Thanks for the info. I was aware of the long term drift and the need for an something else. However, my application is not to fly the plane unaided for an extended period of time. It just need to keep the plane stable for a maximum of 10-20 min. The plane would be controlled to some degree by a pilot, but is intended to keep a plane dosile enought for beginner thumbs and also to prevent certain unwanted scenarios such as stalls. I would say that long term position sensoring would best be done by a GPS device, however the resolution requires that the plane fly high enough that the GPS error does not slam the plane into the ground.
-Q.
Originally posted by Lynx
You can't detect position change with a rate gyro, it's not possible. Even the most sophisticated Heading Hold Gyro's (which use micro controllers to do just what you're talking about) still drift to the point where they become useless after time. It's a law of control theory that this type of thing is impossible because of parasitic loses. The only way to do what you're talking about is to create a custom device or purchase an attitude gyro, I have however never found anything that's suitable for RC work in the attitude gyro department, the soul exception being the FMA co-pilot, which successful stabilizes on two axis' using a special sensor that can detect the horizon based on the infrared signature. You may be able to hack apart this product in order to get the raw artificial horizon data you're looking for. But I doubt that's a project for the faint of heart.