non linear curve fitting

[Maryam hosseini]

Hi

I am using Labview 8.5 to capture data form VNA. I need to do the Curve Fitting On the obtained data.

I am using the nonlinear Curve fitting function at the moment but when I run the  VI it gives me an error.

I was wondering to see if someone can help me on that.

Thanks

[altenbach]

We cannot help without more information.

1. What kind of error? (error popup, incorrect result, labview crash, cannot find a solution, ...?)
2. What is the model you are trying to fit?
3. How many parameters?
4. Do you provide the correct number of initial parameter estimates?
5. Are you using the fitting express VI or the plain nonlinear fit?
6. VI model or formula model?
7. Show us some code!

😄

... and what is VNA???

Message Edited by altenbach on 02-28-2008 10:58 AM

[Maryam hosseini]

Thanks for your quick response.

1. The Error is: 23026 occured at maximum iterations exceeded

                     possible reason(s):

                     Analysis: No optimum Found

2. the model is      a1/(1+4*((x-a2)/a3)^2) which is the lorentzian curve.

3.   3 parameters

4. yes, I do.

5. I am using the non linear curve fitting function which is availbe in the Mathematics Library.

VNA is vector network analyser the measurement eqipment that I am using for my project. it gives me the waveform of S21 against frequency.

and the 3 parameters are the resonator parameters which are Quality factor, resonant frequency and insertion loss.

The code That I have written is attached.

I would be really gratefull if you could help me on that.

Thanks

[smercurio_fc]

Debugging this will be easier if you provide some data, since we don't have the instrument. Your VI writes the X and Y data to file. How about uploading that data?

[Maryam hosseini]

Hi

Here are the data files, one for x and the other for y

thanks

[Maryam hosseini]

Because the first bit of the program is trying to capture data from VNA and then it does the fitting, I think it is better to make those parts disabled and try to do the fitting on the data Files that I have already attached.

If you do not have time to do that I will do it post it in 1 or 2 hours, because I do not have access to labview now.

Thanks

[altenbach]

OK, with a few minor tweaks if fits just fine. 🙂

1. Make sure that the estimates are reasonable, e.g. a1=small negative, a2 near  xmin,  a3 similar to the xwidth.
2. Your data is not very well balanced, so you need to increase the tolerance to e.g. 1e-16.

Message Edited by altenbach on 02-29-2008 09:37 AM

[altenbach]

Also make sure you save with enough significant digits. Your Y data is highly truncated.

Save it with at least 6 digits of precision.

[Maryam hosseini]

Hi,

Sorry, I attached a wrong file.  I have two sets of data for y one of them is used for linear fitting the other for non linear.

This time I have attached the correct one.

The fitting still does not work.

Could you have a look at the code again?

Thanks

 

[Maryam hosseini]

Hi

Have a look at the imgae of fitting that I captured.