Here we check an air bearing worktable of a roundness measuring machine Mitutoyo RA-211 to demonstrate the accuracy and sensitivity of our measuring system. This roundness measuring machine had not been used for some time, and the air cleaner had not been activated for a long time, but we doubt there was any trouble in the air bearing of the worktable. In the interferometer, each fringe shift corresponds to the relative distance of two mirror [6], on other hand, fringe shift has the relation with the vibration isolated effect of the roundness measuring system. The quality of the image of the fringe is not as excellent as can be expected in the standard Michelson set-up but is still good enough for signal processing in the next step. The single mode fiber is difficult to align to behave as a light source of a Michelson interferometer.

We can make it easier if in the pre-process we place an objective lens in front of the laser light without a single mode fiber. The plane mirrors 1 and 2 are mounted in the position of almost equal optical paths under study associated with the suitable hyperbolic fringes.

We must adjust the fringe patterns to what we desire. Then we substitute the objective lens with the single mode fiber, and readjust the light field again. The screen is located on the rotating object. We can use a wireless CCD camera to trace the screen and a video frame grabber to analysis the fringe variations. Although the scale altered a little in the fringe pattern, it would not affect the fringe's phase variation.

When the pressure at the air regulator of the roundness measuring machine reaches the specified value of 4 kgf/cm2, we turn on the worktable rotation motor to see the fringes of the interferometer attached on the rotating worktable. In the meantime, the wireless micro-CCD camera begins to record the fringe variations. The standard speed of rotation is 6 rpm.

Because every fringe line in these images is wide and difficult to calculate, we must thus use a thinning technology to extract the skeleton of the fringe. In Fig. 3 and Fig.4 the video image of the hyperbolic fringes and their thinning patterns are shown. We can obtain the value of the angle inclination from the computer simulation of the hyperbolic fringes in Fig.5. The parameters of computer simulation are listed as follows:

(a). h = 300 mm;a =0.27° ; d = 1mm;

(b). h = 300 mm;a =0.2724° ;d = 1mm;

The unit of the axis scale in the figure is millimeter.

In order to make the comparison for our experiment, we bring the electronic style of the roundness machine close to the worktable, and rotate the worktable by automation observing the results of the recording pen of the roundness measuring machine movement. The stylus is held in a spring clip and comes into contact with the worktable's top surface. Surface displacements of the stylus are converted into electrical signals which are fed to an amplifying stage where magnification up to

´ 5,000 is obtained.

The jump in the fluctuation (Fig.6) occurs in the mean while the interferometric fringes moves may be accounted for by dust and rust in the air bearing mechanism. A reading of the position of the jump in fluctuation reveals it occurs in the mean while the interferometric fringes moves in the modified Michelson set-up.

Although classical electronic sensors have good sensitivity for a few nanometers,

the results of our measuring system are still more reliable because all of the components are on the worktable in our system, but the stylus is mounted on the side of the worktable. It is therefore doubtless that a relative movement exists in the worktable and stylus.

index