Re: Null corrector for a fast mirror
- It looks like Ed's Dall conjugate method can beat other simple single lens nulls easily and subject to the mirror the source doesn't need to be at that great a distance. In this case 4m. The orientation of the null lens gives a better null at 4m than the other way round does at 5m. Tolerances other than the usual mirror rad problem are reasonable probably better than Ross. My thoughts on that is guaranteeing a 1/10 wave P/V null or better. Increase the distance to 24m - I could do that in the garden probably longer and the null is to 1/200 p/v. A sort of final figuring session removing ato meters of glass with rouge on a feather duster.
One interesting aspect is that an F3 mirror works at F2.8 in the test so the mtf is better than any other method I've looked at. Ross comes out at over 3 and normal Dall at over 5
This is the source at 4m set up. Best viewed at original size.
One major irritation with Oslo - does anyone know how to fix the aperture of the mirror when working with finite sources - eg set an object na that more than fills the mirror but with the actual mirror diameter forming a stop? In real life it would but not it seems in oslo physics.
I guessed that there would be problems using Bath with a null set up like any of those mentioned.
- For a complete solution source anywhere I think the sag at h is needed which can probably be derived from that link. :-) Not sketched anything out so a pure guess. From memory Sixtests allows the mirror to source distance to be specified but that may be just to reduce errors from normal Foucault readings. I don't think it allows offsets to be specified which are another source of error apart from any aberration effects.
Where some views of aberrations go astray is that when a mask is used the aberrations of the whole mirror may not be important. Instead it's the size of the circles of confusion produced by a series of small off axis mirrors in relationship to their diffraction disc size.
I also suspect that there is another test that could be used these days but no sums done to check how large a sensor is needed for good say 1/20 wave results. Hartman, if that is how it's spelled. That one would just involve taking an exposure with the lens off a dslr, moving it a specific distance along the axis of the mirror and taking another.
--- In email@example.com, "Richard F.L.R. Snashall" <rflrs@...> wrote:
> On 5/12/2013 6:06 PM, John wrote:
> > So what is the equation for the exact slope of a conic versus height?