| Author |
 Topic  |
|
|
Rollo78
Germany
6 Posts |
 Posted - 07/25/2011 : 7:22:34 PM
|
Origin Ver. 8.5.0G SR0
Operating System: Windows 7
Hi there,
i'am trying to do fits on beta spectra. (with the impuls analyzer) I want to use instrumental weighting and i have the square root of the events per channel in an extra column.
When i do the fits with statistical weighting i get reasonable values for reduced chi-squared (1-1,6). When i switch to instrumental weighting with the same fit parameters (dataset for weighting is [Book1]Sheet1!C which is calculated by sqrt(Col(B))) i get numbers for reduced chi-squared in the range of thousand, exactly the same number which do i get when i switch to no weighting (again same fit parameters).
Shouldn't the instrumental and statistical weighting give the same number for reduced chi-squared in my case?
w_ij=sigma_ij^2=yij
Is the instrumental weighting working if origin doesn't complain about the dataset?
There is no hint in the results report that Origin actually have used instrumental weighting.
What's wrong here?
Any help is greatly appreciated!
Best regards
Marco |
|
|
easwar
USA
1964 Posts |
Posted - 07/26/2011 : 10:20:54 AM
|
Hi Marco,
Yes, if the error dataset is sqrt of the Y values, then instrumental and statistical should give same results.
I tried in 8.5 and 8.5.1, with sample dataset \samples\curve fitting\gaussian.dat where the third column is sqrt of 2nd column, and I get same values for parameters and chi-sqr when using instrumental or statistical (and the values from these two methods are different from the case when weighting is set to None).
Please check if the instrumental dataset is correctly assigned in the control on the fit page of the analyzer. If the problem continues, please send your OPJ file to tech support. You can click on the "Send file to Tech Support" link on top right of the forum page, and then mention this post.
Easwar
OriginLab |
 |
|
|
Rollo78
Germany
6 Posts |
Posted - 08/02/2011 : 3:33:15 PM
|
Hi,
i don't know why but it worked with a new opj file.
Best Regards
Marco |
|
|
| |
Topic |
|
|
|
Wrong reduced chi-squared?