Building+version+R03+of+the+Bath+Interferometer+(Right+angled+layout)

Below shows the latest version of the Bath interferometer. More testing has improved development and understanding of the system.

The system consists of:-
 * Green laser from Maplin (N02KH) of 532nM wavelength (£14)
 * Small 10 x 10mm non polarising beam splitter (on loan, thanks John) These can be purchased from Night Optical UK for approx £60
 * First surface mirror 10mm high x 300mm long mounted on a tip/tilt/rotation mechanism
 * 5mm diameter approx 6mm FL webcam bi-convex lens mounted in webcam lens housing (Maplin/Ebay for a few quid)
 * Sweex HD webcam and small doublet lens to frame mirror under test (FL of lens will vary depending on mirror FL and diameter to test)
 * Linear stage for Y axis (on loan, thanks Ian)
 * Proxxon Micro compound table for X/Z axis adjustment (Ebay £80)
 * Scrap aluminium plate and angle parts
 * Power supply and switch (3 x AA batteries in a box). Green laser cannot be adjusted for brightness, so no need for a pot.

The image below shows a Phillips Toucam used for imaging. This has a doublet lens to frame the mirror under test. This camera has now been replaced with a Sweex HD webcam for higher resolution with an appropriate doublet to frame the mirror under test. The lens is mounted in a modified webcam lens housing. The optical parts of the R02 version have been used, but the mechanics have been rebuilt completely to allow for fine adjustment of the optics. The camera bracket is yet to be completed. The lab jack is ahndy for camera adjustment. The lens was taken from webcam lens. The exact FL I am not sure yet, but it must be quite short (6mm maybe) to diverge the narrow green laser beam.

The lens now allows for the beam to diverge enough for F3 mirrors (see lower sections for testing a sphere). Separation distance is now less than 10mm. The housing of the lens from the webcam was re-utilised and allows for easy replacement with another lens of different FL for other optics if required. It also enables the lens to be made square to the beam. A piece of Foamex was used to hold the lens. A 12mm hole was drilled to hold the lens in place.



The small flat mirror was attached to a tip/tilt mechanism for fine adjustment to align the beams. The beamspiltter can be rotated about it's centre also. The laser is a Maplin code and can be adjusted make the beam parallel wil the optical axis of the system. Odd pieces of aluminium plate and angle were used to connect all the parts together onto the stages. One of the best and simplest layouts I have seen on the web is as below (not sure where I got this from now). If I re-make in the future, the mirror adjustment would look similar to this.

Images still have some dust particles, but have much more contrast! Cleaning the small optics is a challenge in its own. = = =__Testing a Parabolic mirror__=

This test was not meant to be an accurate test but a 'ball park' test to see if OpenFringe would give the numbers somwhere around what I was expecting for the mirror under test. This was the same mirror as used in R01 and R02 and is of good to excellent optical quaility on sky. It is a David Hinds mirror cerca 1980 that I bought to make my first real telescope.

__Videos of the test:-__
The link below shows the mirror at rest, part way through the freezer in the garage kicks in and you can see the fringes move about.
 * [[file:2012_12_27_Bath_Interferometer_R03_010x.avi|2012_12_27_Bath_Interferometer_R03_010x.avi]]**

The link below shows me moving the stages very slightly to adjust the fringes
 * [[file:2012_12_27_Bath_Interferometer_R03_011x.avi|2012_12_27_Bath_Interferometer_R03_011x.avi]]**



Star test images indicated quite good inside and outside definition. At the moment for initial testing, only one image has been used. No multiple images for averaging have yet been taken till we get to grips with the basics of using the software and setting up the system.

How FFT analysis works is still a little unclear to me, but following the procedures on the web seem to work so far. More reading and testing required. The quaility of the image, number of fringes etc seem to be quite critical in obtaining good results.



=__Averaging and rotation__=

A few days later (early Jan 2013) I was aware that I was not using the latest version of OpenFringe! I was using version 10 from the wiki. This had issues with averaging wavefront files. I joined the tech group and downloaded version 13.9. This now works a treat in averaging. Version 10 appeared to be putting in SA when averaging wavefront files, but not when averaging Zenike files? The image below is the result of 5 igrams for 0, 90, 180 and 270 degree rotations (20 in total) for the 221mm parabola. There appears to be some minor treffoil left in the surface, but the mirror looks good.

=__Testing a sphere__=

I was then loaned a good sphere. 200mm in diameter and around F3 FL so quite short at approximately 1185mm ROC. The Bath should provide good data on this mirror and hopfully prove the system including the diverging webcam modified lens. We intend to put this mirror through a number of tests by different people and reset the configuration of the set up each time. We hope that the data will give near consistant results to prove that the system is reliable and good enough for us to start figuring mirrors we have to complete. Initial tests showed that the F3 spherical mirror did indeed appear to be spherical (or close to it) as teh fringes could be nulled straight. I could get 4 or 5 fringes to be quite straight across the mirror. For sampling in OpenFringe, I defocused teh fringes to get more of them for testing. A Sweex HD web cam is now used for imaging to get better sampling.

The images below are for future reference for the manual settings of the Sweex HD webcam (bought from Maplin). The auto settings for teh Amp were no good and oversaturated etc with no visible fringes. Camera exposure was left as auto and seems ok. Keeping the optics clean again is a big issue and not easy to do. Some development on that is required to remove dust etc. I found it difficult to remove the bright laser reflections too in the fast F3 system. One can be seen in the image below near to the left hand side egde.



More fringes are obtained by de-focusing the system very slightly. Apparently this is needed for FFT analysis? Again, keeping the optics clean and no reflections is crutial. The is one reflection and a large piece of dirt in this image below, but each time it is cleaned, a different piece of dirt will appear somwhere else!

__Video of the sphere under test:-__
Link to the video below shows teh sphere under test and me adjusting the stages for finge nulling.


 * [[file:2012_12_28_001_test_sphere_200mm.avi|2012_12_28_001_test_sphere_200mm.avi]]**

=__OpenFringe Reports for the 200mm F3 sphere__=

The two images below show the 'Report' page for the spherical mirror. The upper image has 1st order astigmatism and coma disabled. This is the default method it appears for OpenFringe. The second image I enabled them. It can be seen that the quaility is degraded a little. This may be mainly due to astigmatism inherent within the Bath interferometer set up due the path separation distance. At F3 this would exaggerated and appears to be around 1/4 wave at the wavefront. Playing with the Zernike terms, the largest component is the X axis astigmatism term which happens to be the beam separation axis too.

For the formula OPD = D2 *d2/16*R2 D = Diameter of mirror d = Beam separation R = Radius of curvature OPD = Path difference (divide by wavelength to get the wave error)


 * Focal length || F || 3 ||
 * Diameter of mirror || D || 200 ||
 * Path separation || d || 10 ||
 * Radius of curvature || R || 1200 ||
 * Path difference || OPD || 0.000144676 ||
 * Wave error 550nm ||  || 2.63047E-07 ||
 * Wave error 550nm ||  || 2.63047E-07 ||
 * Wave error 550nm ||  || 2.63047E-07 ||





__Further testing for astigmatism introduced by the Bath method__
Further testing and adjustment of the X axis astignmatism was done. I used 2 different interferograms of different amounts of fringes in each. The images were cleaned up in PSP to remove the dust particles first. Each image was loaded and screen dumps taken with astigmatism and coma removed, then X axis astigmatism enabled. The results from each of the interferograms was similar showing around 1/4 wave difference when the astigmatism was disabled then enabled. Note that the astigmatism is in the X axis. All other terms are relatively small. Therefore this is encouraging that fringe orientation and number of fringes (for a sphere at least) appear to be consistent.

Image 11x results
Image below has X astigmatism disabled Image below has X astigmatism enabled

Image 12x results
Image below has X astigmatism disabled Image below has X astigmatism enabled

=__Conclusions__=

The document linked below is the best document to use for a detailed understanding of building and testing. It has methods for using OpenFringe too.

[]

We need to do more testing and refining of the system. We need to understand FFT and how to use OpenFringe properly. The latest version of OpenFringe resolves some averaging issues I had with version 10.

The tech group has lots to offer including latest version of OpenFringe (13.9 as of December 2012) []