MikroTik

Hello,

I have working setup of poit to point link over 24 kilometers using pair of rb112 + sr5.My setup is bridge with station-wds using nstreme with best-fit and framer-limit 3200.Everything is working fine but there is difference on client side between rx and tx ccq levels.Results from slave:

status: connected-to-ess
band: 5ghz
frequency: 5180MHz
tx-rate: 24Mbps
rx-rate: 54Mbps
signal-strength: -64dBm
tx-signal-strength: -64dBm
noise-floor: -98dBm
signal-to-noise: 34dB
tx-ccq: 77%
rx-ccq: 92%
p-throughput: 23483
wds-link: yes
nstreme: yes
polling: yes
framing-mode: best-fit
framing-limit: 3200
routeros-version: "2.9.38"
802.1x-port-enabled: yes
authentication-type: none
encryption: none
compression: no
current-tx-powers: 6Mbps:16,9Mbps:16,12Mbps:16,18Mbps:16,24Mbps:16,36Mbps:16,48Mbps:14,54Mbps:13
notify-external-fdb: no


rx-ccq is almost over 90% , but tx-ccq vary from 56 to 80 %.So speed in one direction is different.Any suggestions what to check to fix problem?Excuse me for any errors in english.

The one with the lower ccq most lilky as a small amount of interference.

Your link really does look good tho, I'd love those figures

It's just some noise, I'd guess. You could monkey with your pigtails to make sure it's not a wiring problem, but that's pretty much within tolerance as far as I'm concerned, though you might get more throughput if you could get the CCQ numbers up higher.

If you're having throughput problems, try turning off nstream - the RB112 doesn't have a whole lot of CPU and you might be able to get more raw data through the link by turning it off.

Thanks for quick responses.I`m pretty sure in my pigtails cause i have chaned them twice.It sounds more true like it`s an interference cause of position of master.It is placed on a tv tower(best point for this link) where are a lot of antenas on different frequencies and some of them use to work above 1KWatt in spectrum above 13Ghz.I have very strange problems at the beggining when i`ve put antenna cable into N-female of pigtail ping to ruterboard came with losses.I`ve changed hole antenna and the only difference was that new one have a metalic ring.

If you're having throughput problems, try turning off nstream - the RB112 doesn't have a whole lot of CPU and you might be able to get more raw data through the link by turning it off.

There are at least two schools of thought - one says dilution is the solution to pollution, one says concentration ...

We're quite happy with the RIC/522 which is now based on the RB112. You don't mention the gain of the antennas you are using, so it is impossible to calculate your expected signal strengths and consequent operating margin.

However, we've done links with 30 to 40 dB of SOM which needed narrower beam antennas in order to disregard (read: not see) the interference. Well planned point to point links typically do not suffer from interference if the antennas are sufficently narrow in beam width. e.g., 4 degrees of beam width makes it difficult for even an actor to upset your link.

At other times, it seems that in certian cases a wider beam antenna may be helpful for "diluting" the interference, just as is the case with CPE installations.

regards/ldv

One thing we have noticed is that if we have a good signal with little or no noise but we hard set the rate to 6mbps we get a low ccq (under 40%). If we then increase the rate to whatever the best rate the link quality can support we get a much better ccq (between 80% - 100%). Not completely sure why this.