| Author |
 Topic  |
|
|
smburu
USA
12 Posts |
 Posted - 10/24/2015 : 02:03:43 AM
|
Hi,
I am trying to fit a curve to my data in order to calculate the amplitude and the widths (fwhm) of the peaks.
I hope the attached image is viewable. I am using the Peak Analyzer in Origin Pro to fit peaks.
The problem I am having is that my data has asymmetric shape, with peaks of different sizes and shapes. I use the local max criteria to find the peaks and then try to fit it with a gaussian function.
I got the error message: "Fit did not converge - mutual dependency exists between parameters.You may have overparameterized the fitting function. Fixing one of them may eliminate this problem." Therefore, I fixed the all the x-centers of the peak locations. The result was a converged cumulative fit, whose individual fit peaks had a reasonable amplitude to my data. However, all the individual fit peaks had the same fwhm.
Therefore, I unfixed the peak x-centers and fixed the widths which all had similar values. The result was that the peak x-centers shifted slightly. I then unfixed the widths and fixed the x-centers which were slightly off. The result is in the image below, which is by far the best result I have for now.
I have tried various other methods but I have never had results that were good enough to report. It seems that the short peaks are the ones causing a problem. How can I solve this issue?
I would also like a method that I can save to quickly apply to about 30 other data sets that I have with similar but distinct assymetric peaks. I was also having a problem getting similar results using a saved analysis with this particular dataset.
Please help me.
Thanks,
~Sarah
Origin Ver. and Service Release (Select Help-->About Origin):
Operating System: |
|
|
jasonzhao
China
262 Posts |
|
|
smburu
USA
12 Posts |
Posted - 10/26/2015 : 09:52:00 AM
|
Dear Jason,
I had followed the instructions in the link in order to learn how to use the Peak Analyzer. The problem I had were the error messages that the fit did not converge. Therefore, I had to fix some variables.
Is there any other way that you can suggest for finding heights and fwhm?
Thanks,
~Sarah |
 |
|
|
smburu
USA
12 Posts |
|
|
smburu
USA
12 Posts |
Posted - 10/26/2015 : 10:23:20 AM
|
 |
|
|
|
jasonzhao
China
262 Posts |
Posted - 10/27/2015 : 05:48:45 AM
|
Hello,
You can send the data to tech@originlab.com for a further check.
Best regards!
Jason
OriginLab Technical Service |
|
|
| |
Topic |
|
fitting multiple peaks to get fwhm and heights
