Before talking about cross-polarization and how to deal with it, let's consider how the current lines look from the satellite side on the surface of direct-focus antennas
In Figures А, Б and В, I showed the streamlines with purple and green lines when reflecting waves of vertical and horizontal polarization, respectively.
Figure A shows current lines reflected from a "flat" mirror with F/D > 3
Figure Б shows the current lines reflected from a long-focus mirror with F/D > 0.25
Figure В shows current lines reflected from a short focus mirror with F/D <= 0.25
This figure explains why "small" mirrors with a large F/D ratio have less cross-polarization noise than "deep" or short-focus reflectors, and the following figure explains why, with equal diameters and F/D ratio, a direct-focus cutout from a paraboloid of revolution "noises" less than the offset, and which of the two offsets with the same aperture, which are cut from the same rotation raboloid, is better:
H cross-polarization noise 1 at the mirror point, which is symmetrical about the vertical axis, is compensated by H cross-polarization noise 2.
V cross-polarization noise 1 at the mirror point, which is symmetrical about the horizontal axis, is compensated by V cross-polarization noise 2.
Due to the symmetry of the direct focus paraboloid of revolution, the mirror does not generate significant cross-polarization noise.
From Figure Б we conclude that the greater the curvature of the mirror (the smaller the F/D ratio of the original paraboloid), the greater the value of the cross-polarization components. The smallest contribution to the cross-polarization is made by the central parts of the paraboloid, and the largest is made by the peripheral ones. And, although interference is compensated on a symmetrical paraboloid, the phenomenon of cross-polarization introduces a negative effect in the form of a decrease in the value of the vector of the reflected useful component. In other words, we can talk about the "deterioration" of the reflectivity of the mirror (again, I emphasize - with a lower F/D ratio of the original paraboloid).
That is why the F/D ratio of the original paraboloid is important to know in order to evaluate the quality of the offset cut. True, the "hot Estonian guys" claim that a "long-focus" (in their understanding) cutting can be made from a paraboloid with any F/D, and in this they are absolutely right: it is enough to deviate the axis of the "cutting" cone with a constant angular opening more strongly from the axis paraboloid, how do we get an offset cut with a given "offset F/D" ratio (taken in double quotes, because I categorically do not accept this term, but use it only for the sake of understanding me by my opponents). But if we made a cutout from an initially long-focus paraboloid of revolution, then in the section we will see a slightly distorted grid of current lines, and, cutting out from a short-focus paraboloid, in order to get a sufficiently large beam length to the aiming point, we will have to go to the periphery, where the current lines are distorted to the maximum
Since the topic is about adjusting offset antennas, I emphasize that the vertical axis of the notch passes through the plane of symmetry of the vertical polarization current lines, and the resulting cross-polarization interference on opposite sides of the blue dash-dotted line within the notch cancel each other out.
For waves of horizontal polarization, symmetry is not observed, and therefore interference from the side of vertical polarization turns out to be uncompensated. They represent a rather powerful coherent interference, the greater the more the frequency bands of transponders of orthogonal polarizations overlap.
On the spectrum, we see that the SNR of transponders of horizontal polarization in the same beams is somewhat less than for vertical polarization:
For the 10886 MHz transponder, both the frequency band and the modulation parameters coincide - here the effect is maximum - the cross-polarization component of the vertical polarization signal reduced the SNR of the horizontal polarization signal by 1 dB!
If the mirror is suspended on an absolutely vertical azimuthal axis, then for all, except for the vertex satellites, the occurrence of cross-polarization interference will be observed, and the more the vertical polarization plane deviates from the vertical, the more.
In the photo, the vertical axis of symmetry of the antenna A is perpendicular to the axes B and D.
The E axis is parallel to the D axis. The C axis is strictly horizontal.
The lower mirror is installed on the basis of minimizing cross-polarization interference, and the upper one - from minimizing effort during installation.
Two important conclusions can be drawn from the consideration of current lines on the mirror surface:
- cross-polarization occurs when the current lines are not parallel to the plane of polarization;
- cross-polarization is the stronger, the more the symmetry of the installation with respect to this plane is broken.
Cross-polarization interference occurs not only on the antenna mirror, but also in the LNB itself:
Due to the manufacturing technology, the probes are often inserted into the waveguide at an angle of 45°.
It is difficult to imagine a more favorable angle for the formation of cross-polarization interference...
In addition, in the waveguide there is a part of the probe B-C, bent under 135°, and the slot D in the waveguide further violates the symmetry of the setup:
As a result, the V probe receives partially, as interference, horizontal signals polarization, and on the H probe, on the contrary, interference from the vertical polarization signal is induced:
They appear on the spectra as "fits" or "humps", which on a smaller scale repeat the shape of the peaks of opposite polarization, which gave rise to them.
After adjustment to the LNB angle of rotation, they can often be compensated by turning the probe from the vertical to a small angle delta, such that
AB*sin(delta)=BC*sin(45-delta)
As a result, the SNR of the vertical polarization transponder increased by 2.6dB, and horizontal - from 13.5 to 15 dB, i.e. by 1.5dB!
Using the dynamic spectrum of signals of two orthogonal polarizations at once, which my IQmonitor program shows in real time,
you can quickly and accurately set the LNB to the polarization vector of the satellite signal.
On the left - the spectrum of the 12174 MHz transponder before tuning, on the right - the same spectrum after 30 seconds, during which the setting was made.
The increase in SNR was 1.5dB, of which 0.4dB was due to a decrease in the cross-polarization interference from a neighboring transponder of orthogonal polarization.
