Re: [Help-gnucap] Graphics output

[Karel Kulhavy]

On Wed, Oct 10, 2007 at 03:26:27PM +0200, Per Bull Holmen wrote:
> Hey, I'll try again - in hopes that the list is not
> completely dead. My last problem is now solved, though
> I did code some plain c++ (don't know whether that was
> really necessary). Now gnucap turns out to be far
> superior to any other simulator I've tried, for my
> purpose... :) Those other ones include LT-Spice,
> ng-spice, MacSpice and more. Thanks to Al Davis and
> other contributors!!

Can gnucap already do noise analysis?
> 
> Anyway, it seems I don't get graphics output from the
> plot commands? At first I thought there was some shell
> variable thing, and tried setting DISPLAY environment
> variable - but then I thought perhaps gnucap is not
> meant to output graphics, other than characters based
> "graphics" on the command-line?
> 
> What do you others recommend for graphics output,
> installing gEDA? Or have I missed some settings in
> gnucap/shell? I would like to have a simple solution

gtkwave or gnuplot or GNU R. Depends on what you want to do.  If you want to
browse like oscilloscope, gtkwave. If you want calibrated graticules and pretty
graphs, gnuplot. If you want graphing together with further numerical analysis
(frequency spectrum of impulse response, statistical distributions etc., GNU R.

Here is an example page
http://ronja.twibright.com/technotes/how_rx.php

First come some simulations of power filters. I wanted to make sure the LC
filters filter properly and don't have any band where they significantly "ring"
and actually improve passing of unwanted signal through the shielding envelope.
This was done solely with gnucap and gnuplot.

Then there is an example of theoretical analysis I did to predict how is range
degraded if multiple segments of particular optical link are chained with
amplitude limiting the signal in between but not retiming it. It began with
simulating the receiver chain to get the exact RX waveform. Some of the graphs
are produced by gnuplot, some by GNU R.

> so that I can easily get a window with a curve by
> issuing a single command or perhaps pressing a button.
> This is not really scientific engineering... :) I have

The MS Windows-style click-and-clack Lego Duplo approach (warning - unsuitable
for children above 3 years) is not scientific engineering. Because your screen
has no output which you could plug into GNU R to get frequency,
statistics or whatever analysis. The proprietary
system you are probably talking about is
capable just of what's pre-programmed and not a single bit more.

CL<
> gotten quite comfortable with netlists though, so
> graphical *input* is not essential, but might get
> essential later. As you can see I'm totally
> inexperienced with circuitry simulation.
> 
> (BTW I'm doing sims on MacOS 10.4 on intel-mac. That's
> how I've been able to test both Win Spices and
> MacSpice. I prefer to do it the mac way or X-Windows
> way, though... :)
> 
> 
> 
> -- 
> 
> Mvh
> Per Bull Holmen
> 
> "People demand freedom of speech as a compensation for the freedom of thought 
> which they seldom use" - S?ren Kierkegaard
> 
> 
>       
> _________________________________________________________
> Alt i én. F? Yahoo! Mail med adressekartotek, kalender og
> notisblokk. http://no.mail.yahoo.com
> 
> 
> _______________________________________________
> Help-gnucap mailing list
> address@hidden
> http://lists.gnu.org/mailman/listinfo/help-gnucap