Seems there is a vendor who at least has an idea where our problems live:

>Welcome to Multi-Lane RF Technology...
>
>We put our own synthesizer in front that is tied to the Wifi radio and downconvert the carrier to a 374MHz intermediate >frequency (the same IF the old PRISM 2.5 chipset used). By doing this, we have significantly increased the fractional >bandwidth of the carrier to allow near brickwall filtering. After filtering, we upconvert the signal back in to the radio.
>
>The results in the lab have been pretty exciting. With an added 30dB+ of adjacent channel rejection through this active >filtering design, we are able to get 2-3x the performance in crowded environments

Background is the selfinterference problem every WISP fight his battles against as soon as his installations
grow.

Here are the basic numbers for 802.11ac. The minimal requirements for a 802.11a receiver are the same:

http://cdn.rohde-schwarz.com/dl_downloa ... nology.pdf

Search for: Adjacent and Nonadjacent Channel Rejection

I see no numbers regarding adjacent and nonadjacent Channel Rejection on the Datasheets from MT.
So I guess the numbers are like the minimal requirements?

This is a very important parameter which influences the quality of our service a lot. So I would like to
know it and I would like to select the cards in my APs which deliver the best rejection values.