Open ended coaxial stub model

Help
Yves OESCH
2011-06-08
2012-12-15
  • Yves OESCH
    Yves OESCH
    2011-06-08

    Hi,
    I'd like to simulate a filter on QUCS, which is based on open coaxial stubs ( http://www.qsl.net/yu1aw/VHFPower.htm )
    QUCS proposes the coaxial line as component, but how to model the open end. It cannot be left open, as there is an error in the netlist in this case. I think to place a capacitor, but which value to choose?
    Best would be to have a model of the open coax, that models the fringing fields, as this is available for the microstrip lines (MOPEN)
    Anybody has an idea about how to model an open coaxial stub?

     
  •  magneto98
    magneto98
    2011-06-08

    Hi hb9dtx,

    I don't get an  error during simulation.
    Did you mean the "coaxial transmission line" component?
    Schematic file:

    <Qucs Schematic 0.0.16>
    <Properties>
      <View=0,0,1153,800,1,0,0>
      <Grid=10,10,1>
      <DataSet=coax_stub_match.dat>
      <DataDisplay=coax_stub_match.dpl>
      <OpenDisplay=0>
      <Script=coax_stub_match.m>
      <RunScript=0>
      <showFrame=0>
      <FrameText0=Title>
      <FrameText1=Drawn By:>
      <FrameText2=Date:>
      <FrameText3=Revision:>
    </Properties>
    <Symbol>
    </Symbol>
    <Components>
      <COAX Line5 1 480 280 -100 -26 0 3 "2.29" 1 "0.022e-6" 0 "1" 0 "2.95 mm" 0 "0.9 mm" 0 "1000 mm" 1 "4e-4" 0 "26.85" 0>
      <Pac P1 1 280 240 18 -26 0 1 "1" 1 "50 Ohm" 1 "0 dBm" 0 "1 GHz" 0 "26.85" 0>
      <GND * 1 280 320 0 0 0 0>
      <COAX Line6 1 640 180 -26 16 0 0 "2.29" 1 "0.022e-6" 0 "1" 0 "2.95 mm" 0 "0.9 mm" 0 "1000 mm" 1 "4e-4" 0 "26.85" 0>
      <COAX Line7 1 760 250 16 -26 0 1 "2.29" 1 "0.022e-6" 0 "1" 0 "2.95 mm" 0 "0.9 mm" 0 "1000 mm" 1 "4e-4" 0 "26.85" 0>
      <COAX Line8 1 880 250 16 -26 0 1 "2.29" 1 "0.022e-6" 0 "1" 0 "2.95 mm" 0 "0.9 mm" 0 "1000 mm" 1 "4e-4" 0 "26.85" 0>
      <Pac P2 1 1020 250 18 -26 0 1 "2" 1 "50 Ohm" 1 "0 dBm" 0 "1 GHz" 0 "26.85" 0>
      <GND * 1 1020 320 0 0 0 0>
      <.SP SP1 1 250 440 0 71 0 0 "lin" 1 "0" 1 "200MHz" 1 "400" 1 "no" 0 "1" 0 "2" 0 "no" 0 "no" 0>
    </Components>
    <Wires>
      <480 180 480 250 "" 0 0 0 "">
      <280 180 480 180 "" 0 0 0 "">
      <280 270 280 320 "" 0 0 0 "">
      <280 180 280 210 "" 0 0 0 "">
      <480 180 610 180 "" 0 0 0 "">
      <670 180 760 180 "" 0 0 0 "">
      <1020 280 1020 320 "" 0 0 0 "">
      <1020 180 1020 220 "" 0 0 0 "">
      <760 180 880 180 "" 0 0 0 "">
      <760 180 760 220 "" 0 0 0 "">
      <880 180 1020 180 "" 0 0 0 "">
      <880 180 880 220 "" 0 0 0 "">
    </Wires>
    <Diagrams>
      <Rect 600 694 331 254 3 #c0c0c0 1 00 1 0 0.2 1 1 -0.1 0.5 1.1 1 -0.1 0.5 1.1 315 0 225 "" "" "">
    <"S" #0000ff 0 3 0 0 0>
      </Rect>
    </Diagrams>
    <Paintings>
    </Paintings>

    Carsten

     
  • Yves OESCH
    Yves OESCH
    2011-06-11

    Hi Carsten,
    you're right, the error was coming from another stupid reason. To answer your question, Yes I am speaking of the coaxial line component.
    Anyways, it still not clear for me:
    If I leave open a coaxial line, to use it as a stub in a circuit, is the open end effect of the coaxial line integrated into the model or not?
    Any idea or link?
    Best Regards
    Yves

     
  •  magneto98
    magneto98
    2011-06-13

    Hi Yves,

    I guess, open end effects are not included in the "Coaxial line component".
    There would be a model "Coaxial line open" as there is for microstrip and coplanar waveguide.
    A quick look in the Technical papers also gave no hit for open end effects in conjunction with coax.

    The name of this  IEEE paper sounds quite promising for me:
    "Universal formula for frequency-dependent coaxial open-end effect"
    http://ieeexplore.ieee.org/search/freesrchabstract.jsp?tp=&arnumber=803623

    Good luck
    Carsten

     
  • Yves OESCH
    Yves OESCH
    2011-06-14

    Hi Carsten,
    Many thanks for the hint. I've read the paper, and this time the calculations were not that awfull, and almost usable. I wired up a small excel sheet and ended up with a fringing capacitance of about 0.08pF, almost constant in the frequency of interest (up to 1.3 GHz) for a RG-213 geometry. This is almost negligible in fact.

    But it might be fun to add this open end coaxial model to QUCS for those who might need it. But I've no real idea how to do it. I guess first make a subcircuit whose behavior is according to the model… then commiting it to the source code…
    Still some work for me to use and understand QUCS until I will be able to do it.
    Anyways, thanks for the hint and the usefull help.

    Best Regards

    Yves / HB9DTX

     
  •  magneto98
    magneto98
    2011-06-15

    Hi Yves,

    glad to hear that.
    Could you post the formula, that you found, please?
    Unfortunatly, I'm not familiar with subcircuits.

    Carsten

     
  • Yves OESCH
    Yves OESCH
    2011-06-22

    Hi Carsten,
    I'll do it, it's not a simple formula, but a set of formulas in an excel sheet, but I'd like to check it first, and I've no time these days. I hope to do it during July. Sorry for the delay.

     
  • Yves OESCH
    Yves OESCH
    2011-07-17

    Hi Carsten,
    I've found some time to check my excel sheet against the original IEEE paper. I hope I didn't made any error, as there are quite a number of parameters to copy with several decimal places… This paper is based on curve fitting, and polinomial approximations.

    My excel sheet that translates the paper into a directly usable file can be downloaded here:
    http://dl.free.fr/nam1D7Fcv
    Password to get it is simply "qucs" I don't know if the user is empty or also "qucs" tell me if you can't get it.

    It should be stored on the server at least 30 days from the latest download.

    I've added a calculation of the characterisitc impedance of the coaxial line, so that the parameters at input should be the same as the one QUCS asks for when designing with coax lines, so taht further implementation as a new component or sub circuit should be eased.

    I'd be glad to get your comments.

    Yves

     
  •  magneto98
    magneto98
    2011-07-20

    Hi Yves,

    i got the excel sheet but had no time yet to take more then a quick look.
    I hope to do it during July. Sorry for the delay.

    Carsten