Re: [Gnuplot-info] fitting a implicit function with gnuplot

[Hans-Bernhard B. <HBB...@t-...>]

On 05.09.2013 12:52, U Maennl wrote:
> I have the following problem for a school project.
> We collected data from a diode with an resistor in series.
> Now I am trying to fit this data in order to get the value of the series
> resistance.
> The following formula I used:
> I = I0*exp((x-I*Rs)/nkT-1)  (x=V) via I0, Rs, n, T=297
> I know the problem could be solved by change to V = ... but I like to keep
> V on the x-axies.

You'll want to do a "3D fit" of a function that describes your equation, 
i.e., something like this:

	implicit(V,I) = I0*exp((U-I*Rs)/n/k/T - 1.0) - I
	T=297.0
	fit implicit(x,y) 'data' using 1:2:(0.0):(1.0) via I0,Rs,n,T

The drawback is that the z errors are totally bogus, and have no 
relation at all to your data's precision.  So don't put too much trust 
into the result, and none whatsoever into the parameter errors you get.