we have 4.10

I have a poor link that is only running at 3 mbps with 10 Mhz channel bandwidth.
It is a link with no good line of sight, because many trees and bushes are there.
Is it possible to use 802.11n for that link and getting more speed?

Greetings

what setting do i use on n cards if i use chain as i cant seem to get ccq up i played with tx powe as on defualt i get -20

for 3km link please help with a basic setup that works with 20-30 mb that i can play with

I have a poor link that is only running at 3 mbps with 10 Mhz channel bandwidth.
It is a link with no good line of sight, because many trees and bushes are there.
Is it possible to use 802.11n for that link and getting more speed?

Greetings

unfortunately, no. n-draft usable when everything is almost ideal. and after that nobody knows why if not usable. we fighting for bandwidth test of 100Mbps (FDX) @ 5km link already year - unsuccessfully. -55dBm levels, 140Mbps, but only one directional. when direction=both, worst results done.

can anyone help i have lOS 3km

26dbi grids both sides
433ah on both sides and n radios
i have connected one chain0

what cards
what signal levels

has anyone tried n link bonding ?
is 300 mb real throughput achievable or it's just a 2 more years dream ?

it's still only a dream

it's still only a dream

I have a customer trying to accomplish it right now. He's using some high dollar filters to do it. He's got the link working, but he's still working out the kinks to try to get it above the bandwidth of just one fully maxed out 2x2 40MHz N link.

Hi guys,
that's the result of my tests with 802.11n on Mikrotik hardware, keep reading for an abstract of results and images with settings.

I started using the settings suggested by "exe" on the previous posts and make some changes to find the best tradeoff between stability and performance.

SCENARIO
AP and STATION: RB433AH and R52n radios inside outdoor metal boxes with u.fl-N pigtails
Antennas: Jirous JRC-24 DuplEX
Distance: 4km with perfect LOS

RESULTS (see images below)

Highest UDP speed reached: 180Mbit/s stable and simmetric on downlink or uplink, CPU average load @ 70-75%, average ping @ 15ms


Highest TCP speed reached: 90-95Mbit/s stable and simmetric on downlink or uplink , CPU average load @ 100%, average ping @ 40ms





ROUTEROS SETTINGS (while testing at 150Mbit/s UDP fixed speed)




NOTES
- Using HW-retries of 4, 10 or 15 the results didn't change that much
- I made the tests using RouterOS 4.9, 4.10 and 5.0beta2 on both devices with no change on performance
- Each outdoor box has only one RB433AH, one R52n and 2 pigtails. At first I put the test board together with another RB433AH in the same box and the RF interferences caused the link to work perfectly only on tx side: the rx side was interferred by the other miniPCI radio even though their working frequencies were spaced by 400MHz!

I will soon do more tests using RB750G connected after each RB433AH, so the results will be more effective and similar to a real production scenario.

Bye,
mattia

anybody seen one direction bandwidth tests resulting in 346Mbps UDP and 140Mbps TCP? Thats what I was getting yesterday with my sr71's and was curious if anyone else has seen better results?

it's a kind of magic

no, it isn't a magic. no even impressive result.

let iw4eky show result of bandwidth-test with direction=both

iw4eky, show us result of bandwidth-test with direction=both

Ok, below you can find the results of bandwidth-test using "both" directions!

Highest UDP simmetric speed reached: 66Mbit/s stable on both downlink and uplink, CPU average load @ 30%, average ping @ 40ms


Highest TCP simmetric speed reached: 58Mbit/s stable on both downlink and uplink, CPU average load @ 100%, average ping @ 40ms


Highest UDP asimmetric speed reached: 90Mbit/s down and 35Mbit/s up, both stable on downlink and uplink, CPU average load @ 15%, average ping @ 9ms


Highest TCP asimmetric speed reached: 68Mbit/s down adn 35Mbit/s up, both stable on downlink and uplink, CPU average load @ 100%, average ping @ 34ms



And I'd like you to believe me that I post the results just to be of help to the Mikrotik user community: I don't want to challenge anyone nor I desire any prize! This is my job and I'd like to share the results with you, as I was helped by the work of other users on this forum.

Bye,
mattia

wasn't looking for a prize.....I was a little shocked with my results as they were the best I've seen but I felt that it was still less than adequate for 5ghz-N and I was looking to see if anyone had gotten closer to the 200 Mbps mark than I had. I am curious why everyone uses RB800 or RB433AH's since it seems to be better to have more processor. The best I had "both directions" was 170 Mbps/170 Mbps UDP. TCP seems to jump around a lot but mine was hanging around 80 Mbps/60 Mbps both directions and part of me thinks its just that the bandwidth test doesnt like x86 as i have confirmed with a couple mikrotik distributors that there is a problem with x86 and the btest which i have seen with other tests of my own. I want to help out here too but I want to get the best results or make sure I have the best config so that I don't hand out problems with my setup. Thank you for any help or advice that you can share.

impressive results iw4eky
keep us updated if you get better results..
If you can get a picture of your installation too to see what it does look like would be very nice
BTW, i'm using your same antennas and i recommend them too

yeah!

impressive!

yeah.. glad to see your success I think of moving back to 11a-turbo setup already..

.. I'm think of moving back to 11a-turbo setup already..

+1

OK, I'm having... issues with 802.11n and despite of the reading of this forum I can't cope with configuration.

Here's the hardware I've got:
2 x (RB 433 UAH, ROS 4.10, R52n)
The wireless interface is in WDS, which is added to a mesh. Both antennas of R52n are AirMax AM-5G16, connected on pig tails only. The distance between my 2 test MTs is about 6-7m.
Iface config: Now, problems I'm having:

1) During quite a few of my tests, the wireless link on the 2. layer was up, but there was no layer 3 link, when one of the MTs was in a high-speed mode; when the link dropped to 54M (a), everything worked OK. Unfortunately, I'm not able to reproduce this right now...

2) Sometime the reported link was even 300M/54M, but always only one peer of the connection has .n mode, the other is 54M max. Worse, when I run the speed test, it is something like 9-10Mb/s (tcp, both directions) or 12 (udp, both); even in one direction (the faster one, >> 140Mb) for udp it is 13Mb/s!!!

3) Sometime, despite the same configuration and exactly the same physical position, .n modes just don't set up -- I'm ending with 54/54 or even worse...

Any help would be really appreciated

I really hope the official 5.0 ROS rolls out soon. I can't wait to try this on some of my major backhaul links, just can't risk BETA on them!

better don't risk. to prevent some trouble. we test 5.0beta2 on pc before. but it cannot be installed on pc. lot of troubles to rollback to 4.xx

Just curious. What are the max pps people are seeing with various boards. i.e. 411ah, 433ah, 600, 800. Just want to see how the cpu power and ram impact the ability of these types of connections pump out the packets. I gather that on a solid link one could see 90 to even 100 mb/s tcp using the boards with 10/100 ports. Now I would like to see what the max is packets at various packet sizes. (64 on up to 1500). I personally have seen as high as 50k pps (aggregate) using 64 byte packets on an n link using r52n 411ah.

The bandwith in WDS (ap bridge/wds - ap bridge-wds) mode is much less than in the mode of the ap bridge/station. When will it be fixed? VPLS not offer as WDS need to MESH.

The bandwith in WDS (ap bridge/wds - ap bridge-wds) mode is much less than in the mode of the ap bridge/station. When will it be fixed?

I dont think it can be fixed, you are splitting radio time 50/50 between AP and STATION. If you make it station its 100% dedicated to connecting to AP, if it's AP BRIDGE WDS then its 50% station, 50% AP.

I dont think it can be fixed, you are splitting radio time 50/50 between AP and STATION. If you make it station its 100% dedicated to connecting to AP, if it's AP BRIDGE WDS then its 50% station, 50% AP.

So using MPLS bridge setup should eliminate this problem?

Yes. On nearly the same Hardware. But they do not implement DFS so i am
Not allowed to use it.

what max distance you test
with n-draft m5

What about latency and jitter?
I tested rocket m5, but the latency was high and jitter was crazy..

The bandwith in WDS (ap bridge/wds - ap bridge-wds) mode is much less than in the mode of the ap bridge/station. When will it be fixed?

I dont think it can be fixed, you are splitting radio time 50/50 between AP and STATION. If you make it station its 100% dedicated to connecting to AP, if it's AP BRIDGE WDS then its 50% station, 50% AP.

Thank you for your reply. Well, how do I optimize the network, if I use a dual radio - one module to access and one module for backhaul? MPLS does not help here.

Hi , I am new one im mikrotik. Im using RB433 with N wifi card and 5ghz double antena. I use it only for 1 km and i think i need to attenuation signal. Now Its working like before with single 5ghz antena with G wifi card. I dont find where i can attenuation signal, can somebody help me ?


im using rocketdish dual antena http://wificentrum.sk/dual-polalarizacn ... _d734.html

what max distance you test
with n-draft m5

Its written there on the main page 2.3km.

Its worth noting for people having bandwidth issue during testing that doing bandwidth tests from router to router caps out at around 30mbit for anything below RB433AH. During our testing we found that RB433 plain can only receive 30mbit, Changing to a 433AH resulted in more bandwidth but even than a RB433 will cap out at around 70mbit

Now for a question, Is there a reason for a single chain N link not to go into 40mhz? We've set HT extension channel but it wont go past 65mbit-HT and signal only shows a max of HT20-8 even tho the signal is -65

Its worth noting for people having bandwidth issue during testing that doing bandwidth tests from router to router caps out at around 30mbit for anything below RB433AH. During our testing we found that RB433 plain can only receive 30mbit, Changing to a 433AH resulted in more bandwidth but even than a RB433 will cap out at around 70mbit

Now for a question, Is there a reason for a single chain N link not to go into 40mhz? We've set HT extension channel but it wont go past 65mbit-HT and signal only shows a max of HT20-8 even tho the signal is -65

Low Signal or Channel on the Edge of the Band.

Figured it out, the 5.0 beta doesnt like 40mhz. Downgrading to 4.10 and I'm pushing 70mbit TCP thru the link

Hi , I am new one im mikrotik. Im using RB433 with N wifi card and 5ghz double antena. I use it only for 1 km and i think i need to attenuation signal. Now Its working like before with single 5ghz antena with G wifi card. I dont find where i can attenuation signal, can somebody help me ?

Simplest way to do is to turn down the output power to minimum.
If that is not enough, point your antennas a little to the sky.

Maybe some hint what signal to look for? I mean in particular numbers.

If you use the R52Hnm card's with only 1 pigtail but a tech enables both chains are you likely to damage the card in anyway? Have had a tech telling me they enable both chains despite only having 1 chain antenna's

Maybe some hint what signal to look for? I mean in particular numbers.

-55 to -65 is good from my experience

PC-Lan-433AH--------34km--------RB800 TCP Test is done from pc connected with 100mbit lan
2 Chains Nstreme
soon will change 433AH with RB800 also I'll fix Signal Strength (now -66/-67) I hope Ill Get 140Mbit Tcp



1. On RB800/RB433 side I have to enable TxPower TAB all rates fix (in this case is set 20Db) with out this link disconnect/reconnect (3-5min)
2. I have to set manual TX/RX Rates HT MCS MCS10-MCS15 with out setting link after 10-15min TX/RX rates goes down 39Mbit-Ht/52Mbit-HT

please post script with full config

Just TX Power set on 20DB on both sides
HT Extension Chanel:above control and HT AMPDU Priorities 5,6,7.
+Nstreme
Everything else on default

thanks

so its just 1 chain or 2

so its just 1 chain or 2

2

what anttens dual pol or two single ones

Look at this xezen
http://forum.mikrotik.com/viewtopic.php ... is#p178990
This was my old link I made some modification just I update longest part od the Link 34km to 11N and with N it works more better and more stable (also without PC's I put RB's)

I'm debating purchasing an R52Hn over an R52n to use inside an alix board. Which would be better to use for just a single AP inside a home? Are they both equally supported as far as open source drivers go (ath9k)?

I'm looking for the best performance (i.e. near 100mbit tcp bandwidth).

Is the configuration at http://wiki.mikrotik.com/wiki/802.11n_Setup_Guide still valid? The wiki article's ~10 months old at this point and there's been a lot of changes to wireless behavior in 4.x from looking at the changelogs.

I´ve tested some links too and here are my results.
Exe´s config

1. link 600m, -49dB, 110cm offset antenas, dual polarity, 20MHz channel, very long signal cable, very bad enviroment (over 100 ap on scan list). R52n RB800-RB433AH
35Mb/s Full Duplex over iperf

802.11a has only 10Mb/s Full Duplex

great one

2. link 1,9km, -60dB, dual polarity jrc 24dB, 20Mhz channel, noisy enviroment (25-30 ap on scan list) R52n RB411AH-RB433AH
32Mb/s

very good too

but...

3. link to 6. link
1km, dual polarity jrc24 or jrc 29, 20MHz channel, normal enviroment (15-25 ap on scan list)R52n 433AH-433AH

2-8Mb/s Full duplex for no reason. Signal was ok, ccq no...

i´ve spent many hours changing antenas, cables and wireless cards but nothing helped.
So i bought sr71-15 replaced them and had 25Mb FullDuplex.

then i started testing r52n and found something very strange...
all r52n has good signal value on the link, but only a few of them were able to run at 30Mb/s Full Duplex. After 30 hours of outdoor testing i found that 48 r52n were damaged.. speed about 5Mb/s, 16 r52n were able run at 25-27Mb/s, and only last 36 from 100 were able run at 35Mb/s Full Duplex (tested on exe´s config on 1 chain, 40MHz channel)

the worst thing about this is that there isn´t other way to figure out which one is good a which one isnt. (or i didn´t found one yet)

So if you have problems with your link and you´ve tried everything, including replace wireless card, don´t be so sure that new ones are fine.

Hi all,

We too are tinkering around with the N settings with and without N-streme.

Now we are testing a great x86 board with 5.0 beta 6 becauzse that supports our gbits ethernet ports.

What we are looking for is a steady 100 Mbits FDX.
We now have something strange but i am sure it is just a bug.
We have 2 r52Hn cards per board and dual n-streme setup.

We get a lovely 540 Mbits data rate each way:)

And as for TCP speeds we are hitting 120 Mbits in separate directions (send or receive) and only 85 in both. Whick is not much better than a 5 Ghz- A dual n-streme.
Our CPU is not the issue as it is enjoying itself at 55%

Any ideas or should we for now just go back to 1 radio card and try to get 80 Mbits FDX.
I have the printscreen below. it WAS fun to see the datarates:) even though it did nothing for the performance

hope to hear from you

Ec

So now, i understand, why Mikrotik don't write normal drivers for SR71-15
They don't need the clients, who loved to use Mikrotik hard&soft...
So we start t use Rockets, and each days we found that it works really

So now, i understand, why Mikrotik don't write normal drivers for SR71-15
They don't need the clients, who loved to use Mikrotik hard&soft...
So we start t use Rockets, and each days we found that it works really

Yip unfortunately

Hi forum!

We have been performing some tests with 802.11n+R52n cards and although we got excelent results when short distances and good (very good) signals, as the link gets longer and poorer N behaves quite worse than simple A (at least on 5GHz).

This is the scenario:
17km link
BASE: RB433 (v4.10) + R52n + Ubiquiti Dual Polarization (http://www.ubnt.com/downloads/RocketDish_Datasheet.pdf)
STATION: Same equipment than BASE

These are the results:
Frequent disconnections even when not performing bandwidth tests
Decent signal but very poor bandwidth
No improvments as frecquency changes
No improvments either with different configurations (nstreme, wds, 5Gh-only-N, a,n, 5Ghz, etc)
No improvments when using different polarization combinations on HT Chains.
CPU goes to 100% on both BASE and STATION when performing TCP BandwidthTest (Not UDP)

These are the screenshots:
BASE-Wireless:

BASE-HT:

BASE-HT MCS:

BASE-WDS:

BASE-NSTREME:

BASE-Signal:

STATION-Signal:

BASE-BandwidthTest UDP:

BASE-BandwidthTest TCP:


Any suggestions will be welcome.

Thanks in advance.

Hi forum!

We have been performing some tests with 802.11n+R52n cards and although we got excelent results when short distances and good (very good) signals, as the link gets longer and poorer N behaves quite worse than simple A (at least on 5GHz).

This is the scenario:
17km link
BASE: RB433 (v4.10) + R52n + Ubiquiti Dual Polarization (http://www.ubnt.com/downloads/RocketDish_Datasheet.pdf)
STATION: Same equipment than BASE

These are the results:
Frequent disconnections even when not performing bandwidth tests
Decent signal but very poor bandwidth
No improvments as frecquency changes
No improvments either with different configurations (nstreme, wds, 5Gh-only-N, a,n, 5Ghz, etc)
No improvments when using different polarization combinations on HT Chains.
CPU goes to 100% on both BASE and STATION when performing TCP BandwidthTest (Not UDP)

These are the screenshots:
BASE-Wireless:

BASE-HT:

BASE-HT MCS:

BASE-WDS:

BASE-NSTREME:

BASE-Signal:

STATION-Signal:

BASE-BandwidthTest UDP:

BASE-BandwidthTest TCP:


Any suggestions will be welcome.

Thanks in advance.

THe tx/Rx ccq is very bad.

THe signal strength will be better id it is neg 60 - 62 dB

Looks like the distance is to long for the equipment or the antennas could not be properly aligned.

I would first start by aliging the antennas what length.

Send me you google earth place mark then I calculate you the signal it must be and give you a graph of the path

Hey Newbie,

Thanks for your quick response. The antennas might not be well-aligned but still I think bandwidth is too poor for the signal got. Anyhow we'll be pleased you could send those info to us. Link placemarks below:

BASE-Lat: 39.519716°
BASE-Long: -0.466781°

STATION-Lat: 39.381389°
STATION-Long: -0.574501°

Best regards.

Hey Newbie,

Thanks for your quick response. The antennas might not be well-aligned but still I think bandwidth is too poor for the signal got. Anyhow we'll be pleased you could send those info to us. Link placemarks below:

BASE-Lat: 39.519716°
BASE-Long: -0.466781°

STATION-Lat: 39.381389°
STATION-Long: -0.574501°

Best regards.

-0.574501

This is your problem :

there is no clear line of site
please give me your email address I will send you infgo

Sorry I didn't mention we have 18 meters towers at both places. My email: ingenieria@grupoawa.es

Tons of thanks.

Hey Newbie,

Thanks for your quick response. The antennas might not be well-aligned but still I think bandwidth is too poor for the signal got. Anyhow we'll be pleased you could send those info to us. Link placemarks below:

BASE-Lat: 39.519716°
BASE-Long: -0.466781°

STATION-Lat: 39.381389°
STATION-Long: -0.574501°

Best regards.

-0.574501

This is your problem :

there is no clear line of site
please give me your email address I will send you infgo

Hello

I got a problem with r52n cards, If I set the data rates to configure or default, the maximal HT rate which achieved is HT20-8, and its dont achieved any HT40 data rates..so the maximal TCP throughput is 44Mbit FDX. I try to upgarde from 4.10 to 4.11 but still nothing, try to reset the router but it didnt helped. I'm thing that I use the right configuration, couse same place exectly same hardware is working at the moment with more than 108 mbit tcp throughput.. FDX, I also try to change the cards with other r52n-s

Heres my config at point A
Signal level is: -60/-63
noise-floor: -122dBm

Anyone see the same problem? And got any solution for this?

I'm still waiting for things to sort out with 11n in Mikrotik. Currently I have more luck with turbo links.

then i started testing r52n and found something very strange...
all r52n has good signal value on the link, but only a few of them were able to run at 30Mb/s Full Duplex. After 30 hours of outdoor testing i found that 48 r52n were damaged.. speed about 5Mb/s, 16 r52n were able run at 25-27Mb/s, and only last 36 from 100 were able run at 35Mb/s Full Duplex (tested on exe´s config on 1 chain, 40MHz channel)

the worst thing about this is that there isn´t other way to figure out which one is good a which one isnt. (or i didn´t found one yet)

So if you have problems with your link and you´ve tried everything, including replace wireless card, don´t be so sure that new ones are fine.

Anyone from mikrotik care to comment on this? Seems pretty serious!

then i started testing r52n and found something very strange...
all r52n has good signal value on the link, but only a few of them were able to run at 30Mb/s Full Duplex. After 30 hours of outdoor testing i found that 48 r52n were damaged.. speed about 5Mb/s, 16 r52n were able run at 25-27Mb/s, and only last 36 from 100 were able run at 35Mb/s Full Duplex (tested on exe´s config on 1 chain, 40MHz channel)

the worst thing about this is that there isn´t other way to figure out which one is good a which one isnt. (or i didn´t found one yet)

So if you have problems with your link and you´ve tried everything, including replace wireless card, don´t be so sure that new ones are fine.

Anyone from mikrotik care to comment on this? Seems pretty serious!

how was that test done? speed is not an indication of card quality. it will depend on current conditions of the link, how many people use the link, if there is any traffic on it, etc. You would have to do the test in controlled environment with no interference and no connected devices to be even close to good testing conditions.

then i started testing r52n and found something very strange...
all r52n has good signal value on the link, but only a few of them were able to run at 30Mb/s Full Duplex. After 30 hours of outdoor testing i found that 48 r52n were damaged.. speed about 5Mb/s, 16 r52n were able run at 25-27Mb/s, and only last 36 from 100 were able run at 35Mb/s Full Duplex (tested on exe´s config on 1 chain, 40MHz channel)

the worst thing about this is that there isn´t other way to figure out which one is good a which one isnt. (or i didn´t found one yet)

So if you have problems with your link and you´ve tried everything, including replace wireless card, don´t be so sure that new ones are fine.

Anyone from mikrotik care to comment on this? Seems pretty serious!

how was that test done? speed is not an indication of card quality. it will depend on current conditions of the link, how many people use the link, if there is any traffic on it, etc. You would have to do the test in controlled environment with no interference and no connected devices to be even close to good testing conditions.

THis is the frustration part of Mikrotik . They have the same attitude as Micro Soft. Only they know.

Thereality is that if we install a different link on the same frequency using Motoral equipment we can get up to 200 Mb/sec

The reason why I bother to log these issues is to support MY not to critisize. I normaly did not log faults etc just used a different supplier.

Sorry for bothering you guus and go to this trouble of trying to get systems working I will revert to my previous method of working

THanks

Is it so hard to write some details about the issues instead of just complaining that some of your cards are 'bad'?

Is it so hard to write some details about the issues instead of just complaining that some of your cards are 'bad'?

No not at all. The main problem for me is heat. an then dealing with Scoop. WHatr we have is it over heats and then it fails.

The speed getting 80 - 100 Mb/sec is super for me and the software is stable. The links are staying up for days and I am happy with the speed.

I truly want to help and contribute

I do understand that it is not a full duplex that is why we have a radio for rx and on for tx.
We have returned a 411 to poynting direct in Cpt where the heat unsoldered the connector. You can check it with them.
We could have resoldered the connector to the PC board.

We took the radio to scoop they tested it and no fault find. To scared to put it into the network again.

I can give you a login on to our router I can install more equipment that we can test and do what ever you guys need to do so it is not that I don't want to give more info ...
I can get stickers that change coulor with temp it is a good indication of where the temp stopped. THat can proof the temp prob.

My main prob is not getting the heat sorted and as a designer I understand that that is not easy as it comes and goes.

To me 70 M is super

Is it so hard to write some details about the issues instead of just complaining that some of your cards are 'bad'?

No not at all. The main problem for me is heat. an then dealing with Scoop. WHatr we have is it over heats and then it fails.

The speed getting 80 - 100 Mb/sec is super for me and the software is stable. The links are staying up for days and I am happy with the speed.

I truly want to help and contribute

I do understand that it is not a full duplex that is why we have a radio for rx and on for tx.
We have returned a 411 to poynting direct in Cpt where the heat unsoldered the connector. You can check it with them.
We could have resoldered the connector to the PC board.

We took the radio to scoop they tested it and no fault find. To scared to put it into the network again.

I can give you a login on to our router I can install more equipment that we can test and do what ever you guys need to do so it is not that I don't want to give more info ...
I can get stickers that change coulor with temp it is a good indication of where the temp stopped. THat can proof the temp prob.

My main prob is not getting the heat sorted and as a designer I understand that that is not easy as it comes and goes.

To me 70 M is super

moz-screenshot-16.png

Play little with Tx Power put it 18,19 or 20 dBm see what's happen with this signal level of course if everything else is ok u can rich 120-140 TCP

Is it so hard to write some details about the issues instead of just complaining that some of your cards are 'bad'?

No not at all. The main problem for me is heat. an then dealing with Scoop. WHatr we have is it over heats and then it fails.

The speed getting 80 - 100 Mb/sec is super for me and the software is stable. The links are staying up for days and I am happy with the speed.

I truly want to help and contribute

I do understand that it is not a full duplex that is why we have a radio for rx and on for tx.
We have returned a 411 to poynting direct in Cpt where the heat unsoldered the connector. You can check it with them.
We could have resoldered the connector to the PC board.

We took the radio to scoop they tested it and no fault find. To scared to put it into the network again.

I can give you a login on to our router I can install more equipment that we can test and do what ever you guys need to do so it is not that I don't want to give more info ...
I can get stickers that change coulor with temp it is a good indication of where the temp stopped. THat can proof the temp prob.

My main prob is not getting the heat sorted and as a designer I understand that that is not easy as it comes and goes.

To me 70 M is super

moz-screenshot-16.png

Play little with Tx Power put it 18,19 or 20 dBm see what's hapen

Will do

So you are running the BT test from the router itself?

So you are running the BT test from the router itself?

Problem is his TX/RX rate is too low

no, it's completely normal. signal is good, this means card is fine. the environment should be at fault for not so good speed.

and playing with tx-power is not a good idea also

This is not btest this is actual internet traffic (not using Nv2 just nstreeme)

this is exactly the type of action that will cause the mentioned card problems. on lower rates this is fine, but on higher rates this power is too much and could damage the card.

only use "card rates" when changing power.

So you are running the BT test from the router itself?

Problem is his TX/RX rate is too low

Yes I do enough CPU power to do it

normis But if I put TX Power on default TX/RX rates drop to HT-20

get bigger antenna instead. or at least use card rates, not all rates fixed.

Erastus, please stop sending screenshots to support. Use the "Upload attachment" form below your post text.

get bigger antenna instead. or at least use card rates, not all rates fixed.

I send you a google earth .

The link is 8.5 Km I have two 30dB dishes and running ht1 and ht2. THe distance is 9 km I also send you an screen with the freq ussage and the scan stats.

then i started testing r52n and found something very strange...
all r52n has good signal value on the link, but only a few of them were able to run at 30Mb/s Full Duplex. After 30 hours of outdoor testing i found that 48 r52n were damaged.. speed about 5Mb/s, 16 r52n were able run at 25-27Mb/s, and only last 36 from 100 were able run at 35Mb/s Full Duplex (tested on exe´s config on 1 chain, 40MHz channel)

the worst thing about this is that there isn´t other way to figure out which one is good a which one isnt. (or i didn´t found one yet)

So if you have problems with your link and you´ve tried everything, including replace wireless card, don´t be so sure that new ones are fine.

Anyone from mikrotik care to comment on this? Seems pretty serious!

how was that test done? speed is not an indication of card quality. it will depend on current conditions of the link, how many people use the link, if there is any traffic on it, etc. You would have to do the test in controlled environment with no interference and no connected devices to be even close to good testing conditions.

i´ve tested od 400m link. Link si working fine for 3 months nonstop (my house) so i´m testing there.
Speed was tested by bandwidth test between rb433ah in tcp. I know that speed is not indication of card quality, but thats the only differance between them. CCQ is fine until full load si running.

Another part of this funny story are reclamations of these cards

get bigger antenna instead. or at least use card rates, not all rates fixed.

I send you a google earth .

The link is 8.5 Km I have two 30dB dishes and running ht1 and ht2. THe distance is 9 km I also send you an screen with the freq ussage and the scan stats.

why not post here?

get bigger antenna instead. or at least use card rates, not all rates fixed.

I send you a google earth .

The link is 8.5 Km I have two 30dB dishes and running ht1 and ht2. THe distance is 9 km I also send you an screen with the freq ussage and the scan stats.

why not post here?

THought it must go through you guys first as some other support channels

I see you use 4900-5200 frequencies. Most antennas have been made for 5800 or 5400, look at the antenna radiation patterns. You will also see that the wireless card is calibrated to run at it's full power at 5800.

If you can't change antenna, and you have no choice but to use such frequency, maybe you can upgrade to v5beta and try Nv2 which is what the mentioned motorola uses (TDMA)

I see you use 4900-5200 frequencies. Most antennas have been made for 5800 or 5400, look at the antenna radiation patterns. You will also see that the wireless card is calibrated to run at it's full power at 5800.

If you can't change antenna, and you have no choice but to use such frequency, maybe you can upgrade to v5beta and try Nv2 which is what the mentioned motorola uses (TDMA)

Thanks I did not know the centre freq of the card is 5800 MANY THANKS!! THe dish is quit linear from 4900 but I will do that immediately !!!

When I attempt to adjust down the power levels of the r52N cards, the interface stops registering clients and often requires a reboot of the router to get it back. This sounds like a bug to me. I need to bring down the power as the other radio units in the mesh are seeing signals in the 40's and 50's from each other. This is resulting in a ccq of less than 10%. not real impressive.

Since this is a mesh, the radio units are all running dual omni's and are about 1500 feet apart in wooded terrain.
Any suggestions on how to raise the ccq would be appreciated.

We never experianced that are you using 2.4 or 5G ?

Howmany stations how far appart?

This is strange we can def lower the power with out any problems. But then why use an Hn and lower the power?

We are using 2.4 only. We are lowering the power because the levels are in hte -30 to -45 range. Absolutely too hot. Radios really like to see signals in the -60 range. The terrain is a highly wooded area so we used the higher capable radio cards.

There is one thing we are changing today however. When we set this up we used a single radio to carry the mesh and to connect the clients. We are replacing the MT nodes today with dual radio units so that the mesh can have its own channel and the customers can connect on their own AP. Maybe its just too many things talking in the same space.

Has anyone had any experience using these enclosure/antennas:WX572003 from wispmax? Seems like another option for mimo client antenna.

It looks like a reasonable enclosure except for the fact that it includes mmcx connectors rather than the u.fl type that the mikrotik R52 series cards come with. That would be a show stopper to me

Hello

I got some time to, investigate the problem, so I climb up and see that one of the cables are full of water on the connector side. When I replace it, the signal didnt change but the ccq goes up to 99/99% and the bandwidth back to 160mbit/tcp/one way

Thanks

Hello

I got a problem with r52n cards, If I set the data rates to configure or default, the maximal HT rate which achieved is HT20-8, and its dont achieved any HT40 data rates..so the maximal TCP throughput is 44Mbit FDX. I try to upgarde from 4.10 to 4.11 but still nothing, try to reset the router but it didnt helped. I'm thing that I use the right configuration, couse same place exectly same hardware is working at the moment with more than 108 mbit tcp throughput.. FDX, I also try to change the cards with other r52n-s

Heres my config at point A

Code: Select all

/interface wireless
set 0 ack-timeout=dynamic adaptive-noise-immunity=none allow-sharedkey=no \
    antenna-gain=0 area="" arp=enabled band=5ghz-onlyn basic-rates-a/g="" \
    basic-rates-b="" comment="" compression=no country=no_country_set \
    default-ap-tx-limit=0 default-authentication=yes default-client-tx-limit=\
    0 default-forwarding=yes dfs-mode=none disable-running-check=no disabled=\
    no disconnect-timeout=3s frame-lifetime=0 frequency=4920 frequency-mode=\
    superchannel frequency-offset=0 hide-ssid=no ht-ampdu-priorities=0 \
    ht-amsdu-limit=8192 ht-amsdu-threshold=8192 ht-basic-mcs=mcs-0 \
    ht-extension-channel=above-control ht-guard-interval=any ht-rxchains=0,1 \
    ht-supported-mcs=mcs-0,mcs-1,mcs-13,mcs-14,mcs-15 ht-txchains=0,1 \
    hw-fragmentation-threshold=disabled hw-protection-mode=none \
    hw-protection-threshold=0 hw-retries=15 l2mtu=2290 \
    max-station-count=2007 mode=ap-bridge mtu=1500 name=\
    wlan1 on-fail-retry-time=100ms periodic-calibration=default \
    periodic-calibration-interval=60 preamble-mode=both \
    proprietary-extensions=post-2.9.25 rate-set=\
    configured scan-list=default,4920 security-profile=default ssid=\
    test station-bridge-clone-mac=00:00:00:00:00:00 \
    supported-rates-a/g="" supported-rates-b="" tx-power=10 tx-power-mode=\
    all-rates-fixed update-stats-interval=disabled wds-cost-range=50-150 \
    wds-default-bridge=bridge1 wds-default-cost=100 wds-ignore-ssid=no \
    wds-mode=dynamic wmm-support=disabled
/interface wireless nstreme
set wlan1 comment="" disable-csma=no enable-nstreme=yes \
    enable-polling=yes framer-limit=3200 framer-policy=none

and the config at point B:
/interface wireless
set 0 ack-timeout=dynamic adaptive-noise-immunity=none allow-sharedkey=no \
    antenna-gain=0 area="" arp=enabled band=5ghz-onlyn basic-rates-a/g="" \
    basic-rates-b="" comment="" compression=no country=no_country_set \
    default-ap-tx-limit=0 default-authentication=yes default-client-tx-limit=\
    0 default-forwarding=yes dfs-mode=none disable-running-check=no disabled=\
    no disconnect-timeout=3s frame-lifetime=0 frequency=4920 frequency-mode=\
    superchannel frequency-offset=0 hide-ssid=no ht-ampdu-priorities=0 \
    ht-amsdu-limit=8192 ht-amsdu-threshold=8192 ht-basic-mcs=mcs-0 \
    ht-extension-channel=above-control ht-guard-interval=any ht-rxchains=0,1 \
    ht-supported-mcs=mcs-0,mcs-1,mcs-13,mcs-14,mcs-15 ht-txchains=0,1 \
    hw-fragmentation-threshold=disabled hw-protection-mode=none \
    hw-protection-threshold=0 hw-retries=15 l2mtu=2290 \
    max-station-count=2007 mode=station-wds mtu=1500 name=\
    wlan2 on-fail-retry-time=100ms periodic-calibration=default \
    periodic-calibration-interval=60 preamble-mode=both \
    proprietary-extensions=post-2.9.25 rate-set=\
    configured scan-list=default,4920 security-profile=default ssid=\
    test station-bridge-clone-mac=00:00:00:00:00:00 \
    supported-rates-a/g="" supported-rates-b="" tx-power=15 tx-power-mode=\
    all-rates-fixed update-stats-interval=disabled wds-cost-range=50-150 \
    wds-default-bridge=bridge1 wds-default-cost=100 wds-ignore-ssid=no \
    wds-mode=dynamic wmm-support=disabled
/interface wireless nstreme
set wlan2 comment="" disable-csma=yes enable-nstreme=yes \
    enable-polling=yes framer-limit=3000 framer-policy=dynamic-size

Signal level is: -60/-63
noise-floor: -122dBm

Anyone see the same problem? And got any solution for this?

My results:
800m link
x86 775 Celeron 1600mhz
r52n
dual polarity 31dB parabolic
4m LMR-400 pigtails
mpls bridge
In future will try to upgrade 12km 802.11a nstreme turbo link )

Hello,
We have p2p link, rbb 433 ah + r52n on both sides. We use 4.6 firmware with exe config (nstreme + hw retries 15 + wds) it works prety good, stable CCQ average 97-99% connection speeds 270/270 mbps real speeds also pretty goods. We try to upgrade to 5.0rc1 and we got poorly results with nv2 CCQ not working, connetion rates striking from 160 mbps to 270 mbps very unstable, ping on idle link 3X !!!!!! higher 3-4 ms on idle link, on 4.6 firmware we have stable ~1ms link, speedtest also shows not stabdle results. Configuration i same only we changed in 5.0rc1 to use NV2.

I want ask somebody have success results from 5.0rc1? Or better today use 4.6? With 4.11 also we got bad results, ccq striking, connection rates droping and so on. I think 4.6 is best for 802.11n with mikrotik.

Thanks for reply

ping on idle link 3X !!!!!! higher 3-4 ms on idle link, on 4.6 firmware we have stable ~1ms link,

Throughput and CCQ in separate issue than PING.

All of you be aware that frames in NV2 are time-based (2ms default) so if you send a echo request reply could be replied in next frame (2-4) ms + additional delay from buffers, link etc. so 3-4 ms Delay using NV2 with default settings is good result.

Hello,
Thanks for suggestion. Can you tell on 500 meter link i should use tdma-period-size: 1? or lower? To archyve best ping latency? Maybe someone have config with best latency and throuput on very short link with 5.0rc1 and nv2? Maybe some recommendation on short links?

Thanks

I have done many WDS bridges (802.11a) which bridge an 802.1q tagged network. The 802.11a mikrotik config did not know anything about 802.1q. It just passwd the vlans and mac address back and fourth with no problem.

I am trying to setup my first mikrotik 802.n-only wds link. I am pretty sure i have the same configuration procedures as when I set up my 802.11a wds bridges. My problem is than when I activate the 802.11n wds link the remote network takes a dump and stops talking.

(Point A) cisco switch#1 --802.1q--> Trango bridge -->------------microwave link------------> Trango bridge --802.1q--> Cisco switch#2 (Point B)
(Point A) cisco switch#1 --802.1q--> Mikrotik N wds bridge -->-----microwave link-----> Mikrotik N wds bridge --802.1q--> Cisco switch#2 (Point B)

This did work well when my Mikrotik wds bridge was an 802.a configuration

Notes: I have verified MAC address and interfaces and bridge settings
Notes: My cisco spanning tree on my switches is normally configured to prefer the Mikrotik link.

Is WDS using a 802.11n configured exactly the same or is it different from 802.11a ?
Does WDS using 802.11n pass the 802.1q traffic or do I need to modify a packet size ?

This was working prior to replacing both mikrotiks with 802.11n cards

help

Tom Jones - A WISP up here in North Idaho

Hi to all,

as I am new in 802.11n, I am trying to setup R52n to work with on sinlge poular feeder. Is this possible and how?

Thank you in advance.
Costas.

Long time reader, first time poster. You know how it goes

Been playing with 802.11n for the past year or so, but I thought i'd drop some info on the latest test link I setup.

Using RB711A's with 26dB Grid Antennas.

We're yet to fully align as our weather has been horrible, but the single stream 802.11n is amazing!

Routerboard 711A's on both ends
RouterOS 5.0rc5
3KM link (Yeah a short one, I know)
10MB/s throughput TCP.
NV2 enabled.
So far, very happy. I've previously tested R52Hn's with 60cm Solid Dual Polarity dishes and not gotten anywhere near these results (however, also not had ROS5.0 and NV2 to help.)

Only thing to note is, you can't do an onboard BTest (TCP) as the 711's run out of CPU before they run out of bandwidth space so the data was FTP traffic from PC to PC and HTTP download from PC to PC, both produced ~100mbit throghput!

So, very happy so far

you can also try the Btest tool from PC to PC, more control over the kind of traffic you test:
http://www.mikrotik.com/download/btest.exe

you can also try the Btest tool from PC to PC, more control over the kind of traffic you test:
http://www.mikrotik.com/download/btest.exe

Yeah, it needs an update though so you can specify multiple streams instead of just running 10 instances

We also aligned today and now top 11.7MB/s

Need Gigabit 711A's

Well since 711 uses only single chain, I wouldn't expect more bandwidth than Ethernet can handle. Still, it's very good achievement with RB711 using 11n. Currently we are investigating possibility to upgrade our networks to 11n. Since most of CPEs are now RB711, and those that aren't can easily be upgraded simply by swapping 11n card in them. Only concern is the performance of 11n in P2MP mode. We are choking at current 11a speeds and todays bandwidth demands

For months I have tried to get laptops with wireless N cards to associate to Mikrotik router using 802.11n - but it seems only possible to connect using b/g.

I have tried endless configurations without success. Is Mikrotik 802.11n only designed to connect to other MTs? I can see no other threads other than people discussing point to point links.

What if we want to use MT exclusively for campus and building APs? Fine if we stick with b/g but otherwise we will have to go with UBNT or Ruckus kit for wireless N clients.

Has anyone managed to solve this? Please post configs!

just make sure you didn't enable nv2 or nstreme.

For months I have tried to get laptops with wireless N cards to associate to Mikrotik router using 802.11n - but it seems only possible to connect using b/g.

I have tried endless configurations without success. Is Mikrotik 802.11n only designed to connect to other MTs? I can see no other threads other than people discussing point to point links.

What if we want to use MT exclusively for campus and building APs? Fine if we stick with b/g but otherwise we will have to go with UBNT or Ruckus kit for wireless N clients.

Has anyone managed to solve this? Please post configs!

Have you enabled the extension channel?

I would also be setting manual rates and disabling all the B and G channels so only the HT-MCS rates are allowable to run, then see if they can connect.

I've not played with any 2.4ghz N gear as its just too noisy where I live.

No I didn't enable NV2 - it wasn't available until recently anyway and certainly not using Nstreme.

I have tried every combination of extension channels, HT-MCS settings but I cannot get PC or Mac wireless clients to connect on N only.

If this won't work we will abandon MT for future hotspot deployments.

No I didn't enable NV2 - it wasn't available until recently anyway and certainly not using Nstreme.

I have tried every combination of extension channels, HT-MCS settings but I cannot get PC or Mac wireless clients to connect on N only.

If this won't work we will abandon MT for future hotspot deployments.

I've a centrino lap. It scales up to 11n rates for downloads but stays at 11a Rates (54Max) up.
Seems there are some incompatibilities.

If I'm not mistaken on some centrino laptops that is the max TX rate. They only use N for RX. Why i don't know...

I've a centrino lap. It scales up to 11n rates for downloads but stays at 11a Rates (54Max) up.
Seems there are some incompatibilities.

And how long has N been available on Mikrotik? Seems like other manufacturers can get it right. Why not MT?

We are working on a Point to Multipoint project with R52nh cards and have been having trouble getting 802.11n to work. 802.11a works great, even with 5Mhz channels. Would would like to try N to see if there is an improvement in range and/or speed, but have had trouble configuring (we have used 4.13, 4.11 and 4.9, with the same result)

We have also had issues getting nstreme or NV2 (on 4.13) working on either 802.11a or n. Any pointers we can use to help us out?

Trisc - I'm not far from you at the moment - if you wish to, could you contact me direct? See below. Might be able to help?

- Each outdoor box has only one RB433AH, one R52n and 2 pigtails. At first I put the test board together with another RB433AH in the same box and the RF interferences caused the link to work perfectly only on tx side: the rx side was interferred by the other miniPCI radio even though their working frequencies were spaced by 400MHz!

I used RouterBoard 433ah r52n 2 wireless cards for this test.One wireless card I made as AP and and the second one as a client(without minipigtails and antennas) and the signal strengh was -56 db. I wrapped the wifi card with a layer of antistatic film, another layer of aluminum foil and the last layer of an antistatic film again. Then I checked the signal and I got to -72 db

This really helped.

I´ve also tested this on cards i use on normal links (wnc cm9) and nothing happened. Signal was always the same - around -71dB

I've considered doing this as well. Could you post photos of the setup.

- Each outdoor box has only one RB433AH, one R52n and 2 pigtails. At first I put the test board together with another RB433AH in the same box and the RF interferences caused the link to work perfectly only on tx side: the rx side was interferred by the other miniPCI radio even though their working frequencies were spaced by 400MHz!

I used RouterBoard 433ah r52n 2 wireless cards for this test.One wireless card I made as AP and and the second one as a client(without minipigtails and antennas) and the signal strengh was -56 db. I wrapped the wifi card with a layer of antistatic film, another layer of aluminum foil and the last layer of an antistatic film again. Then I checked the signal and I got to -72 db

This really helped.

I´ve also tested this on cards i use on normal links (wnc cm9) and nothing happened. Signal was always the same - around -71dB

It seems like you could have some heat related issues with the cards wrapped up like that. Is this something you have ran in production?

dont know exactly how long i am using it, but at "Thu Sep 09, 2010 2:39 pm" i´ve written a post here with a few problems in 802.11n, so it has to be only few days after that. So lets wait for Summer

I do not have much of these cards.... maybe 20, but so far they are working just fine....

I just taked a photo of that.... sorry for the poor quality... cellphone photo...
http://img573.imageshack.us/img573/3596/19022011097.jpg

Capture of RB SXT-5D in action.

224Mbps reached, but too bad this is in UDP test :p
TCP, still testing, which one gone bad, my laptop ethernet or SXT ethernet.
Cause i only able to transfer 60Mbps full duplex

And just how long was that linke? -55 signal so what, 100m max?

t3rm,
any possibility of posting your configs here?
I have a R52n link setup here:
http://forum.mikrotik.com/viewtopic.php ... 85#p250885
But struggle to get to these speeds.

And just how long was that linke? -55 signal so what, 100m max?

I have a 25 mile link that's supposed to be -55.

please post config SXT 5Hnd ! Cant get them worked more than 135Mbps ... connected !
upgraded to 5.0rc10 but no result !
I wonder how to get 270Mbps and at least aprox 100Mbps duplex at least on the table.

one side
[admin@MikroTik] > interface wireless print
Flags: X - disabled, R - running
0 R name="wlan1" mtu=1500 mac-address=00:0C:42:84:66:1B arp=enabled
interface-type=Atheros 11N mode=station-bridge ssid="MikroTik"
frequency=5240 band=5ghz-a/n channel-width=20mhz scan-list=default
wireless-protocol=nv2 wds-mode=disabled wds-default-bridge=none
wds-ignore-ssid=no bridge-mode=enabled default-authentication=yes
default-forwarding=yes default-ap-tx-limit=0 default-client-tx-limit=0
hide-ssid=no security-profile=default compression=no
other side
[admin@MikroTik] > interface wireless print
Flags: X - disabled, R - running
0 R name="wlan1" mtu=1500 mac-address=00:0C:42:84:46:99 arp=enabled
interface-type=Atheros 11N mode=bridge ssid="MikroTik" frequency=5240
band=5ghz-a/n channel-width=20mhz scan-list=default wireless-protocol=nv2
wds-mode=disabled wds-default-bridge=none wds-ignore-ssid=no
bridge-mode=enabled default-authentication=yes default-forwarding=yes
default-ap-tx-limit=0 default-client-tx-limit=0 hide-ssid=no
security-profile=default compression=no

Hello,

I have this configuration:

2x RB411 + R52n-M
1x RB433 + 2x R52n-M
4x GRID 5GHz 27dB

RB411 client ->((( ~14km )))->[5640MHz] AP RB433 [5540MHz]<-((( ~15km )))<- RB411 client
Mode is 5GHz-only-N (MCS 0 - 6)
Signal ~ -70

When I try TX bw test from ap to both of clients it works fine, but when I try RX to clients it goes up to 1 Mbps and link drops...

Does anyone have any idea what could it be?

ottoshr, in order to get to 270mbps data-rate you need to use both chains on the SXT device (check if you have enabled them for TX and RX), and you need to enable the ht-extension-channel support.

m4rk0, what software did you use to make those link drawings?

m4rk0, what software did you use to make those link drawings?

Hello, we use this website for profiling http://www.alphimax.com/

Do You have any idea about my problem?

Hello @ all.

We are a small wisp in Germany. Im reading for a long time here in the forum. We have problems with N, too. The same Problem as chvdr.

Our Equipment at both Sites:

RB433AH + R52N + 2*30dBi Mars Antennas Flat Pannel (No DUAL!!) but in Dual-Polarisation (1 * hor + 1 * ver).

The 2 Antennas (H + V on both Sites) are mounted in a distance of 1 meter. The Distance of the 2 Points (AP and Client) are 11 Km.

The Line is TOP.


Thanks for help.....

joban

well you have the dual 29dbi for some time now
anyway what's of interest for me is a case study for use of MARS antennas with MK boards
what's the distance/ throughput that you where able to get

Hi again,
I have 2 SXTs on 240m, no obstacels.

tx power set to all rates fixed - 0

frequency 5765
ht channel below control on both sides

route attenuation - 95dB
antena gain 16dBi
wifi card power 0dBm

so 0+16-95+16 gives mi signal strengh about -63dBm

so why i have these?
Ch0 -78/-87dBm
Ch1 -87/-87dBm

If i set tx-power to all rates fixed = 10

10+16-95+16 gives mi signal strengh about -53dBm

then i have
Ch0 -64/-66dBm
Ch1 -66/-67dBm

so 10dB on a power gives me this differences:
Ch0 14/21
Ch1 21/20

how is this possible?
ROS 5.0rc11

What is difference between SXT 5HnD and SXT-5D?

no difference, internal numbering was upgraded.

and can you tell me why 10 more dB in wireless card gives me 20 more on signal?

If you want i can give you access to them.

signal level in RouterOS is not a very exact number. if you increase power by one unit, it will not increase signal by one unit. It doesn't work that way.

Hi Normis,
relating to questions/issues mentioned in this thread and others regarding N...
Are you guys aware of issues and working on it, or do you believe MT N support (+for legacy devices) is working 100%

Just like to know if we can expect changes/improvements or is this as good as it gets?

Ekkas

i just dont get it. It seems that sxts has a really bad wireless cards.
I´ve measured EIRP and got this:

EIRP power measurment
RouterBoard SXT
Measurment on 2m with 17dBi antenna
1dB power loss on connector between antenna and sond
Sond: R&S measuring up to 6GHz
Power loss on 2m: 54dB



note: Prijmaci signal = Signal level on sond.

That means, that sxt´s output power is ok, but recieved signals are really weird

how did you measure this?

what is "Signal level on sond."?

please provide more detailed description how did you measure it and make some pictures from the setup and the configuration.

used RB SXT with ROS 5.0 rc11 - level 4, ap bridge, tx power - all rates fixed from -30 to 30
2 meters between the SXT and the antenna with known gain (54dB - environment attenuation)
antenna gain 17dB, 1dB power loss on connector between antenna and probe
probe: http://www2.rohde-schwarz.com/en/produc ... RPZ92.html


EIRP = antenna gain - cable loss + wireless card power
EIRP - environment attenuation + antenna gain (on sond) - cable loss = received signal
EIRP = cable loss + environment attenuation + received signal - antenna gain on sond
EU limit for EIRP is 30dBm, or 27dBm without DFS

for example - tx power +10
EIRP= 1dB + 54dB + (-8dB) - 17dB = 30dBm

So in other words...
if I set tx power to 10, i have 30dBm EIRP
environment attenuation = 32.4 + 20 x log frequency (MHz) + 20 x log range (km)
environment attenuation = 32.4 + 20 x log 5800 + 20 x log 0,23 = 32,4 + 75,27 -12,76
environment attenuation = 95dB

so it should be something like this
30dBm - 95dB + 16dB (other SXT antenna gain) = -49 dBm

but thats the problem....
ch0 60/62
ch1 58/60

~10dB is missing

If i set tx-power to 0
EIRP = 20dBm

so it should be something like this
20-95+16=59dBm

conlusion?

please note that for the receive signal values the noise floor levels also are taken into the account. And that is why maybe the rx signal is not so high.

Noise floors that change wildly between ROS point releases and even packages!

Is there a reason RX signal is adjusted by NF?

Matess you have made many mistakes in your tests:

1. signal level in RouterOS is not precise, and can't be used for exact measurements, it's meant as a guide only, to compare better and worse situations, when aligning antenna

2. you said you used a "17 dbi known antenna" without specifying the manufacturer, and whether it's a certified antenna to be used for testing. many regular and cheap brand manufacturers specify incorrect gain and other parameters, this is not to be trusted unless it's a certified antenna made specifically for testing.

3. to be able to get any kind of accurate reading, the SXT should be fully transmitting 100% of the time at maximum possible duty cycle. otherwise, measurement is inaccurate

It was measured in the college laboratory - University of Technology Brno Czech Rep.
They have much more better equipment, then i have.

And as you see... tx power regulation is good. But thats all....

please note that for the receive signal values the noise floor levels also are taken into the account. And that is why maybe the rx signal is not so high.

Yes, it would explain why after changing wifi cards for cards from Mikrotik signal deteriorates.

There are a few good cards for miniPCI express.

Are you planning some kind of a powerful little router with a miniPCI Express interface? RB800 is not very cheap. Something like 411AH should do the job.

Hello guys,

I have questions on what kind of antennas are you all using for your tests and what can I expect when matching 802.11n with various antenna options?
(hardware being 433AH and R52n radio card, 5Ghz band)

1) sector antenna single polarization
2) sector antenna dual polarization (tx/rx vs tx,rx / tx,rx chains)
3) MIMO antenna (single or dual polarization)

I am using dual polarization jirous parabolic antenas
http://en.jirous.com/antenna-5ghz/jrc-24-duplex up to 2km
http://en.jirous.com/antenna-5ghz/jrc-29-duplex up to 4km

longer distances, or huge interference on offset antenas (110cm) with nn-12 duplex
http://antenna.cz/nn-12_DUPLEX/nn-12_DUPLEX.pdf (sorry only in czech)

offset antenas costs half price then jirous 29, but it is really hard to install them (antenna focusing)

i´ve tried to use ubnt dual polarization dish.... waste of time.

Hello,

I have this configuration:

2x RB411 + R52n-M
1x RB433 + 2x R52n-M
4x GRID 5GHz 27dB

RB411 client ->((( ~14km )))->[5640MHz] AP RB433 [5540MHz]<-((( ~15km )))<- RB411 client
Mode is 5GHz-only-N (MCS 0 - 6)
Signal ~ -70

When I try TX bw test from ap to both of clients it works fine, but when I try RX to clients it goes up to 1 Mbps and link drops...

Does anyone have any idea what could it be?

You share the program in which considered profiles

Thanks all for answer.

I think maybe using one chain can get little more bandwidth compare to 5Ghz turbo nstreme. Add one more antenna is little complicated need more space, stronger tower more cables is not option for now on that place. Also dual polarity antenna is usually solid parabolic antenna need stronger tower sometime time wind is very strong.
May be tested on some other place.

One more thanks to all.

Hi,

My name is carlos, i implemented a 80211n + MIMO 2x2 Wireless Network with diversity only, I used: RB 433AH, R52Hn, Routeos license 4.x, 60 mts. tower, in the Amazonic Jungle in Peru.

The results are 60 a 64 Mbps (data throughput)

A MIMO System, with STBC, Beamforming and Spatial Multiplexing are a very important solution for improve my network.

I am continue my research (PHD) about MIMO System.

I attach documents on round,