What is the theoretical and actual cross polarization attenuation from horizontal to vertical polarization?

Can anyone tell me what the theoretical and actual cross polarization attenuation from horizontal to vertical polarization is ?

The reason i ask is that i have the idea that for receive a vertical yagi has little gain for horizontal signals, and this idea seems to be backed up by the fact that a vertical 3 element yagi i have built does indeed seem to null out vertical signals much better then another 4 element horizontal yagi that i have. So i'm trying to work out what's going on and knowing the cross-polarization attenuation is a good start.

Theoretically, cross-polarization loss is infinite.

In practice, both the receive and transmit antennas aren't exactly linearly polarized. They will have fields in the cross polarization to some extent due to imperfections in their construction, for example misaligned elements, or less than perfect isolation from common-mode currents on the feedline, mast, or tower.

Furthermore, some paths can alter polarization, for example by reflecting off slanted surfaces. Real environments thus introduce some degree of "leakage" into orthogonal polarizations.

In practice, cross polarization is probably between 15 and 40 dB, depending on the antennas and the environment.

A yagi is a purely linearly polarized antenna.

Since H and E fields are orthogonal to each other, every antenna has a theoretically infinite cross-polarization attenuation.

The actual attenuation depends on factors like material thickness, production accuracy, and the reliability in which the signals are actually polarized in either direction, so there's no general formula that we could tell you for your 3 or 4 element yagi – but as a rule of thumb, if it's a good one, it's going to be high.

Another rule of thumb that I basically just came up with from my tummy: If it has many directors (i.e. non-driven elements), then chances are that if a wave incides even slightly slanted from a theoretically suppressed cross-polarization, the effects of the non-orthogonal part will be easier to measure – simply because it's harder to align four elements exactly than 3.

• "every antenna that has a linear polarization" could be simply "every antenna". Circular polarizations have infinite attenuation in the opposite sense. Mar 2, 2019 at 1:09
• That's true, editing. Mar 2, 2019 at 9:51
• I agree except for the fist sentence. 1) Every antenna has a degree of elliptical polarisation, it's just a question of how small it is. No antenna has infinite cross polarisation attenuation. 2) Cross-pol discrimination is not caused by or related to the fact that they're at right angles. Also in an elliptically polarised wave, which is all practical waves, there is some E and H in every plane you choose. Mar 2, 2019 at 14:22
• @tomnexus I heartily disagree. My first sentence clearly is about theory; and yes, the Hertzian dipole has perfect cross-pol attenuation. Mar 2, 2019 at 14:24
• @tomnexus re 2: well, the point about circular pol is it's linear pol with the polarization rotating along distance, or if you will, a set of two superimposed linearly polarized wavefronts with a phase offset between them; thus: you'd get theoretically infinite cross-circ-pol attenuation exactly due to the fact that you're constantly rectangular to the antenna's polarization. Mar 2, 2019 at 14:28