Problems loading modified sch files

[pedrochemgla]

Hi everyone

I've just started to take impedance spectra and to use MEISP. I wanted to modifiy the Randles circuit from the library to get a better fit for my data. However, after modifying and saving the circuit under a new name in the same folder as data and project I get this error message (see picture attached). What can I do use the modified file?

Cheers

Pedro

[Yevgen Barsukov]

Could you send your shematic file to understand what

it might need? It might be using a dll or something else.

Regards,
Yevgen

[pedrochemgla]

Yevgen, thanks a lot for your prompt reply. Please see the attached file.

Kind regards

Pedro

[Yevgen Barsukov]

Pedro,

this schematic is using a user defined function "nud" which

is located in /cirlib/functions and is refered to as /functions/nud.fun

To make it work, place your *.sch file under /cirlib folder

(because that is its original file before modification was probably located), then the reference to function file will

work.

Regards,
Yevgen

[pedrochemgla]

Yevgen, many thanks for your reply. The file worked fine.

However, I cannot fit my data. It's early days for me in the EIS world! My experimental conditions should fit a model based on a modified Randles circuit in which the capacitor has been changed for a CPE. I attach the data file and the circuit I'm using. Any hints in the right direction will be greatly appreciated.

Kind regards

Pedro 

[Yevgen Barsukov]

Pedro,

did you make assignment to time-constants order before using prefit to get

intial guess values? For example Rser would be 0, Rct and Cdl 1 (e.g lowest time

constant order), then diffusion would be 2 etc.

Please refer to MEISP user manual on details of using pre-fit. In most cases

fit fails due to bad initial guesses, although it can also be that you have too many parameters.

Btw what particular problems did you observe? Maybe you can send a screenshot.

Regards,
Yevgen

[kashif farooq]

As a new user for curve fitting i m facing also same problems its very hard to get the fitting and some tie 1 model give the fitting and for other measurements of same case need other model its really need too much hard work and i m using zview or fitting is there any other good software that can be used for curve fitting.

--

Muhammad Kashif

010-2430227

PhD Student (Nano Electronic Engineering),

Institute Nano Electronic Engineering (INEE),

Universiti Malaysia Perlis (UniMAP),

Lot 106, 108 &110, tingkat 1,

Blok A, Taman Pertiwi Indah,

Jalan Kangar-Alor Star,

Seriab, 01000, Kangar,

Perlis, Malaysia.

[Kur@]

The basis of interpretation of research results is the the identification structure model that determines the presence and connection of elements in the the sequence the scheme. Structural identification of the model is the most important stage in its construction. Since the parametric identification (determining face values of circuit elements) is not always allows to determine correctness of the model. Sometimes a computer simulation of the model gives a precise match with experiment, but the processes described by the model can not take place in this system. Thus results obtained do not correspond to the actual process.

Therefore there is no difference what program to use, sometimes for not correct model all programs show better matches than for the correct model.

Субота, 21 липня 2012 р. 05:45:51 UTC+3 користувач kashif farooq написав:

[pedrochemgla]

Yevgen, I have a much better fit now. I think I was overcomplicating my circuit. According to the fit, the capacitor C1 does not seem to be doing much so maybe I should replace/remove it (attached screenshot Fitting_no_CPE). According to the literature, my data can be fitted by replacing the capacitor with a CPE. However, if I try that I get a rubbish fit (see screenshot Fitting_CPE). Any suggestions for this?

Thanks again for your time and comments. I'm an absolute beginner in the EIS field and finding it quite challeging and very interesting.

Regards   

[Yevgen Barsukov]

Pedro,

the reason why the second fit succeeded is that you eliminated multiple elements that have similar effect on the spectrum. For example CPE has similar effect as diffusion element placed in parallel to a capacitor at least in some frequency ranges. It is the key to use absolute minimal model that describes the spectrum, otherwise parameter values will have too high standard deviations to be meaningful.

They way your spectrum looks like, it can be represented by one limited-length diffusion element (it is 45o line that curls up into vertical at low frequencies). It could be used without anything in parallel, and only serial resistance.

You have to find what physically relevant to your system model that would correspond to such functional relation. It could be a finite length porous layer with some surface capacitance, without any electrochemical reaction (at least in this voltage range). 

Regards,
Yevgen

[pedrochemgla]

Yevgen,

Thanks again for your comments.

I've tried to fit my data to a circuit with the function "double layer capacitance at a porous electrode" (Image 6) and the fit is slightly better to the ones tried with all the other "planar finite (or infinite) length diffusion" functions. In any case, it's not better than the Randles circuit (Image 8) which is the best one so far. Maybe this is one of those cases where there is a good fit to a meaningless system. Back to the fit to the  porous electrode, I changed the weighing method from modulus to unit and I got better SD for the Warburg elements but worse SD for the series resistance (Image 7). Is that a correct thing to do? I have followed all the advice given in section 5.1.1 of the manual but still I can't improve my fit.

Is there anything else I could try? I'm still confused about the fact that I got a much better fit just by using the Randles circuit (Image 8)

even though the values for the capacitor were too small.

Regards

Pedro

[kashif farooq]

Hi,

   How are you all the group members. I did the simulation of my data using the ziew. Please can you guide me how to compare them b/c its much confusing for me. Especially the values of R2 and CPE. I will wait for positive suggestions.

Thanks