Tyler and David,
Let me see if I can come up with a clearer example.
Linear to Linear
H to H = -50 dBm
V to V = -50 dBm
The V and H are orthogonal, so > 20 dB apart. This means that H only sees H and V only sees V.
Linear to Slant
H to -45° = -53 dBm
H to +45° = -53 dBm
V to -45° = -53 dBm
V to +45° = - 53 dBm
In this case, the two signals I receive from the H polarization on the transmitting radio are split between the -45° and +45° polarized antennas. So if looking just at the case of H to -45°, then it is correct that I lose 3 dBm vs linear to linear. However I pick that 3 dB back up from the +45° polarized antenna when the chains are combined.
So the logical question remains, why do slant at all? The answer is that it really is a decision at the time of building the product as to what we think is going to best performance in that individual product. There is no real RF benefit of us doing linear over slant as I’ve shown above. There is also the case of circular polarizations transmitting to linear/slant which I will cover in my white paper.
Thanks for the questions guys, helps me come up with better ways to describe this.