Chronology Current Month Current Thread Current Date
[Year List] [Month List (current year)] [Date Index] [Thread Index] [Thread Prev] [Thread Next] [Date Prev] [Date Next]

Re: optics

On Mon, 03 Nov 1997 13:27:59 Gary Karshner <karshner@STMARYTX.EDU> wrote:

I hate to disagree with you, but the advantage of the matrix method in
optical design is not only that you can "easily" computerize it, but that
you can abandon the paraxial assumption, and follow "individual" rays
through the lens system. This has made possible the rapid advances that have
occurred in optical systems in this half of the century. An optical designer
used to measure the number of rays he could trace through a system per day,
now it is measured in microseconds.

It is true that programs for exact ray tracing, even for non-spherical lenses,
are used extensively by practioners. These programs can be highly realistic
in showing all aberations, etc. But matrix methods in common undergraduate
texts, such as Meyer_Arendt or Pederotti, are based on paraxial approximation.
This is explicitly stated by each author; I just checked it. The approximation
is made when Snell's law is written as n1*alpha1=n2*alpha2.

And rays which are not confined to meridional planes (planes containing the
optical axis) are ignored. That is why I was saying that the results are not
more realistic than what would be obtained with old methods. High computing
speed is due to computers. Same assumptions, same results.
Ludwik Kowalski