MultiWii

Like most, my FC (v2.1)was prone to significant vertical wandering in altitude hold mode. Many users add some open cell foam over the baro sensor to shield it from rotor turbulence. I decided to try something more controlled and built a mechanical low pass filter. This is a large sealed air chamber with a small bleed hole. Mine is constructed of sheet plastic and is about 10mm x 8mm x 25mm high with a 0.8mm bleed hole. The interier volume of my chamber is about 1.5cm^3. This 'box' is placed over the baro sensor and sealed to the board with rubber cement. I have a board protection plate over the FC that keeps the chamber pressed to the board.

I have only one test flight so far, but it is clear that the altituce hold is much improved. Getting the right throttle setting is still a bit of a challance, but once achieved, the quad holds altitude quite well. I'll do more testing, put I'm very pleased.

Mine was more low tech, rubber hollow plug, top cut off, with cotton on the too hole. Simply slides over the Baro

While it's clear that open cell foam, or cotton, over the baro sensor reduces high frequency noise, the FC can't actually read high frequence noise. The FC cycle time is about 0.2-0.3 seconds. Mathamatically, the digitized baro reading can't have a frequency higher than half the sample rate. When my quad 'hunts' in altitude, it does so at a 3-4 second period, thus about 0.3 hz. If the altitude hold is trying to follow a 0.3 hz noise, open cell foam, or cotton, can't filter that low of a frequency. A software filter can do that, but it consumes program memory and cycle time doing it. Hence my attempt to construct a hardware filter that perhaps can.

I don't know if my filter's orfice/volume ratio is small enough, but it's got to be better than just foam or cotton. I suspect I need a smaller orfice. I will continue my experimentation.

After more experimentation, I've discovered the orfice in my filter is actually too small. The filter is a bit too low pass. I've opened the orfice up to about 0.08" and most of the oscilations have gone.