Reply-To: "Kevin Aylward"
From: "Kevin Aylward"
References: <3E0C8548.A5FBAFBC@n.zgs.de> <3E0C8EE4.F5C1362E@ieee.org>
Subject: Re: pspice problem
X-Newsreader: Microsoft Outlook Express 6.00.2800.1106
X-Inktomi-Trace: pc2-rdng3-5-cust66.winn.cable.ntl.com 1041019337 6906 18.104.22.168 (27 Dec 2002 20:02:17 GMT)
Date: Fri, 27 Dec 2002 20:02:13 -0000
NNTP-Posting-Date: Fri, 27 Dec 2002 20:02:18 GMT
> Reiner Goetz wrote in part:
>> I want to simulate a resistance depending of frequencies Rx=Rx(f).
>> example: Rx=R/f
>> I can't find any way to do this with spice.
> Anything that is realizable is generally possible to model in spice.
However, this particular function is not available in basic
Spice3/XSpice. I think this is doable in PSpice which has some
> Troubles building a model can sometimes be a red flag that the concept
> for the model is too unrealistic. And if you are not at least
> attempting to model reality then what's the point?
> Specifically, what real world device has the characteristic you are
> trying to model?
>> I tried it with the current behaviour sources, but "f" as an Input
>> (variable) is not possible.
> To construct a non-linear resistor, one can set up a voltage dependent
> current source as a non-linear function of its own terminal voltage.
> I've never tried it, but believe a behavioral source's describing
> function can include frequency. But even if you could, why would you
> wish to model a non realizable component?
Its not a non-linear resistor. This is a resistor which varies with
frequency, and as you note, is not really physically realisable , except
as an approximation.
SuperSpice, a very affordable Mixed-Mode
Windows Simulator with Schematic Capture,
Waveform Display, FFT's and Filter Design.