My data cannot be fitted by exponential fit vi

[altenbach]

Well, a polynomial fit is meaningless if you have an exponential curve and are interested in the damping paramter. Right? 🙂

And, as we found out, you cannot use the old style exponential fit, because it does not include an offset (C).

Here's a quick rewrite using the LabVIEW 2009 exponential fit. As you can see the best coefficients are:

A=10.888

B=-0.000659

C= 16.83

(... and C!=0 as I already predicted!)

 

Have you tried the curve fit express VI? It should work just fine, giving reasonable initial estimates.

Message Edited by altenbach on 11-24-2009 12:27 AM

[J_Sham]

Wow~ fitted perfectly! So.....does it mean that i have to upgrade to LabView 2009 in order to do it? or can u send me the block file if it can be solved by Fitting Express? Thanks!

[altenbach]

---

J Sham wrote:
So.....does it mean that i have to upgrade to LabView 2009 in order to do it?

---

No, you can do everything in 7.1 and you have several possibilities. (Of course there are many other reasons to upgrade ;))

I no longer have access to LabVIEW 7.1, so I cannot give you any code. All fitting VIs have received a significant overhaul in LabVIEW 8.0.

Fortunatly, the default function for the old version of the "Nonlinear Lev-Mar Fit" (search the palette!) is y= a0*exp(a1*x)+a2 if I remember right, meaning you should be able to use it right out of the box. Simply give it the x, and y data and some reasonable estimates for the three parameters. Attach your code here if you get stuck.

Have you even tried the Curve Fitting express VI? Should be trivial to use.