Bath-interferometer

At the end of 2019 a workshop Bath-interferometer was announced by the Almere observatory. Earlier that year I had attended a Foucault and Ronchi workshop and wanted to expand my knowledge of optical testing methods by attending this workshop as well. In contrast to the first workshop, this time the workshop would include the construction of a Bath-interferometer.

This device (see figure 1) is slightly, although not much more, complicated that the Foucault-test device. A collimated laser-beam is used as a light source (red lines in figure 1). The beam passes a diaphragm and then enters a beam-splitter. Half of the light goes straight on and passed a biconvex lens of short focal length that initially converges the collimated beam, but after a few millimetres causes it to diverge (still drawn in red in figure 1). The other half of the light makes a right turn (shown in orange), is deflected left by a first-surface mirror and hits the mirror-under-test at the centre. That beam is reflected (still collimated, now shown in green) and ends up at the biconvex lens of the Bath-interferometer, which makes it to diverge again. This latter diverging beam is the reference-beam, which behaves as if it originated from a perfect spherical mirror. The first beam (the one that went straight on from the Bath-interferometer) was reflected from the whole surface of the parabolic mirror (shown in yellow) and therefore carries the information of that whole mirror-surface. It is aimed at the first-surface mirror of the interferometer and both are combined, creating an interference-pattern as a result (dashed green/yellow lines).

The idea was to build the Bath-interferoneter in Almere, but sadly enough COVID19 intervened and the workshop was postponed several times. Finally, in October 2020, it was decided to give the workshop over the internet. All attendees were sent the parts that were ordered for the workshop and the construction could be done at home. The basis of the interferometer was 3D-printed and partially constructed by the observatory, all that the attendees had to do was soldering a resistor, a potentiometer and a 9V lead to the laser and to mount the whole contraption on the XYZ-table that came with the set. To allow the interferometer to be placed on a photographic tripod I created a baseplate with a 1/4" threaded hole in it. Finally, as I wanted it to work with a fixed camera, I created a T-bar with a camera mount for my ZWO ASI290MC and ZWO 2.8-12mm zoom-lens. (see figure 2).

As the external surfaces of the square beam-splitter could potentially cause reflections, it is common to have the beam-splitter mounted at an angle. This, however, requires the camera to be rotated to an angle twice the orientation of the beam-splitter. So soon after I had created the first camera-mount, I modified it to make the camera orientation 10° to allow a 5° rotation of the beam-splitter (see figure 3).

With the 10°-offset camera-mount the Bath-interferometer was ready for use, but like with the Foucault-test set-up, the position of the camera is quite critical and so is the parallelism of the laser beams.

First the beams are made as properly parallel as possible. Being not dissimilar to the design of a sextant, aligning the optics of a Bath-interferometer can be done in the same way. Just aim the device at a distant object with vertical and horizontal lines and adjust the first-surface mirror until the direct and double reflected (i.e. the image as seen through the beam-splitter and first-surface mirror) images coincide.

Getting the camera in the correct location is done by looking at the reference beam, which has to cover the direct image of the mirror-under-test (see figure 4). It is simply a matter of shifting the camera sideways and up- or downwards until the reference-beam completely covers the mirror. Depending on the camera lens it may happen that the reference-beam remains too small. The size of this beam depends on how close the lens can be placed to the beam-splitter. In case of the ZWO zoom lens, this meant that I had to remove the most forward ring of the lens, in order to allow it to get just a few millimetres closer to the beam-splitter.

Once first-surface mirror and camera are properly adjusted, the camera will show an interferogram (igram) of the mirror (see figure 4 where the mirror has a slight offset to the left). These igrams can be recorded and processed in software like DFTFringe.

In order to avoid instrument-induced errors, the images are shot with the mirror in eight rotational positions (i.e. with the mirror rotated in 45° steps) and with the igrams rotated every 45°. Multiple images can be taken of each igram- and mirror-position in order to average the results.

So far two Newtons in my possession have been analysed, both sadly enough showing astigmatism.

Like the Foucault-test device the current set-up delivers images at a low scale, something I hope to improve upon by replacing the ZWO lens by a 25mm or 35mm fixed focal length lens.