Research and development of fiber optic gyros began in the mid 1 970s and focused on improving the
gyro's sensitivity to rotation and reducing noise. Next, bias performance was addressed. By the early
1980s, fiber gyros were achieving bias errors of 0.01 O/ in a laboratory environment. Scale factor
performance was initially addressed at McDonnell Douglasi in the late 1 970s by the use of a closed-loop
fiber gyro which employed acousto-optic frequency shifters. Good performance was achieved but this
approach was not productionized.
In the mid 1980s, Thomson CSF developed a fiber gyro which used a double closed-loop technique
employing a digital phase ramp and an electro-optic phase modulator2. Derivatives of this approach
have been adopted by most gyro producers in the world. This technique has enabled fiber gyros to have
high scale factor linearity and has significantly improved scale factor stability and repeatability.
Today closed-loop fiber optic gyros using derivatives of this technique are in production for many
tactical applications3 . These include tactical missiles, smart bombs, and attitude and heading reference
systems (AHRS) and require gyro bias performance of 1 to 1 0 O/ and gyro scale factor performance of
1 00 to 1 000 ppm. Closed-loop fiber gyros using derivatives of this technique are presently in
development for future inertial navigation systems4 (INS) which require bias performance in the 0.00 1
to 0.01 0/hr and scale factor performance in the 5 to 50 ppm range.
This paper will examine how this double closed-loop, digital phase ramp technique functions. An error
source unique to this type of closed-loop gyro, the deadband error, will be examined along with a
technique for eliminating it.