Re: ATM Loveday Secondary

[Anthony Stillman <atmer@flash.net>]

>...for a flat detector field curvature
>is by far the dominant aberration...

>Coma and astigmatism cancel out in the first two reflections, but there's a
>small contribution from the 3rd reflection...

>Mike Peck


Mike, others,

I agree that for a sufficiently large and sufficiently curved field,
defocus becomes a considerable and troubling issue, on a flat detector.
For this particular design, while the field curvature is a possible
concern, the field size are tiny.  Hence, as I noted in my previous post,
the amount of defocus is not a consideration.  In fact it is at least an
order of magnitude smaller than that needed to create a defocused spot as
large as the Airy disk.

I must disagree with your assertion that there is neither coma of stig in
the wavefront presented by the secondary to the tertiary.  Admitedly, given
the field size, these abberations are small, but they are not negligible as
they are amplified by the tertiary, which as you correctly point out also
adds its own coma.  Though there are a number of ways to convince yourself
that these abberations must be present in a confocal biparabolic reflective
afocal beam de-spander(1), it can be most easly done by simply replacing
the tertiary in your OSLO LT model with a perfect lens of identical focal
length.

A further note regarding field curvature.  First order theory is an
excellent starting point for optical analysis, but by definition it does
not consider the effect of third or higher order abberations.  Hence, the
field curvature derived from the paraxial model is not ideal.  Again, given
the field size, it makes no difference.

I should own up to a mistake I made, but only recognized today.  I created
several different Loveday models and the field curvature I presented came
from a model different from that previously described.  I can say as a
general observation, that for all the models I examined, the radius of my
perfered field curvature differed from that derived by OLSO by no more than
5%.  I would add, that the Loveday field is not actually spherical, but
rather a general asphere which is very closely modeled by a sphere.


Anthony



1)  ;-)