Using 'sddsemitproc'

[tennant]

Hi,

I am trying to use 'sddsemitproc' but running into various problems. All the examples that I have seen - in this forum or in the documentation - assume that elegant has generated the data. Suppose I have data in spreadsheet and want to analyze it. If I understand the documentation correctly, I generate an sdds file of the data (R11, R12, R33, R34, Sx, Sy) and I'm good to go.

When I try something like: sddsemitproc ScanData.sdds OutData.sdds

I get the following error: 'Error: number of error sets must be >2'. I thought specifying errors was optional. Any help would be appreciated!

--chris

P.S. I am attaching an input file, which is itself curious because according to 'sddscheck' the file is corrupted. However, I can still plot columns with 'sddsplot' (?!).

[michael_borland]

Chris,

sddsemitproc requires that you perform error analysis. It allows the logic of the program to be significantly simplified. If you aren't interested in errors, then use a command like
% sddsemitproc ScanData.sdds emit.sdds -nErrorSets=3 -errorLevel=1e-9,1e-9

Also, your file has two problems:
1. You don't have a row counts value before the table, nor do you have no_row_counts=1 in the data statement.
2. You need to add a column called Step to the file. sddsemitproc requires this as an independent variable. You can add this as follows:
% sddsprocess ScanData.sdds -define=col,Step,i_row,type=double

--Michael

[tennant]

Hi Michael,

I was wondering if you could briefly explain the method used to extract emittances in sddsemitproc. I have been analyzing some data (simple, single quad scan) and comparing with the results of a multiple regression fit of the data. I get good agreement. But when I remove some data points, it seems that sddsemitproc is more sensitive to those changes than a simple regression fit. I was just curious about the method you use.

Thanks,
chris

[michael_borland]

Chris,

The method is described in my Ph. D. Thesis (SLAC R-402). The relevant pages are attached.

--Michael