Long time networking hobbyist. I have setup many PTP links outdoors using ubiquity gear and I did own a few of the first wireless 751U-2HnD mikrotik devices.
I want to setup some PTP links inside of a house. I am currently using Unifi Access Points in a mesh configuration but they are not 100% stable. Have weekly dropouts that last for up to 10 minutes and probably a bunch of smaller drop outs. This is most likely due to a 5 device mesh with only 1 upload AP, therefore all sharing the same channel. Performance is actually adequate for my needs, at around 60 mbit for the furthest client. I’d like to make the setup more robust by adding some PTP backbone links inside the the house for the existing Unifi aps but don’t have line of sight.
I’m debating between using omni directional antenna Mikrotik devices like hAP ac lite TC to make the wireless PTP bridge or directional high gain devices like the SXTsq-5ac. I might be over thinking this but I’m worried that directional antennas might not work like I want them to when the signal has to bounce off of and through walls.
So I’m comparing
hAP ac lite ------ indoor short distance no LOS walls stucco ceilings— hap ac lite
sxtsq5ac ------ indoor short distance no LOS walls stucco ceilings —sxtsq5ac
I have spots to mount the sxtsq5ac out of sight on top of tall cabinets so that isn’t a big factor. I guess in essence will directional antennas help or hurt performance through walls and ceilings in a house? Granted my goal is 60 mbits and stability.
Powerline was used before the Unifi mesh and the Unifi mesh is superior so far. AFIK they have the same problem as mesh because they are operate on the same channel so they can’t speak at the same time. The house also has like 5 electrical subpanels so that can’t help. I have been doing a ping test and getting 99.0% uptime, so I’m trying to improve upon that.
The mesh is working with omni antennas so I know the signal can get through, just wondering if directional will help or make it worse. I know being able to use more than 1 channel will help. The Audience looks nice but its an expensive solution that I don’t want to try at this time since I already have Unifi APs. I will probably end up trying a pair of directional antennas but my distributer was out of stock of the SXTs but not the mounts, now it is the opposite. Covid shortages or demand I guess?
I’ve used a Disc Lite 5GHz dish to connect to the 4x4 MIMO 5GHz radio on an Audience through thick walls in a very old, large house. It is actually two houses with additions to connect them into one large house, with stone and cement outside walls (covered in plaster) now inside the new building envelope, making traditional mesh impossible.
I was able to establish a solid 5.8GHz 2x2, 20MHz wide, ~80Mbps connection from the Audience with the Disc5 Lite in a room where phones and laptops could not stay connected to the Audience on 2.4GHz. I used NV2, not 802.11 for the ptp link. I then wired the Disc5 to a cAP AC2 for normal WiFi distribution on that side of the house, using PoE.
No drops, no issues, works perfect under load. SINR of about 36dB. The Disc5 receives more bandwidth than it can send, due to the Audience’s omni antennas. Speed tests are usually ~15Mbps up and ~50Mbps down at the interface of the Disc5 Lite.
Putting another Disc5 in place of the Audience would obviously give more bandwidth, but the WAN speeds are only about 20u and 60d, so there’s no point.
My personal experience is that Omnidirectional performs better than directional. eg. hAp ac2 and hAP ac Lite, perform better than Omnitik ac, and even better than SXTsq 5 ac. In my indoor situation! Units on different floors, path has to pass several brick walls also.
The only explanation I have is that the signal propagation is totally unknown (and is not the straight path through the walls and brick/concrete ceiling), but passes through several spots by reflection.
The EIRP of all systems being the same (that value is low in Europe/ETSI regulatory domain), the total energy transmitted by the omnidirectional is larger than the energy transmitted by the directional units. So the lower the (physical) antenna gain, the higher the TX power (= EIRP-antenna gain), and the higher the received signal. (But the CCQ is below 90%)
All of this better/worse outcome is unpredictable for any other setup, as it fully depends on the attenuation of the direct path.
Putting another Disc5 in place of the Audience would obviously give more bandwidth
Not my expectation. Antenna gain is reciproque. The gain at reception if often of more value than the gain while transmitting, because the legal EIRP limit will reduce the TX power for high gain antenna.
So a Disc5 with high gain will receive a strong signal from an (omnidirectional) Audience from any direction. Replacing the Audience with a Disc5 will reduce the transmitted power to be available at the max power in a small beam only. The strength in the beam is the same as the omnidirectional strength if setup to a legal value. The “I” in EIRP is causing this. “I= isotropic equivalent”, the EIRP is calculated and legally limited as if the beam was omnidirectional.
In my setup the SXTsq received a stronger signal form a wAP ac than from the Omnitik ac. But the Omnitik ac received a weak but stronger signal than the wAp ac from the SXTsq.
(That experiment is where my “pseudo” theory comes from.)