I. Introduction

Fiber optic displacement sensors have played an increasing role in a broad range of industrial applications. There are two types of the intensity modulated methods usually adopted in fiber optic displacement sensors: the microbending methods [1-2] and the light reflection methods [3-4]. The light reflection method can use a single fiber or a bundle of fibers for transmitting and receiving light signals. As shown in Table 1 the advantages of a fiber bundle intensity sensor, which are described in detail in the body of this paper, are the simplicity of construction; and the precision and high frequency response.

The simplest type of displacement sensor probe involves measuring the amount of light which is transmitted from a bundle of fibers and reflected back into another bundle of fibers from the tested object. It is reasonable to expect that the total amount of light reflected depends upon the incident light power, objective surface reflectivity and the configuration of the bundle of input fibers probe and receiving fibers probe.

These fiber sensor systems can create a response curve which compares the amount of the reflected light and the amount of the displacement. We find a problem with the existence of an apex of the response curve because when the reflected surface is approached, light will not be reflected into the receiving fibers probe. The initial linear rise in the response curve is called 'front slope'. The portion of the response curve descending from the apex is called 'back slope'. Therefore, at close target distances it would have the highest sensitivity, and the lowest sensitivity at long distances. Due to the apex of the response curve, there are two different values of displacement for the same intensity level, and therefore the real distances can be confused.

Many devices have been intended to improve the response curve. For examples, in Kissinger's device [5] a signal beam separator / splitter providing a control for the optical signal to compensate for changes in reflectivity and increasing effective distance away from the target that sensor can operate. In Shigeo Moriyama and Fumihiko Uchida's device [6], three signal detectors perform the automatic correction of the reflection coefficient, then the apex of the response curve remains in one of three detected signals, but disappears in the others. But these devices are still difficult to set-up and align.

It should be noted that in a single-fiber arrangement the front slope disappears because when the reflected surface is approached , light remains to be reflected into the same fiber. This gives us the concept to use a bundle of fibers to replace a single-fiber. Using this idea, with the light source passing through and being reflected back through the same fibers probe, we can obtain a wider

measuring range which is more easy to align than a single-fiber and no peak exists in the response curve. We also refer to the idea of Gary P. Bertollini's device [7] : a cube direction coupler is used to couple the incident and reflected light. This cube direction coupler is a PBS ( polarized beam splitter) which solves the back reflecting light field when light is coming into the bundle of fibers confuses the signal light field.

According to the generalized requirement of a fiber sensor[8, we specified the fiber bundle displacement sensors as follows:

(i) Compatibility with system and installation requirements.

(ii) accuracy better than 0.5 %, desirable of 0.1 %

(iii) diameter less than 10 mm

(iv) absolute position encoding and inherent 'zero' position

(v) MTBF larger than 10[5h

Optical fibers in the mechanical field have many applications[9], and in order to verify this type of fiber optic sensor can apply to monitoring the mechanical behavior; e.g. pressure, vibration, and surface displacement with high precision and sensitivity, we use a plate excited by a vibrator with a tunable amplitude and frequency for testing.

nextindexback