuffed into the NSL-74 box.

Flight B: Pi-in-the-Sky-type system (PITS) with, GPS, camera, and LoRa
and then further down the bridle, the NSL-73/73 box with an AP510 and Mate808 camera.

Flight B included a re-assembled the PITS board to test out 3 changes:

• Previous use of this system saw GPS reception issues due to radio frequency interference from the Pi camera. Assembling the parts in a long, thin shielded case helped isolate RFI previously. So it was tried again in a slightly more compact design, plus using fancy copper tape instead of trying to box-tape household aluminum foil.

• Boost circuits for LiPo batteries failed in the past. So this flight used just three lithium AA batteries to provide the required 5V supply voltage. New lithium AAs deliver 1.8V each instead of the 1.5V of normal AA batteries. Together these delivered 5.4V and simply fed into the Raspberry Pi without any boost converter. Energizer Ultimate Lithiums were known for excellent performance when cold, but would 3 AA cover an entire flight?

• Glitches in previous power supplies caused computer reboots. These can occur when the violence of burst or landing causes batteries to bounce within their spring-loaded battery packs. To prevent this, Paul added a 5V 4F super capacitor to the 5V rail. The Pi now operated for at least a minute without batteries!

With the Hurricane Florence flight under his belt, Paul felt comfortable that the flight predictions within the storm were still accurate... as long as the balloon could be filled safely, protected from wind gusts.

He and Chris G. showed up at Horseshoe Park in North Raleigh around 5am to try Flight A. They found the gates closed; and even worse, had to abort entirely due to failure of the low-light camera. They decided to try Flight B around 9am. This gave Paul a chance to take a brief nap and allowed Aurora T. to join them.

Regrouped in the same parking deck as the Florence flight, the three prepped the Flight B gear for launch. Almost immediately, groggy Paul broke-off the simple wire Lora antenna on his PITS box. Doah! Without a soldering iron handy, he could only reattach the wire by crimping it with pliers and applying crazy glue. It seemed to work.

Would the antenna fall off during flight? That could jeopardize some flight images. Better add a backup camera!

Just before balloon fill, it was decided that the flight could use another camera. But how? The payload box from Flight A was quickly cannibalized. The faulty low-light camera was torn-off, the AP510 removed, and the parachute snipped. The remaining plastic box held just a Mobius Mini camera and was quickly tied on to Flight B, hanging about 3m below. The balloon fill then proceeded normally. With everything ready, the team then walked the gear to the top of the parking deck.

The weather up to this point had been wonderful. No rain and only occasional wind gust from the nearby hurricane. Just as the team reached the top of the parking deck, a strong sustained gust arrived. Chris held-on as the big balloon tried to wrench itself free in the wind. After what seemed like minutes, the gust subsided enough to allow launch.

The flight proceeded well. The PITS continued to operate with strong radio signals, delivering telemetry as well as images.

As the flight continued, it reached 8km and a high cloud layer comprised of ice. The on-board cameras recorded what looked like a falling snowstorm as they passed up through the stationary ice particles.

After breaking through these clouds, the cameras were greeted by a beautiful solar halo and sundogs. All cameras also recorded some sort of subparhelic bright spot. This was too far from the sun to be a "dog", but too close to be a "glory".

The PITS system suddenly stopped sending live images. The telemetry continued fine, so it wasn't an antenna failure. There must have been a software bug, or maybe the camera just froze. Otherwise the flight continued as expected. Eventually the flight rose to a height where Hurricane Dorian was seen.

Chris had to go back to work, but Aurora and Paul drove south of Raleigh to chase. They waited at I-40 & US-42 for burst. While there, they spotted the faint white dot of the balloon as it drifted 33km above. The fast-moving low-level storm clouds would occasionally block their view of the balloon. They joked that the balloon will pop while it was behind one of those clouds. It did.

The landing projections for the trackers all agreed that the payload would land nearby -- at first. Unbeknownst to the chase crew, the parachute had become tangled in the balloon remains shortly after burst.

Paul monitored telemetry while Aurora drove north to intercept. They realized to their dismay that the flight wasn't slowing down like it was intended to while under parachute! As it fell though, it offered video images seldom seen on flights: the view of one payload taken by another.

The payloads continued to scream in at 15m/s. The predicted landing zone soon updated to show Lake Benson. The chase crew noted "at least that would be safer than landing on someone." Paul had packed waders, but he might have to swim.

Instead of splashdown, telemetry packets continued to come in, and eventually showed a landing site only a couple dozen meters from the shore. Google maps revealed woods.

The payload draped across the top of a very tall pine tree at least 28m up!

Chris met back up with Aurora and Paul and looked at the situation. It was a loooong way up, and there was no clear opening to make a bow and arrow shot. The payload was doomed to stay up there until it rotted and fell.

Recovery: The second half of the story

Steve W. was down at the coast inspecting the storm damage when he heard of our plight. Steve always 'knows a guy' and this time was no different. He knew Steven Epps, the remarkable owner of Epps Arboriculture LLC of Clayton, NC. The two Steves got chatting and before we knew it, had a rescue plan set for Sunday Sept 8.

But how to get up the remaining 5m to the payloads? At 23m, he tied on an additional set of ropes and climbed all the way to the very crown of the tree.

From there he was able to reach down to the payload and untangle it. Then it was just a matter of reversing his procedure down to 23m and then quickly repelling down from there. Wow!