In the late 1980’s CCD cameras became small and affordable enough to fit on-board an RC airplane. I combined one of these early cameras, a 120 line b/w security model made by Uniden, with a hand built 250 mW Amateur TV kit from North Country Radio and a 1/4 wave ground plane antenna from the ARRL antenna book. The ground end consisted of the same antenna connected to an ATV downconverter from PC electronics and a VCR for tuning and recording. The display was a Gorilla brand green screen monitor from my Commodore 64. The system worked surprisingly well since there was little RF interference in those days.
My first video airframe was a Carl Goldberg J-3 cub powered by a K&B 61. The camera looked out the front window. My first video flight ended in a crash when the elevator servo quit. It was the first flight on a brand new radio and I should have tested the radio better. I built another J-3 and had several successful video flights with that airframe.
While the J-3 worked OK I realized that slower was better especially for flying by video, now called first person video or “FPV”. I had learned to fly on a Sig Kadet Mark-I in 1976 and knew the Kadet was a stable and slow flying airplane. I built a Kadet Senior with space for a larger fuel tank and installed an OS 48 Surpass 4 stroke engine for power. I installed a better board type camera from Supercircuits and mounted the camera on a tilt mechanism on the bottom of the aircraft. I used a dipole antenna embedded into the fin which reduced antenna vibration. My early experiments were published in the Louisville Courier Journal and the Frankfort State Journal.
I spent the summers of 91 and 92 with my head inside a cardboard box. I was eventually making entire flights from takeoff to landing from video. It was fun to hear other pilots react as the plane disappeared behind a treeline only to see it re-appear through a clearing in the trees. I started testing the limits of range of both control and video systems. This was years before there was any guidance from the AMA or FAA on FPV flying since very few people were doing it.
About this time a friend and fellow RC pilot Bill Dickerson and I started experimenting with aerial photography using cheap point and shoot 35mm cameras. We took pictures out the side window of Bill’s Goldberg J-3 and vertical shots out the bottom of the Kadet. Bill was also a Part 61 pilot so we took photos from his Cessna 172 that we could compare with our RC photos.
It became apparent that aerial photography from radio controlled airplanes was not just feasible, it was in many ways superior to aerial photographs from a full size aircraft. Photos from a full size plane are shot through hundreds of feet of hazy atmosphere. Long lenses are needed to capture smaller subjects but long lenses are more affected by vibration and aircraft motion. A fast, vibration tolerant wide angle lens can be used in an RC airplane flying very close to the subject.
We also discovered that RC aerial photography has huge advantages in time, cost, and safety. When you consider the time required to drive to the airport, pre-flight the aircraft, make the flight, land, refuel, and post-flight the aircraft it is often comparable to driving directly to a site and shooting the subject with an RC airplane. Full size aircraft are expensive to operate and flying low and circling to take photos is less safe than cruising at higher altitudes.
In the summer of 1994 I pitched the idea for an RC aerial photography company to Bill and we agreed upon the name New Perspective Aerial Photography. With a concept and a name we felt were ready to turn our hobby into a successful business in a vast untapped market of eager customers.
We initially though that RC helicopters would be the best platform for RC aerial photography so we mounted our video equipment and a point and shoot camera on a Kyosho Concept 30. We quickly found that it is very difficult to hold a helicopter steady at altitude. The vibration, smoke and our poor helicopter piloting skills resulted in poor quality images. After tinkering around with these finicky flying machines for a few months we decided that fixed wing aircraft were a better platform.
We knew we needed a better camera so we purchased a Cannon EOS Rebel with a 70-210 mm lens. We modified a Kadet Senior to allow for vertical or side photography. A quality camera combined with a stable and slow flying airframe resulted in consistently good pictures.
The camera was a bit heavy for the Senior so we built a custom airframe which was approximately a 113% scaled up Kadet Senior powered by an OS 51 Surpass. The SLR camera was mounted on a tilting mount driven by a large, hand cut plywood gear and plastic chain drive. The camera could be rotated from straight down to a point where the lens was inside the fuselage. We switched to a 28mm aspherical lens with the focus locked on infinity for nearly every job. I built a PIC based control board that monitored missing servo frames and battery levels and displayed this information on LEDs in front of the camera. A second PIC board near the camera provided tilt control with hall effect limit sensors as well as camera trip and auto retract functions. The two systems were electrically isolated with independent batteries and were connected with fiber optics to eliminate RF interference.
Despite our tendency to design and build rather than market our business, we successfully completed several jobs with the Kadet. This airframe performed so well we framed up a backup airplane that we could quickly finish off in the event we crashed the original. We never had a need to complete the backup plane. The original has been converted to electric power and is now a relaxing plane we fly just for fun.
Our next airframe was a 10′ span 133% aircraft. It was similar to the Kadet but had a lifting stab like the Telemaster. We were considering carrying a medium format camera or building a retractable pan-tilt mount for the EOS. In hindsight, we should have built the camera mount first and then designed an airframe around it rather than building an airframe that could carry anything. This plane initially flew on an OS 91 4 stroke and was later fitted with Bill’s OS 120 4 stroke. While it could be hand launched it was just too large and cumbersome to operate in the spaces we normally had available. We outfitted a 5′ x 10′ enclosed cargo trailer with our ground equipment and by 1998 we had an impressive although cumbersome system. Fortunately we never purchased a medium format camera. Our giant airplane never flew a job.
About this time I started experimenting with various stabilizing systems. We flew a BTA autopilot which had two mechanical gyros and a barometric sensor for vertical control. The BTA unit worked well but was a bit heavy. I attempted to build my own autopilot using one of the first available chip accelerometers from Analog Devices. The device used two PIC microcontrollers and two single axis sensors. The device could stabilize the aircraft as long as you turned very gently. However without the benefit of gyros, centrifugal forces would overwhelm gravity resulting in a “death spiral”. I was in one of these downward spirals and was instinctively pulling up elevator when I switched to manual control The elevator servo immediately went to full “up” and I earned the distinction of someone that was able to break an airplane in mid-air. Everyone at our field looked up when they heard the bang of MonoKote exploding as the wing folded on the Kadet Senorita.
Bill and I evaluated our trend towards large complex aircraft and we decided that it was time to simplify.
Smaller is better
As we encountered jobs that required flying from small and obstructed areas we realized that our large and complex aircraft were limiting where we could fly. We needed smaller and lighter aircraft. We also recognized the safety benefits of lighter aircraft constructed out of styrofoam. The Cannon EOS was too large and heavy for a smaller airframe and we didn’t really need the variable focal length or autofocus capabilities of an SLR lens. We started designing around point and shoot cameras. We were inspired by the aerosonde with its inverted V tail design. The pusher engine kept the smoke and oil away from the camera. Another big benefit to pusher aircraft is that we could belly land them without breaking the prop. The early versions used an OS 40 FP glow engine and a cheap plastic lens point and shoot camera for testing. This model also carried the FMA Co-Pilot stabilizer although it really didn’t need it.
As brushless electric motors started becoming available we switched to electric power to avoid the vibration of glow engines. Doug Allgeier, our local electric airplane expert, was a wealth of information as we switched to electric power before it was “easy”. This version was powered by an Aveox motor on an Astro Flight gearbox turning a 12×6 prop. It used 16 nicad cells and carried the fantastic Yashica T4 35mm point and shoot on a tilting/retracting camera mount that completely retracted inside the fuselage. The T4 was hacked so it could be remotely tripped by a PIC microcontroller. By this time we had transitioned from balsa and MonoKote construction to foam and fiberglass. We constructed the tail from vacuum bagged blue foam with an aluminum arrow shaft embedded as the pivot tube. The full flying tail surfaces had enough deflection to maintain control at very slow landing speeds. The fuselage was blue foam and fiberglass. The tail booms were carbon fiber We had many successful shoots using this airframe.
As digital cameras reached usable resolutions we simplified the airframe to further reduce weight and reduce build complexity. We used the same Eppler 422 high lift, high drag airfoil from our previous design since it allowed for slow flight and payload carrying capability. This design used a smaller, geared Mega motor and carried our first digital camera, a Sony DSC U20 (2 MP). The camera sat right up front on a removable tilting mount. This airframe was very light with a high thrust/weight and could steeply climb out of and be spiraled into the smallest landing sites. This was also our first plane that used lithium polymer (lipo) batteries. We built a couple of these from Dryvit insulation cut with a hot wire.
We finally had a system that we could safely fly in most areas. While we had to turn down a few jobs that would have required flying in urban areas we were more commonly pleased to find an open field near the subject that allowed for a safe launch, flight and recovery.
Our tendency to focus on design and development rather than marketing was one of our fundamental problems. We realized this but forged on with a new design anyway. A major problem with the camera on the front of the aircraft was shooting a subject in crosswind or downwind. If the camera could be mounted on a pan-tilt mount under the aircraft and have a 360 degree unobstructed view, then the pilot could park the aircraft into the wind and the camera operator would have the freedom to compose the shot. Learning from our previous mistake we built the camera mount first and then built the airplane to carry it. Since we could not belly land with the camera underneath some sort of landing gear was needed. We solved the problem with a “deployable” landing gear that dropped by gravity helped by a rubber band. The gear did not need to retract since we always hand launched the aircraft. The gear did not need a mechanism to lock it in place due to the rearward angle when deployed. This design, our last fixed wing type, had a 360 degree pan/tilt mount for a Pentax Optio S6. We chose the Optio because of its small size and it was the only miniature type available with a remote infra-red shutter release. I hacked the OEM remote and built a PIC based IR LED driver that eliminated the need for a mechanical trip mechanism. The S6 was our first digital camera with video-out so we no longer needed a separate alignment camera. This system was the best and last fixed wing solution used by New Perspective. The aircraft is extremely lightweight, has an excellent climb rate, can hover in a light wind and can be mushed into the smallest landing areas.
In 2008 the FAA banned this type of work and combined with the real estate crash we decided it was time to give up on the idea of making money with drones.
In 1994 I could only dream of a stable, hovering multirotor platform. I knew the concept was possible and often discussed building “a helicopter that’s not a helicopter”. In 2001, I built a coaxial twin rotor, 24″ diameter flying “tube” with four control vanes on the bottom driven by four servos. I fabricated my own foam core vacuum bagged rotor blades to save weight and drove them from two Astro Flight brushless 020 motors on high ratio gearboxes built by Doug Algeier. While it could lift it’s own weight consisting primarily of 8, 2000 maH, Nimh cells it pulled over 30 amps doing it. I shelved that project due to the high current consumption and lack of efficient rotor blades.
By 2006 outrunner motors and lipo batteries had largely solved the thrust to weight problems so I knew a quadrotor was feasible. It seemed like a no-brainer to replace the mechanical complexity of a traditional helicopter rotor head with a microcontroller and piezo gyros but I was frustrated by the lack of suitable tractor/pusher propeller sets.
Some of the first multirotor aircraft started to appear in the news but the first commercial units were extremely expensive. In 2009 I purchased a few inexpensive gyros from Hobbyking in preparation to build my own multirotor controller but almost immediately discovered there were brilliant folks around the world with the same idea. KK multicopter, Arducopter, MultiWii and Betaflight were rapidly refining firmware and taking advantage of increasingly accurate sensors. The open source projects benefit from the synergy of enthusiasts around the world as they design, develop and test new hardware and software at a rapid rate. These systems can do things that would have been considered magic just a few years ago.
In 2010 the multirotor craze really started to pick up worldwide thanks to amazing websites like DIY Drones. I built my first quadrotor using the APM 1 in the fall of 2010 and have built dozens since. Maybe it was time to put a camera on this thing…