Show your support for community science. Donate to Public Lab »

Public Lab Research note

ideal mechanisms & silly putty camera rig

by mathew | July 01, 2016 23:25 01 Jul 23:25 | #13255 | #13255


@Ranon is working on a silly putty and rubber band powered automatic pan and tilt rigging for cameras. The pan functionality was easy, but the tilt has been hard to implement. I wanted to look at some ideal mechanisms and see what guidance I could find on improving/designing a better tilt system.

I started by browsing KMODDL (Kinematic Models for Design Digital Library) at Cornell, a library of classic engineering teaching models of idealized mechanisms. I found that the tilt mechanism is a version of a 4-bar linkage, well described on KMODDL with mathematics tutorials, example videos, and CAD models for making a demonstration model.


Lessons learned and paper prototyping

I learned that one issue with Ranon's 4-bar linkage was that the steady and moving bar were equal lengths, which lead to sticking points in the rotation. The moving bar should remain smaller than the stable bar.

I didn't feel the need to get too far into the mathematical modeling of the system, and instead built a paper prototype to determine the lengths of the different bars in the linkage. Its just a few strips of paper held together with pins. By pinning it in different places, rotating it, and then measuring the distance between the pins, an idealized version of the mechanism can be quickly adjusted and tried out.

I set some goals: I wanted approximately 30 degrees of angle change for the camera, and I wanted the mechanism to sit below the camera platform.

In the paper prototype, the camera platform is on the left. Here it is at its maximum and minimum extension:




A weakness of spring powered, egg timer powered, or rubber band powered KAP rigs is that they don't stop panning to take each photo. As long as the motion is slow and steady this weakness won't compromise photo quality too much, although the slower it is the longer it takes to complete one rotation and the more likely it is that the kite has moved the rig so much that all the photos will not stitch together.

When a similar constant motion is used for tilting the camera while the rig pans, a new problem emerges. If the tilting is not coordinated with the panning, photos will be taken in a spiral instead of in multiple parallel loops. These photos could stitch into a spiraling panorama, which may or may not be satisfactory. So an improvement would be to allow constant panning motion but incremental tilting motion. Ideally, the camera would tilt only once per 360° panning rotation. This can be implemented with an escapement or cam (I can't find the video of the cam tilting example). More detail of the escapement solution is here.

This type of tilting solution might be more trouble than it is worth. If the camera is mounted in portrait mode, the vertical coverage is increased enough that only one row of photos is needed for many projects. Panning can happen indefinitely so 360° coverage is possible. As long as the camera is in portrait mode and the lens focal length is adequate, very good coverage is possible. That is the route I took with the Aerobee Rig. So a desirable feature of any self panning rig is to allow portrait mode mounting of the camera. Then camera tilting could be added (or not).


Is this a question? Click here to post it to the Questions page.

Reply to this comment...

Your suggestion is to continuously pan, and then increment the tilt? That is sensible. One reason why we were doing a separate pan and tilt mechanism, however was the idea that instead of repositioning the camera within the rig, it would always point the same position in the rig, and we'd move the attachment point/pan mechanism hole-hog, and shooting for a narrow panorama, rather than a full sphere.

The portrait mode is also a good solution, that is how Jim Day's rig worked most of the time. he would detach the horizontal mounting L and mount directly to the verticla bracket: Jim Day's rig

Is this a question? Click here to post it to the Questions page.

Reply to this comment...

Login to comment.