# Why are most antennas in cellular networks +/- 45° polarized?

I've just been asked a strange question that I cannot find an answer to (even on the internet it seems I can't find any explanation for this) and I ended up wondering why most of the antennas which work as "base stations" are nowadays +/- 45° polarized. While I do understand the meaning of polarization diversity, it is still not clear to me why antennas' producers chose those 2 polarizations instead of horizontal / vertical or other 2 orthogonal angles whatsoever. Moreover, I am sure that the radiators inside an antenna cannot perfectly have a +/- 45° polarization. What happens then if those are "rotated" a little bit (let's say +55° / - 35°)? Is the orthogonality the only important thing, or is a correct orientation (+/- 45°) important too and why is that?

Bonus question: I think that the polarization of mobile phones has nothing to do with the polarization of base station antennas (because of reflections and multi-paths their transmissions' polarizations could be received "rotated" in comparison to how it was sent), is that a correct assumption?

EDIT: I can't really remember the source, but I read somewhere that we don't use H-Pol, because the ground greatly attenuates the field in that case.

• As for polarization and reflection: Yes, multi-path reflections blur linear polarization (circular polarization is more stable, and is sometimes used to allow two channels on the same carrier frequency, but as far as I know mainly for systems in the direct line of sight as microwave point-to-point transmission). Jul 16, 2015 at 8:02
• Interesting doc, but I can't quite follow the conclusions. kathrein-scala.com/tech_bulletins/DualPolarized.pdf Seems to suggest that you wouldn't want to broadcast horizontally, but with 45/45 antennas, you can broadcast on both. Jul 16, 2015 at 8:31
• @SebastianRiese Yes in fact C-Pol are used for satellite comms. Jul 16, 2015 at 8:57
• @BowlOfRed I read the document and am absolutely on the same wavelength (lol, we are after all talking about e.m. here), but the point I don't understand... What would really change when pol are rotated slightly (let's say 5°-10°)? Would an attenuation in one pol. be compensated by a better reception in the other? What about Co to Cross pol levels? Jul 16, 2015 at 8:57
• I know that MIMO antenna setups estimate the mixing matrix of the channels and then use this to separate the signals, this should also work for slightly offset polarization. Jul 16, 2015 at 9:13