Answer to my question by skogsgurra

[CoCra]

Hi Skogsgurra

your answer does not really help. Of course I do know that impedances are normally complex.

But assume I am calculating the complex impedance of an RC-circuit. The parameters of R and C will be real. If I then use the complex impedance of the RC-circuit thus obtained as "data" to be fitted, the mcd fitting procedure should find the parameters I originally used and they should be REAL. However, this is only the case if the guess values are already very close to the original values.

This causes problems if I want to apply the fitting procedure to actually measured data.

 

[Skogsgurra]

Sorry. I couldnt resist the "complex real world" paradox.

I still do not see your problem. You can make up any complex impedance from real components - even negative impedances can be built.

I think that it is a bad idea to start a new thread. We should keep this discussion in the original thread.

 

[CoCra]

Sorry, for opening a new thread. I am using this forum for the first time. Just after having submitted the new thread I realized that I did wrong.



I surely know that one can make complex impedances from real components. My only problem was to tell mathcad not to use complex fitting values instead of real ones for I always obtained complex solutions which might be correct from a mathematical point of view, but not from a physical.

I have solved this problem now.

 

[IRstuff]

This should be posted either in the Mathcad forum here or the Mathcad collaboratory website:

http://collab.mathsoft.com/~Mathcad2000

Numerical fitting, by its very nature, is imprecise. However, if you apply the constraints that R and C >0, it should force the solver to dump any complex solutions.

TTFN