From: Walter Stafford <STAFFORD@bbri.harvard.edu>
  To  : PETER.ROSSMANITH@MSM.BASF-AG.DE
  Date: Mon, 1 Jun 1998 7:33:38 -0400 (EDT)

RE: wierd fringe displacement

Dear Peter and RASMBers,

	This "problem" arises from the properties of the optics (less than 
one fringe vertical jitter) and the fourier analysis itself (integral 
fringe shifts).

Both of these effects are taken into account in all three of the available 
software packages that I know about. The fix for this has been well known 
and was always in the original dcdt software. I started this analyis on the 
Model-E Rayleigh system where both effects had to be taken into account.
It was not necessary for absorbance data.


The MS-DOS versions and the MacIntosh versions both allow you to align the 
fringes above the meniscus. This removes the vertical optical/mechanical
jitter that is less than 1 fringe. Then it lets you select a region in the
solution column in which to remove the remaining intergral fringe shift
that may occur across the meniscus. 

The very latest Beckman version allows both of these corrections.

It is important for my dcdt version that the entire array of all 2048
pixels be taken each time. The software, by the two operations just
mentioned,  takes care of the problems described by Peter. 

I urge you *NOT* to take the data the way Beckman originally described; you
will get unuseable results with that method. The data must include the
meniscus region at the very least. Remember, you have to able to select the 
meniscus from the screen.

Another *VERY* important point: the Interference Optics, in its current 
incarnation, cannot be used reliably above 50,000 rpm. There are timing 
problems above that speed that result in random spurious deviations in the 
fringes. Beckman is close to having a solution to this problem as well.


Hope that helps. Download the latest version of the software and you should 
be OK.

Walter Stafford