After the success of our Grand Canyon mission, members of the team set out to do more ambitious missions and founded the group Night Crew Labs. In October of 2015, Night Crew Labs, consisting of Bryan Chan, Ashish Goel, Tyler Reid, Corey Snyder, and Paul Tarantino, set out to capture the northern lights from a high altitude balloon in the high latitude regions of Alaska. To do this, they quickly realized that large aperture cameras with good low-light performance would be needed to capture the beauty of the aurorae on a moving platform.
To achieve this ultimate goal, several smaller goals had to first be attained. This included building a rig that can handle low light conditions as well as multiple camera angles, improve our launch and recovery skills, and improve our overall mission planning. To test each of these, we set out to build this rig and test it in the San Francisco Bay Area where we live. The goal here was to do a full test and identify any problems in the design before attempting a night launch in the extreme cold of the remote Alaskan wilderness. We were excited to see where we live and share the beautiful Bay Area from high altitudes.
The payload was split into two boxes – the first stage (the bottom box) and a smaller second stage (the top box). The first stage contained 3 separate cameras. The first is arguably our primary payload, our selected low light capable camera, the Sony A7S mirrorless camera pointed towards the horizon. The Sony A7S was outfitted with a Sony 10-18 mm f/4 wide angle lens. That specific lens was chosen for its light weight, auto-focusing capability, and cost as it was much less expensive than its full-frame, E-mount, wide angle lens counterparts. Using Canon or Nikon lenses were not an option due to extra weight of the lens adapter. Next, we had a GoPro Hero 4 Black donated by the GoPro team also pointed towards the horizon. Finally, we had a Samsung Galaxy S5 provided by Broadcom taking pictures at 10 second intervals pointed straight down (nadir).
The Galaxy S5 served as the brain of our payload. It was equipped with a GPS + GLONASS + BeiDou + QZSS receiver used for navigation of the balloon. We worked with Dr. Frank van Diggelen, V.P. of GPS Technology at Broadcom as well as a Consulting Professor at Stanford University, as well as Sergei Podshivalov who is a Senior Technical Director at Broadcom, to achieve this part of the mission. They provided a Broadcom phone for this and our other missions as many GPS devices do not work above an altitude of 60,000 feet. This is due to export control on these devices which are regulated to cease function above 60,000 feet if travelling faster than 1000 knots. This limits the applications of the GPS receiver, though it is not a violation to work above 60,000 feet if you are moving slowly, as would be the case on a balloon. For simplicity, many manufactures just stop operation at 60,000 feet regardless of speed, so not all GPS devices will work for tracking balloons. However, since our first mission in 2011 Broadcom has made it standard for its chips to work above 60,000 feet (as long as you are sub-sonic!). In addition to providing navigation and cell communications, the phone was also taking pictures at 10 second intervals. In all, the phone logged latitude, longitude, altitude and battery temperature throughout the course of the mission. While inside the cell network, the phone would text us its GPS location through the where’s my droid app. When there was sufficient data coverage, the phone would also upload its position on the internet where we could follow via an android real time GPS tracking app. For our future Alaska mission, we will be incorporating a USB radio module to work with the phone to serve as an additional recovery mechanism in remote areas.
The phone was not our only GPS link. Our Grand Canyon mission showed us that cellular links can be unreliable. As such, we also flew a SPOT Trace GPS tracker. This device sends us position every 5 minutes while below 20,000 feet through the Globalstar satellite phone network. This network is very reliable, even in remote regions, making it this device a good recovery unit but it is not great for tracking the entire flight which is why we also fly the Broadcom phone.
In order to mount all these cameras and associated electronics in the right orientation, we designed a custom chassis that Stratasys was kind enough to 3D print for us very short notice. We used a 3D printed solution due to the custom nature of the design as well its proven light weight and durability. We had much success with a Stratasys 3D printed structure in our Grand Canyon mission. The chassis was printed in red ABSplus material using sparse fill to save on weight while maintaining strength. Many thanks to both Mac Cameron and Laurie Carr of Stratasys for reaching out and working with us on this project and providing both 3D printing and valuable technical expertise.
The second stage contained two cameras – one GoPro Hero 3+ looking up at the balloon and the parachute, and a GoPro Hero 4 Black looking down. The second stage also contained external batteries for these GoPros as well the SPOT Trace GPS tracker.
The week of the launch brought two major storms to the west coast of the United States. On top of that, the trajectory simulations predicted that we would be landing somewhere between the Big Sur mountains and the high cliffs of Pinnacles National Park. So we had to thread the needle in terms of timing the launch between the two storms and ensuring that the payload lands in a flat, narrow stretch 5 km wide between two mountain ranges.
In terms of launch site, we wanted to launch from somewhere close to the Golden Gate Bridge and upon consultation with Federal Aviation Administration (FAA) officials, we decided to launch from an open field in the scenic Presidio of San Francisco on the 12th of December, 2015 at 11:30 am. Due to the larger payload mass compared to previous missions, we needed almost three 110 cu. ft. tanks of Helium to provide the adequate amount of buoyancy. The launch went smoothly and thanks to Timothy Brodsky, and Danny Morris we had some amazing footage of the setup from the ground and thanks to Adrien Perkins we had some great footage from the air taken by a DJI Phantom UAV.
The balloon reached a height of approximately 91,470 feet and burst a little earlier than expected due to some overfilling on our part. Thankfully, it landed in a rather benign area close to agricultural fields, south of Salinas, very close to the 101 freeway. We all drove to the landing spot and after a short hike, were relieved to find that there was no damage to any of our equipment. After returning to the cars, we immediately pulled out the SD cards from the cameras and to our delight, found that all the cameras had captured some amazing footage. We were literally jumping with joy when we saw the amazing views of the Golden Gate Bridge, San Francisco downtown, Bay Bridge, Monterey Bay and a giant rainbow across the bay area. We also captured some spectacular high frame rate footage of the balloon bursting and the parachute deploying. One can see how the rip travels along multiple longitudinal lines, leaving behind a cloud of talcum powder. The parachute goes through some transient oscillations before deploying in a stable configuration.
We learned some important lessons regarding moving the location of the smartphone in the first stage to improve the Signal to Noise Ration (SNR) for the GPS signal as well as monitoring the amount of Helium carefully to ensure the balloon reaches maximum altitude. Our method of balloon filling relied on achieving neutral buoyancy with a predetermined ballast weight. This sets the ascent rate and also determines the burst altitude. We filled the balloon a little too much with helium and is why we burst at a lower altitude than expected. As for placement of the GPS, it appears that the a7S was causing some interference with our Galaxy S5 GPS unit. This is likely because the camera was placed on top of the GPS antenna in close proximity, physically blocking some of the sky and some GPS signals, as well as causing some electronic interference.
Overall, we consider the mission a massive success. We cannot wait to do this in Alaska and in other places around the world!
Many thanks to:
Timothy Brodsky for helping with logistics, cinematography and his countless hours in producing our video.
Jill Merrigan for design and implementation of our website and for her marketing advice.
Adrien Perkins for his expert UAV piloting skills and cinematography / photography.
Danny Morris for his help with cinematography / photography.
Dr. Frank van Diggelen and Sergei Podshivalov for their continued support since our first mission in 2011 in providing GNSS enabled smartphones for balloon navigation as well as software, analysis, and much technical advice.
For donating a Hero 4 Black camera for the mission.
Mac Cameron and Laurie Carr for 3D printing our custom housing for this mission as well as for much technical advice.
And to all those who came to show their support on launch day:
- Stephanie Kainec
- Aimee Kerr
- Sara Knox
- Devanshi Kukadia
- Jerami Martin
- Jill Merrigan
- Danny Morris
- Adnan Raja
- Hakim Raja
- Sarrah Reshamwala