SXTsq 5ac as a PtP link - looking for suggestions for improvements

Hi,
First post so please be gentle.
Based in the UK.
I am setting up a short Point-to-Point link between two buildings which are about 75 meters away from each other.
Initially, I was trying to do this ‘network bridge’ using a 60GHZ pair of nRAYG-60adpair but this was problematic as the link was dropping out every few minutes - most likely because of the fact that one of the aerials is mounted indoors and is connecting to the other arial through a double glazed glass window. Because of that, I have moved onto 5GHZ and a pair of SXTsq ac 5

Building 1 - Master - SXTsq ac 5 mounted on a bracket on the external wall, about 8 meters off the ground level, aligned towards the building 2/slave
Building 2 - Slave - SXTsq ac 5 mounted indoor, behind a glass window, about 9 meters off the ground, aligned towards the building 1/master
A clear line of sight between the two aerials (except the glass window on one end)
Currently, both aerials are showing 2-3 bars of signal strength. Master showing 3 bars of 5, Slave showing 2 bars of 5, sometimes 3rd LED with flash up.
Here are the screenshots of the setup :

Master view:

Running a speed test from Master :

Running Ping :

Slave view:

My problems :
Tx/Rx between the two aerials is not great.
When I had it on the bench while setting it up - the link was about 600Mbps both ways.
Just now it’s 10% of that and fluctuating from 6Mbps to 100Mbps.
I was hoping to see at least about 200-300Mbps of solid throughput at this distance and at this location.
When trying to set up the wireless profile with 80Mhz to make the whole connection ‘wider’ - the link gets established for a few seconds and then drops off. This is why I had to lower the width of the channel to 40 MHz.

Would you be able to give me any suggestions on how to improve this ‘network bridge’?
I’m a little bit overwhelmed with the number of options and settings on Mikrotik and need some help to fine-tune it.

At first glance the signal is much too low for that distance.

I would set the frequency to 5500 and start with a 20MHz channel width.

The signal on the chains should be mid- 60s to be useful, and at that distance I would expect better.

Have you tested at the distance but not behind glass ? even temporarily, if the glass has any metal coating on it then that is probably the issue.

Hopefully as you are in the UK, you bought it from us, we are happy to provide advice, just email or call.

Nick
LinITX.com

Did you have a double glazed glass on the bench between the two devices? :open_mouth:
If not, speed is not comparable, let’s say that you lose 1 Mbps because of the 75 m distance and 499 because of the glass.

Can you try (briefly) what do you have with the window open?

Are you sure-sure that you have no way to put the now indoor device on the window sill?
Like (only as an example) a flat ethernet cable?
CAT6/PoE suitable ones do exist (even if they are not exactly cheap) example:
https://www.waveform.com/products/ethernet-window-entry-cable

In any case, rather than screenshots, post your configurations, following this:
http://forum.mikrotik.com/t/forum-rules/173010/1

Thanks for the replies.

Yes, I bought it in the UK from LinITX.com on 27/09/24.

In case of testing it outside - unfortunately, the office in building 2 where the glass ‘issue’ is is on the 1st floor and the massive double-glazed window panes are fixed into the position and cannot be opened. I have no way of putting the aerial outside :frowning: Access from outside is also not possible as I would need a scaffold tower or a cherry picker to mount it.

I will carry on with testing this on different frequencies and 20Mhz for a start. Will keep you posted.

Here’s the edited config from the Master aerial as per the instructions.

# oct/10/2024 15:04:21 by RouterOS 6.49.17
# software id = edited
#
# model = RBSXTsqG-5acD
# serial number = edited
/interface bridge
add name=bridge1
/interface wireless
set [ find default-name=wlan1 ] band=5ghz-onlyac channel-width=20/40mhz-XX \
    country="united kingdom" disabled=no frequency=5640 mode=bridge \
    nv2-preshared-key=edited nv2-security=enabled ssid=edited \
    wireless-protocol=nv2
/interface wireless security-profiles
set [ find default=yes ] supplicant-identity=MikroTik
/ip hotspot profile
set [ find default=yes ] html-directory=hotspot
/interface bridge port
add bridge=bridge1 interface=wlan1
add bridge=bridge1 interface=ether1
/ip address
add address=192.168.10.196/24 interface=bridge1 network=192.168.10.0
/ip dns
set servers=192.168.10.1
/ip route
add distance=1 gateway=192.168.10.1
/system clock
set time-zone-name=Europe/London
/system ntp client
set enabled=yes server-dns-names=uk.pool.ntp.org

If one of you could check this config and tell me if its OK as it is for me to start testing other frequencies/channel widths?

Contacted LinITX.com and they helped me tweak the settings on both APs.
Now I am getting solid 400-500Mbps link between the aerials and I am very happy with the bridge!

And, unless they are a trade secrets, can you share these tweaks? (I am assuming that they are “generic good practice” that may be useful to some other forum user)

First of all, using a camera tripod - I have tried multiple different locations for the aerial in the office, the one behind the glass.
There, even 10-15 cm of a difference was making 5db of a change in signal strength.
Once I’ve identified the best location for the aerial I followed the suggestions for the settings :

  • setting your client/slave SXT to 20/40/80MHz XXXX & Frequency: Auto so it relies on the AP which channel selection it does.
  • setting On your AP, set to 20/40/80MHz Ceee on 5500 or 5600 Frequency- I went with 5500 as it had much better stats than 5600

in total, it went from -80db to -65db
The sync speed went from 100Mbps to 580Mbps
Now I have a solid, uninterrupted connection between aerials.
The ping average between aerials is now 5-6ms

Hope this helps.

Very interesting, thanks.
All the tweaks you did make a lot of sense to me, what I find surprising Is the relevance of the positioning.
5 db difference in 10-15 cm displacement seems a lot, I wonder what the reason could be for such a dramatic change at such a short distance, I believed that there wasn’t the need of such an accurate positioning on short links, maybe it is some effect of the glass?

Interference (in this case a constructive one) can explain big differences in signal strengths. And at 5.5GHz, the mentioned displacement (10-15cm) is “many wavelengths” (which are characteristical for distance between “interference bands”).

I know that @OP wrote that there’s LOS between antennas. And antennas have nominally something like 25° beam width. But: antennas have quite strong side-lobes at 45° off main direction (+6.2dBi compared to +15.8dBi of main lobe) and it’s possible to have some reflections of side lobes inside Building 2 … which can interfere (either constructively or destructively) with main lobe (which passes between both antennas directly, without any reflections).
I can imagine the window frame to act as a reflector … and if SXT antenna sits in the center of window around 1/2 of window width/height inwards, that would create 45° direction (give or take) from antenna to window frame. Fresnel zone diameter for 5.5GHz link over 75m distance is 1m (1.4m for 150m link) and of the window size is anything near 1m, then amount of RF energy, hitting window frame, is quite considerable.

And I guess there are other features between both antennas, which could reflect the 45° side-lobe. But I can hardly imagibe such a large effect.

Glass could interfere as well as it’s not entirely “transparent” … but I can’t imagine positive effect as easily as the frame reflectins described above.