Reflector SXT live test

Define what we want to achieve maximum throughput based on the capabilities of the equipment, but they do not perform satisfactorily.
As used in this case, equipment SXT-5Hn firm MikroTik, without the use of our proposed grating reflector “Reflector-SXT” able to work steadily at a distance of up to 2 km (meaning the use of the RB / SXT on both sides of the radio link). Just have to remember to line of sight and noise of the ether as the one and on the other hand, that can greatly affect performance and operation as opposed to ideal conditions, even under similar terms. There have been attempts to organize a radio link at a distance of 4km. As a result, received signal levels - (78-80) dBm at one polarization, while the other polarization does not even start.
Further tests were conducted at a distance of 6.5km from the use of “Reflector-SXT”.
The first thing you should realize that at speeds of information to which we aspire, iron is built on the processors Atheros (MIPS) should not be used for routing and speed restrictions in the construction of trunk radio links due to low productivity. The only exception is the model RB/800 processor PowerPC.
The necessary adjustment and optimization. Tests were conducted on MikroTik OS version 5.4.
The devices must operate in the bridge. To do this, the radio interface must be configured in a mode of WDS. We draw on the side of the base configuration:

  • Create the bridge interface

  • Add a network card interface

Proceed to setting up a radio interface

Now we perform the settings on the side of the station.

Finish setup. After that wound up on the link speed 300Mbit/300Mbit.
Signal levels -57/-57dBm
Throughput testing tool bandwidth-test should be done not from the base to the station and into the aisle. Otherwise, you will see a rate less than 40Mbit / s - will affect the performance of processors. We ran three tests simultaneously with the PC-based router from the base of Q6600 to three routerbor located at the client station.

looks good, but what speed did you get :slight_smile: ?

link speed 300Mbit/300Mbit.
Signal levels -57/-57dBm

I was asking about throughput. 300Mbit is datarate. Your ethernet port is only 100Mbit. What about bandwidth test with UDP/TCP

I’ve made a link with two SXT over about 5.5 km.
RX level on both sides is about -74dB.
No reflector dished had been used.

Data Rates have been manually limited to MCS10 80 Mbps using 40MHz, two chains and a long guard interval.
I reach 30/30 Mbps TX/RX simultanously with TCP Bandwidth test using the connected RBs of each site.
Testing in one direction only, I can reach 60Mbps TCP througput.

MCS11 with 108Mbps worked, too. But I prefer stable connections over speed.

Automatic rate selection didn’t work at all.
(Which means that data rates were completely unconfigured)
The SXTs were constantly trying to sync between higher rates, but then falling back to lower rates.
This caused very flaky connections. After setting the data rates manually, I had a very stable link with the above written measures.

The configuration is plain simple. Just the new station-bridge mode, which happily circumvents all the weirdnesses of EoIP or WDS.

So I think, that the report of SergeySavchuk would have been even nicer, if he wouldn’t have used WDS at all.
(If I remember correctly, wireless-N doesn’t work good/at all with WDS)

reply please how do you set MCS10 only ?

just set datarates to ‘configured’ then go to ht.msc tab an select you preferred MCS.

where is long guard interval ?

Some more info from the field.
Distance 30km speed 10 Mbit
At base station sector antenna + 802.1a equipment.
With ReflectorSXT25 + 10dBi

cool