| Author |
 Topic  |
|
|
Ecki
Germany
Posts |
 Posted - 11/09/2004 : 06:33:50 AM
|
Hello...
...i have one question...my data are coming from an hplc measurement...i get a peak which actually consists of two peaks overlayed...and another single peak...if i do the following: analysis --- fit overlayed pulses ( in german: fit überlagerte impulse) everything is ok and i can calculate the percentages of both overlayed peaks, but the problem is, that absolutely symmetrical peaks are required for this analysis...so my problem is: i have peaks with a little tailing! they look like the example curve which you can find under the menu analysis --- not linear kurve fitting ( nicht lineare kurvenanpassung) --- category --- peak functions --- extreme! so the peak is not 100% symmetrical! the problem is that if i do the analysis i mentioned before just for the peak which is really just one single, but tailed peak (all peaks tail in the graph, due to hplc) i find out that 25% is due to tailing...but the final question: CAN I DO THIS ANALYSIS WHITH RESPECT TO THAT THE TWO OVERLAYED PEAKS HAVE THIS TAILING! the background for that is that the ratios of the two overlayed peaks change by time and when there is just one left, the analysis i mentioned is wrong!
Thanks for help and excuse my english  |
|
|
easwar
USA
1965 Posts |
Posted - 11/09/2004 : 09:49:01 AM
|
Hello,
If I understand you correctly, you wish to fit your data with multiple instances (more than one) of asymmetric peaks such as the Extreme function.
If the above is what you want, you can do this with the NLSF Advanced Fitting Tool. You can open the tool from the Analysis->Nonlinear Curve Fit->Advanced Fitting Tool... menu item. Once the NLSF tool is open, select your data, then select the Extreme function from the Peak category. Then go to the Options->Replicas menu item in the NLSF tool. You can then specify the number of replicas. For example, if you want to fit with two Extreme peaks, set the num replicas to 1.
Then when you go to the Fit page in the NLSF tool, you will get an extra set of parameters for the 2nd peak. Note that in the Replicas page you can also share parameters between your peaks - in which case some parameters will be common between the peaks. You can then assign appropriate initial/guess values to the parameters and fit your data.
Easwar
OriginLab
|
 |
|
|
Ecki
Germany
Posts |
Posted - 11/09/2004 : 1:12:28 PM
|
Hello..
...thanks for your fast reply... i figured out how to do this...unfortunately it seems that the decay of the function is a little to slow!...so my question is: How can i modify the equation of the EXTREME according to my problem!...or do you have any other suggestions?
Thank you very much...
Markus Löweneck |
|
|
|
easwar
USA
1965 Posts |
Posted - 11/09/2004 : 1:36:02 PM
|
Hi Markus,
So it sounds like you would like to fit with your own user-defined peak function, and would further like to use replicate feature...
Currently replicates (multiple instances of the function to fit one dataset) is supported only with built-in functions, and not with user-defined functions.
However, there is a workaround. Say you want to only fit two peaks in your data with your own function. You could then define a user-defined fitting function using either Origin C (supported in 7.0 and later) or Y-script (supported in all versions). For example, let us assume you want to fit with gaussian function. You can define your function as below:
Parameters: y0, xc1, w1, A1, xc2, w2, A2
// for Y-Script
peak1 = A1 * exp( -(x-xc1)^2/ (2*w1^2) );
peak2 = A2 * exp( -(x-xc2)^2/ (2*w2^2) );
y = y0 + peak1 + peak2;
or
// for Origin C
double peak1 = A1 * exp( -(x-xc1)^2/ (2*w1^2) );
double peak2 = A2 * exp( -(x-xc2)^2/ (2*w2^2) );
y = y0 + peak1 + peak2;
Replace the gaussian expression above with your own function (and your own set of parameters). Then assign initial values, and if necessary/desired, set bounds on parameters so that each peak stays separate etc, and fit your data.
The limitation is that if you then want three peaks, you need to add another
peak3=
statement in your function definition, and so on for more peaks.
Easwar
OriginLab
Edited by - easwar on 11/09/2004 1:45:02 PM |
|
|
| |
Topic |
|
|
|
Integration of a doublepeak with tailing????