| Author |
 Topic  |
|
|
jateengandhi
USA
Posts |
 Posted - 06/10/2011 : 5:49:51 PM
|
Origin Ver. and Service Release (Select Help-->About Origin):
Operating System:
Hello,
I have tried Lorentzian and Gaussian Multi-peak fit and both the attempts were unable to ascertain the peak at @65.9. This was the case in spite of me specifically selecting that peak during peak selection step in both the attempts. Could this be solved without Peak Analyzer tool, since I am using 7.5 SR0 and it does not include it? Appreciate your help.
JG |
|
|
jateengandhi
USA
Posts |
Posted - 06/13/2011 : 5:27:23 PM
|
So I am guessing nobody has seen or knows what could be the reason for this? I am attaching Lorentzian fit as well. It misses the 3rd peak just like Gaussian (pic in original post). Entire dataset for the 3rd peak has negative values. I wonder how.  |
 |
|
|
larry_lan
China
Posts |
Posted - 06/14/2011 : 01:22:18 AM
|
Hi:
May be you can try the following:
First, Origin perform fitting on RAW DATA. Even you set Y scale as log scale, it will still fit on raw data. And if you turn your graph back to linear scale, you can see the curve is quite flat on X(65.9). So, I suggest you log your Y data first, and then fit.
Second, although you can explicit see there five peaks, some hidden peaks may also exist, and these hidden peaks may affect the shape of the cumulative curve. Actually, it's very very difficult to have a good fit by 5 peaks only. So, adding more peaks can make the fitted cumulative curve smoother.
For example, I fit 9 peaks at the following position:
And the final result looks better.
Besides fitting the peaks by "Fit Multi-Peaks", you can also try PFM if you are using OriginPro 7.5. Drag the file "PFM.OPK" from "OriginEXE\PFM\" folder into Origin, and then you will have PFM.
Thanks
Larry
OriginLab |
|
|
|
jateengandhi
USA
Posts |
Posted - 06/14/2011 : 10:23:17 AM
|
larry_lan,
Thanks a bunch. It worked much better than before. Although from attached images you can see that there is significant difference between Lorentzian and Gaussian 'Fit Multi-Pick' (I do not have Pro). First one is Lorentzian and the second one is Gaussian. Looking at your data it looks like you did Gaussian, did you? Also I am amazed by the fact that you were able to analyze using a 20kb jpg file. I guess I still have a lot to learn.
Thanks again,
JG
 |
|
|
| |
Topic |
|
|
|
Multi-Peak fit for 7.5