Minimum Measurable astigmatism with Shack-Hartmann WFS

[David Miller]

Hi,

I just found this forum, and it seemed like the perfect place to ask about this. I have an older Shack-Harmann WFS from WaveFront Sciences. I was using it to check colimation for a holography set-up, and no matter what, the minimum measured astigmatism was ~lambda/10. All other aberations (defocus, coma, spherical) were ~lambda/50. THe RMS wavefront deviation is ~lambda/30, mostly due to astigmatism, it would seem.

After some digging, I found a Blue Sky Collimeter in the back of a cabinet in the lab, and used this for an independent measure. The beam *blinks* on and off - so it would seem the phase is very flat.

I tested the calibration of the WFS by placing it several meters from a pinhole, and measuring the radius of curvature. It was within the error of my tape measure.

Is there a minimum phase error a WFS, or its software, can measure correctly? As in, if the error is too small, numerical noise or something else becomes an issue? Could it be that after all these years, the hardware no longer matches the calibration file?

Thanks for any feedback,
David

[Phil Hobbs]

Post by David Miller
Hi,
I just found this forum, and it seemed like the perfect place to ask
about this. I have an older Shack-Harmann WFS from WaveFront
Sciences. I was using it to check colimation for a holography
set-up, and no matter what, the minimum measured astigmatism was
~lambda/10. All other aberations (defocus, coma, spherical) were
~lambda/50. THe RMS wavefront deviation is ~lambda/30, mostly due to
astigmatism, it would seem.

I don't know that exact unit, but in general Shack-Hartmanns are the
pits except for adaptive optics, where speed is more important than
accuracy. Fixed shear plates are also crap, even the shiny models with
two plates and two cameras. I had one some years ago that showed every
wavefront as a potato chip, even when I put a piece of transparent tape
halfway across the beam. Pure crapola.

The problem with S-Hs is that the actual data are low-quality
measurements of local wavefront slope, which have to be integrated
across the field to get the P-V numbers. It's super easy to get that wrong.

Post by David Miller
After some digging, I found a Blue Sky Collimeter in the back of a
cabinet in the lab, and used this for an independent measure. The
beam *blinks* on and off - so it would seem the phase is very flat.

I have two, and have relied on them for 25 years. Highly recommended.

Post by David Miller
I tested the calibration of the WFS by placing it several meters from
a pinhole, and measuring the radius of curvature. It was within the
error of my tape measure.

Dunno.

Post by David Miller
Is there a minimum phase error a WFS, or its software, can measure
correctly? As in, if the error is too small, numerical noise or
something else becomes an issue? Could it be that after all these
years, the hardware no longer matches the calibration file?
Thanks for any feedback, David

Cheers

Phil Hobbs