# Correct number of elements in 4nec2 antenna simulation

I am designing a 1.25-wavelength dipole antenna with the 4nec2 software. What I noticed is that the input impedance of antenna is changing greatly when I change the number of segments.

How many segments should I choose?

• If your dipole is modeled as a single length of wire, which is common for a simple dipole antenna, then you will want an odd number of segments if you are choose a center feed via the excitation card. Nov 5, 2015 at 21:13

Generally in NEC2, 10 segments per wavelength is good, so 12 or 13 segments depending on where you want your feed.

But you are doing better than just following the rule, investigating the effect of the number of segments!

I'd suggest trying 4, 6, 8, 10, 15, 20, 30 segments per wavelength and comparing the results. The impedance should be pretty constant from 10 to 30. If not,

Watch out for all the other rules too. Segment length to diameter - they must not be too fat; Length to length ratios - joining long and short segments; radius step changes, joining at too narrow an angle so they overlap too much; small loops are difficult in NEC, etc etc.

• Also, the excitation point (driving the antenna) is usually chosen by naming the segment number on a wire of your geometry. To properly place the excitation point where you want it, you need to be mindful the number of segments and lengths of segments. NEC2 uses the middle of the segment number specified to insert the excitation. If this is not properly done then the resulting impedance can change in ways that you may not expect. Even though you may be using 4nec2, you should be familiar with the geometry cards and excitation as described in the NEC2 user guide. Nov 5, 2015 at 18:17

As I understand it, up to the point where the elements are too "fat", as mentioned previously (around 3 diameters in length, which the NEC2 complains about if you go under), the more elements you have the more accurate the calculation will be. The tradeoff is in computation time vs. accuracy, and as you increase elements the computation time will increase faster than the accuracy will. Once you start getting in to 1/10ths of a dB or ohms impedance you might be better off stopping there or going with fewer elements.

Back when computers were large, slow and expensive, getting the right tradeoff was much more important than it is now. The code runs very fast on a modern computer, although if you're doing iterative searches for an optimal solution it still might be useful to think about using only as many elements as you need.

Choose correct segments number for Electromagnetic waves simulation at NEC2 by checking results convergence of N simulations:

https://github.com/rezamarzban/correct_segments

• Hi Reza and welcome to Ham SE. There are a few of us that have grown up with NEC2; I still use it as it's simple and I trust it. Please will you edit your answer to include all the relevant information here, without a link. Link-only answers can rot and become less useful over time. Dec 30, 2023 at 16:11