..on the lookout for a ballpark figure of SNR I should be aiming for in an effort to optimize a few test MT PtP links I’m in the process of deploying.
Currently the AP comprises RBSXTsq5ac 16dBi, 23º and clients RBSXT5nDr2 16dBi, 25º using a 20Mhz channel with SNR standing at ~50dB* with the Rx rate fluctuating a little between MCS HT13 115Mbps to MCS HT15 144Mbps on 2 streams, short guard interval selected. Chains 0 and 1 both appear somewhat close to -67/-70dBm, giving a combined Strength of -65dBm on a noise floor of -115dBm, with CCQ standing mostly between 75-90%. The links push ~80-90mbps TCP data with success, however to account for worst case scenario I’m considering replacing the aging RBSXT5nDr2 16dBi, 25º with RBDisc5ac 21dBi, 12º in an effort to tryout the increased antenna gain (+5dBi) to further stabilise the links towards the highest MCSs and possibly aim for 173.3mbps on MCS VHT8 (256-QAM 3/4) supported by the 802.11ac protocol.
On a sidenote this being an urban deployment 40/80Mhz channels are out of the question I’m afraid for these specific links as the number of other AP / interference / somewhat obstructed freshnel zones takes this out of the equation. I’d surely like to at least try VHT9 on 40Mhz for a theoretical speed of 400mbps, but I guess we can’t have everything.
- Noise floor is the somewhat arbitrary -115dBm, which is the default for most of these devices. So I’d take the 50dBm mentioned above with a grain of salt.
update: having used alt. vendor’s equipment already and measured noise at apprx. -97dBm, the 50dBm figure drops down to a more realistic 32dBm - 
Surely a spectral scan could provide more realistic values, but for the time being this not possible using AC gear due to chipset limitations presumably?