Two theoretical optimal antenna questions

Discussion in 'Antennas, Feedlines, Towers & Rotors' started by M0AGP, Jul 28, 2020.

ad: L-HROutlet
ad: l-rl
ad: Left-2
ad: L-MFJ
ad: Subscribe
ad: abrind-2
ad: Left-3
  1. M0AGP

    M0AGP Ham Member QRZ Page

    Yep the horizontally stacked yagis are a contender for sure. I seem to remember there was a rule of thumb for the minimum separation when you stack yagis - some percentage of the boom length I think. I wonder if you could horizontally stack 3 of these yagis in a 20m space? If we assume 1/8 wave radials, that leaves 15m to stack 3 yagis, or a separation of 7.5 meters between yagis. I wonder if that is too close?
     
  2. N3OX

    N3OX Ham Member QRZ Page

    I'm up to 11dBi free-space on a parasitic 9m radius 11-circle dipole array with lumped reactances at the feedpoints just messing around by hand.

    upload_2020-8-9_15-42-41.png

    upload_2020-8-9_15-43-13.png

    upload_2020-8-9_15-44-13.png

    upload_2020-8-9_15-44-35.png

    This would be an interesting one to throw into an optimizer with realistic feedline stubs.

    If it's pattern switched with the same component set, you probably don't really want the beamwidth to get much narrower. Needs more elements I guess.

    upload_2020-8-9_16-0-10.png
    It actually reduces the gain from the double stack but ends up being a clean pattern.

    upload_2020-8-9_16-22-53.png
     
    M0AGP likes this.
  3. AC6LA

    AC6LA Ham Member QRZ Page

    Rudy Severns, N6LF, did something similar with one driven element in the center of a square and four additional elements at the corners. Two opposing corner elements act as reflector and director and the other two opposing corner elements are made inactive by being open circuited at their feedpoints. In Rudy's case each element also had some top loading. Looks like this:

    [​IMG]

    With this pattern:

    [​IMG]

    Full details here (scroll down to the "Parasitic Elements" section) including links to the N6LF article and related articles.

    The associated model file "Vertical 3-el Parasitic.weq" currently has 130 segments total including the top loading wires. If the top loading is removed and the number of segments in each vertical is changed to 6 then the model could be used with the free Demo version of AutoEZ.

    Dan, AC6LA
    https://ac6la.com
     
    N3OX and M0AGP like this.
  4. M0AGP

    M0AGP Ham Member QRZ Page

    Thanks for posting this Dan - very relevant and very interesting! Yes open circuiting is a simpler way of switching off parasitics than de-tuning them. Smart! I wonder if you could also switch in/out some extra electrical length so you could reverse the direction of the Yagis?
     
  5. M0AGP

    M0AGP Ham Member QRZ Page

    Wow 11 dBi without optimization. How stable is that configuration to noise in currents or phases? Nice.
     
  6. N3OX

    N3OX Ham Member QRZ Page

    Seems pretty stable to everything. Perturbed a bunch of capacitors 5% and one of the currents 5% and got this (blue trace is original)

    upload_2020-8-9_19-4-18.png
    I'd guess if optimization can give it excellent F/B or quite a bit higher gain it'll get less stable to perturbed components.

    Will let you know if I try to optimize it. I don't have Excel otherwise I'd see if I could try it with AutoEZ :)
     
    M0AGP likes this.
  7. AC6LA

    AC6LA Ham Member QRZ Page

    All kinds of back-level versions of MS Office are available on eBay for very reasonable prices. Snoop around a bit and you can probably find something for $10 or $20. The other day I saw "Home and Student" 2016 from a seller in Sweden for $7, instant download.

    And I forgot to mention earlier: With the free AutoEZ Demo the calculation limit is 30 segments. But larger models can be opened, saved in EZNEC format, and then calculated directly in EZNEC rather than via AutoEZ. That way you can experiment with all the variables that control a model, seeing what changes what. The drawback is that all the various AutoEZ charting options won't be available, just what EZNEC can do itself.

    Dan, AC6LA
     
    M0AGP and N3OX like this.
  8. M0AGP

    M0AGP Ham Member QRZ Page

    Gents - here is an iteration on the above ideas. I am assuming we can switch each element into being a driven element, a reflector (add inductance at base), a director (at capacitance at base) or "off (open circuit). Here is the geometry:
    upload_2020-8-10_14-42-36.png
    except imagine the diagonals are the same lengths as the horizontals so it looks like a sort of snowflake.

    All elements are the same physical length.

    Say we wanted to point this beam northwest. Then we throw switches so we get this:

    upload_2020-8-10_14-36-18.png
    We have switched off all the grey ones and the one in the lower left is switched into inductance at the base so is a reflector. The next one diagonally is a driven element, then the rest are all directors (capacitance to shorten electrically).

    This should act like a 9 element Yagi - though not optimally spaced.

    And by throwing switches we can steer it to the 8 major compass points.

    So it should act like a steerable N element yagi, where N is the number of elements that both fit into 20m by 20m and that give the best yagi gain.

    Given this has to be a "reversible Yagi" (!) I wonder what the optimum spacings would be so that you can reverse it?

    The distance between the outer group and 2nd from outer could be adjusted, but all the rest would probably (?) have to be the same spacing (no?) - what is optimum?

    So 2 questions:
    1) optimal reversible yagi = ?
    2) how many elements would best fit in the 20m by 20m square assuming 1/8 wave radials?
     

    Attached Files:

    AK5B likes this.
  9. M0AGP

    M0AGP Ham Member QRZ Page

    Umm - that would be “northeast”, not “northwest” - oops (I even typoed my correction jeez)
     
  10. M0AGP

    M0AGP Ham Member QRZ Page

    To make the reversible optimal yagi question more concrete: the "boom length" is about15 meter long and it is a 20 meter band yagi.

    The thing to optimize is the average of gain and "reversed gain" when the elements are swapped over but the spacings remain the same but in reverse (yes, those solutions aren't the best due to side lobes, F/B etc - but this is to make it more specific).

    One simple solution would be to assume a 16m boom and uniform 4 meter spacing (0.2 wavelength) giving a long 5 element beam.

    I have the feeling that this would have much less than 11 dBi gain however so this antenna will not be the "winner" for the best parasitic structure!

    Might be easier to build though?
     

Share This Page

ad: elecraft