Eznec Question

Discussion in 'Antennas, Feedlines, Towers & Rotors' started by K5WZI, Jul 23, 2019.

ad: L-HROutlet
ad: l-rl
ad: Left-2
ad: FBNews-1
ad: L-MFJ
ad: Subscribe
ad: Left-3
  1. UA3TW

    UA3TW Ham Member QRZ Page

    I have an antenna EZNEC model and I'd like to scale it for a higher frequency band (twice as high for example) and to put the scaled model within the initial one to see if I can have a 2 band model.
    Is it possible without tedious printing all over again?
     
  2. K7TRF

    K7TRF Premium Subscriber QRZ Page

    I've done that with nested Moxons and other similar antennas. I don't recall the exact keystrokes but the basic idea is to first save your model for one frequency band. Then use the wire scaling function in the wires dialog box to scale the antenna to a new frequency and use Save As to give it a new name. Do that for each new frequency band you want to build a scaled model of the antenna. Then from one of those models use the Import function to bring in the additional wires from one of the saved models. When you import you can either replace or add to existing wires, in this case you want to add wires.

    There are also ways to export wire lists to ASCII files and then import those saved text files. I've done that before as well when I wanted to merge multiple wire lists.

    It's been a while since I went through that exercise so I don't recall the exact menu items to use or sequence of events and do remember it took me a few trials to figure it out but you can definitely scale a wire list in just X and Y axes or in all three axes if that makes sense for your antenna (e.g. incorporates elements that extend using Z as well as X and Y) save the wire lists and then import saved wire lists into one of the scaled versions.
     
  3. UA3TW

    UA3TW Ham Member QRZ Page

    Many thanks, Dave. I know, how to scale and save but do not know, how to import from other file. Will do another try. In MMANA it is easy.
    I'd like to nest those models or the like
     

    Attached Files:

    Last edited: Feb 6, 2020
  4. K7TRF

    K7TRF Premium Subscriber QRZ Page

    I just checked it out and it's pretty easy in EZNEC as well. Just go to the File menu in the upper left on the main control panel, on the pull down menu select Add Description to bring in the wire list from a saved file and then use the file browser to select the correct file to import. You'll bring in the entire saved EZNEC description so you may have to sort out multiple Sources, transmission lines and the like but it's a quick way to bring in an entire saved file.

    You can also bring in ASCII files from that same file menu but if you have saved EZNEC files then the Add Description method is pretty easy.
     
    UA3TW likes this.
  5. UA3TW

    UA3TW Ham Member QRZ Page

    Yes, it happened! Thanks a lot, Dave!
     
    K7TRF likes this.
  6. UA3TW

    UA3TW Ham Member QRZ Page

    I got AutoEZ a mounth ago - Sergej UA2FW has gifted me one. Roughly can work with it but still have to get used to it. Now trying to pacify some of my MMANA models with its EZNEC counterparts.
    The MMANA models have real life antennas that work according to their models. Will report later, when I will have my mind made up.
     
  7. UA3TW

    UA3TW Ham Member QRZ Page

    Up to now I converted the MMANA model that Gennadij RU7L has made (photo). For simplicity I made it of constant diameter tubes. Let me know please what drawbacks I made.
     

    Attached Files:

  8. AC6LA

    AC6LA Ham Member QRZ Page

    I don't see much of anything wrong with the conversion. I guess it all depends on what you were trying to resolve/investigate.

    A few comments:

    1) I see you replaced .maa wires 6/7 and 13/14 with NEC transmission lines. Those .maa wires have a diameter of 20mm and a spacing of 40mm. With an assumed VF of 0.97 that works out to Zo 153. Your Zo 160 is certainly close enough.

    TLZo.png

    2) Wires 38/39 and 40/41 in the .maa model would cause problems with a NEC engine so you changed them slightly in the .weq version. Another good thing to do. Perhaps a small improvement would be to break .weq wire 3 into three collinear wires with the middle one having the same length as wires 4/5 (80mm) and having 3 segments. That would keep the segment boundaries aligned. Then set the length of the interconnect TLs (TLs 1 and 2) to something like 0.0001m rather than "Actual dist" (which is 0.05m). That would better simulate a direct connection. But perhaps I've misunderstood why you chose Zo 500 ohms for those TLs.

    I ran sweeps in AutoEZ (with model 5_BAND_2017_7_for_AutoEZ_tune.weq) and MMANA (with model 5_BAND_2017_7.maa) from 13 to 30 MHz. Here's a comparison of the SWR values:

    CompareSWR.png

    And here's a comparison of several other metrics at five spot frequencies.

    CompareOther.png

    In addition to the different engines there is also a fairly large difference in the segmentation. With the AutoEZ .weq model: 560 segments at 14 MHz and 676 segments at 28.6 MHz. With the MMANA .maa model: 827 pulses at 14 MHz and 1393 pulses at 28.6 MHz. Note that when AutoEZ opens a type .maa file the segmentation for all wires is set as "20 segs/WL, odd". That definitely will not be equivalent to the MMANA "-1" automatic segmentation.

    Dan, AC6LA
     
    KF6A likes this.
  9. UA3TW

    UA3TW Ham Member QRZ Page

    Many thanks, Dan. One thing still bothers me. You compared the .weq model and initial 5_BAND_2017_7.maa. They meet very well. There is 5_BAND_2017_7_AutoEZ_tuned.maa where I have input the dimensions of the elements after AutoEZ optimisation into the original model. It does not coinside so well - 17m and 12m bands want driven elements lengthening. Where it comes from I wonder.
    Now I began to do the same with this model https://forums.qrz.com/index.php?threads/small-tribander-model.620299/#post-5040767 and it seems to be much more difficult. Up to now I only succeded in making .weq model that do not send segmentation check warnings. Is it possible to improve it any better?
     

    Attached Files:

    Last edited: Apr 29, 2020
  10. AC6LA

    AC6LA Ham Member QRZ Page

    I'm afraid I don't know.

    Wow. That will be tough to do with NEC. The first thing I did was take a look at the AGT results. Pretty terrible with both the NEC-2D and NEC-4D engines.

    AGT.png

    I imagine a large part of that is due to the multiple wire junctions near the source. W59 (with the source) has seg len 0.027 m and diameter 10 mm. Collinear wires W4 and W60 have seg len 0.505 m and diameter 30 mm. Adjacent segments should differ in length by no more than a 2:1 ratio and should have a smaller jump in diameter.

    Also, wire pairs W61/W63 and W62/W64 intersect very close to the source. That would be a good place to use the Cebik "almost zero length transmission line" trick to make the connection, rather than a direct wire-to-wire connection.

    Finally, if you look at the EZNEC Wires window and click Other > Show Stepped Dia Correction you'll see this:

    SDC.png

    Because of the various bends and connected wires, EZNEC is unable to apply the usual substitution of one constant-diameter wire in place of a set of stepped-diameter wires. That will lead to even more inaccuracies in the calculations. See "Stepped Diameter Correction Overview" in the EZNEC Help.

    If you really, really want to model this antenna with a NEC engine you might try breaking it down to just the active wires for a particular band. Then try combining two bands, then three bands, etc. Just a thought.

    BTW, I used this tool to compute the AGT for the MMANA model "Main_pentabander_comment.maa" from post #10 in the thread linked above. After removing all load loss and changing the material to "No Loss" here's what I got at 14.15 MHz:

    AGTmaa.png

    Again, pretty terrible.

    Dan, AC6LA
     

Share This Page

ad: vanity