NSL-74

The early morning of 25 June, 2019 marked the flight of the SpaceX Falcon Heavy STP-2 mission from the coast of Florida. This launch carried multiple payloads to space, including a collection of cubesats and the Lightsail-2 project.

Meanwhile, up in Apex, NC; Paul L. flew nighttime test flight NSL-74. Items to be tested included:

• 915MHz LoRa T-Beam tracker

• Very low-light camera and video recorder

• Anti-spin sea-anchor boom

• Customized hand-warmer heater

• A combined APRS/LoRa ground station, remote near NC State Univ.

Since he wanted to test a low-light camera in the stratosphere, why not try to film STP-2 from 30km over Raleigh?

He reused the simple plastic payload box from NSL-69. On board was the AP510 tracker and Mate808 camera pair from many previous flights, including NSL-72 & 73.

New to the flight was a 915MHz version of the T-Beam tracker with Paul's software loaded. Code for this 1st generation hardware can be found here. The 433MHz version flew successfully on several flights in the spring, but this was the first for the 915 model. The higher frequency does not propagate as far as 433MHz, but being within the United States ISM band, 915MHz allows use by HAB fliers that lack an Amateur Radio license (like most students and teachers). The unit looks identical to the 433 version; except for (counter-intuitively) a longer antenna.

Also new aboard; a RunCam Night Eagle 2 Pro low lux camera. Intended for the nighttime drone flyers, this camera sports a 0.00001 lux (@1.2F), low light CMOS sensor. It only produces 800x600 black-and-white video. Paul used it for astrophotography, but added a small RunCam mini DVR board to enable recording of the video. Both camera and DVR came to only 20g.

The RunCam, like many small cameras, runs warm. The DVR chip though, runs very hot. So the DVR was positioned bottom-center within the foam payload box. Air passageways carved in the foam would, in theory, help carry this heat up through the payload and past the three LiPo batteries. This would hopefully keep the batteries from freezing during a night flight.

To test the affect of offset drag upon payload spin, Paul built a very light weight boom (35cm bamboo skewer) and parachute (10cm square paper towel). This 4g addition should have the affect of adding 10g of drag to one side of the payload during ascent. Would a "sea anchor" dangling to one side of the payload affect spin? The contraption, constructed of biodegradable materials, may not survive the violence of burst.

Due to the lack of solar heating of a night launch, Paul chose to add additional heating to the top of the payload. This would help keep the T-beam battery warm. So he added a small pouch of lightweight cloth, sewn and filled with 5g of iron/vermiculite (mix common in air-activated hand warmers).
[Spoiler: he needn't have worried.]

Night rocket launch

SpaceX had a several hour launch window to operate within. As Monday evening progressed, the STP-2 launch time slipped from Mon 2330EDT to Tues 0230EDT. To synchronize NSL-74 with STP-2, Paul pushed his launch time to Tues 0120EDT. This pushed the estimated landing time from a reasonable 0030EDT to 0330EDT. During those three hours of delay, the predicted flight track moved closer to Raleigh and away from Falls Lake -- that was nice.

At 0045, the process of activating the gear for flight began. The two trackers were turned-on to get GPS lock. Then the Mate808 and RunCam camera/DVR were activated. For a quick video test, lights around Paul's house were turned off to provide the sample image below. The glow from the refrigerator night-light inside the house provided the only source of light. The Mate808 recorded only black, but the RunCam showed objects in detail.

Within seconds of the completion of the payload assembly, both trackers lost GPS lock. Integration testing days before showed that simply placing the GPS antennas away from each other worked. Several sheets of aluminum foil shielding had already been placed between the cameras and the two trackers.

With no tracking and a deadline, Paul scrambled to disassemble the payload, move antennas, and add more shielding. He re-positioned the AP510, which accidentally knocked its microSD card loose. It also placed the top of the AP510 up against the top lid of the payload. This would become an issue later in the flight. Finally, the trackers started to behave and the payload was secured for flight.

During the fill of the balloon, the small foam-filled payload box began to overheat. The AP510 reported 45C inside, and the RunCam's video faded from the edges and eventually greyed-out completely. The flight finally launched, 18 minutes later than intended (Paul later remarked that had he been fully awake, he would have added more lift gas to make up for this lost time). In the scramble, the anti-spin sea-anchor boom was abandoned. It will have to be tested on a future flight.

The flight initially proceeded well. APRS and LoRa data streamed in. Paul's Arduino/Windows LoRa receiver (http://www.daveakerman.com/?p=1982) sported a small 915MHz Yagi antenna. This didn't need to be pointed in any specific direction in order for data to be received. It even worked well inside his house while he caught a quick nap. Occasionally the receive frequency needed to be manually adjusted as the payload's temperature changed.

Both APRS and 915LoRa receivers tracked very well throughout the flight. The station registered LoRa packet loss only as the transmitter drifted DIRECTLY 6.3km overhead (the top mounted antenna creates a radio 'shadow' below it). The LoRa gateway software used Automatic Frequency Control and locked onto the tracker's signal as the frequency drifted with the temperature changes.

Around 4km in altitude, the internal temperature started to drop. The AP510 reported 39C as the RunCam video slowly started to return. The resulting image, though, was full of noise.

At 14km, the AP510 tracker suddenly fell silent. Upon review afterwards, video recorded that the payload lid had pressed down on the power button and, during a particularly bumpy ride, turned the tracker off. Now the payload only ran on the experimental 915MHz LoRa tracker.

SpaceX flight images

NSL-74 was only around 15-18km during the launch of STP-2. This puts the launch pad below line of sight, but some of the upper levels of the flight should be visible. After reviewing the video, the following was seen:

STP-2 T+

0:01:13 First see bright light in the direction of Florida coast. It begins to rise.
0:02:33 Light gets dimmer when side boosters shutdown and stage
0:02:54 Two lights as center stage continues while the side boosters reignite (boost back)
0:03:35 Back to a single light as center stage shuts down. Boost back still going.
0:04:10 No objects seen. Boost back ends.
0:07:23 Large bright lights seen descending over Florida as the side boosters reignite (entry burns)
0:07:35 Dark again at end of entry burn
0:09:22 Light further out to the east as the center stage entry burn begins
0:09:52 Light disappears at end of entry burn

Basically the camera caught the major high altitude phases of the first stage flight. The second stage Merlin engine burn was not visible.

Mystery flash

Around STP-2 T+7:04 (2:37:04EDT) the RunCam recorded a very brief flash off the the Georgia coast. This lasted only two 1/30sec frames and was not seen again. There doesn't seem to be any associated STP-2 event at that time. Were there thunderstorms off the coast of Georgia?

NSL-74 flight continued ascend with one working tracker and two cameras.

Meanwhile, as the payload approached 30km, Paul, now caffeinated, loaded his receiving gear into his chase car and drove off towards north Raleigh. He was in the vicinity when the payload fell below the altitude that could be clearly picked up by the remote gateway. His portable receiver didn't immediately pick up a signal, which he inferred as the payload was not in a tree -- good news maybe. It should be on the ground somewhere, in one, or several pieces. Slowly driving around a neighborhood, with the small yagi antenna sticking out his car window, Paul finally received a packet. The tracker map reported that it was on a grassy front lawn two blocks away. As it was about 4am, he drove quietly to the payload and innocently snatched it from the curbside.

Upon opening the payload, the smell of melted foam filled the car. The electronics were shutdown carefully, so as not to burn fingers.

The new 915MHz version of the TTGO T-Beam tracker worked well. It could be clearly received at distances of 10km.