lev mar input constant

[altenbach]

---

NicolasMoua wrote:

I have the same probleme as you with my labview file and the Levmar function: When I run the vi it returns the exact parameters that I used as inputs, and the Fit Y values are all NaN.

---

This is not enough information to troubleshoot the problem.

There is no Levmar function, Are you talking about the nonlinear curve fit? Which one? (constrained, vi model, formula model, express, etc)  

What is the model? How many parameters? If you calculate the model with the guess parameters and graph it, does it look anything like your data?

What is the range of the Y values. Is it e.g. very small?

Do you get an error and If so, what is the error number?

Show us your code and some typical data and we will figure it out... 🙂

(notice that you added to an old thread that is already marked as solved. You should have started a new thread instead)

[NicolasMoua]

Thank you for the quick answer.

Are you talking about the nonlinear curve fit? Which one? (constrained, vi model, formula model, express, etc)  

Yes, I'm talking about the nonlinear curve fit: formula model.

What is the model? How many parameters? If you calculate the model with the guess parameters and graph it, does it look anything like your data?

The model is: sum of Ri * (exp (-t / Taui) with i  from0 to a number which I can fix.

So there is 2*i parameters: Ri and Taui.

Results are the same as datas.

What is the range of the Y values. Is it e.g. very small?

Yes, the range of Y is very small. 

Do you get an error and If so, what is the error number?

The error number is 23087

[altenbach]

---

@NicolasMoua wrote:

What is the range of the Y values. Is it e.g. very small?

Yes, the range of Y is very small. 

---

 "Very small" can mean different things to different people, for example 1m is very small for an astronomer, but very large for an atomic scientist 🐵

Why can't you be a bit more quantitative?

Does it work if you multiply the y data with e.g. 1e6 and adjust the parameters accordingly?

Alternatively, try to wire the weight input, e.g. as described here.