Pls help me with my problem
Have link 27 km There are RB 912 5 + Rocket DIsh from both sides. MCS set manual because otherwise work unstable. So than set MCS 12 ccq link is 100/100 and works very good but cannot receive more that 120 Mb. But when set MCS 13 or MCS 14 ccq link start 99/98 or 97/96 but sometimes go down to 37/70 or less - so have problems with pings and speed. When i reboot the AP and everythinh work. But the situation with foll down ccq is not good and impossible to use MCS 13 or more.
Try selecting MCS12,13,14 at the same time.
If the error rate is low is it will try to stay at MCS14 and if the error rate gets high instead of low ccq it will drop to MCS12 but not lower, because you fixed it to those three.
Look at the rx signal levels at both end. The higher MCS rates the higher signal or SN they need to work stable. Maybe you just don’t have enough signal to noise ratio to get a stable MCS 13 or 14 rate…
Compare the real signals received for each level with the sensitivity needed for each of these mcs levels in the radio.
If there is also some interference source nearby you even need more signal to get better stable link…
Ok. MCS12 is 16-QAM, where MCS13 & 14 are 64-QAM. To have 13 & 14 working stable you are looking to achieve signal levels in the range of -35 to -45. With an average general background noise level of -105 you are looking for 60 to 70dB S/N.
Actually I have a 9km link with two 30dBi Jirious H+V pol. antennas on both ends and rb911G at both ends (same board basically as your 912).
I have the same issue as you; MSC12 is nice stable with a connection rate of 180Mbps (40Mhz channel and SGI) and a CCQ of close to 100%.
The moment I force both to work with MSC13 or 14 the CCQ drops considerably and the net result is I can ‘pump’ less traffic over that link (and have more unstable ping time) than in MSC12. So I set 12 fixed and a fall back at MCS10.
My tx/rx signals are -51/-49 seen from AP-bridge unit and only the tx has a fixed CCQ of 100%. The rx which in fact has slightly better signal is continuously running between 97 to 100. (This is actually the ‘upload’ stream, which usually has less traffic and links with low traffic are always prone to lower CCQ levels.
Ping times are pretty stable around 3ms with an occasional max of 10ms.
So, my link is much shorter than yours. But although my frequency is relative free, there are other frequencies in use just outside the band I set for these units.
I don’t know about your link, but I’ll guess you have a much more free spectrum that would explain why your link performs as good (or as bad) as mine while you have to bridge a 3 times larger distance… I don’t think you are doing so bad after all!
Like me, if you are really thinking of getting a higher throughput, or a more stable link, I think you might start considering to replace your boards for the new ac boards. I am even thinking of converting it in a triple chain link. This link for me has to start transporting way over 100Mbps in the near future which presently is not achievable.
You have to realize that given speeds (and link stability or latency) are a product of the distance, antennas, radio output and receivers sensibility, free fresnel and state of the used spectrum. Your issue is the sheer distance, mine is the noise level and interferences from other links.
One other think you can do to improve the link is to get yourself bigger antennas. Rocket dish come in 30 and 32dBi models. And the new 30dBi models of MT also look very promising.
No, although you don’t want it any stronger. But I have several (client) units connected even at -30 and one or two even at -25. They still work fine…
But yeah, I always try to get somewhere between -45 and -60.
Sot that’s the highest you can get in present configuration. If you want more you do have to change some of the variable you can change.
There is not a lot, power output of radio, gain of antenna (both on both ends), other frequency. And make sure fresnel is free.
One other option I can think of is creating a 3 chain (3 spatial stream) antenna setup and see if you can still get 16-QAM to work, it will give you at least higher throughput (theoretically).
i swapped 912 board out for 411gl with dbii card and fixed my problems. the 912 board has problems with timing i believe. i had to remove all 6 of them i had installed and havent looked back since.
“I believe”? Explain yourself better. Why do think timing is an issue? What does MT say about that? Is this just a problem on some 912 series boards or does it effect the whole production line?
I have closer link with SXT+ additional antennas - 10 km.
So I decided to tested this one and change manually MCS - so now works at MSC 14 stable ccq is 100/100 - 6 days.
After that all want to say that possible the problem are in RB 912 - they are not improved
It just seemed like there was something wrong with NV2 to the point it would have excellent ccq until you passed traffic thru it and my gut feeling was the timing of TMDA was off to the point it couldn’t sync up well. No scientific data but lots of testing and history behind mikrotik gear - never seen anything like it. Same exact dish pair using other boards fixed the issue every time.
I use to run from 2 years the mcs fixed at 10 and 12 on 40 km link. 100% ccq. 120 Mb peack.
I want to change it with AC board, but 6.19 seems changed the way to set the fixed mcs rate.
Hello. Pick up my topic. Have another ptp link. Distance is 17 km. Use the same rocket dish + Rb 912 uag. This link is shorter than my first (27 km) . So have the same problem. Only great work st MSc 12 ccq is 100/100 . When pick up to MSC 13 or MSC 14 ccq link bavome not stable and in fact have less trought ( only about 80 mb)
So in what is the problem??
