Webb Space Telescope

[dtvmcdonald]

The Webb Space Telescope finished unfolding today. Next step
is aligning all those mirrors.

Speaking of mirrors ... have you noticed that NASA has a habit of making
it impossible for someone to find that actual specs of their things.

I wanted to find the specs of the mirrors in the main telescope (there are three). Official values are just not there. I did find one set of values
that somebody cooked up to make a published drawing, but in
a simulator they were not accurate enough to give the specced performance. I wanted to see what it was like .... optics used to be
the heart of my business, including ones cooled to liquid nitrogen
or liquid helium temperatures like the Webb.

This really ticks me off. I think its a conspiracy. So I got playing around.

They do give some good large scale drawings that all look alike.
They give the numbers for the exact diameters, and the overall focal length, so I have the scale. I measured the drawings and fed that into
my cheapie (i.e. free) simulator. I've had expensive (by academic standards, that is, like $10,000) optics made up using designs from that program and the devices work as specced.

I used the auto-optimize to get the curvatures and asphericities and tilts.
This is very highly overconstrained and resulted in a unique solution.

The results when simulated match their specs nicely.

Its interesting to note that the design is NOT diffraction limited at wavelengths shorter than 2 microns, at least not over
most of the field.

At long wavelengths the resolution is, of course, abysmal due to diffraction.

I hope this gizmo is more of a success than the European (CERN) LArge Hadron Collider, which really really disappointed lots of high energy physicists, as it works perfectly but showed that the particles they expected do not exist.

[old_tv_nut]

Could you explain a bit more about "not diffraction limited?" What is the limitation at the shorter wavelengths?

[dtvmcdonald]

I'm not sure exactly why they say its not diffrection limited. My design has a max RMS wavefront error about 0.09 micron over a 200 mm diameter.
Its less than 0.03 micron over a 100 mm diameter fileld. But they are using Strehl ratio of 0.8 as criterion. That probably means that they are including things like surface roughness (I think they quote 38 nm typical value),
figure errors (unspecified) , the impossibility of spending forever tweeking the mirror positions, plus diffraction due to the spokes and the hexagon separations. Adding all that up say 0.07 micron and
.12 micron for the two fields. You multiply by 12 to get the wavelength for
a Strehl ratio of 0.8. This gives about 0.84 micron and 1.44 micron
the the diffraction limited wavelength for the two field sizes's edges.
That's reasonably close to their spec of 2 microns.

Given that the thing has lots of edge differaction problems, its
gonna give images for stars that look awful.

The various detection devices are fixed in the fields, with the
longer wavelength spectrometers in the outermost areas. The images of
galaxies with the shortest wave imager will look nice.