[Help-gnucap] non-linear dependent sources

[John Vodden]

Hi

I've recently discovered gnucap and have very little experience of
similar simulators. I like the look of gnucap because of the open source
licence and because of its inherent interactivity. I am using gnucap
0.35 from the debian package.

As a learning exercise (and partly because I am interested in the
results) I am trying to analyse a simple circuit. The circuit derives from the 
average model of a power electronic converter and is very much an equivalent
circuit, containing passive components and dependent sources. I wish to
use gnucap to perform dc analysis to determine a suitable bias point and then
obtain a linearized small signal frequency response at that bias point with ac
analysis.

The problems I am having arise from my need to have a non-linear voltage
dependent voltage source. This is non-linear in the sense that the gain
is set by a separate input signal not just from a non-linear function of
the control input. This is required because the control input to the
source is a capacitor voltage and the separate input signal represents
the average switching function.

For the d.c. analysis I modelled this device as follows so that the
generator signal controlled the gain of the controlled source. There is
a capacitor between nodes 2 and 0.
        Ed (8 0 2 0) gen(3)

This worked very well and (with the help of gnuplot) I was able to
select an operating point that kept the capacitor charged to the correct
value. I added a voltage source in series with Ed so that when I
performed "op" analysis it would find this bias point.

I had hoped that running op and ac would then produce a linearized
frequency response but this didn't work and I got no response (despite
the fact that op found the correct point.) In order to check that I am
using the syntax correctly I manually linearized the circuit such that
"Ed" became two voltage sources, one dependent (Vc*gen0) and one
independent (gen*Vc0) where Vc0 is the operating capacitor voltage and
gen0 was the operating input value found through dc analysis. This new
ac analysis worked and produced a believable frequency response.

I would like to know if there is a way to model a dependent source like
this with variable gain in such a way that gnucap is able to linearize
the circuit and perform ac analysis. Linearizing circuits by hand is a
long process for all but the simplest circuits. Documentation for other
simulators suggests that it could be worked around with "value={}"
notation or a "B" element but I don't think either of these are
implemented in gnucap. Is there another approach or way of thinking
which would enable me to perform this analysis?

Thank you for any help you can give, and sorry for quite such a long
email for such a small question!

John

--------------------------------------
Please ignore the following disclaimer

This message has been checked for viruses but the contents of an attachment
may still contain software viruses, which could damage your computer system:
you are advised to perform your own checks. Email communications with the
University of Nottingham may be monitored as permitted by UK legislation.