3
$\begingroup$

I've been trying to get acquainted with the 4NEC2 antenna modeling program. Perhaps this is obvious, but it's not clear to me that wires with one end in the same place are connected. I'm attempting to model a simple quarter-wave antenna with four radials, but I'm getting results that seem suspect. I shouldn't be seeing this symmetric pattern (picture below) along the $z$ direction. It makes no sense. It's as if the radials weren't even there.

I've compared my results to those of a vertical with four radials from Dick Reid, KK4OBI (NEC file here). Those show a strong asymmetry, as I would expect. There seems to be some deep magic going on here in the NEC file that I don't understand.

-Rod AD0YX (call sign updated too!)

Here's the NEC file I'm using (updated)

CM High-altitude balloon 70cm antenna
CM Rodney Price AD0YX July 2017
CE
SY freq=434.65             'In wideband simplex band
SY len=71/freq             'meters/sec over MHz
SY rlen=len                'length of radials
SY angle=30                'angle at which radials droop
SY rdroop=rlen*cos(angle)  'droop along x or y axis
SY zdroop=-rlen*sin(angle) 'droop along z axis
SY height=0                'height of bottom of vertical element above radials
GW  1   11  0   0   height  0       0       height+len    #8
GW  2   11  0   0   0       rdroop  0       zdroop        #8
GW  3   11  0   0   0       0       rdroop  zdroop        #8
GW  4   11  0   0   0      -rdroop  0       zdroop        #8
GW  5   11  0   0   0       0      -rdroop  zdroop        #8
GE  0
GN  -1
EK
EX  0   1   1   0   1      0   0
FR  0   0   0   0   freq   0 
EN

Here's the radiation pattern from 4NEC2:

enter image description here

Here's what I get when I run the KK4OBI model:

enter image description here

Improve this question
$\endgroup$
5
  • $\begingroup$ Looks like a great NEC file. Wire radius is reasonable. Yes, wires are considered connected if their ends are close enough together. But your antenna is too long - a full wavelength? Try L=70/F. Why the LD card on the driven element? $\endgroup$
    – tomnexus
    Jul 19, 2017 at 5:48
  • $\begingroup$ You are using full wave-length radials. $\endgroup$
    – Dick Reid
    Jul 19, 2017 at 21:38
  • $\begingroup$ Oops. That was an oversight. And I don't know how that LD card got in there. $\endgroup$
    – Rodney Price
    Jul 21, 2017 at 0:09
  • $\begingroup$ The KK4OBI graph looks exactly like what would happen if you mounted your model one or two wavelengths above a perfect ground plane. Try it and see if you like. Is your VSWR ok? The pattern is so neat it looks like there might still be a connection problem. $\endgroup$
    – tomnexus
    Jul 23, 2017 at 5:20
  • $\begingroup$ I tried what you suggested, putting in a ground plane. It gave me a pattern like the KK4OBI pattern above. Then I took the ground plane out of the KK4OBI file, and got a pattern like mine. So my intuition seems to be wrong, and I guess I have to believe what the program is telling me. The VSWR is reasonable, and varies as expected with the angle of the radials. So it's working! Thanks for your help, -Rod AD0YX $\endgroup$
    – Rodney Price
    Jul 27, 2017 at 0:26

1 Answer 1

Reset to default
3
$\begingroup$

Rod,

Ends of wires are connected if they are at the same location. Your model is correct in this regard. However, the wires need to be about one-quarter wavelength long rather than the full wavelength shown.

The 4NEC2 software provides the Automatic Gain Test (AGT) check box when you Generate (F7) the results of your model. Use this in the early stages of modeling. If the Gain correction is too large the results will be in red indicating the model needs to be corrected.

Height for a model over ground needs to be one-half wavelength for optimum conditions. After your model is satisfactory, adjust height according to your real conditions and see what changes. Your model at one-tenth wavelength will show most of the radiation being reflected upward. Use the free space option if you wish to eliminate ground interactions.

The radial angle downward will need to be 40-45 degrees to get a minimum SWR around 50 Ohms. The 30 degree setting you show will be an impedance under 50 Ohms. Since you use Symbols, you can use the Optimizer function to find the best (L)lengths. Similarly, you can optimize for best angle if you use Sin()*L and Cos()*L of the element Length.

Here is the 4NEC2 model information about elevated radials.
http://www.qsl.net/kk4obi/Elevated%20Radials.html

Your model should be able to reproduce the results of the 4-radial model included in the study. (That model coding is not directly comparable because it uses spherical geometry and adjustable center-fed/off-center feed at any angle for all elements).

The Load card (LD) is an unnecessary complication at this stage of development.

Improve this answer
$\endgroup$
4
  • $\begingroup$ Doesn't GN -1 mean No ground? That's appropriate for a VHF simulation. $\endgroup$
    – tomnexus
    Jul 20, 2017 at 18:56
  • $\begingroup$ @DickReid, thanks for your thorough answer. This antenna is going on a high-altitude balloon, so I'm not worried about the ground. I am a bit worried about metal bits of other things near the antenna, and I'm curious about E-field levels near the antenna, so I thought I'd try out the antenna modeler. (I'll add those bits in once I trust the results of the model without them.) $\endgroup$
    – Rodney Price
    Jul 21, 2017 at 0:14
  • $\begingroup$ @DickReid, I had a look at your NEC file, and I'm afraid I'm still stumped after several hours. Any further suggestions? Thanks. $\endgroup$
    – Rodney Price
    Jul 23, 2017 at 2:44
  • $\begingroup$ Turns out the file is correct - I just didn't believe what it was telling me. Thanks for your help. -Rod AD0YX $\endgroup$
    – Rodney Price
    Jul 27, 2017 at 0:28

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .