My Biquad doesn't work

Discussion in 'Wireless Internet' started by Knight, Sep 6, 2007.

  1. Knight

    Knight Guest

    Hey guys,

    What's happening? I thought everything came out great! but the
    BiQuad I just built
    doesn't seem to show any increase in signal strength while using
    Netstumbler!
    I can't see where I went wrong can you?
    http://img.villagephotos.com/p/2003-8/361947/IMG_6728(Large).JPG
    I don't have a piece of wire down in the center stub but I don't think
    that matters do you?

    How can the center pin be connected to one side of the element
    and the other side of the element be connected to the board? Seems
    like a short to me.
    Thanks guys!

    Pete
     
    Knight, Sep 6, 2007
    #1
    1. Advertisements

  2. Knight

    mike Guest

    I've read on some URL that if you don't use wire down the center stub,
    it will skew the pattern.
    Also what is the spacing between PCB and Biquad, the picture makes it
    look to close.
    At dc it would be a short, at 2.4Ghz there's a lot of inductance and
    capacitance.

    Mike
     
    mike, Sep 6, 2007
    #2
    1. Advertisements

  3. Yep. Your construction in the photo is totally wrong. It won't work
    as shown. The ground ends of the quad elements must be soldered
    directly to the ground sleeve of the coax connector, or to short coax
    extension. In addition, that looks like a BNC or F connector, which
    is a different problem. The center conductor is far too long, the
    connector is lossy at 2.4GHz, and it might be a bit flimsy. You'll
    find that most construction articles demand an "N" connector or coax
    cable extension.

    I scribbled a detailed rant on the subject at:
    <http://groups.google.com/group/alt.internet.wireless/msg/63d3438fe861bbaa>
    It should give a list of biquad construction articles that work, along
    with a list of one's similar to yours that will not work.

    My favorite are the ones that do not use any connector, such as:
    <http://martybugs.net/wireless/biquad/>
     
    Jeff Liebermann, Sep 6, 2007
    #3
  4. Knight

    Knight Guest

    Mike,
    The spacing is about 15 mm like the article I followed to build it
    suggested.
    I think I have it working now. Don't think the card was connected!
    duh.
    But that wire down the center just doesn't seem like it would matter
    after
    all that center casing is part of the wire if in fact you do solder
    one in!

    Pete
     
    Knight, Sep 6, 2007
    #4
  5. Knight

    Knight Guest

    Jeff,

    It is a "N" connector. A "N" bulkhead connector.
     
    Knight, Sep 6, 2007
    #5
  6. Knight

    Knight Guest

    Guys, heres a screenshot of what's happening while the quad is
    connected.
    Does it look OK to you? Doesn't look bad to me!
    http://img.villagephotos.com/p/2003-8/361947/biquad.jpg

    Let me know what ya think. This is the same one as pictured above,
    didn't change a thing.

    Pete
     
    Knight, Sep 6, 2007
    #6
  7. Ok, I couldn't tell the size from the photo.
    Looks like about 8dBi of gain over what you were using as a reference
    antenna. That's about what I would expect from a Biquad. I guess
    it's working.

    A while ago, I got into a debate over the construction of biquads and
    how much creativity can be tolerated. So, I ran some models and got a
    suprise. See:

    A proper biquad:
    <http://802.11junk.com/jeffl/antennas/Biquad/>

    A creatively built biquad-junk:
    <http://802.11junk.com/jeffl/antennas/biquad-junk/>

    Despite the miserable 7:1 VSWR mismatch, the junk antenna still
    manages to get about 11dBi of gain, the same as the proper biquad. So,
    if your radio can handle the crappy VSWR, feel free to throw it
    together any way that seem workable.
     
    Jeff Liebermann, Sep 7, 2007
    #7
  8. Knight

    Knight Guest

    Jeff,
    Earlier you wrote "Your construction in the photo is totally wrong"

    I guess it really doesn't matter then does it? Seems to be working
    ok!
    Can that be? Would it work better the other way?
    Pete
     
    Knight, Sep 7, 2007
    #8
  9. Yep. It's totally wrong. However, it appears that such creativity
    doesn't have as much effect as I original assumed. Look at it this
    way: Anything is better than the stock rubber ducky antenna.
    Yep. My simulation of doing it the wrong way shows that it works well
    enough. You decide if it's good enough for your purposes and if it's
    worth your time doing it the right way.
    Yes. The rather high VSWR (impedance mismatch) will certainly have an
    effect on gain. If I plug the estimated antenna impedance from:
    <http://802.11junk.com/jeffl/antennas/biquad-junk/slides/main.html>
    as 58.7 -j106 ohms into a mismatch calculator at:
    <http://chemandy.com/calculators/return-loss-and-mismatch-calculator.htm>
    I get about 3dB of loss with this mismatch. In theory, you should be
    able to recover most of that by proper construction.
     
    Jeff Liebermann, Sep 7, 2007
    #9
  10. Knight

    miso Guest

    http://www.lazygranch.com/images/wifi/wifi_bq_1.jpg
    http://www.lazygranch.com/images/wifi/wifi_bq_2.jpg

    Not the greatest looking construction, but it works. Loops are the
    same length as the diamond.
     
    miso, Sep 7, 2007
    #10
    1. Advertisements

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.