Quote:
Originally Posted by Ether
The datasheet for the ADXRS150 says the null has a linear acceleration effect of 0.2 degrees per second per g.
So if the gyro were mounted, say, 38/2 = 19 inches away from the center of rotation of the robot, and the robot were rotating at 180 degrees per second, you could calculate the effect:
180 degrees/sec = 3.14 radians/sec
(3.14^2)*(19) = 187.5 inches/sec^2 = 15.6 ft/sec^2 = 0.488 g's radial acceleration
So (0.2 degrees/sec/g)*(0.488 g) ~= 0.1 deg/sec
|
Ouch! I had never really considdered how much the rate of spinning dominates the terms as it rises. So if you have a very fast spinning robot (you know who you are

) even mounting the the gyro pretty close to center will lead to more accumulated error than you'd want if trying to go with a field-centric control, say. (and of course you need to either scale the output or go with a gyro that won't hit 5v before you hit maximum spinning speed)
Thanks for the analysis!