Re: [ATM] A Null Test

["Richard F.L.R. Snashall" <rflrs@rcn.com>]

Richard wrote:
>
> RFLRS> It looks like the major gain
> RFLRS> here is the second direction reversal in the longitudinal error.
>
> Not sure I understand that Richard.
A picture's probably easier for me at this point.  At:

    http://users.rcn.com/rflrs/offner01-01.jpg

I put a plot of the longitudinal error of the original two
lens solution.  At:

    http://users.rcn.com/rflrs/offner01-02.jpg

I put a plot of the longitudinal error of the three lens
solution.  It doubles back twice instead of once.  Not too
sure, though, given the performance levels we've seen so
far, that it really matters much in real life.
>
> RFLRS> Interesting test.  The lenses are small and the final cone angle is
> RFLRS> also small.  Makes me kind of wonder why someone didn't recommend
> RFLRS> that as an alternative (on the "other" list).  Is there any price
> RFLRS> to be paid?
>
> It would be interesting to do a thorough tolerance analysis of
> both methods. My gut tells me that three lens have more chance
> of contributing errors than two, BICBW
>
> I kept the lens parameters fixed and put a V solve on image to
> first lens distance, inter lens sparation and second lens to
> mirror focal point. I've started making a table, but the
> extremes of f/2.4 to f/3.5  and K=-0.8 to K=-1.3 optimised
> easily. Worst case null seen was 1/250th wave and typically
> better than 1/1000 wave.
>
I've been getting the same kind of numbers (f/2.5 .. f/5.0 and
K=-0.8 .. -1.2).  By switching to the Offner form, I've now been
able to get a single solution sheet for the entire range, even
though not all the solutions have two direction changes.
_______________________________________________
ATM mailing list http://www.atmlist.net/