levenberg marquardt

[altenbach]

Whenever you get an error code, you can go to "menu...help...explain error..." and paste the number. (see image)

I don't really know what "wrong letter" means, but I would guess you have a syntax error somewhere.

It is impossible to guess without seeing the code. What is your model formula?

Message Edited by altenbach on 01-27-2009 01:11 AM

[jmhf]

Hi,

error solved. As you said, it was a syntax error in the model.

I think it is now working well.

Thanks a lo lot

jm

[jmhf]

And the fit is now a little bit strange. Instead of pain a continuous cuadratic line (like the model)

it paints some traces, creating a discontinuous curve.

Do you know why?

Thx

[altenbach]

---

jmhf wrote:

Do you know why?

---

No, not without seeing the VI! 😉

There could be many reasons. For example what you currently show in the picture should not even work, because you only give starting values for a, but not for b and c.

In any case, if you want to fit to e.g. variable order polynominals, you definitely don't need levenberg-marquardt, there are much better ways.

[jmhf]

That´s right!

The problem was solved. It came from a bad function definition.

Thanks for everything

jm