New routers with better wall penetration coverage?

[dlandiss]

In this case i would consider using a pair of cheepo 2.4Ghz yagi's from DX

Keep in mind that antennas with gain are directional, and the more gain they have the more directional they are. So yagis, parabolic dishes, etc can work over astounding distances but at the expense of beamwidth.

 

[Truls]

Keep in mind that antennas with gain are directional, and the more gain they have the more directional they are. So yagis, parabolic dishes, etc can work over astounding distances but at the expense of beamwidth.

Yes, but the ac68 has 3 antennas, so use one yagi to cover the kitchen and one yagi to cover the dining room and use the one dipole antenna as general coverage. And those low gain yagis still have good beamwith.

 

[dlandiss]

Yes, but the ac68 has 3 antennas

That suggests another question: which radios in the AC68 use which antennas and when?

 

[sfx2000]

That suggests another question: which radios in the AC68 use which antennas and when?

All of them... they're shared after the WiFi radios, so both 2.4GHz and 5GHz use the same set...

 

[dlandiss]

OK. I got the impression that they were fed in different combinations and phase angles to optimize the signal strength to active clients.

 

[sfx2000]

OK. I got the impression that they were fed in different combinations and phase angles to optimize the signal strength to active clients

Beamforming happens in the baseband, not in the RF front-end... there is a dependency on the antennas being matched as a result, same with spatial multiplexing - so always keep the antenna's matched...

 

[dlandiss]

Beamforming happens in the baseband, not in the RF front-end

Thank you. Probably easier to do it there where everything is still digital instead of dealing with multi-GHz RF.

So given a hypothetical situation of three very-directional antennas, pointed in different directions, and only an active client in one direction; would the other two radios simply not transmit that particular data stream?

 

[sfx2000]

Thank you. Probably easier to do it there where everything is still digital instead of dealing with multi-GHz RF.

So given a hypothetical situation of three very-directional antennas, pointed in different directions, and only an active client in one direction; would the other two radios simply not transmit that particular data stream?

I wouldn't recommend it, as the baseband is taking sounding measurements, at least for 11ac and 11n radios, so the three streams are combined at some point, whether implicit based on feedback from the radio, or explicit based on clients...

If one wants the best throughput based on a single AP, keep the antenna's balanced and omni...

It's different than old-school 11g/11a...

 

[dlandiss]

I wouldn't recommend it, as the baseband is taking sounding measurements, at least for 11ac and 11n radios, so the three streams are combined at some point, whether implicit based on feedback from the radio, or explicit based on clients...

If one wants the best throughput based on a single AP, keep the antenna's balanced and omni...

It's different than old-school 11g/11a...

Thank you. I really appreciate the new knowledge of engineering factors that influence the router's behavior.