What I meant by compare for yourself was:
"Copy" the text in the highlighted Code box . "paste" the text into a new text file (like notepad) and save as "righttest.nec" (or wrongtest.nec ) save/move them to the "4nec2\models" folder.
Open 4nec2, click on the folder icon near the top left.
Select the file you want (righttest.nec or wrongtest.nec)and click open.
Find the calculator icon (or F-7 ) and select far field pattern and generate.
The most obvious difference is 2.79 dBi max gain at the horizon vs 1.49 dBi at 10 degrees.(on a free space model).
Wires should be constructed where "end one" is closest to the feed point.
That's my understanding of it. Though I can't seem to find this rule in the users manual.
I think it was a 3 wire dipole model I used originally to prove it to myself.
24, I don't get the same gain and angle numbers when I model your L antenna using Eznec and I can't explain that. Does 4nec2 produce model integrity like Eznec does with its Average Gain test? Our gain numbers and angle are not even close, and my results for both models are the same regardless if I connect the wires...one way or another.
Said another way, when using Eznec the only different results I see when I fix the end connections for the two wires differently is the currents and phase distribution as noted in the output for the Tabular Currents Log, and the red line indicator for the phase noted for the models in the Antenna View.
Except for the phase and currents differences I see, I was surprised to see your models showing different results in gain and angle when you switched the end connections for your models. Can you also compare the currents between models. I just compare the first segments for the two wires in this case.
I also added a dipole model to my attachment below by making your L model straight like a dipole. That changed the match a bit, but I did not correct the dimensions in order to return the model to resonance like my L model shows. So, the model shows some reactance by moving the element into the vertical plane. However, the model is still pretty close to what theory predicts, and this is why I did this...just to show that my Eznec model is pretty close to what theory predicts for gain and angle in free space for a 1/2 wave dipole. You will also note that I too used a split feed point in order to get closer to the center of the radiator.
I have a feeling that in modeling we should always connect wire ends as follows. This also considers that the actural wire number used here or for example only...and is not significate to this issue. We can connect any wire to any other wire as long as the ends follow what I suggest below. I can be wrong on this however, if someone can explain this better or tell us why this doesn't matter. I tried to talk to Roy about this, but did not get to first base.
Code:
wire 1 end 1 connected to wire 2 end 2
and we should not connect as follows:
Code:
wire 1 end 1 connected to wire 2 end 1
I will add, that when I create radials or a top hat using the create radials feature in Eznec...all the elements look to connect in error to me as noted below:
Code:
wire 1 end 1 connects to wire 2 end 1 and so on with the other wire in the radial set
As a result the models with even number of radials always shows all radials are in phase including opposing radials, and as I understand this aspect of antenna theory, the two opposing radials with currents flowing out toward the open tips should be out of phase with each other...and thus cancel each other out into the far field.
Beyond this I cannot explain further why Eznec does what it does when the wires are connected the opposite way. I also can't explain why we are seeing different results in the gain and angle. By chance did your match also change when you changed the connections?
View attachment GHZ24's wire connection idea.pdf
