To help the potential helpers, you should specify, in how far you did not succeed. Did the fit not match your data or didn't it work at all?

Dear Dr. Juttu:

Since in the list of topics-words appears the word Impedance-spectroscopy, perhaps would be possible stating from classical relation:

x^2 + y^2 =1

since a non-decentralized semicircle or a semicircle centralized on the x-axis (real parte of complex impedance) obey the function above.

Kind regards

Marcos Nobre

 Dr. Juttu

Your circle is not usually perfect . The center is not  in the real axis . it is  above  it . you can use Zview software. it is more appropriate for a such refinement .

Dr. Madani

Zview software is very good for fitting Impedance data. But we can fit only one semi-circle in a single plot. If my plot has got two semi-circles, I can not represent the data with two semi-circle fits.

I need any function for fitting so that I can fit the impedance data in Origin software.

I once used a fitting software called THALES, which provided the possibility to simulate virtually every thinkable equivalent circuit diagram. However, the software was not freeware and tricky to handle.

Maybe you have to go to the sketchboard yourself. Two semi-circles can be represented by two randles cells in series; with basic circuit analysis you should be able to obtain the Z(w) formula. Maybe, you can do the fitting more easily using the Bode plot instead of Nyquist? 

Hi, to compare your plotted Zs" vs. Zs' data with a perfect semi-circle you can use the following equation (works in qtiPlot): Y= (y0+(((r^2)-((x-x0)^2))^(1/2)))

You can modify the following variables:

y0: Y-axis offset (Zs", useful for depressed semi-circles)

r: radius of the semi-circle

x0: X-axis offset (Zs', useful for multiple or depressed semi-circles)

It works well if you work with small scale. If the plotted data is somwehere in the MOhm range, I usually divide the data by a factor of 10^n for easier fitting. In short: Its easier to do the fit if the axis goes from 0 to 10 instead from 0 to 10^9.

I found the equation in some forum. Unfortunately I forgot the original source. (Shame on me). You would do me a big favor, if anyone would tell me some kind of source for this equation.

Cheers

Edit: Also here is the equation if the data on the Y-axis is negative:

Y= -(y0+(((r^2)-((x-x0)^2))^(1/2)))

hi David,

I think you got that equation from Origin forum. http://www.originlab.com/forum/topic.asp?TOPIC_ID=7914

I tried doing that but I failed. He says that we need to put reasonable initial values for the parameters. But what is the reasonable values. I havent tried shrinking the data by dividing it. 

try to refer this youtube video:

https://youtu.be/LJdrmKi8nOc