Re: Kohler illumination question...



justbeats wrote:
Given that the final step is to hide it behind frosted glass, why is it
necessary to get a focussed image of the lamp filament when setting up
for Kohler illumination?

I tried to answer this myself by taking images at different powers with
the filament defocussed to varying degrees. I couldn't detect any
difference in final image quality for any of the settings (as long as
the filament remains nicely centered of course).

After unsuccessfully Googling for the answer, I've another (maybe
related) question. Is critical illumination considered an old and
inferior technique, or a modern and superior one?

    Thanks
    Beats

There is a very practical consideration when you design a microscope illumination and that is to weigh performance vs cost. Usually the level of correction for the illumination is just as good as it is absolutely required - a frosted glass helps a lot in hiding imperfections. But this frosted glass should be placed as close as possible to an aperture plane and be optically as "thin" as possible to diffuse the light source but not to cut out too much.
A second reason, as already mentioned by another contribution is the shape and radiation uneveness of the light source itself. If you take a filament, then the source is pretty uneven, that's why in the past you got special Halogen bulbs from some high-end manufacturers which had a denser filament structure. If you use a small light source like an arc or a diode, then the optical aberrations of the lamp condenser and the subsequent optical elements are usually far too much to produce a proper image of that light source.
Yes, a perfect microscope would also have a perfectly designed illumination system, but this is much too expensive and people would not pay for it. And even if they would pay for it, the marketing aspect of it - selling it as a competitive advantage - would not work, because customers would say - why would I want to pay $500-1000 more for that illumination system if I can have a similar performance at much lower price by just placing a frosted glass into the lightpath.
A last but not least reason is that the frosted glass increases the nA of the illumination which helps "improving" the imaging performance of the microscope.
Regards George
.




Relevant Pages

  • Re: Kohler illumination question...
    ... In one of the prior posts you said that "Kohler illumination ... Since the filament is not a flat surface some portions of the filament and its reflection will not be at the same plane and therefore will not be in the same focus further on. ... The easiest way to see if the illumination optics are well-designed is to set the optics for critical illumination and to examine the image of the filament with the objective lens to determine how completely focused the filament is. ...
    (sci.techniques.microscopy)
  • Re: =?windows-1252?Q?K=F6hler_illumination_question=2E=2E=2E?=
    ... If you look at the illumination design of larger research microscopes, especially those which have a relay for illumination then you find these devices implanted there. ... Rene wrote: ... also not in theory (when applied with field iris). ... iris, instead of the filament. ...
    (sci.techniques.microscopy)
  • Re: Kohler illumination question...
    ... necessary to get a focussed image of the lamp filament when setting up for Kohler illumination? ... the filament defocussed to varying degrees. ... Critical illumination, for all practical purposes, is functionally ...
    (sci.techniques.microscopy)
  • Re: Kohler illumination question...
    ... If the geometry of the coiled filament were the problem, ... me think that in order to create a truly Koehler illumination, ... Why the situation might improve with a frosted glass, ... >I believe the problem originates with the design of the source lamp ...
    (sci.techniques.microscopy)
  • Re: Low cost vision system, thoughts
    ... i.e. illumination with an incandescent ... and color response of the imaging device. ... or another type of light source without a different ... ratios, if you have more yellow light, your objects look yellow. ...
    (comp.robotics.misc)