Open Source Pan/Tilt Head: Part One - Mechanical Design
Starting this project at the beginning of the COVID-19 pandemic meant I wanted to find an erector-set-like mechanical platform that would suffice to prototype the electronics and programming. Normally I would have access to sheet metal tools and a vertical mill at my local makerspace, but until I can visit again safely, all parts would have to be off the shelf. Luckily the maker community and robotics competitions such as FIRST have inspired the development of some pretty robust mechanical products, and I eventually settled on Actobotics by ServoCity, based on the wide range of structural and motion components that were available. The cost is reasonable, and even better, almost every Actobotics component has a .STEP file so I could easily model the assembly before my purchase.
Note that Actobotics components can also be purchased from Sparkfun and other retailers, but I found ServoCity to be the most convenient considering they also carried motors, gears, belts, and other electronics, all of which I would need for the project.
The overall size and shape of the pan/tilt head was something I spent a large amount of time researching. Many DIY examples offer a substantially large bracket, an excessively tall pan axis, or have motors sticking out from the sides - all making the footprint unwieldy for remote installation. ServoCity offers a few solid pre-made solutions, but they are still too bulky, and trying to add the sensors reading out the pan and tilt positions would be tricky. Generally, many of these examples are made for time lapse photography, so obviously the design constraints are quite different. An excessively large bracket weighs more and has more inertia when moving, and longer pieces of material flex more. This adds up to an assembly that isn’t agile enough for video and gives shaky results.
For inspiration, I spent a lot of time looking at this compact solution with internal motors from Alibaba, and I came across this other great build as I was writing this part up. From here, I came up with a short list of goals for the yoke design:
The motors should be mounted in-line with the horizontal and vertical beams of the yoke so the ends don’t stick out
The pan axis bearings need to be integrated into the horizontal beam so they don’t take up space under the yoke
This means that the pan motor rotates with the yoke, instead of being fixed and driving the pan axis
Both the pan and tilt axes need to integrate some sort of position sensor to limit each axis’ movement
The output of the tilt axis needs to support the camera in a way that supports balancing the camera
One of the great features of the Actobotics platform is the compatibility with the pillow block bearings that ServoCity also carries. As it turns out, these bearings are very nice quality and lend themselves to precision motion with essentially no ‘slop’ that would make for an unstable platform. The tilt axis uses a 1/2” stainless steel shaft, while the pan axis uses a 1” stainless steel tube. The larger pan axis diameter I think does a better job handling the combination of axial (thrust) and radial loads from having the yoke shaped like an “L” instead of a “U”. Here’s a view of the Actobotics U channel stripped away to show the pillow blocks:
Although I’m skipping ahead a bit, the gears on both axes are mainly to engage with a potentiometer which will read out an absolute position value. They also function to clamp the shafts on one side of the bearings. The other side is secured by a timing belt pulley. More on this in Part Two!
Another challenge was sorting out how to support a camera. The platform needed to be adjustable vertically and also along the axis of the lens to balance out both small and large cameras. At this point, I started to consider using 15mm rails and clamps, and so I quickly mocked this in to see how it would fit. Eventually, I came up with a solution that works well using SmallRig parts and a cheap QR plate which I will show in a later post.