Hardware tests
Bath interferometer feedback Sunday, 21 March 2010 14:47 |  Written by chonum |  |  | 
I've been using the Bath interferometer quite a few time now, so i write down this article to give back my overall feedback. I don't come back on the Bath design, it has been thouroughly commented by Charles Rydel and Dale Rowe, the author of Open fringe. Mecanicals The most tricky part when building a Bath interferometer is about adjustement needed on the different mecanical and optical components. Here is the exit beams layout on a classical 90° layout :
Here are the adjustments i have perceived as mandatory or strongly advised :
Height axis (Y) in red. Interferometer X and Y translation for the tilts, and Z for the focus. Good sensitivity needed (some µm). Scale on Z useful for chromatism analysis. Laser Rotation X for the beam to be center on the biconvex lens. Some ArcMin sensitivity. Beamsplitter (cube) Rotation X to remove secondary beams from the Igram. 1° sensitivity. Angle Mirror Rotation X & Y to make both beams parallels. Some ArcMin sensitivity. Biconvexe lens Heigh translation to center the laser beam. 1mm sensitivity. 90° rotation or the ability to remove the lens. Note that the cube and mirror settings may be fixed when correct. Take care of possible vignetting for very open beams. A lens should be used on the camera to avoid diffraction issues. Bath interferometry Mirrors Mirrors are tested at their radius of curvature which can be either the center of a spherical mirror, or the local center (twice the focal length) of a parabolic mirror. In this case (parabolic), the test introduces a spherical aberration due to the fact we are not working on a flat wevefront. This SA can be calculated this the radius, the ROC and the conic constant of the mirror, and then removed from the Zernike Polynome. Nevertheless of the mirror is large and/or very short, this SA becomes important regarding the precision we are looking for (some L vs some L/100s). In that case, the test should be performed in autocollimation or this a Null lens. That limitations is true for ALL testing methods (Fizeau, bath, Wavefront analysis...). The astigmatism induced the the off axis reference beam is to be taken into account as well. It is negligeable though with average mirrors with not to short F/D ratios. Open fringe can calculate this error. Subtracting it directly from the X 3rd order term of the polynome is tricky though as we must be sure of the igram alignment on the X axis. Summary : + very goog precision + Zone defaults can be seen + Zernike interpolation + Tests are very reproductable + Fringe tracing and change of operator doesn't change much results. + PTV, RMS, FTM and PSF - Induced astigmatism can be an issue on large/short F/D mirrors. - Spherical aberration when testing asphericals at the ROC. Refractors Mirror are autocollimated on themselves, but in the case of refractor the collimation is done on the flat mirror. So we are never sure to be on the optical axis of the instrument ! That means that when lookink at the astigmatism, we have three components :
- From the instrument : lens alignment, lens defaults and mecanical constraints. This is this part we would like to measure. - Induced by the beam separation of the bath setup, the reference beam being off axis. It can be calculated (185nm here), but it's difficult to remove it directly due to the lack of precision in the DSLR alignment on the X axis. - The fact that we are never sure to be on the optical axis with this double pass setup. So astigmatism may become important on small and/or short refractors (below 100mm). At the contrary of parabolic mirror, we don't have any SA when testing refractors as they generate spherical wavefront. + Very good precision excellente for 5th order and more, and on 3rd order spherical. + Infinite-focus combinaison
+ Zernike interpolation + Tests are very reproductable + Fringe tracing and change of operator doesn't change much results. + PTV, RMS, FTM and PSF + possibility to analyse chromatism and sphérochromatism. - Impossible to separate instrument astigmatism from test induced astigmatism. - need a good flat mirror - difficult to know when on optical axis Conclusions Bath is a very good metrology method for the amateur. It's cheap, easy to set up and it shows much more than the Foucault. It's reproductable and shows few sensitivity to the operator. Nevertheless some limitations have to be taken into account such as the induced astigmatism for short F/D, or the ROC limitation when testing an aspherical mirror and therefore it cannot replace professional tests.
Last Updated (Tuesday, 30 March 2010 16:53) |
Two Sky Watcher 80ED facing the interferometerSunday, 21 March 2010 13:22 | Written by chonum | | |
We had a little Bath with the Chatenay Malabry astroclub named Aphélie. We tested a couple of classical 80ED skywatcher refractors. One was 3 years old, the other from 2009.
As usual now, the tests have been done at three different wavelength : 405, 532 and 635nm. For each lambda, three igrams have been done with different tilts. With this double pass setup, fringe are separated by 1/2 L.
The main issue with this double pass Bath setup is regarding with the astigmatism, coma being negligeable both by the optical design and in the numbers. Astigmatism comes from three different sources : - From the instrument : lens alignment, lens defaults and mecanical constraints. This is this part we would like to measure. - Induced by the beam separation of the bath setup, the reference beam being off axis. It can be calculated (185nm here), but it's difficult to remove it directly due to the lack of precision in the DSLR alignment on the X axis. - The fact that we are never sure to be on the optical axis with this double pass setup. This one won't show when testing a mirror of course. This is an important drawback of the Bath setup when testing a refractor, especially with small lenses (below 100mm). Overall performances are good. But it's impossible to know where they stand exactly between the two values with and without astigmatism. We can just notice that the 3rd order astigmatism is quite the same on all results : X being 0.1 L and Y 0.03. The X includes the off axis induced astig component. The old one is slightly less good than the other one. Nevertheless the optical design looks similar. It's representative from a ED doublet with a very low dispersion Crown component (FPL53 here). It's close from the FS-60 for example. Be aware that during the test, when changing of laser source, the focus point shift from several tenth of µm, that shows that with telescope presents some chromatism.
Last Updated (Tuesday, 30 March 2010 16:55) A Starblast 113/450 surrounded by Zygo, Haso, Bath and FoucaultFriday, 12 March 2010 09:49 | Written by chonum | | |
Hi guys, after the quite contreversial testing of both the SkyWatcher and the Celestron, Jean Luc Dauvergne from Ciel & Espace magazine lend me a small starblast that has been tested with a HASO and a Zygo. Before he gave me those results, i tested this very starblast with the Bath and get a picture with the Foucault.
The Shack hartmann has been performed on a real star. Even if it is not as biased as with a Roddier for the integration is done on discrete measurements and not on one exposure blurred by the seeing, that leads to better results than on a breadboard. Still, the difference with the Zygo is slight and the overall wave front characteristics are pretty much the same with some astigmatism, and a spherical aberration we can guess under the central obstruction. The Bath is performed at the radius of curvature using a software nulling. With such a small and opened mirror, the astigmatism induced by the reference beam being 10mm off axis isn't negligeable. As i don't know how to remove it (i've got a module, but no angle), i present results with and without 3rd order astigmatism. And without : Results are pretty close from the Haso and Zygo's, except that as i'm testing a complete mirror, the spherical aberration we can see on the central part decrease the overall RMS. A FFT test (more fringes, no tracing but Fourrier transform) gives a better view of the medium frequencies errors. We can see the central bump, and another circle at mid aperture. Note that this test requires a super clean Igram free from speckles and diffraction artefacts ! Which is not absolutly the case here (that gives two peak artefacts) : Taking into account the central obstruction, we get close from the professional gears results : Here is the simulated Foucault, smoothed by the Zernike interpolation : And the real onel : This beginner's instrument is quite good. It looks like as for the two other mirrors Syntha doesn't handle the central part, but in that case it's with very small consequences. The mirror is really parabolic with a conic constant of -1 Last Updated (Friday, 28 May 2010 22:56) Interferometry test : Takahashi TOA 130Friday, 05 March 2010 09:48 | Written by chonum | | |
Hi all, the happy owner of a Takahashi TOA130 came by to inflict some torture tests to his baby. Fat, heavy and very cold baby by the way as it spended the whole day in his car trunck... After one hour lost trying to align the full test bed, we could finally start !
As before, i remove third order astigmatism and coma as we never know for sure if we are right on the optical axis when testing in autocollimation a refractor. All values are normalized to 550nm. Precessing is done with Open Fringe 11, and cross checked with Atmos Fringe.
Conclusion : That is an amazing refractor, able to swallow from red to purple and to deliver exceptional performances. But that is also a big cat and it requires a long temperature adaptation as for a reflector. Who said that refractors were always ready ?
Last Updated (Sunday, 21 March 2010 13:19) Celestron 150/750 testedThursday, 04 March 2010 09:31 | Written by chonum | | |
As done previously with its cousin Skywatcher, i tested the Celestron 150/750 beginner's product. Let's see if it's handling better than the blue one which got awful result. As before, test is done with three interferograms from a Bath setup processed with two different softwares. Test is done as the center of curvature and induced spherical aberration is nulled by software. Wavelength for test is 532nm, and normalization at 550nm. At the first sight when looking at the igrams, we can see there is an issue on the central poart of the mirror. The synthetic Foucault shown that as well as the wavefront 3D map. The real Foucault image shows it as well : a severe bump in the middle :
Simulated Foucault :
This mirror is slightly better than the SW, and it's surface is much smoother. But it's still quite poor this one lambda PTV and L/6 RMS. This kind of important defect should have been seen during QC.
Simulated PSF :
Last Updated (Sunday, 21 March 2010 13:19) |
It gets a bit better. Eccentricity closer to -1. RMS a bit higher, but still low in absolute.