The intention is to try both. I have started with an assumption that servos will provide the best “connected” feeling among other advantages, but the DJI FPV drone uses a brushless gimbal for camera pitch so my concerns may be invalid, regardless of that here are my initial thoughts on it.
Flight Experience and Control
The biggest concern I have for a brushless setup is that it might lag behind slightly when changing the pitch of the aircraft. For this gimbal sub-system to be flyable you need to have some form of reference to the angle of the aircraft, in the case of a servo gimbal you can feel confident in the angle of the aircraft based on the position of your head (after some practice). I have concerns that producing the same feel with a brushless gimbal would be difficult, especially for a ‘general solution’ that anyone can implement.
Reliability and Durability
This is an obvious win for the servos, at least for a big servo, the current setup I am using consists of a very strong gimbal printed from TPU and a 20kg/cm servo, connected by a 1mm steel linkage, which basically means I’ll never damage the servo, as the linkage will bend long before the servo was strained, the linkage can be bent back in a few seconds and in my experience that’s usually the only thing that takes damage.
In the case of a brushless gimbal, it is far less durable and less reliable, but it does have some notable advantages.
Natural Movement/Video Quality
A shaky camera is almost never what’s desired out of drone, we like smooth movement in video and this will be the ultimate deciding factor for some applications, unless both could produce natural movement…
Servos are twitchy, I didn’t realize how twitchy they were until I started using them for this application, I was told by my engineering consultant (friend) that it was due servos using potentiometers, which have margins or error, and that servos are basically innately twitchy due to that. This is true, however I hoped there were ways around this. Early Testing and designs didn’t show promising results, I tested various methods of dampening to keep the twitches of the servo from reaching the camera, and had some success with this, in fact I was able to get about a 90-95% success rate for smooth servo gimbal movements. Still, that 10-15% of failure stung, especially when considering filming applications, no one likes a shaky camera.
A brushless gimbal is a proven method for stabilizing a camera and producing smooth movement, in the case of a FPV drone it could even serve a secondary purpose of stabilization. If it is possible to make a brushless gimbal that could provide a “connected” feel, be durable, reliable, able to be generally and simply applied and to existing systems, and be cost effective, then I started down the wrong path.
Before starting to build and test brushless gimbals I intend to thoroughly explore solutions that utilize servos, as even the most complex setup will be far more simple than a brushless solution. If only there were servos that weren’t twitchy… I did a lot of testing with a lot of expensive servos (consider Supporting Medlin Drone) and made an important discovery.. check out Finding the Right Servo in the Wrong Places.