Super Cub camera experiment

After the success of the SkyFly video flights, I decided to try some still photography. The cameraship in this case was my Hobbyzone Super Cub, which has a lot more lifting capability that the SkyFly.

The Super Cub was also well-suited for this use because it came with Hobbyzone's 'X-Port' hookup, which allows a single-channel on/off accessory to be added to the plane with no modification. I bought their aerial drop module, which is basically an electromagnet that stays on until the button on the tranmitter is pressed, which turns off the magnet and allows whatever is secured by it to fall away from the airplane (they send some weighted streamers and a doll with a parachute with the module).

Following some advice on an RC forum, I pulled apart the module and replaced the electromagnet with a 5VDC SPDT relay from Radio Shack. I connected the normally-closed contacts of the relay to the shutter switch lines of an Agfa ePhoto 780c digital camera. When the X-Port unit is plugged into the plane and the power is on, the relay is kept energized until the trigger on the transmitter is activated. That releases the relay, closing the normally-closed contacts and triggering the camera to take a picture.

I mounted the camera to the belly of the Super Cub with some rubber bands:





This mounting allowed me to use the stock wing hold-down points as anchors, and kept the aircraft's center of gravity the same distance from the wing leading edge. There was adequate clearance between the camera and the ground for the plane to taxi, but the addition of a 12 ounce camera to a 20 ounce airplane made ROG takeoffs a very iffy proposition. After several aborted takeoff attempts I ended up hand-launching the airplane. When running on a 9.6V 1000 mAh NiMH battery it had adequate climbing capacity.

On July 14, 2007 I flew the airplane/camera combination for the first time. Flight conditions: 75F, sunny, winds WNW at 5 mph. The results weren't too spectacular (see bottom of this page). This is mostly due to the fact that the old Agfa camera is quite limited when compared to modern digital cameras: it takes an 8 MB max memory card, requires four AA batteries, and only has a 640 x 480 sensor chip (about 0.3 megapixel). I used it because it was available and expendable, and I wanted to see if the control circuit I built would work. I plan on using a better camera in the near future, and installing a mount point for the Audiovox video camera I used with the SkyFly as well.

One other issue with picture-taking is how the X-Port itself works. When the transmitter button is pressed, the motor shuts off, and the control surfaces center themselves. This is apparently due to how Hobbyzone implemented the control system, and since there is a half-second delay between pressing the button and when the camera actually takes a picture, the plane can pitch and roll a bit, making aiming the camera challenging. A different plane with a conventional radio, using a normal servo to trigger the camera would give better control.

Click here for July 14, 2007 images.

Update: July 17, 2007
Installed velcro mount points for the Audiovox DC400 camcorder. First video turned out very well.

Update: July 18, 2007
I upgraded the camera to a Digital Concepts 4.1 megapixel unit, which when equipped with two lithium AA batteries tips the scales at only 4.6 ounces. This made a huge difference in the flight characteristics of the airplane, allowing ROG takeoff and less fear that the plane would fall from the sky. However, the resulting images showed some severe artifacts compared to the original, less capable camera. Since the X-Port cuts the power to the motor when it is triggered, I suspect this is due to vibration of the camera caused by the propeller windmilling.

Update: July 19, 2007
I tested my vibration hypotheses by putting a double-layer of bubble-wrap between the camera and the airplane. The images do indeed show less of the distortion the first set exhibited, but it is not completely gone except for a few pictures shot vertically.

More videos, this time backward-looking: Video 1 (with a somewhat sub-optomal landing in a wheatfield during a stunt. No damage! That little plane is surprisingly sturdy). Video 2 went much better.

Update: March 27, 2008
An experiment with video frame extraction. A flight was made recording video of an observatory site, then individual frames were extracted and stitched together via Autostitch. The results are not bad, considering the quality of the original video.