[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: c2d is a Stub that only Produces Error Messages
From: |
Lukas Reichlin |
Subject: |
Re: c2d is a Stub that only Produces Error Messages |
Date: |
Tue, 22 Feb 2011 19:52:02 +0100 |
On 22.02.2011, at 14:26, Doug Stewart wrote:
>
>
>
> On Tue, Feb 22, 2011 at 7:33 AM, Lukas Reichlin <address@hidden> wrote:
> On 22.02.2011, at 13:16, Doug Stewart wrote:
>
> >
> >
> >
> > On Tue, Feb 22, 2011 at 7:04 AM, Lukas Reichlin <address@hidden> wrote:
> > On 22.02.2011, at 05:59, Thomas D. Dean wrote:
> >
> > > Octave 3.4.0, control 2.0.0
> > >> From the Converting from S to Z domain thread,
> > >
> > > octave:1> num = [1 0];den = [1 0 4];
> > > octave:2> sys = tf(num, den,1);sysz=c2d(sys,1);sysout(sysz,'tf')
> > > error: tf: c2d: not implemented yet
> > > error: called from:
> > > error: /home/tomdean/octave/control-2.0.0/@tf/__c2d__.m at line 31,
> > > column 7
> > > error: /home/tomdean/octave/control-2.0.0/@lti/c2d.m at line 46,
> > > column 7
> > >
> > > Looks like __c2d__ is a stub that only displays error messages, either
> > > "not implemented" or "invalid method".
> > >
> > > # cat /home/tomdean/octave/control-2.0.0/@tf/__c2d__.m
> > > ...<snip>
> > > function sys = __c2d__ (sys, tsam, method = "zoh")
> > >
> > > [p, m] = size (sys);
> > >
> > > switch (method)
> > > case {"zoh", "std"}
> > > error ("tf: c2d: not implemented yet");
> > >
> > > otherwise
> > > error ("tf: c2d: %s is an invalid method", method);
> > >
> > > endswitch
> > >
> > > endfunction
> > >
> > > tomdean
> >
> > Yes, @tf/__c2d__.m is only a stub right now. @ss/__c2d__.m offers
> > zero-order hold only. There is a list of missing features in the PROJECTS
> > file:
> >
> > http://octave.svn.sf.net/viewvc/octave/trunk/octave-forge/main/control/devel/PROJECTS
> >
> > Lukas
> >
> > _______________________________________________
> > Help-octave mailing list
> > address@hidden
> > https://mailman.cae.wisc.edu/listinfo/help-octave
> >
> >
> > Lukas , are you saying that c2d does not work any more???????
> > Doug Stewart
>
> Actually it never did for transfer functions. Internally, control-1.0.11
> computed everything in state-space. As for now, you need to convert to
> state-space by sys = ss (sys) and discretize afterwards. Note that there is
> no d2c at all and that helping hands are always welcome ;-)
>
> Regards,
> Lukas
>
>
> How does one revert to the old control tool box?
> Doug
This was my post on 2011-02-08 from the octave-dev <address@hidden> mailing
list, please read the third paragraph:
FYI: control-2.0.0 released in package forum - please upload
> Hi all,
>
> I've just published control-2.0.0 in the package release forum [1]. It
> requires Octave 3.4.0 and it offers better compatibility (than control-1.0.x)
> with the "control system toolbox" and parts of the "robust control toolbox"
> of the "dark side". Since it uses the OOP features introduced with Octave
> 3.2.0, you can write things like sys3 = sys2 * sys1 for a series connection
> of two LTI systems.
>
> Furthermore, the new control package is based on Fortran routines from the
> Slicot [2] project. Reliable solvers for AREs (linear-quadratic regulators
> and filters), Lyapunov and Sylvester equations are included. Controller
> synthesis includes pole placement techniques as well as H-2 and H-infinity
> methods.
>
> I hope you like my package. For those who don't and/or have m-files written
> for the old version of control: there's still the old control-1.0.x package
> available under extra/control-legacy. Please note that it is not possible to
> use both control packages at the same time. Compiling the Fortran code to
> oct-files can take up to several minutes, so please be patient when using
> "pkg install".
>
> Enjoy
> Lukas
>
> [1]
> http://sourceforge.net/apps/phpbb/octave/viewforum.php?f=1
>
> [2]
> http://www.slicot.org
There's a link (older versions) on OctaveForge [1]. Additionally, you can grab
any revision from the SVN repository at any time. SF even creates a tar.gz file
automatically, you just need to remove the leading "octave-" in order to
install it by "pkg install".
Regards,
Lukas
PS: I tinkered with control-1.0.x until I realized that I was riding a dead
horse and therefore the best strategy would be to dismount. Don't whip
control-1.0.11 too much ...
[1]
http://octave.sourceforge.net/control/index.html
Re: c2d is a Stub that only Produces Error Messages, Thomas Weber, 2011/02/22