1.1

General principle

The longitudinal sensor used was previously developed in the frame of a CNES contract to control relative positions of space telescope mirrors, with a resolution of 100 nm. The specified working and displacement range were 0 to 3 m, for a maximal speed equals to 150 mm/s, needing an electronic bandwidth of 200 kHz. The goal was to design small sensor heads, allowing their integration directly onto the structures whose deformations are measured.

The interferometric measurement is based on the use of a Michelson interferometer. To gain place, it was decided to realize it in integrated optics. Waveguides are created using a double thermal ion exchange of silver and sodium in a silicate glass. More informations about their realizations are given in [3].

The interferometric circuit made in integrated optics on glass is the following (Fig. 1):

Fig. 1 :

Layout of the optical circuit

The optical circuit includes:

To get the information about the direction of the displacement, it is necessary to work with two interferometric signals having a π/2 phase shift. In such a system, this is realized by locating precisely the four output single-mode waveguides into the interference pattern [2], [3].

After electronic conversion in the detection unit, by equalizing two by two amplitudes signals and then subtracting them, we obtain the wanted two signals, with the same amplitude and without DC level, independently of the optical contrast value. Measurements have shown phase shifts equal to 90° ± 5°, excellent values that don’t need to be corrected by software.

In this way, we obtain an interferometer insensitive to optical misalignment compared to a bulk optics device, alignment being done by the mask design itself.

Added to the Michelson circuit are a collimator to focus incoming light into the interferometer, two lenses to collimate the measuring beam and focus the return signal, 4 photodiodes with their pre-amplification electronics (Fig. 2).

Fig. 2 :

MOUSE I sensor head

The overall optical head obtained called “MOUSE I” (Metrologic Optical Unit, generation 1) is a light and small sensor, as shown in the Table 1,

Table 1:

MOUSE I sensor head characteristics

*

: with the IASI-type laser source emitting 2 mW + an 1x3 coupler

In interferometric systems, displacement informations come from fringe counting. So, spectral characteristics of the laser are of high importance.

Sensor heads are optically linked by fibers to a prototype of flight models of frequency stabilized laser source developed for the IASI instrument and a detection unit is dedicated to deliver the displacement information.

1.2

Adaptation to the specific needs

The specific need for the HSOB was to be able to “see” displacements of several nanometers at a maximal working distance of 700 mm, in a vacuum environment. That means that resolution has to be better than 1 nm.

Resolution depends on the noise to signal ratio. To improve it, we reduced the noise by limiting the bandwidth using digital filters and we increased the signal by using an update of the MOUSE I laser source. This upgrade was made for MOUSE II development [4]. This has consisted in changing the DFB diode by a more powerful one, compatible pin-to-pin. The laser output power was then 48 mW instead of 2 mW.

To be useable in vacuum, some electronics components of the sensor head were changed, due to their type of package. An other criterion was also to reduce power dissipation.

Interferometric signals treatment is performed in a digital way, using a PXI system, allowing performing high-resolution fringe treatment.

The setup architecture is shown in Fig. 3.

Fig. 3 :

Architecture of the setup

1.3

Resolution measurements

The difficulty of this test is to create accurately very small optical path differences, in a controlled way. The principle is to use the tilt of a thin glass plate through both arms of the MOUSE, the tilt being determined by a piezo actuator displacement and controlled by a standard SAGEIS-CSO interferometer (HC 500).

We use a spherical reflector in front of the control interferometer. Its property is that any beam entering the effective aperture will be reflected by the rear of the sphere, with opposite direction of propagation. The typical translation tolerance in accordance with the beam axis is ± 400 μm (in fact, with small angle approximations, the generated translation during tests will be quite null).

The arm length is the distance between the spherical reflector and the rotation axis.

The value of the displacement can then be recorded during the measurement integration time and erratic movements of the piezo actuator, if any, can be taken into account.

A view of this test bench is given in Fig. 4.

Fig. 4:

photograph of the test bench

Measurements have been made in a covered box in order to avoid any optical perturbations due to air refractive index variations.

The Fig. 5 and Fig. 6 represent displacements values delivered by a MOUSE I sensor head, with and without a 15 Hz pass band filter. Theoretical values are also given by the green curve. These results show that a resolution of 18 pm is achievable, well below the requirement.

Fig. 5 :

Results on resolution tests

Fig. 6 :

Results on resolution tests (part of the figure above)

Resolution measurement on a second sensor head has given a value of 15 pm, making sure of such results.