NSL-67

A payload had been put together in early October in hopes that it could be flown around Hurricane Michael. Downgraded to a tropical storm, this passed over NC on Oct 11th. The remnants of the eye were still intact as the storm passed Charlotte -- which looked intriguing on satellite images. But this soon collapsed as the storm traveled north and then was forced to the east over the RTP. The timing and conditions were never right to launch.

The winter jet stream was descending down to North Carolina, and within 24 hours, conditions to fly easily would disappear for at least several weeks. But there was a short window in the weather and scheduling to allow for a flight the morning of Oct 12, 2018 -- a beautifully clear day. With about an hour's forethought, the Paul and Matthew L. swapped payload modules around and opted for a flight to test some new hardware and software.

The 700g payload consisted of the battered NSL-65 & 58 box with the following:

• Raspberry Pi Zero with GPS, camera, LoRa and modified PITS (Pi-In-The-Sky) software

• Mobius Mini Action Camera, looking horizontally

• AP510 APRS tracker

• Outside Hung Compact Rain-Activated Pull-down

• Directed Arboreal Recovery Node

The payload had a simplified experimental 92cm chute and was hauled-up by a 600g cell with $10 worth of H₂.

A few software changes had been made since the previous PITS flights. A tweak to a LoRa register hoped to increase its transmission output by 3dBm from 50mW to 100mW. Would that make a difference in reception?

Code was also changed that would allow for the ground crew to more easily see the projected landing site on the main Habhub map. This would present itself as a fake balloon appearing at the predicted landing site with the call-sign XXNCNS.

During payload activation, it was discovered that the Pi was unable to correctly talk to its camera. Prodding and rebooting didn't fix things, but it was decided to go with launch anyway as the downloading of images was just a secondary nice-to-have. The payload was sealed up and launched from Apex around 0923EDT. The flight path would take east, following US-64, towards Tarboro.

The chase crew followed along US-64 east. The performance of the PITS telemetry was much improved over previous flights. Using a simple mag-mount antenna (length of coat hanger cut to the right length for 434Mhz) on the chase car roof, distances of 40-50km showed packet loss of only around 20% (RSSI -110). Using a spiffy 7-element Yagi antenna, the packet loss dropped to 0% (RSSI -103). Tests showed that the Yagi could even be pointed in the wrong direction and it still receive strong signals. The chase crew varied their distances from the payload and took signal readings. As the payload approached 30km altitude, Matthew started a Dave A's Python script that would plot the predicted landing site on the Habhub map.

The chase crew arrived at the landing area about 25 minutes before landing. To make things more challenging, the storm-swollen Tar River bisected the landing area north-south. There were only two bridges to cross the river and these were some distance apart. If the new chute design was too effective, then the payload would drift "long" and land to the east of the river. If not, to the west. Three landing predictions (one on-board and two Habhub web-based) showed landing to the east, so the chase crew parked there to wait (see map below).

At 15km, the chase crew started to think that they stood a chance to see the landing, or even catch it.

But then things started to get concerning. The on-board prediction track began indicating a landing directly in the river, or at least the inaccessible swamps surrounding it. Packet after packet, the system kept pointing to a very wet landing. The chase crew prayed for a parachute failure while watching the projected Time-Till-Landing counter come in via telemetry. At TTL=270 (4.5 minutes until landing) things changed. During these last two kilometers of descent, the payload was suddenly carried northwest by two whole kilometers, parallel, but thankfully away from the river! Letting out the collective breath they were holding, the team sped back on to the highway to go west across the river. 90 seconds before landing, all three landing predictions started to converge on the map. It was away from the river, but would it land on the highway? Pesky speed limits and stop lights prevented the crew from seeing the landing, but they were overjoyed that they didn't have to swim to recover the payload.