Since it was first proposed in 1975, the fiber optic gyroscope (FOG) has steadily improved in performance and manufacturability. Now a mainstream, high-volume manufactured product with performance approaching Ring Laser Gyros (RLG), FOGs offer substantial advantages over competitive technologies in terms of reliability, cost, and complexity. Optical fiber splicing and related processes are at the core of this achievement.
Production techniques to meet FOG performance requirements for strength, optical loss, and polarization extinction ratio (PER) include fiber preparation, splicing, and post processing. Looking ahead, miniaturization will become increasingly important in FOG design, and manufacturing automation will be critical for improved consistency, throughput, traceability and quality, and cost reduction.
FOGs are interferometric devices utilizing the well-known Sagnac effect. Figure 1 presents a simple example of such a device. Figure 1a shows an optical fiber loop of radius R with N loops. The device has an input/output (I/O) coupler that launches two counter-propagating light waves. When the waves are recombined at the detector photodiode, they will add coherently. If the loop is rotating, then a phase difference develops between them, which manifests as an intensity difference at the detector.
To understand this, imagine a clockwise rotation of the loop at a given angular speed ω. Between the time the light enters the loop and exits the loop, the I/O coupler will have moved a distance ΔL (Figure 1b). The clockwise wave will travel a distance L+ ΔL, and the counterclockwise wave will travel a distance L-ΔL. The path length difference between the two is 2ΔL, which corresponds to a phase difference between them of:
where R is the radius of the loops, N is the number of loops, λ is the wavelength of the light in the fiber, and c is the speed of light.
Assuming an exact 50/50 split of the two waves and no power loss in the fiber, the light intensity at the detector is given by:
where I0 is the input intensity.
Given the large number of splices required to produce a FOG—typically 10 to 20 per device—splicing plays a key role in optimizing performance. Splicing an optical fiber requires stripping the fiber buffer, then cleaning, cleaving, aligning, splicing, recoating, and proof (strength) testing the fiber. Using the best proven technology is critical for obtaining the required splice performance.
FOG Requirements and Splicing Techniques
Due to the nature of optical fiber, various transmission modes can be excited and there can be crosstalk between them. Since these modes will have different propagation speeds in the fiber, any crosstalk will degrade the interference pattern, decreasing sensitivity. To avoid this effect, single mode polarization-maintaining (PM) fiber is generally used. The primary parameter of interest for PM fiber is crosstalk or Polarization Extinction Ratio (PER). PER is one of the most critical splice performance metrics.
PM fiber has a “fast axis” and a “slow axis” (birefringence). Typical fiber specifications for polarization crosstalk between the two axes are on the order of 25dB/100m and multiple kilometers of fiber are typically used in a FOG. It is important that splices do not add significantly to the intrinsic fiber PER.
To maintain the PER through a splice, the fast and slow axes of the two end faces to be spliced together need to be aligned (Figure 2). End view imaging combined with Advanced Image Processing (AIP) is used to line up the stress rods in the PM fiber. If the stress rods are not symmetric about the core, the AIP will calculate the optimal rotation alignment for maximum PER and minimum loss. Using this technique, PER values >35dB are routinely achievable. These values are generally not significant compared to the intrinsic fiber PER.
For maximum performance, active feedback alignment is an option. Fiber pairs can be aligned and rotated relative to each other while observing PER on a properly calibrated power meter. When the PER is maximized, the fiber pair is fused. Since access to the ends of the fiber is required for the measurement it is not always possible to utilize this technique, so reliable passive alignment is critical.