I put together a tutorial for connecting two SXT’s together in bridge mode.
Here are some Screens shots of the final bandwidth tests…
Notice the RX interface actually shows - 239.2 Mbps…that’s the fastest UDP result I’ve ever seen!
TCP Test through SXT - Remember there is only 100 Mb Ethernet Port 
More at the Blog:
http://www.wispforum.net/entry.php?5-How-to-Connect-two-Mikrotik-SXT-5D-s-In-Bridge-Mode-Part-I
I’ve followed your instructions and couldn’t get SXT’s work in Station Mode.
When I configure it this way SXT A and SXT B could comunicate between, but I couldn’t reach Station device from switch.
Even if SXT A was in Station Mode directly on the switch, and SXT B was remote in Bridge Mode I could ping SXT B but not SXT A.
Until I switch SXT A in Station wds Mode. Then it worked.
I was wondering why I need to configure my P2P devices as WDS?
if you have bridge station mode then that was the problem - you can’t bridge station mode.
Try to use station-bridge mode or station-wds. Here are some info that might help you:
http://wiki.mikrotik.com/wiki/Manual:Wireless_Station_Modes
Thank you uldis,
It work now with Station WDS mode but please, correct me if I’m wrong, reading this, for P2P link I should use station-pseudobridge?
Quote: if only one end-user device must be connected to network by means of station device.
Also, is there a way to manualy enter static MAC’s?
Hi,
I want to get pair of SXT for a bridge in a short range. I have to connect 2 buildings in the central area of the city and my questions are?
thanks for your time.. IF i buy this toys I will send a feedback
Station-pseudobridge is intended for use where you are needing to bridge to a non-mikrotik device, it uses what you can think of as “Mac address Nat” to allow your device to connect out but no traffic will be able to start from the outside and pass in.
For proper bridging you use the other methods.
120M is no problem, simply turn down the TX power on both devices as needed.
I need that bridge only for internet sharing.. so there will be more traffic towards them.
Do I need to configure something specific in that case?
120MB TCP? half duplex?
thx for your time
Sorry, I was meaning 120 meters distance between the 2 is fine.
Speed you will get (pending good line of sight and clear wireless channels) will be able to max out the 100Mbps ethernet port, same as the tests from BrinkNetworks
This tutorial is excellent and eliminates alot of the confusion for those of us that aren’t really good at using the CLI.
Note that one SXT is in BRIDGE mode.
The second is in Station Bridge mode.
Using this method you don’t have to use WDS.
Well, this tutorial is ok, clearly says how to config everything step by step. But it happily demonstrates achievable bandwidth using only btest tool (both UDP/TCP). Could someone explain to me please, how this compares with real life applications? I can get same results in my sxt p2p setups using mikrotik btest tool, but I couldn’t reach those speeds while downloading a file from a http server for example.. I’m really struggling to replace my 11a turbo nstreme links with NV2 11n.. The later one performs a lot worse than the first.
Just Great 
downloading a file from a http server
this will just make one TCP connection and will not be able to load the link to it’s maximum. this is a shortcoming of TCP. to give you more speed, more TCP connections are needed, that’s why “download accelerators” were invented, and that’s why peer to peer programs use many connections, or use UDP for high speed transfer.
I’m aware of that. But how would you explain than even with a single TCP connection in that same test you can reach alot more in turbo 11a, or let’s say by using ethernet alone? What is different in wireless link?
Is the same tutorial used for point to point links with other equipment (as 411 or 433)?
the only difference is that some other devices will only have one wireless chain active. the SXT has dual-pol antenna
this is a lab test, i realized this all depends on the environment the units are being installed, factors like Noise, adjust radios etc would have an adverse effect on performance/ throughput. I Got 300Mbps/300Mbps lab test but when installed in the field i could hardly get 10 Mbps though the tx/rx show 270/270 Mbps
Just to add swimming in a swimming pool is different from swimming in the ocean (we have waves currents etc) so you best swimming pool techniques might not apply in the Ocean. So the tutorial should add an NB: This is a lab Test and in the real world tweak around with the frequency, protocols (Not always that nv2 will give best performance), alignment, tx power etc
On Mars it will be faster..
Great tutorial!
Did you test also on some real distance … let say 5 km ?