Attached are three photos.

Transcription

1 Attached are three photos. First, I'm a retired engineer and you can probably see that machining is my hobby. I don't know if you are familiar with Shack interferometers, but assuming you are, here are some details: The interferometer has two main configurations: One uses a home-made Shack cube which puts a right angle bend in the main beam. This was constructed first, and is shown in the "home" photo. The aluminum tube extending to the back of the photo holds the two Offner null lenses. Later in my experiments I acquired a very good quality professional, or "custom", Shack cube on long-term loan. In the "custom" photo, the cube (which is actually a cylindrical beamsplitter with a spherical reference surface) is partly seen at the left end the large aluminum housing. The housing also holds the microscope objective. This setup has a straight-through main beam that exits to the left. Viewing is done from the side, perpendicular to the mirror axis. The overall platform has a surplus HeNe laser directed to the right and then steered by "fold" mirrors at the far right. The beam then passes right-to-left in front of the laser and through a negative spreading lens on its way to the microscope objective, which is hidden in both cases. Both cubes have 5 micron pinholes. Alignment is extremely tricky because the objective's focused beam must go through the pinhole. I hope this was helpful. When I get accustomed to the yahoo group I'll post some photos if people want to see them. Paul Lind Here is a bit more detail. The "custom" cube is on loan to me by an optician friend. It was made several years ago by a different optician, David Anderson. I don't think the cubes are for sale.

2 The "homemade" cube was assembled from three parts. First is a 20mm beamsplitter cube from Surplusshed.com, $25. Second, an uncoated 50.8 mm FL plano-convex lens, 25.4mm diameter, RoC=26.35mm, and Thickness =5.2mm, an $18 stock item 01LPX113 from Melles Griot. It is cemented to one side of the cube. The combination of 20mm and 5.2mm places the center of curvature (CoC) of the lens about 1mm outside the opposite side of the cube. (The important thing is that the radius of curvature of the lens is slightly greater than the sum of the lens thickness and cube thickness). The third part is the 5 micron pinhole, which is more expensive. It is NT from Edmund Optics, $42 for the unmounted type, which I glued to a little spacer about 1mm thick to place the pinhole at the center of curvature of the convex lens surface. I believe a larger pinhole would be best for most amateur work. A common complaint about Shack interferometers is alignment difficulty. That is, placing the pinhole at the precise CoC of the lens surface (reference surface), and focusing the laser beam onto the pinhole. Two brief references are: Optical Shop Testing by Daniel Malacara, p 35. The first chapter of this book provides an overview of interferometers types. A.pdf gives brief description in paragraph Alignment was a real problem for me. Another interferometer that's worth studying is the Peter Ceravolo design. It is for mirrors slower than about f/3, and it is easier to align. However, it uses two fairly expense parts: A tiny GRIN lens for $160 produces a point source, and the reference element. Ceravolo sells these for $410. Ceravolo's article about building a "Spherical Wave Interferometer" can be downloaded in PDF form at:

3 I also believe his interferometer could be improved if the cube beam splitter were replaced by a pellicle beam splitter, but those are more expensive. Regards, Paul About the lens question: The lens surface should be better than the mirror under test, but only over the diameter that illuminates the test mirror. Note that the mirror's f-number at the center of curvature (CoC) is double the f-number at focus (see Ceravolo paper). Therefore an f/3 mirror has an f/6 light cone at it's CoC. To test this mirror, a "reference surface" with 25mm radius or curvature (distance from pinhole) must be accurate over a 25/6 = 4.2mm central spot. Inclusion of a null lens raises even more questions about surface accuracy. Unfortunately, manufacturers don't specify lenses this way. They indicate something like 1/4 wave per inch. About the Pinhole; I don't know what determines the maximum pinhole size. An expert told me that diffraction from the pinhole is not desirable. Rather, the focused beam should pass through the pinhole mostly concentrated in the center without too much energy hitting the pinhole's edge. Then it's possible to produce wide beams for testing f/1 mirrors, which is common today. Otherwise, if the pinhole is evenly illuminated, it produces an Airy disk pattern and the disk is too small for reasonable sized pinholes. So, apparently, the beam coherence across the mirror surface is determined primarily by the laser's focusing lens. The pinhole apparently has other purposes such as helping with alignment and blocking outer diffraction rings from the source. Regards, Paul