Dipole over perfect ground, directivity discrepancy

[johan]

Hi,

I was playing with openems and put together a simulation of a dipole over a perfect ground plane. The dipole is placed 0.5 wavelength above the ground plane, which should result in a maximum directivity of about 8 dBi. However, the result I'm getting is 5.3 dBi.
I do not understand this discrepancy, am I missing something in the simulation?

/Johan

[jonsen]

Why are you expecting 8 dBi? I'm new to this but that seems high for a dipole.

[johan]

8 to 9 dBi seem to be the generally accepted gain of a 1/4 wave dipole at 0.5 to 1 wavelength above perfect ground. A quick NEC simulation shows:

Screenshot_20220102_101423.png (24.87 KiB) Viewed 1971 times

Now I don't expect this simulation to reach that high, because NEC is using an infinite perfect ground. The PEC in this simulation will only be as big as the simulation is. But I expected the result to come closer than it did, and extending the simulation in X and Y to 10 wavelengths did not give a higher directivity, so I don't think it is the limited ground size that is the problem here.

One thing I noticed is that the value I get is almost exactly 3 dB less than expected (5.34 dBi, compared to NECs 8.38 dBi).
I know that when using a PEC mirror in CalcNF2FF I have to manually divide Prad by 2, because CalcNF2FF does not internally compensate for the apparent doubling of the power the PEC mirror is causing. Could this also cause the directivity to be normalized to a value 3 dB too high?

Removing the PEC and simulating the same model in free space result in a maximum directivity of 2.18 dBi, which is close to the theoretical 2.15 dBi.

[johan]

I modified the simulation to use a perfectly conducting sheet instead of the -Z PEC mirror. This results in a directivity of 8.2, very close to what NEC shows for the same antenna. So it doesn't look like the NF2FF calculation take the PEC into account when it calculates the directivity.
Is this intentional?

Screenshot_20220106_151934.png (29.5 KiB) Viewed 1830 times