[Author Prev][Author Next][Thread Prev][Thread Next][Author Index][Thread Index]

Re: Re[2]: [ATM] A two way null test for convex hyperboloids



Richard,

On Apr 6, 2005 12:22 AM, Richard <cnc@cncservo.co.uk> wrote:
> Hi Mauritz,
> 
> Yes, that's more or less what I'm ray tracing now. It seems that
> the null for those specific raidii and K is quite dependant on
> the glass type. The 10 wave abberation I saw is with BK7. If I
> change the glass to fused silica I can re-optimise focal
> distance to get down to about 1/2 wave, but with generic quartz
> I can't get any better than 15 waves. What exactly is the
> "QUARTZ C" that oslo is using?
> 

Yes, it does depend on the index of refraction.
The choice of a quartz subtrate was arbitrary, but perhaps a bit stupid.
Here is the prescription for a BK7 substrate instead,
*LENS DATA
Wavelength 632.8 nm
 SRF      RADIUS      THICKNESS   APERTURE RADIUS       GLASS  SPE  NOTE
 OBJ       --        674.432508 V    1.000000             AIR     
  1    1.4140e+03     20.000000     78.097034 S           BK7 C    cc  -6.35
  2  -1.3746e+03    -20.000000 P   79.230497 AS   REFL_HATCH   cc -5.907
  3    1.4140e+03 P -674.432508 P   78.096979 S           AIR   *      cc -6.35
(The other surface should now be 1414 mm RoC instead and converges to a
conic konstant of -6.35. Still the 1374 mm surface converges to the same conic
as before -5.907 as we want.)

The sensitivity to testing wavelength is not too big, this example would give an
ok null also for Na light 589 nm with a source/knifeddge distance of 673.4 mm.

> Personally, although quite an original idea, I don't see the
> benefit over using a plano back to the mirror and getting the
> null from a ross or offner null setup. At least that keeps the
> number of aspheric surfaces down to one, and you can be sure of
> surface that null does represent. With the bi-convex idea, I
> would be very worried by the interdependance of the two radii
> and conic constants - maybe if the refractive index of the glass
> is slightly different, the two surfaces may converge to
> something altogether different.

This interdependence is what I'm using :-) The point is that if you
can measure the RoC for the two surfaces with a spherometer, one can
calculate to which aspheric surfaces the testing and polishing
procedure converges to.
If it is simpler to polish and verify a flat and use an aux lens for
nulling instead, I'm sure you can jugde better than I. I'm just
tossing out an idea for discussion and feedback.

Thanks,
Mauritz
_______________________________________________
ATM mailing list http://www.atmlist.net/