I am rebuilding old house and preparing internet installation to every room using Omnia WIFI 6. Besides ethernet sockets, I would like to prepare good WIFI in every room. The house has thick walls and WIFI signal does not penetrate well. So I used coax cables with SMA connectors and extended external WIFI Antennas from technical room directly to the rooms where it ends with small antenna. I do not need those 2.4/5Ghz multiplexers, so can I remove those multiplexers from omnia replace it with additional separate SMAs? This way I could have 6 SMA ports and could connect it using 6 coax cables to 6 different rooms.
You shouldn’t do that.
I tried something similar (used 4 antennas to two different rooms and attached all to one AW795-NP1), result is no suitable connection at all.
I hence connected instead 2 antennas each to a separate AW7916-NPD and now have perfect signal and throughput in both rooms.
So better foresee 2 coax-cables per room and connect them each room to a dedicated WiFi-card.
Are those 6 rooms on one floor? If no - why not add one access point per floor? Unfortunately TO can only handle 3 WiFi cards and thus (in my scenario above) 3 rooms. I decided to dedicate 2 rooms per floor with excellent 5GHz (or maybe later when having more clients 6GHz) and one central 2,4GHz WiFi-card. So if I need throughput, I just walk to one of those rooms, in other spots the 2,4GHz-network delivers the basis connectivity (for WiFi-calls, music streaming, IoT-devices and alike).
This is really not good idea, because losses occur in the cable and the longer cable is, the greater are the losses. The rule of thumb here is the shorter the cable, the better.
If the walls are really really thick, you should use some AP connected to the Ethernet socket for each room. I believe, that many known manufacturers offer central control over them from one place.
1 Like
apart from the cable loss @michalko58 mentioned, this is not how the modern wifi-cards work.
they use all connected antennas at once and try to calculate the signal run time and strength for each antenna seperately to combine it into the best overall signal. if you place the antennas too far apart, or in seperate rooms even, you completely disable the card´s ability to do that. the result will be a very weak signal, or, most likely, it´s not working at all.
so, all antennas of one wifi-card should be relatively close to each other. two rooms, two cards is possible, but then again, the cable loss. seperate APs connected via LAN is optimum, but wifi “mesh” should also possible, depending on your hardware.
This is how MIMO works (or does not if the antennas are far).
But this is definitely not how MIMO works. The result, as I get it, should be a “1x1 MIMO” stream in each room. Which should still be able to provide strong signal and high speeds (I guess at least 300 Mbps should be doable on MIMO stream).
Also, using iwctl or similar tools, I think you should be able to completely disable MIMO stream grouping, so the card would not even try to group them.
But this is definitely not how MIMO works. The result, as I get it, should be a “1x1 MIMO” stream in each room. Which should still be able to provide strong signal and high speeds (I guess at least 300 Mbps should be doable on MIMO stream).
Due to the signal loss, over X meters of the cable, you will be glad if you have at least some signal 
That’s possible. But it is not due to the fact that the antennas cannot “hear” each other, which is what @n8mahr suggested.
To fight the signal loss, you could try increasing TX power. But it probably won’t help in the receive direction. I know that for GNSS antennas, there are some with pre-amps integrated directly into the antenna, which basically cancel the RX cable loss (the term used is LNA = Low Noise Amp). Not sure if something like that is possible in the Wifi world, too.
What you described in your first post is something called DSA - distributed antenna system. It is mainly used by mobile carriers, not sure if wifi can support it. The gear required for this is far from cheap. Try to look up DSA
Well - I installed four coax-cables from one AP in the basement to two rooms in the ground floor. Cable length is somewhat 10-15 meters.
If I attach those four coax-cables to one single AW7915-NP1, I get no working connectivity - tested with several mobile devices and laptops. So MIMO does NOT work in that case. Difficulty is (but MIMO should perfectly work around that): two of those coax-cables need to pass 2 x 5*25mm² power cables running through the wall, which will for sure weaken the signal strength.
But if I use 2 AW7916-NPD instead, I get perfectly clear signals below -60/-50dBi when 1m away from the antennas (+3/+5 dBi omnidirectional antennas). Iperf shows for both rooms a throughput of beyond 1Gbit on 5GHz-frequency.
Why did I test that? I planned my house in the beginning from the perspective of maximum throughput and thus chose 2 AW7916-NPD (to switch to 6E later when I have enough mobile devices supporting that). But in the meantime I wanted to reduce power consumption and thought I could get rid of one of the cards but as mentioned above that failed…
2 Likes
Actually it somehow works, I tested it and I reach about -40/-60dB nearbe antennas which are inside brick wall 1 cm under surface (not visible because of esthetic purpose). On the other hand I can reach only 35Mb/s in those rooms. There must be something wrong because even I am nearby Turris Omnia, I can still reach maximum speed about 40 Mb/s while over cable I can reach 50Mb/s which is limit of my VDSL provider.
I will need to start over and create some realiable way of measuring speed in local network. I cannot use speednet because my internet is slow. I expect that I could have optic fibre in 1 or two years (it is already two streets from me), so I would like to prepare for that. Measuring speed of MIMO will be difficult.
I just used Coax 50ohm RG58A-U, 5mm, because this cable is usually used for Wifi antennas. Now as I am going to more details and thinking why it cannot reach that speed, I found that this cable suppress high frequencies (over 2Mhz), so maybe I will rework it with better high frequency coax with really low suppression.
I don’t want to have a house full of visible blinking network devices, so this approach looks still good and promising to me.
Antenna inside the wall is really not a good idea. You should better use some ceiling AP.
Or, you can find out, if there is some ceiling lamp with AP function.