[RASMB] interference optics

[David Hayes]

Hi Ute,

For clarification, it is tradition in AUC circles to call "jitter" the small variations in y axis offset which is in interference data.  The huge offsets you see are "jumps".  There is no valid interference data in the meniscus position, so the fourier transform that is following the position of the fringes can "jump" one or more whole fringes when the algorithm crosses the meniscus.  

As you note, SEDFIT can fit the data and fit the noise and then subtract the noise out of the data set.  The jitter and jump are just the RI, or radially independent noise.  The time independent noise is the possible slant and the rough baseline of the optical system and the cell windows which is most evident when you have  a spec of dust on your cells (this data set has observably very clean cells and a basically flat interference baseline) Of course, this works best when the data is able to be well fit by SEDFIT.  

The program SEDANAL by Walter Stafford uses two simple algorithms to subtract out jitter and jump before fitting the data.  Since it does this to the raw data in two steps, it is easy to distinguish the two causes of noise.  In SEDANAL the time independent noise is subtracted out by using the dcdt method where two scans are subtracted from each other and then the dcdt transformed data is fit.

As a scientist interpreting AUC data, you can not ever get rid of noise.  The jitter can be reduced by mechanical improvements to the optical system:  such as the planned improvements for next generation of AUC devices.  The fringe jumps are trivial because they really are exact integers and by fitting or just by an inspection algorithm they can be corrected.  You can't get rid of the meniscus with interferometry data.  Likewise, no lens or cell window is going to be perfectly flat, so the interferometer will measure that noise.  The time independent noise can be reduced mostly by better radial reproducibility in the raw data so that fitting or dcdt subtraction will work better.

There are more subtle sources of noise that affect boundary shape and are unavoidable, like the bending of the incident light because of the concentration gradient, buffer mismatch, and the precession of the rotor so even when the major sources of noise are minimized, then other sources of noise will start to be noticed.

My advise is that this data sets look good:  I see no obvious problems worthwhile to try fixing.  You may have a bit of a wobbly baseline due to buffer mismatch, which might be minimized by using meniscus matching cells and more extensive dialysis -- but even with these improvements, your curve fit probably wouldn't be noticeably better.

David Hayes


________________________________
 From: Ute Curth <curth.ute at mh-hannover.de>
To: rasmb at list.rasmb.org 
Sent: Tuesday, January 15, 2013 12:41 PM
Subject: [RASMB] interference optics
 

Dear all,

I hope you had a good start in 2013 and some of you can possibly give me advice on a problem with our Xl-I machine.The raw data of  our interference optics show (and always showed) such an enormous vertical jitter that it is hard to judge whether the sample sediments at all and where the meniscus is located (see attached file). Whereas, after subtraction of TI and RI noise the data look quite nice, I am wondering what the reason of this jitter is. Has anyone an idea about its possible source and how I can get rid of it? 

Thank you in advance

Ute
 
-- 
PD Dr. rer. nat. Ute Curth
> Medizinische Hochschule Hannover
> Strukturanalyse
> OE 8830
> Carl-Neuberg-Str. 1
> 30625 Hannover
> Germany
> 
> Tel.: +49-511-5329372
> Fax : +49-511-5325966
_______________________________________________
RASMB mailing list
RASMB at list.rasmb.org
http://list.rasmb.org/listinfo.cgi/rasmb-rasmb.org
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://list.rasmb.org/pipermail/rasmb-rasmb.org/attachments/20130115/0a2eda7e/attachment.htm>