Fiber Pulling Research Articles

Identification of an Optical Fiber Pulling Process

Abstract The optical fiber pulling process presents an interesting case study of a relatively simple nonlinear system which may be identified with classical techniques. This article deals with such a project carried out in cooperation with the Nordiske Kabel-og Traadfabrikker A/S (NKT). An optical fiber is pulled at high temperature from a solid quartz rod called a preform. The fiber must be pulled from the preform in such a way that its diameter is constant within very close tolerances. Such tight tolerances require automatic control but implementation of such a control strategy requires that the dynamics of the process are known. An identification of the dynamics of the pulling process has been carried out using a variety of methods. These methods have given consistent results and have resulted in a model for the process which allows a simple and effective control strategy to be implemented. A controller has been designed and tested with the actual pulling apparatus. This controller is much simpler and gives much better results than the phase locked loop control used earlier. It is now being used in the daily production of optical fibers at NKT.

Axial stress and its effect on relative strength of polarization-maintaining fibers and preforms

A method is given for the calculation of axial stress in polarization-maintaining optical fiber and preform. Detailed calculations are presented for circular side-lobe (CSL) and bow-tie fibers and preforms. The axial-stress results explain why preforms are more fragile than fibers in these designs. On the other hand, the stress analysis also shows that these fibers are more crack resistant at the surface than conventional single-mode fibers. To reduce the fragility of the fiber, it is suggested that a larger drawing tension be used during the fiber pulling process.

Water impurity in low-loss silica fibre

Water contamination in silica based optical fibre waveguides has been studied. The height of the 950nm water peak was found to vary with fibre pulling speed and reduced sharply from 64 dB/km to 20.5 dB/km as the speed was increased from 0.12 m/sec to 1.47 m/sec. The results have been fitted to a diffusion curve indicating that the water diffuses into the core region from the support tubes used in preform manufacture. Deposition inside a water free Spectrosil tube gave a water peak of only 4 dB/km. A B 2O 3 doped cladding layer was used to provide a barrier to OH migration and the 950nm water peak was reduced from 24 dB/km to 3 dB/km by increasing the B 2O 3 content in the layer from 2 to 4 mole per cent.