On Wed, 22 Nov 2000, Shough, Dean wrote:
> > > What are DO-elements?
> >
>
> Diffractive Optical Elements.
>
> They may look like Fresnel lenses, but optically they are quite different.
No, and yes. True diffractive optical elements do not look anything like
Fresnel lenses!! That's the problem! The Canon DO elements (or the
schematic pictures of them) looked very much like Fresnel elements. Of
course it might have just been a marketing photo to explain something that
is far harder to explain... but at least it confuses me to doubd what they
are saying/doing. I've done some calculations around difractive optics a
few years back... while you can do many strange things with diffractive
optics, but one thing stuck to my mind - because I thought about making
camera lenses with them - was that they work for only one wavelength at a
time. So making a camera lens with pure diffractive elements seemed to be
out.
> A Fresnel lens looks like a very coarse version of a similar DOE. Both the
> Fresnel lens and the DOE look like the equivalent thick lens that has been
> compressed into a series o rings. In the case of the Fresnel lens, each
> individual ring is indeed a section of the original lens and completely
> independent of the neighboring rings. The width and thickness of each ring
> does not matter, only the cross section of each ring.
What you are talking about is true about fresnell length, but has not much
to do with pure diffractive optics... which brings me to my another
thing... maybe what Canon is doing is using very small scale Fresnel lens
with somekind of diffractive optics stuff (at least loosely) to correct
for the problems (that are caused by diffraction, but very much unwanted)
that arise in the borders of the "rings". This is purely guessing as this
technology was not really clear from the Canon web-site (which I can't
find anymore to check this stuff from :-( ).
Sorry, I wrote my answers before I read on:
> In the case of a DOE, this is reversed - the width of each ring is most
> important. The spacing between rings is such that it acts like a
> diffraction grating, directing the light in a particular direction. The
> required ring spacing can be computed from the diffraction grating. The
> profile of each ring affects is the efficiency of the grating, i.e., the
> profile acts the same as blazing a classical grating and determines how much
> of the light is diffracted into the various diffraction orders.
You have much better thoughts here... although I really can't put now my
finger on it... there is something that does not feel right... But you are
probably very close to what Canon has done. Still I don't feel it's really
purely diffractive element. I can't quite understand it - or which part of
it is classical optics and which is diffractive.
Anyways, the lens was painted white... but that doesn't yet mean it has as
good optical quality as L lenses. It's not that I would not trust the
quality of diffractive optics, quite contrary, I don't quite trust this
classical+diffractive element yet to handle the whole spectrum well...
also I could see this happening:
> I expect that the big problem with the 400 mm DOE lens will be scattering
> of light into the wrong diffraction orders. I will be very interested in
> seeing how well this lens performs when there is a bright light source in or
> nearly in the FOV. Canon's use of two DOE almost in contact with each other
> may be an attempt to reduce this problem.
Or maybe one of it's main reasons is to correct the handling of different
colors. (You know, the first one makes a rainbow out of the colors and the
next one tries to correct it... same as with classical elements.)
Actually, I shoule probably shut up and dig up some Canon patents about
this... I just don't have the connection (nor the time) just now... ;-)
Best regards,
Hugo.
************************************************************
** Hugo G�vert **
** [EMAIL PROTECTED] http://www.hut.fi/~hugo **
************************************************************
*
****
*******
***********************************************************
* For list instructions, including unsubscribe, see:
* http://www.a1.nl/phomepag/markerink/eos_list.htm
***********************************************************
