ad: elecraft

Need help modeling a fan dipole in EZNEC

Discussion in 'Antennas, Feedlines, Towers & Rotors' started by W0BTU, Aug 27, 2016.

ad: L-HROutlet
ad: l-rl
ad: Subscribe
ad: Left-2
ad: L-MFJ
ad: Left-3
ad: MessiPaoloni-1
  1. W0BTU

    W0BTU Ham Member QRZ Page

    I've wasted the past few hours trying to model a fan dipole --two dipoles fed from the same feedpoint-- in EZNEC+ v. 5.0.67. It's been an exercise in frustration! I keep getting source errors. And Google searches haven't helped. I've seen suggestions on how this might be done, but nothing that's been tested and proven.

    The antenna needs to have a 75m and a 20m dipole fed from a single feedline. (Once I get that working, I need to add a 40m dipole to it.) I've tried many things, such as adding more wires to separate the dipole feedpoints at the source , but nothing I've tried seems to work. :-(

    Does anyone have a good EZNEC fan dipole .ez model file that they can post here or send me? Please? :)

    Thanks!
     
  2. K7TRF

    K7TRF Premium Subscriber QRZ Page

    There's a few ways to handle this including use of split current sources but for parallel dipoles or inverted-Vs a simple way is:

    - Introduce a very short feeder wire perhaps just a couple of inches long and place a normal current source at 50% of this feeder wire
    - Connect the rest of your dipole wires to each end of this feeder wire.

    So your wire list might look something like:
    parallel dipole wires.jpg

    That's for a 40m/20m set of parallel dipoles fed from a common feed point.
     
    W0BTU likes this.
  3. W0BTU

    W0BTU Ham Member QRZ Page

    Thanks!!! :)
    I created a new model like you suggested. However, I still get the error "one or more sources are incorrectly placed". Screenshot attached.

    What am I doing wrong?
     

    Attached Files:

  4. E73M

    E73M Ham Member QRZ Page

    Change source type to "I" instead SI

    73 Danny N4EXA aka E73M
     
    W0BTU and K7TRF like this.
  5. K7TRF

    K7TRF Premium Subscriber QRZ Page

    Yup, as Danny points out you've got a split current (SI) source inserted mid conductor.

    Split sources are used when you must place the source at the junction of multiple wires and then they're placed at either 0% or 100% of the conductor length or IOW at one end or the other. If you went with multiple conductors emerging right from the source and didn't use the tiny center feed conductor trick you'd want to use a split source but then you sometimes run into troubles in converging dipoles where the thickness of the conductor and angle of the various elements means that segments intersect each other due to their diameters. The center feed conductor trick often helps you avoid that kind of physical intersection issue depending on the specific angles, segment lengths and conductor diameters. Many ways to skin this cat.
     
    W0BTU likes this.
  6. W0BTU

    W0BTU Ham Member QRZ Page

    Thank you, Gentlemen!!! :)

    I changed the source to "I" and there were no source errors when I saved the file. The FF plot looks good on both bands.

    I don't have anymore time right now to continue this, but I will soon. Model attached in case anyone wants to play with it.
     

    Attached Files:

  7. W0BTU

    W0BTU Ham Member QRZ Page

    I've been playing with this again. I added a third dipole, and something is just not right. There is no resonance on 20 meters, and I cannot fix the segments.
    The model and screenshots are attached. I'm probably just going to build the darn thing and see what happens.

    Thanks for the help!! :)
     

    Attached Files:

  8. K7TRF

    K7TRF Premium Subscriber QRZ Page

    I'm not sure what's going on with your model, but a couple of things jump out. The angles are pretty extreme with the low end points so it's more like nested inverted-Vs which may be impacting things a bit. Some antennas like Cobwebs (which are basically a variation on parallel dipoles) can be tough to model if you don't include a sufficient number of wire segments in the high current regions near the feed point. You can tackle this either by just using a ton of wire segments which slows processing and may not be supported by the free demo version of EZNEC or you can get creative with wires so that you have a lot of segments where you need them, close in to the feed point but fewer segments out further from the feed point. Those are more complex models to build but can be helpful for certain antenna designs.

    Here's an 80m/40m/20m set of common feedpoint dipoles with more modest angular separation and heavily segmented. The SWR dip is still pretty high on 20m compared to the other bands but there is a definite dip below 2:1. Since actual multi-band dipoles like this that I've built tend to show better matching I also added 50' of transmission line in the LMR-240 class to emulate feed line losses and how that moderates SWR. Anyway, here's a screenshot showing the wire list, the antenna and the 20m SWR curve including the transmission line.

    But yes, in some cases when you know the design isn't way out there and you have reason to suspect it should work fine it can pay to just go ahead and build it. But from the standpoint of gaining confidence in modeling skills it's sometimes worth the effort to figure out why models don't seem to give the expected results even for relatively common antennas.
    80-40-20-parallel dipole wires.jpg
     
    W0BTU likes this.
  9. AC6LA

    AC6LA Ham Member QRZ Page

    Concerning model "Fan dipole4.ez":

    There are some modeling guidelines to always keep in mind: 1) The segments on either side of a source should be the same length and diameter as the source segment. An easy way to guarantee that is to make the wire with the source have 3 segments and put the source in the middle. 2) The length of a segment should be no shorter than 0.001 WL at the lowest frequency of interest. 3) The length of a segment should be no longer than 0.05 WL at the highest frequency of interest.

    At 3 MHz, 0.001 WL is about 0.328 ft. So the smallest segment should be no smaller than that. And since the wire with the source should have 3 segments, that means that Wire 1 should have a length of 3 times 0.328 or 0.984 ft. Half of that (half on either side of the center) is 0.492 ft. So in your model, every place you have +0.1 change that to +0.492 and every place you have -0.1 change that to -0.492.

    At 22 MHz, 0.05 WL is about 2.235 ft. No segment should be longer than that. In your model the longest segment is over 65 ft. To make the segments shorter increase the number of segments in each wire. So where you have 6 segments, change that to 30. Change 3 to 16. Change 2 to 10. That will make the total number of segments 115 (was 23), still very modest.

    Now the shortest segment is no less than 0.328 ft and the longest segment is no more than 2.235 ft. After making all those changes you will no longer get any EZNEC segmentation check warnings.

    Here's a frequency sweep.

    Fan-sweep.gif

    Modified model attached. Not bad for 80-40-15m. If you wanted 75-40-20m it looks like you have some tweaking to do! :)

    For advanced modelers: If you run an Average Gain Test on this model you'll find that the results are not particularly good. About 1.18 dB at 3.5 MHz and about 1.13 dB at 21.5 MHz. Ideally you want the AGT to be less than ±0.2 dB. For a trick to solve that see the first bullet here:

    http://ac6la.com/aecollection4.html

    Dan, AC6LA
     

    Attached Files:

    W0BTU and KF6A like this.
  10. AC6LA

    AC6LA Ham Member QRZ Page

    I couldn't help myself.

    The use of variables along with the AutoEZ "Resonate" feature makes short work of getting this model to play nice on 75, 40, and 20 meters.

    Fan-sweep2.gif

    And using the L.B. Cebik "zero length transmission line" trick (mentioned in the previous link) makes the Average Gain Test a lot better.

    Fan-AGT.png

    Models attached, in both EZNEC (.ez) and AutoEZ (.weq) formats.

    Dan, AC6LA
     

    Attached Files:

    W0BTU, N7EKU and KF6A like this.

Share This Page

ad: AlphaRF-1