Testing an 68cm F3000mm paraboloïdal mirror during figuring using a quantitive startest.

A first image shows the computer controlled motorized portable mounting (Dobson, Bartels control system with encoders) during the VVS convention in Belgium during which Roger Sinnott of S&T was giving a talk and inspected the monster carefully.

Some pictures of the instrument in the observatory of Johan Coussens in Harelbeke :

mounting the secondary mirror :

first light on august 28, 2002 :

toasting on first light success :

Back to work now :

In the early and cold mornings of november and december 2002 we ( Antoon Maeseele, Johan Coussens and myself Johan Vanbeselaere), mounted a ToUcam Pro and made avi-files of 30 seconds of a defocused star, both in and behind focus at the same distance of focus. This files were stacked with Astrostack to eliminatie seeingeffects. So far we did not correct for defects in the chip itself ( flatfield, noise) as we did expect the mirror still having fairly large errors that would be bigger than the CCD-chip shortcomings. However, as this is a photometric method, our next step will be taking into account this using the webcam or another CCD-camera, having an astronomical friend making the images for us as he did before in the first try-out on his own telescope some years ago.

After stacking, the following images are obtained and fed into a professional program used mainly in adaptive optics to deduce the wavefront ( wavefront curvature sensing).

intra-focal image ------------------------------------ extra-focal image

Some preliminary remarks about details in the images : note the 4 safety clips at the edges, that's also the place of the side-supports. The bigger shadow on the edge is a piece of paper to mark the upside vane in front of the mirror. Images are not exactly of the same size, something to do better in the next session. There are concentric zones visible and the intersection of the vanes does not coïncide with the center of the mirror: if in both images it's in the same direction this means : coma as collimation was far from perfect, done only by pointing a laser at the geometrically center of the mirror and taking care the beam did return in the output of the laser hole of some mm diameter. But we do not know how well the laser is hitting the center and if the mechanical center is near the paraboloïdal center ...
Be not misled by the big offsett of the secondary to be centered in the focusser. The secondary is also a octogonal and it's way of supporting is also obscuring a bit : reason of the strange form.
Right now we're busy finding out how to collimate better. It's a long way of falling, getting up and learning. We did never already collimate such a big and fast mirror, it's diffraction limited field being less than 2mm diameter !!!
Another detail to look for is the relation between outer image diameter and diameter of the obstruction and comparing this ratio between both images : a difference points to spherical aberration, but with this images nothing could be stated with certainty.

What does the curvature sensing program produce ?

First an image of the wavefront :

for sections throught te wavefront :

van Noord naar Zuid

van NW naar ZO

van W naar O

van ZW naar NO

using a free tool found on the net one can convert a GIF file into a heightmap :

An image of a point source at infinity :

Note the very low strehl ratio and the Airy pattern that's totally destroyed. Image scale is in arcseconds.

Same image, scale 2x :

A stacked image of a moon of Jupiter :

A PSF image in different directions :

An image of a synthetic interferogram 8° tilt:

orientatation : 0°

Without words ...

interferograms with different orientations

30°

60°

90°

120°

150°

180°

Interfergrams with 0° tilt and in light of 0.125 micron wavelength :

in light of 0.250 micron wavelength :

These must be lines of equal height with a difference of half the used wavelength but of the wavefront. So the surface is only half that heigh. That means these lines show a height difference of the uses wavelenght/4 ?

Have to verify ...

But these are the most practical images to retouch the mirror surface, but right orientation and height value and sign still needs to be verified !

Finally and perhaps the most interesting : a table of the first 22 Zernike coëfficiënts :

revealing more detailed the errors in RMS nm

and a graph showing relative importance :

As collimation is not yet ok, coma and astigmatism should appear, coma being stronger but both in the same direction !

Coma is indeed stronger, but not pointing in exactly the same direction as astigmatism.

This results are however to incomplete to deduce hard facts about the mirror surface and only an intermediatie step into getting acquainted with such a professional way of testing. Big advantage is a quantitive test with the sensitivity of a startest. In the near future we'll try to compare it to interferometertests indoor, another test we're trying out in the mean time..

At the time making the images, Jupiter did challenge us, so we couldn't resist webcamming, stacking and unsharpmasking it also with the following result :

An image not worth an 68cm, we know, but we try to make it better ... the mirror, not this image ;-)

In still a former stadium we did take some snapshots of Saturn :

no comments, for future reference ...

first light image of the projected sun on a cardboard :

I'd like to thank here my buddies : Johan Coussens and Antoon Maeseele for sharing the excited moments but also for leaving in the middle of the night their warm bed and family, to get out in sub-zero temperatures and hard wind making unsharp images of some star with a webcam : some are doing less crazy things to be my friend ...

Johan Coussens, alias Wilfried , met al die Johans ...

Anton aus ..., alias "den oeden"

I don't want to forget the Astronomical Contact Group, of whom we are members and who is the owner of the mirror, and its president Philippe Vercoutter giving us permission, to take the mirror away and "play" with it somewhere else.

And now back to work : better testing and trying to better the mirror itself ...

Vanbeselaere Johan, december 2002

to be continued ...