My test link:
distance: 4kms
tower height: 25mts
line of sight
frequency :5300 mghz (40mhz) upper control channel

using nanostation M5 (output power 27dbm antena gain 16dbi)
Tower A noise level -93 / signal level -65
Tower B noise level -96 / signal level -67
Airmax Enabled
max transfer
total transfer 57mbits (24 mbits upload / 33 mbits download)

now using sxt 5hnd (output power 26dbm antena gain 16dbi)
Tower A noise level -112 / signal level -66
Tower B noise level -113 / signal level -67
protocol nv2
total transfer 143mbits (65 mbits upload / 78 mbits download)

sxt wins by more than twice the bandwidth using the same frequency in the same towers

good work mikrotik!

was turn on aggregation of packets on NSM5?

was turn on aggregation of packets on NSM5?

yes, agregation packets on and defaults values

i try version 5.2.1, 5.3, 5.3.2, 5.3.3rc2, 5.5b2 all versions, many configurations, auto data rate, fixed data rate, auto ack, fixed ack, with and without airmax, ptp no ack, etc, when link is up airmax quality is 70-80% , when traffic is excessive, as the test of bandwidth, airmax drop to 30% and datarates drop to 45.5 /81mbps,

on sxt negotiate 243/243 and remains the same all the time tx/rxccq at 85/90%.

just remove nanostations and put sxt, same tower, same ftp cable, same frequency, but very different performance

I have links with nsm5 2 or 3 kms 120mbits running test, but in this scenario did not work, however the sxt works perfectly

nsm5 try using more than 1 month for all possibilities to improve it, when you install the sxt in 15 minutes I had everything perfect
(google translator)

How did you make the bw test ?

If you have initiated it from the radios, the test might give you wrong results due to the lack of CPU power on NSM5s.

While I don't doubt your results, the bandwidth test methodology plays a big BIG part is testing real-world capabilities.

My company has had to replace 6 Mikrotik RouterOS backhaul links this year, sure at 3 AM I could push >100 Mb/s through the links, but during daytime they were all over the place, and disconnects were getting out of control. Some of the links have been in operation since the SR71-5 was released. Speed is nothing if the reliability isn't there. We tried fixing the links, replacing Routerboards, radios, even antennas and LMR400 cabling, but the results were always the same, RouterOS is really bad for dealing with interference even when the amount of interference appears insignificant using spectrum analyzers.

We replaced them with Rocket M5s, and yes just like your test I have yet to be able to hit the same speeds I was getting on Mikrotik (I had one link doing >200 Mb/s). BUT and the big BUT is the links uptimes are now months instead of hours, and during the day time the speeds and response times are consistent.

I really like RouterOS the features the tweakability, and the amount of control is unsurpassed. And it hurts since I've got over a thousand RouterOS based units deployed, but I need more then just the capability of speed.

Cheers

testing using rb450G, one in each tower, udp 1500 bytes.

but it is not only the performance of the test, I had to change that link because it had no performance with real traffic, generated latency when was smuggling more than 30mbs, now with 45mbs to 50mbs sxt traffic without latency, and I know I can increase the traffic on the future.

I repeat, I have other links with Ubiquiti running ok, but this site probe all the settings and the solution was sxt

I wisp for over 5 years, with extensive experience in mikrotik, more than 15 towers, 10 links 15 to 40 kms and also with Ubiquiti, over 1000 CPEs, I wanted to bring my experience for all to take advantage, but when Ubiquiti remains unanswered in this case mikrotik has almost the same cost a much more robust radio, the solution
(google traductor)

MyThoughts
here only compare low cost equipment, probe no longer links (30kms) with rocketM5, I have everything with mikrotik, but if I suffer from disconnections once or twice a week, I'll make test with rokets, thanks for the contribution

max rate 300mpbs? or 150mpbs on NM5?

frequency :5300 mghz (40mhz) upper control channel

What do you think?