S-parameters export to touchstone file

2012-03-23
2012-12-15
  • Does anyone know how to export the S-parameters simulation results to a touchstone file?
    This would be a nice feature to have if not already possible.

    Thanks
    Maurizio

     
  • Frans
    Frans
    2012-03-23

    I think it is possible using the command line tool qucsconv.
    also if you select the graph of the plotted s parameters you can at least export the simulation results as csv.
    using a text editor you need to put the touchstone header.
    i agree that a graphical one click way to expport the touchstone file would be really nice to have.

     
  • Yes, the CVS export works, but kind of slow, the data need to be formatted in column one by one. I will look into the command line, maybe I can create a script which can export all Sxy graphs.
    By the way I am having problems with  touchstone files. very simple 2 ports files, but the s-parameters and transient simulation give me a 0V straight line, looks like the port 2 is not even connected. Does anyone know of compatibility issues of QUCS with normal touchstone files? I am not getting any error, so hard to tell what is going on.
    Cheers
    Maurizio

     
  •  magneto98
    magneto98
    2012-03-23

    I had never problems using touchstone files  (up to 8 ports) in simulations

     
  • consider a simple  circuit with only a file component, i.e. a 2 port s-parameter file  . I can simulate the S-parameters,  I see the expected  graph so the touchstone file is ok; but  if I try to simulate a transient response the  output voltage is just a line, as if nothing is connected from P1 to P2. Have you ever tried any simulation with a file component? Ibelieve there is a problem in the way QUCS is using  the S-parameters file. Any touchstone file I use has zero level output with transient or AC analysis simulation. The S-parameters are correct, were created with a quite expensive instrument based on TDR. In this pecific case I have used a  6 GHz low pass filter. Putting a pulse generator at P1 I am expecting to see a signal at P2, not a straight line…..If Instead I use a discrete component filter  the AC or the transient simulation work. The question is then, why a file component based on a touchston file can not be AC/transiet simulated in QUCS ?  Any clue?

     
  •  magneto98
    magneto98
    2012-03-26

    I had no problems with AC simulations and the "s parameter file" component. Please post your schematic / touchstone file (or better use pastebin.com because of the line wraps.)

     
  • Hello magneto98
    I have posted the sparamas_test.sch and the filter s-parameters  on the pastebin website. look for my acount mmastro59.
    I tried several setting of the duringDC prameter of the filter, but the results are not as eexpected. Esentially the filtere is either a short or an open…I am using a 10ps rise time pulse, and the output at the 2.5 GHz is 0 or identical to the input stiulus.
    Thanks

     
  •  magneto98
    magneto98
    2012-03-26

    Hi mmastro59,

    please post the Qucs schematic file. The one you posted is the already converted file (something like SPICE). I don't know how to use that one.
    the touchstone looks correct.
    bye

     
  • Ok I posted the sch file, let me know if it works.
    I was able to have some results in AC analysis. I get a correct analysis setting the duringDC  to  the correct value. Instead still nothing with transient simulation.
    Bye

     
  •  magneto98
    magneto98
    2012-03-27

    I'm not surprised, that the transient simulation doesn't work. There were other people reporting problems. Sometimes (or always?) it can be solved with tweaking the settings of the transient simulation block. If this is a bug or whether your simulation is too complex I cannot say. I guess it is a good idea to use a very simple touchstone file with only little frequency points when testing the simulation parameters. Sometimes it is hard to say, if your simulation is in a infinite loop or just needs some more minutes to finish.

     
  • Thanks Magneto98 for confirming the problem with the transient response. I will try with fewer point.