Re: [ATM] Luri-Houghton prescriptions.

["Bob May" <bobmay@nethere.com>]

----- Original Message -----
From: Richard <cnc@cncservo.co.uk>
To: Bob May <bobmay@nethere.com>
Cc: <atm@atmlist.net>
Sent: Thursday, October 13, 2005 4:42 AM
Subject: Re: [ATM] Luri-Houghton prescriptions.


> Hi Bob,
>
> Thursday, October 13, 2005, 1:52:15 AM, you wrote:
>
> BM> I've been playing with the Lurie-Houghton prescription on
the two
> BM> lenses and seem to find that the equal radius pairs of
curves
> BM> doesn't seem to do the best job.
>
> No, obviously not, but the point of the the LH design is to
> minimise the amount of work and complexity to achieve a
> reasonable result, but with only 2 degrees of freedom for
> optimisation. If you allow 4 radii rather than 2, an air gap,
> and different glasses for the two parts, then an optimiser now
> has 7 degrees of freedom to get its teeth into. Once you allow
> those DOF's, the thickness of the elements come into play and
> you have 9 DOF's.
>
I'm more interested in finding the best set of curves for the
lenses of the same glass.  I might get a little bit better color
correction with different glasses but that I don't feel the need
for as the spot size hardly varies due to the color traced.
Thickness and air gap are variables tho.  I wouldn't consider
thicknesses to be something that needs to be equal as they are
basically going to be close anyway and the grinding process can
easily produce the differences needed while in the coarse
grinding as needed.
I'll also note that the thickness is relatively unimportant to
the prescription as slight differences in the ROC of the lens
will correct for the thickness.
> Best regards,
>   Richard in the UK
Bob May
bobmay at nethere.com
http: slash /nav.to slash bobmay
http: slash /bobmay dot astronomy.net
Addresses have been munged up to keep spammers from bothering me.
replace the obvious words with the proper character.
Thank you for your understanding!


_______________________________________________
ATM mailing list http://www.atmlist.net/