802.11n Slow

greenieofdubbo · Thu Oct 15, 2009 2:34 pm

Hi,

I have an R52N in a RB333 running 5GHz-N-Only mode, I'm using it as an indoor access point for my home. When sitting next to the router with my notebook it connects at 130mbits (Max of my notebooks Intel 4965AGN card apparently) my problem is i cant seem to get more then about 30mbits throughput, i've tried different channels, different notebooks etc...

Any ideas?

Thanks.

Thu Oct 15, 2009 2:35 pm

how did you test it?
Directly to the board or through the board?
Try to use some speed testing program. check the signal, maybe the signal is too strong.

greenieofdubbo · Thu Oct 15, 2009 2:39 pm

how did you test it?
Directly to the board or through the board?
Try to use some speed testing program. check the signal, maybe the signal is too strong.

I'm testing to an FTP server located on the Ethernet side of the router. Signal is -35 with a 98% CCQ when next to the router, i've also tried moving away. No difference.

Thu Oct 15, 2009 2:44 pm

Try to do a multiple connections using some download accelerator. Check the CPU usage on the router when you are doing the transmit.

greenieofdubbo · Fri Oct 16, 2009 11:07 am

Try to do a multiple connections using some download accelerator. Check the CPU usage on the router when you are doing the transmit.

Hi,

CPU is only around 70-80%, i tested with multiple streams and same result.

Please see attached settings, does they look right?

Fri Oct 16, 2009 11:08 am

attach JPG files please

greenieofdubbo · Fri Oct 16, 2009 11:13 am

attach JPG files please

Done

greenieofdubbo · Tue Oct 20, 2009 11:10 am

I've upgraded to ROS 4.1, done a config reset and same result, cant get more then 25 - 30mbit. I also tried another notebook.

Is anyone else running a similar setup? if so what speeds do you get?

I have another R52N on the way for another job, i will test it before it goes out...

blinderix · Mon Nov 09, 2009 1:07 am

I've upgraded to ROS 4.1, done a config reset and same result, cant get more then 25 - 30mbit. I also tried another notebook.

Is anyone else running a similar setup? if so what speeds do you get?

I have another R52N on the way for another job, i will test it before it goes out...

I have the same problem using 493ah + r52n and I'm using 2 laptops with Intel 4965agn, and no more that 30mbps. It connects at 130mbps for a moment, after that slows down. Tried changing frequencies, tried at 2.4 ghz with Ralink 2860 Eeepc-901 and same speed. I'm using 2 8dbi antennas and RouterOS v4.2

cmarsot · Tue Nov 10, 2009 11:23 pm

Hello,

I have same problem.
I run Router0S 4.2 on a RouterBoard RB411A with a R52N Wifi Card.

I have tested connection and network wifi bandwith.
Both clients are running Windows XP.
One has an Intel 4965AGN Network card.
Other one has a TPLINK TL-WN861N.

Both stations connect at 130Mbps (Winbox says 130/54 ?).
But doing a FTP session with my server will not go better than 25Mbps.

CPU of routerboard is around 10% during FTP transfert.

Another user has same problem : http://forum.mikrotik.com/viewtopic.php?f=7&t=36031

semakka · Wed Nov 11, 2009 11:14 pm

I've got 50M using UBNT SR71 15 and RB 433AH

steven

samjan · Sat Nov 21, 2009 10:08 pm

Hello people,
i have the same problem with 2 RB600 and 2 SR71-15 with upgraded 4.2rel.
I try it on the table with 1/4 dipole, and not have more than 25mbps duplex.
May be it will start normal work if i set it on the link? Link is more than 50km with clean LOS, but without testing in office, i don't want to setup it.

I play with all configuration menus, i disable Nstreme and WDS, try only the station-to-apbridge mode with setting the IP to wlan interfaces, and check the TCP speed with internal Bandwidth test.

What the problem with hell?
May be the SR71-15 is not full compatble with Mikrotik? And how we can used 40mhz channels in 4.2rel?

SSD · Mon Nov 23, 2009 7:21 am

I play with all configuration menus, i disable Nstreme and WDS, try only the station-to-apbridge mode with setting the IP to wlan interfaces, and check the TCP speed with internal Bandwidth test.

Use MPLS/VPLS.

boriss · Mon Nov 23, 2009 9:26 pm

Normis

Your answer is always "read this"! That is only what you know ?! Admit that N doesn't work yet so we will not be waisting our expensive time.

Boriss

greenieofdubbo · Mon Nov 23, 2009 10:48 pm

Indeed, its obvious there's a problem here it would be nice to have a response from MikroTik

Thanks.

GLR · Mon Nov 23, 2009 11:00 pm

THG · Mon Nov 23, 2009 11:56 pm

The IEEE 802.11n Draft prohibits using High Throughput with WEP or TKIP as the unicast cipher. If you use these encryption methods (e.g. WEP, WPA-TKIP), your data rate will drop to 54 Mbps. The client drivers will connect using a legacy IEEE 802.11g connection rather than failing to connect altogether, which complies with the IEEE 802.11n draft.

Configure both the WiFi client and AP device's profile to use Wi-Fi Protected Access (WPA2-AES or WPA2-TKIP). You may also choose to configure an unsecured profile, but this option is NOT recommended.

GLR · Tue Nov 24, 2009 12:48 am

Blablabla...
WPA2-AES used.
Encryption disabled already tested !

THG · Tue Nov 24, 2009 2:28 am

Blablabla...

After that post, I retract any attempt at assistance and regret putting forth any effort to help you out. Good Luck in your future attempts to get help from strangers with that attitude.

greenieofdubbo · Tue Nov 24, 2009 3:11 am

I have also tried with no encryption, speeds are still the same.

Thanks for the suggestion THG

GLR · Tue Nov 24, 2009 10:37 am

Sorry, but there is clearly a problem with 802.11n in RouterOS 4.1 and 4.2...

Tue Nov 24, 2009 10:41 am

802.11n is not some kind of a turbo button that you push and get magic speed improvement, it's optimized for reflections and specific environments. if it doesn't work in your situation, use turbo, nstreme etc.

Many people have achieved great speeds with it, if your setup doesn't - it's not because there is a problem with 802.11n

GLR · Tue Nov 24, 2009 10:55 am

I'm sorry but as far as I know, nobody was able to get correct performances with a 802.11n interface in AP mode.
That's not related to the 802.11n standard : Other products work correctly as 802.11n APs.

greenieofdubbo · Tue Nov 24, 2009 11:29 am

I'm sorry but as far as I know, nobody was able to get correct performances with a 802.11n interface in AP mode.
That's not related to the 802.11n standard : Other products work correctly as 802.11n APs.

I agree. I use the same network, same location and same laptops with other 802.11n Access Points and i get 10 - 12MB/s, with Mikrotik i get 3 - 4MB/s.

cmarsot · Tue Nov 24, 2009 11:43 am

I agree too.
I specify same Wifi configuration on an Apple N Spot. Using same client, I have something like 100/120Mbits using Apple Wifi N Spot.
Using Mikrotik, I cannot go further than 25Mbits.

It seems that there is clearly something with RouterOS4.2 and Wifi N cards.
Drivers ? Wifi N Card (R2N in my case) ?

aaa · Tue Nov 24, 2009 10:22 pm

Guys you are wrong. 802.11n is not perfect but not so bad. I have 75mbit real customer traffic right now on 3,5km link.Other thing is wireless experience, right antenna location.

cmarsot · Tue Nov 24, 2009 10:40 pm

Ok, what is your configuration ? AP ?
What is your wifi card in your RB600 ?

What is your customer connection ? What is the customer configuration ? Using laptop to connect on your AP ? Using another RB ?

By the way I would like to have 75Mbits !
So perhaps we are wrong, but give us more information.

GLR · Tue Nov 24, 2009 10:42 pm

Hello,
We are discussing about AP mode...

THG · Wed Nov 25, 2009 5:35 am

It looks like that many of you with poor performance use an Intel® Wireless WiFi card. Intel says that these cards requires a Connect with Intel® Centrino® processor technology certified wireless n access points. Wireless n access points without the Connect with Intel Centrino processor technology identifier may require additional firmware for increased performance results.

A simple search with Google reveals that there is not only problem with the MikroTik RouterOS. I guess that many people buy them on online auctions, just search for the card 4965AGN on ebay and see for yourself. Someone are flooding the market with those cards, and they are not only selling just one card, many of them sells large quantities. Are they really genuine Intel® cards, I just wonder?

Wed Nov 25, 2009 8:58 am

Why don't you just get a good card instead?

THG · Wed Nov 25, 2009 9:11 am

Why don't you just get a good card instead?

What, you mean all the other guys? I have been using Atheros based cards since the early beginning of this century.

Wed Nov 25, 2009 9:12 am

I was talking to those using the cheap questionable intel cards from ebay

THG · Wed Nov 25, 2009 9:32 am

Okey, I misunderstood you. I can understand why people put those cheap cards in their laptops, they cost only around $15. The problem is that they don't work well everywhere.

aaa · Wed Nov 25, 2009 10:11 am

I use two similar RB600 with R52n, wireless configuration is default except ack-timeout. It's fixed and adjusted with two direction max true output.
gbit fiber optic on AP side and 100Base-T network on station side with aprox. 100 customer. IP address is static use simple routing , no bridge, wds or any tunnel.
There you can find installation photo.
http://forum.mikrotik.com/viewtopic.php ... &start=350

Ok, what is your configuration ? AP ?
What is your wifi card in your RB600 ?

What is your customer connection ? What is the customer configuration ? Using laptop to connect on your AP ? Using another RB ?

By the way I would like to have 75Mbits !
So perhaps we are wrong, but give us more information.

cmarsot · Wed Nov 25, 2009 11:09 am

Are they really genuine Intel® cards, I just wonder?

Yes my Intel card is a real one, not one bought from a suspicious Ebay Store.
This is an original one with a Thinkpad T61, bought at Lenovo.

By the way, have same problem with an original Intel in a Thinkpad X200. Wifi car is not a 4XXX but a 5XXX one.
I have also test with another Wifi card, TPLink WN861N. So not Intel one, but I think this is an Atheros chip.
Same result. Saw my previous post.

So I do not want to troll, but it is very easy to say poor result because of an Intel Card. What about users (THG for example) that also have poor results with Atheros wifi card on laptop ?

In the same way, having very good results using Apple HotSpot (Using Atheros Or Intel in Laptop) does not seem to be a wifi laptop card problem.

cmarsot · Tue Dec 01, 2009 1:00 pm

Nothing else ?
Does that mean that we are only very few users using RouterOS as a real AP N indoor ?
I am very surprise that Gurus/Experts are not more involved to solve problems.
As I said, I do not have such issues with Apple AP Extreme.

I am going to see if I encounter such issues with Ubiquiti products.

greenieofdubbo · Tue Dec 01, 2009 1:02 pm

I solved my problem by ordering a Linksys WRT610N

13 - 14MB/s to my notebook!

cmarsot · Tue Dec 01, 2009 1:11 pm

Effectively, as normis suggested, there is clearly a problem with Intel wifi card in notebook.

Well thanks, one more point that clearly says that RouterOS/Routerboard and N are not a good solution for an indorr AP N.

I hope we will have a future RouterOS version that will support correctly N in 2.4Ghz or 5Ghz.

THG · Tue Dec 01, 2009 3:34 pm

Note that the Intel Centrino 4965AGN chipset requires dual band to be able to use the full 300Mbps bandwidth, otherwise it can only achieve 140Mbps (or whatever) on the single band.

Also note that that the Intel Centrino 4965AGN chipset supports 802.11n standard 40MHz channels only in 5GHz spectrum to minimize negative impact to legacy devices/networks nearby.

cmarsot · Tue Dec 01, 2009 3:39 pm

@THG, thanks for the clarification.
Yes with this f*cking card, we connect à 144Mbps. Dual Bonding is not active.
But doing a FTP on a Gigabit server, we have a throughtput on Wireless at 2.5 or 3.0MB/s, very far from 144Mbps or 100Mbps. This could be amzing if we have 100Mbps troughtput with a routerBoard and R2N card.
We do not want to have 3000Mbps using 2.4Ghz with Intel, but 100Mbps would be great. But this is not the case.

blinderix · Thu Dec 03, 2009 12:54 pm

Why don't you just get a good card instead?

OK, I have ThinkPad t61 with mini pci-e Intel, which I'm ready to change with a good card. What card do you recommend?

tonym1975 · Mon Feb 07, 2011 11:07 pm

has anyone ever got this to work???? im useing 5.0 OS and RB411AH SR71-12 card. still sucks on throughput. my d-link blows it away and i hate my dlink!

calman · Wed Feb 09, 2011 12:30 am

Here the old post about N and notebooks.

http://forum.mikrotik.com/viewtopic.php?f=7&t=44701

znet · Fri Feb 11, 2011 9:34 am

802.11n is not some kind of a turbo button that you push and get magic speed improvement, it's optimized for reflections and specific environments. if it doesn't work in your situation, use turbo, nstreme etc.

Many people have achieved great speeds with it, if your setup doesn't - it's not because there is a problem with 802.11n

After that post, I retract any attempt at assistance and regret putting forth any effort to help you out. Good Luck in your future attempts to get help from strangers with that attitude.

Kudos to both of you
This obvious total misunderstanding, and proliferating the assumption 'just it just doesnt work' is beginning to show a trend among users.

Recent frustrating experience with configuring 802.11n has resulted in my conclustion, as others have mentioned, that it does actually work, and in most cases, works very well. It took a good amount of time to understand how the modulation works, how MIMO, multiple streams, power levels, and a host of other quite esoteric and highly technical issues play into proper configuration of an 'N' link. Having used other 'canned' solutions convinces me even more that the MT implementation is much more desirable.

So whats the problem? I have concluded that the technology is naturally separating the wheat from the chaff. The configuration of these highly complex technologies is bringing the 'Peter Principle' into the wireless deployment world. The release of 802.11n technology requires complete understanding, at close to an engineering level, of how these technologies like 'spatial diversity', Multi-In-Multi-Out factor into successful deployment of these radios. The guilty parties and whiners have reached the level of incompetence that has frustrated the above quoted experts who are only trying to help. If you dont understand the difference between MCS7 and MSC15, cease and desist. Let the experts deploy this equipment so the others just dont trash the technology, let alone the bands they operate in. There certainly are glitches in the release of this technology, but if you understand the fundamentals of simultaneous streams and spatial diversity, you can debug these deployments, and give MT the feedback it needs to improve the systems. The demand for speed has given too much horsepower to drivers with learners permits. If those who are having trouble listen to the folks who have the patience to help without condemning the vendor, they might realize that all is not as bad as they purport. Normis, you must be at your breaking point, and I commend you for being as gentle as you have been. However, I can sense that you are doing all you can to not reach into your monitor and straighten out the guilty ones....

Fri Feb 11, 2011 9:41 am

congratulations on raising a 2 year old topic from the dead

but a good post anyway

znet · Fri Feb 11, 2011 9:51 am

Guys you are wrong. 802.11n is not perfect but not so bad. I have 75mbit real customer traffic right now on 3,5km link.Other thing is wireless experience, right antenna location.

Had to comment since this thread is getting way out of hand. I dont have the screenshot, but seeing ~115Mbps tx throughput over a 20mile 802.11n link is a thing of beauty, and additional proof that it works. You are right it isnt perfect, but even the 75Mbps rate you show should make any sane wireless operator, and for that matter user of the system, very pleased to get to what I call 'fractional fiber'. After all, at ~100Mbps, thats replacement enough for the non-existent fiber the customer cant get, but I can deliver at 20 miles. Yes, wireless experience, right antenna location have always worked for me. Have both, always works...Good work on your link!

tonym1975 · Mon Feb 14, 2011 11:22 pm

i dont think that anyone is really reading anything. first and formost. ACCESS POINT MODE INDOORS!!

that is why i posted on here. i have yet seen anyone post true access point configureation that said it works great.

i admit that i dont know everything and that is why i asked does anyone know how to make it work?

i bought "the best" (supposidly) wireless card etc... still didnt work for me correctly. i have used not just laptops i have used a vast amount of different cards, computers, manufactures and all of them have the same result. MIKROTIK access point mode has slow throuput.

that is useing the same internet, same routers, same computers, same switches. the only difference is the access points.

I have used either speakeasy speed test or speedtest.net for my testing. they both have the same results. my dlink is over 3 times faster then the mikrotik.

i want to get away from the box manyfactures. so if anyone can actulay help please do. this is not an arguement. this is a problem that alot of people are having. I KNOW that mikrotik to mikrotik works fast. been there done that. this is a indoor access point.

GLR · Mon Feb 14, 2011 11:26 pm

i dont think that anyone is really reading anything. first and formost. ACCESS POINT MODE INDOORS!!
[...]
I KNOW that mikrotik to mikrotik works fast. been there done that. this is a indoor access point.

+1

calman · Tue Feb 15, 2011 11:06 pm

i dont think that anyone is really reading anything. first and formost. ACCESS POINT MODE INDOORS!!

that is why i posted on here. i have yet seen anyone post true access point configureation that said it works great.

i admit that i dont know everything and that is why i asked does anyone know how to make it work?

i bought "the best" (supposidly) wireless card etc... still didnt work for me correctly. i have used not just laptops i have used a vast amount of different cards, computers, manufactures and all of them have the same result. MIKROTIK access point mode has slow throuput.

that is useing the same internet, same routers, same computers, same switches. the only difference is the access points.

I have used either speakeasy speed test or speedtest.net for my testing. they both have the same results. my dlink is over 3 times faster then the mikrotik.

i want to get away from the box manyfactures. so if anyone can actulay help please do. this is not an arguement. this is a problem that alot of people are having. I KNOW that mikrotik to mikrotik works fast. been there done that. this is a indoor access point.

+1

tonym1975 · Wed Feb 16, 2011 1:14 am

well i see that no one seems to want to help? why is this? i love mikrotik products that is why i want to use them. im not here to lay blame to anyone! i just want to see if we can resolve the issue for many people that either have or had the same problem. looking on other forums on the net i see that others have also gave up and went elseware. i think that is not an acceptible answer.

can someone please help.

i dont know if i posted this before but i had upgraded to the OS version 5 and it help the connection speed reported by the router board but the throuput still sucks.

thanks to anyone that wants to help

Wed Feb 16, 2011 10:03 am

tonym1975, contact support@mikrotik.com with your problem description and include the support output file which is made when you see bad throughput.

tonym1975 · Wed Feb 16, 2011 4:32 pm

i will do this soon as i get a chance. thanks

pakjebakmeel · Thu Jul 14, 2011 1:55 pm

EXACTLY the same problems here, RB493g and r52hn.

I am liasing with MikroTik support now, if something magical happens I will let you know. Did anyone manage to get some decent throughput in the meanwhile?

pakjebakmeel · Mon Jul 18, 2011 11:42 am

Got the 5.6 pre-release version which seems to have some improvements in rate selection.

The clients connect on 150Mbit/s now but still fluctuating link speeds. A lot better but not ideal yet. I manage to get 60Mbit/s through the line on a good day. Sometimes the link speed goes up to 240Mbit/s but never quite reaches 270 or let alone 300. There are definately improvements but I think there is still alot to improve when it comes to wireless in AP mode. When I set the AP to use 802.11g or 802.11a I cannot get more then 1Mbit/s or 2Mbit/s through. My old Alix board running PfSense could get 29Mbit/s on 802.11g, so it's nothing to do with the ether being crowded.

On the plus side; I did find why most Apple mac's cannot connect to MikroTik, it seems that most broadcom cards do not support short preamble mode. When I set this to "any" the Imac and the MacBook Pro connect without any issues.

calman · Thu Jul 21, 2011 7:25 pm

I had performance issues but in my case it was only set hardware retries to 4 and ack to indoors, now i'm installing mikrotik 802.11n coverage with very good results, better than meru systems or ubiquiti unifi (throughput compared ).
what routeros version are you using?

Rivera · Fri Jul 22, 2011 11:25 am

rb493g, 2x52Hn... Same problem, lol.

scampbell · Thu Jul 28, 2011 12:02 am

I have an R52HNm running on RB493G (RoS 5.5) with Laptop connecting at 130Mbps.

The secret to getting this to connect at speeds above 54Mbps for me was to disable TKIP in the security profile and use AES-CCM.

WPA2-PSK must also be enabled.

I am using dual aerials and have enabled both chains in the HT Menu.

/interface wireless security-profiles
add authentication-types=wpa-psk,wpa2-psk group-ciphers=aes-ccm \
group-key-update=5m interim-update=0s management-protection=allowed \
management-protection-key="" mode=dynamic-keys name=secure \
radius-eap-accounting=no radius-mac-accounting=no \
radius-mac-authentication=no radius-mac-caching=disabled \
radius-mac-format=XX:XX:XX:XX:XX:XX radius-mac-mode=as-username \
static-algo-0=none static-algo-1=none static-algo-2=none static-algo-3=\
none static-key-0="" static-key-1="" static-key-2="" static-key-3="" \
static-sta-private-algo=none static-sta-private-key="" \
static-transmit-key=key-0 supplicant-identity="" tls-certificate=none \
tls-mode=no-certificates unicast-ciphers=aes-ccm wpa-pre-shared-key=\
secret01 wpa2-pre-shared-key=secret01

/interface wireless
set 0 adaptive-noise-immunity=none allow-sharedkey=no antenna-gain=0 area="" \
arp=enabled band=2ghz-b/g/n basic-rates-a/g=6Mbps basic-rates-b=1Mbps \
bridge-mode=enabled channel-width=20mhz compression=no 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 distance=\
dynamic frame-lifetime=0 frequency=2462 frequency-mode=regulatory-domain \
frequency-offset=0 hide-ssid=no ht-ampdu-priorities=0 ht-amsdu-limit=8192 \
ht-amsdu-threshold=8192 ht-basic-mcs=\
mcs-0,mcs-1,mcs-2,mcs-3,mcs-4,mcs-5,mcs-6,mcs-7 ht-guard-interval=any \
ht-rxchains=0,1 ht-supported-mcs="mcs-0,mcs-1,mcs-2,mcs-3,mcs-4,mcs-5,mcs-\
6,mcs-7,mcs-8,mcs-9,mcs-10,mcs-11,mcs-12,mcs-13,mcs-14,mcs-15,mcs-16,mcs-1\
7,mcs-18,mcs-19,mcs-20,mcs-21,mcs-22,mcs-23" ht-txchains=0,1 \
hw-fragmentation-threshold=disabled hw-protection-mode=none \
hw-protection-threshold=0 hw-retries=7 l2mtu=2290 max-station-count=2007 mode=ap-bridge mtu=1500 name=\
wlan1 noise-floor-threshold=default nv2-cell-radius=30 \
nv2-noise-floor-offset=default nv2-preshared-key="" nv2-qos=default \
nv2-queue-count=2 nv2-security=disabled on-fail-retry-time=100ms \
periodic-calibration=default periodic-calibration-interval=60 \
preamble-mode=both proprietary-extensions=post-2.9.25 radio-name=Mikrotik_N \
rate-selection=legacy rate-set=default scan-list=default \
security-profile=secure ssid=Mikrotik_N station-bridge-clone-mac=\
00:00:00:00:00:00 supported-rates-a/g=\
6Mbps,9Mbps,12Mbps,18Mbps,24Mbps,36Mbps,48Mbps,54Mbps supported-rates-b=\
1Mbps,2Mbps,5.5Mbps,11Mbps tdma-period-size=2 tx-power-mode=manual-table \
update-stats-interval=disabled wds-cost-range=50-150 wds-default-br
none wds-default-cost=100 wds-ignore-ssid=no wds-mode=disabled \
wireless-protocol=any wmm-support=disabled

catfish · Wed Oct 19, 2011 12:00 pm

Hi!
A have this problem too

rb493g + R52nM + Router0S 5.7

ACCESS POINT MODE INDOORS enabled, but the speed no more ~20-22 Mbit/s

Wed Oct 19, 2011 12:12 pm

please contact with the support output files attached to support@mikrotik.com

DBob · Sat Oct 22, 2011 9:18 pm

I've got exactly the same error.

The PC connects at 130 Mbps (why 130? why not 125, or 135?), and no more then ~25Mbps through the wireless.

The board is a brand new RB751U.

An another noticed thing, see the attached file.
I don't think antenna mode should be there. N can use multiple streams, not just one.

Sun Oct 23, 2011 7:56 am

I have noticed that Windows 7 PC's with Intel 802.11n cards perform poorly when connected to Mikrotik access points, where the same PC connected via Atheros USB or RA-Link 802.11n will perform perfectly.

DBob · Sun Oct 23, 2011 10:19 am

In my case, it's a Windows XP with PCI TP-Link wirless card (I don't know the exact type right now, but with the 3 antennas, and they got only one afaik). Sadly, I cannot try out with another connection right now.

scampbell · Tue Oct 25, 2011 2:29 am

Make sure you are running latest RouterOS and also in Winbox go to System.RouterBoard and make sure the RB751 FW is also updated.

I did a BTEST from a Laptop (Acer Aspire 5920G) connected via Wireless to an RB751U-2HnD which said 130Mbs connection and I could achieve approx 50Mbps throughput in either Send, Receive or Both.

I am using WPA2 PSK and AES-CCM Encryption.

I did note too much RF power reduced the throughput - mine was set to 17dBm max as we cannot use full power here. (1W EIRP Max).

I did another test using a different Acer Laptop and a ZyXEL NWD2205 Wireless N Adaptor and this connected at 300Mbps which is excellent but the throughput was lower at 25Mbps. I did note the CPU was also maxed out at 100% on the RB751U on this test.

Moving the B-test from the RB751U-2HnD to another router further into the network improved throughput to 35 MBps so there may be a config issue at play here or it may be simply a performance limit.

I'll run the same test shortly with an RB493G and R52Hn card and see what I get

Tue Oct 25, 2011 10:30 am

Lastly, I did a BTEST from a Laptop (Acer Aspire 5920G) connected via Wireless to an RB750U-2HnD which said 130Mbs connection and I could achieve approx 50Mbps throughput in either Send, Receive or Both.

Throughput means not from laptop to router, but through the router. You need to do a large number of parallel connections through the device to some other powerful machine. Also make sure that the bottleneck is not your laptop. Also it sounds like you did one connection test. That doesn't adequately load the router, you need tens or hundreds of connections, like from a very high popularity torrent, or download accelerator program with 50 connections.

So perfect test is: "PC --> router --> PC"

DBob · Tue Oct 25, 2011 2:27 pm

The Rb751U has internal antenna (Antenna A) and an external Antenna socket (Antenna B)

About that: "Antenna gain 2x2 MIMO PIF antennas, max gain 2.5dBi; external MMCX option".
The website says that there are two internal antennas, and the optional external antenna only takes over the the second antenna, if plugged in. This is also written at somewhere

Also make sure you are running latest RouterOS and also in Winbox go to System.RouterBoard and make sure the RB751 FW is also updated.

Yes, I've taken care of it already.

scampbell · Wed Oct 26, 2011 7:09 am

Throughput means not from laptop to router, but through the router. You need to do a large number of parallel connections through the device to some other powerful machine. Also make sure that the bottleneck is not your laptop. Also it sounds like you did one connection test. That doesn't adequately load the router, you need tens or hundreds of connections, like from a very high popularity torrent, or download accelerator program with 50 connections.

So perfect test is: "PC --> router --> PC"

By testing as you suggest (pc-router-pc) the best i can get using the Mikrotik Bandwidth test in TCP Mode is 25Mbps (send or receive) and around 40Mbps (20 Mbps send & 20Mbps receive).

The laptop CPU's do not exceed 55% at any point and if the two are both on Ethernet they can do the test at around 250-300Mbps.

Wed Oct 26, 2011 9:37 am

that tool you mention has only one parallel TCP connection, which will give you similar results. you need to use something, that can make 100 or more parallel connections.

basically you are not loading the link at all.

zervan · Wed Oct 26, 2011 9:43 am

I didn't see this discussion and I have written about similar problems with RB751U here: http://forum.mikrotik.com/viewtopic.php ... 48#p286748
I was measuring receiving only (combined receiving and sending was worse) with two different devices (other routerboard / computer) on side A and two different computers (other computer / laptop) on side B with two different network cards on laptop - everything comparing to old Canyon WiFi router.

zervan · Wed Oct 26, 2011 9:46 am

that tool you mention has only one parallel TCP connection, which will give you similar results. you need to use something, that can make 100 or more parallel connections.

basically you are not loading the link at all.

I don't agree - when I set TCP packet size to 60 kB, there was no (or just a little) difference between one connection and two connections (to the same end-point or to other end-point, I've tested both)

scampbell · Thu Oct 27, 2011 4:39 am

that tool you mention has only one parallel TCP connection, which will give you similar results. you need to use something, that can make 100 or more parallel connections.

basically you are not loading the link at all.

OK.

A change in testing method made a big difference.

I have a NAS box with large files on the LAN so I used Filezilla Client and initiated a download of up to 10 simultaneous files.

The result - 80-90Mbps download speed. Much better thank you.

RB751U.jpg

I confess to not understanding why the Bandwidth test which is running at 250-300Mbps on Ethernet does not give similar results under Wireless ?

In what circumstances should we use the Mikrotik Bandwidth Test (if at all) ?

DogHead · Thu Oct 27, 2011 5:33 am

We have been testing a variety of MIMO configurations and are seeing very consistent and disappointing results.

We started with a pair of RB433UAH with RB52Hn cards. Running with two chains, all AMPDU priorities enabled, and all MCS enabled, with AES encryption or with no encryption. We used NS2 in point to multipoint with only one client (will test more later). Tried using WDS, station-psuedobridge, and VPLS. We tested first with MT bandwidth test from router to router. Achieved best performance of about 70Mbps TCP using receive direction only. Then we test with iPerf from computer to computer and get about 60Mbps. Then we test with filezilla and 10 files of 100MB each and get about 5.4MBps or about 45Mbps. Best performance is with station-psuedobridge, which does not make much sense to us. We were expecting VPLS to be best. Not good. BTW computer to computer at wire speed we get upwards of 100MBps file transfer or 899Mbps. So we have a baseline that we know.

We have used ROS 4.17 and 5.7. No difference. We have tried every conceivable test combining MTU sizing, multiple sessions, wireless protocols, timing changes, etc.

Wireless links report speeds of 270Mbps and CCQ of 100/100. We are running on 5GHz with 20/40 below on 5805Mhz.

Then we moved to RB435G with RB52Hn. The results were identical. We were expecting that the Gb Ethernet was the bottle neck. During the tests we looked at CPU on the routers and we were at less than 50%.

Then we moved up to RB435G with SR71a. The results were identical. This is really strange, since we had 3 chains going. The link reports 300Mbps, but the throughput never exceeds 73Mbps.

We also tested RB411U with RB52Hn and the RB711G5UAH or whatever it is called. The high power 711 Gb card with 5GHz radio. In no case could we successfully exceed 73Mbps TCP throughput end to end testing with the MT Bandwidth Test, iPerf or filezilla.

Seems not to matter whether we use 802.11, Nstreme or NV2.

We went head to head against Firetide and Aruba and they are getting 120 to 150Mbps TCP throughput using exactly the same testing methods on 3x3 MIMO.

My question is: Has anyone actually ever achieved TCP end to end throughput that exceeds 73Mbps with Mikrotik gear on a single MIMO radio in point to multipoint configuration? If so, what are your exact system settings? Using Ubiquiti or other radio modules is fine.

It doesn't look like the issue is CPU, wireless reported link speed, etc. We are stumped. We are also frustrated that our competitors seem to be passing us by in regard to performance on even just a simple point to point link. We are about to lose a big opportunity (thousands of routers) because of this.

Help is appreciated.

DBob · Sat Oct 29, 2011 5:44 pm

The Rb751U has internal antenna (Antenna A) and an external Antenna socket (Antenna B)
About that: "Antenna gain 2x2 MIMO PIF antennas, max gain 2.5dBi; external MMCX option".
The website says that there are two internal antennas, and the optional external antenna only takes over the the second antenna, if plugged in. This is also written at somewhere
Also make sure you are running latest RouterOS and also in Winbox go to System.RouterBoard and make sure the RB751 FW is also updated.
Yes, I've taken care of it already.

I still haven't find any solution. Any advice?

calman · Sun Oct 30, 2011 10:21 am

ina testing today rb800 2 different cards, the r52n and the new MiniPCI-e ar9380, I find myself in something like that, the performance of the R52 is perfectly normal (about 54mb) but the new ar9380 (the same config) gives a very low throughput about 30mb and no stable link...
I'm testing with laptop (65mb) 1 chain 20mhz

calman · Sun Oct 30, 2011 10:28 am

We went head to head against Firetide and Aruba and they are getting 120 to 150Mbps TCP throughput using exactly the same testing methods on 3x3 MIMO.

I disagree, testing aruba with 2 chain laptops 20mhz (130) the throughput is about 50mb
similar to meru systems.

The better performance with the tp-link 1043 dd-wrt

TrollMan · Sun Oct 30, 2011 10:35 am

We have been testing a variety of MIMO configurations and are seeing very consistent and disappointing results.

We started with a pair of RB433UAH with RB52Hn cards. Running with two chains, all AMPDU priorities enabled, and all MCS enabled, with AES encryption or with no encryption. We used NS2 in point to multipoint with only one client (will test more later). Tried using WDS, station-psuedobridge, and VPLS. We tested first with MT bandwidth test from router to router. Achieved best performance of about 70Mbps TCP using receive direction only. Then we test with iPerf from computer to computer and get about 60Mbps. Then we test with filezilla and 10 files of 100MB each and get about 5.4MBps or about 45Mbps. Best performance is with station-psuedobridge, which does not make much sense to us. We were expecting VPLS to be best. Not good. BTW computer to computer at wire speed we get upwards of 100MBps file transfer or 899Mbps. So we have a baseline that we know.

We have used ROS 4.17 and 5.7. No difference. We have tried every conceivable test combining MTU sizing, multiple sessions, wireless protocols, timing changes, etc.

Wireless links report speeds of 270Mbps and CCQ of 100/100. We are running on 5GHz with 20/40 below on 5805Mhz.

Then we moved to RB435G with RB52Hn. The results were identical. We were expecting that the Gb Ethernet was the bottle neck. During the tests we looked at CPU on the routers and we were at less than 50%.

Then we moved up to RB435G with SR71a. The results were identical. This is really strange, since we had 3 chains going. The link reports 300Mbps, but the throughput never exceeds 73Mbps.

We also tested RB411U with RB52Hn and the RB711G5UAH or whatever it is called. The high power 711 Gb card with 5GHz radio. In no case could we successfully exceed 73Mbps TCP throughput end to end testing with the MT Bandwidth Test, iPerf or filezilla.

Seems not to matter whether we use 802.11, Nstreme or NV2.

We went head to head against Firetide and Aruba and they are getting 120 to 150Mbps TCP throughput using exactly the same testing methods on 3x3 MIMO.

My question is: Has anyone actually ever achieved TCP end to end throughput that exceeds 73Mbps with Mikrotik gear on a single MIMO radio in point to multipoint configuration? If so, what are your exact system settings? Using Ubiquiti or other radio modules is fine.

It doesn't look like the issue is CPU, wireless reported link speed, etc. We are stumped. We are also frustrated that our competitors seem to be passing us by in regard to performance on even just a simple point to point link. We are about to lose a big opportunity (thousands of routers) because of this.

Help is appreciated.

I too get this, link is at 300mbps but I can never push more than 1/3 of that tcp/ip single session. I'm guessing that there is something in routeros thats adding latency to the packets when wifi is used.

DogHead · Thu Nov 03, 2011 2:42 am

More details on slow performance...

1. Using RB52Hn with 2 chains we were able to get up to 178Mbps actual TCP throughput testing with Ixia. It was not easy...As far as we can figure out:
a. you need to have almost noise or congestion of any type on channel
b. Use NV2
c. Use RB52Hn at both ends
d. Use RB435G at both ends
e. Use exactly the same antenna on both ports or use a cross pol with good isolation
f. Use only 24,36 and 48 Mbps as supported rates
g. Use only 24Mbps for basic rate
h. Use advanced rates
i. Use all HT-MCS supported rates
j. Use only HT-MCS basic rates 0 to 7
k. Use 10km distance
l. Use 2ms timing
m. Adjust transmit gain so you have a SNR of less than 45db
n. Use long interspace
o. Use short preamble.
p. Use adaptive noise immunity
q. Use N only with appropriate above or below HT

It makes no performance difference between WDS or psuedobridge.

We saw no performance difference between SR71 with 3 chains and RB52Hn with 2.

There was no performance difference between 4.17 and 5.7

We disabled all unneeded services.

All of this, if set correctly, will yield good performance. But does it really have to be this hard? We went through extensive trial and error before coming up with the above.

Any noise or congestion in the band will kill performance immediately.

Seems that in point to multipoint, you will probably still be better to use SISO rather than MIMO. Real world there is just too much outdoor noise.

DBob · Sun Nov 20, 2011 12:25 pm

i dont think that anyone is really reading anything. first and formost. ACCESS POINT MODE INDOORS!!

that is why i posted on here. i have yet seen anyone post true access point configureation that said it works great.

i admit that i dont know everything and that is why i asked does anyone know how to make it work?

i bought "the best" (supposidly) wireless card etc... still didnt work for me correctly. i have used not just laptops i have used a vast amount of different cards, computers, manufactures and all of them have the same result. MIKROTIK access point mode has slow throuput.

that is useing the same internet, same routers, same computers, same switches. the only difference is the access points.

I have used either speakeasy speed test or speedtest.net for my testing. they both have the same results. my dlink is over 3 times faster then the mikrotik.

i want to get away from the box manyfactures. so if anyone can actulay help please do. this is not an arguement. this is a problem that alot of people are having. I KNOW that mikrotik to mikrotik works fast. been there done that. this is a indoor access point.

+1

+1

TrollMan · Sun Nov 20, 2011 3:54 pm

i dont think that anyone is really reading anything. first and formost. ACCESS POINT MODE INDOORS!!

that is why i posted on here. i have yet seen anyone post true access point configureation that said it works great.

i admit that i dont know everything and that is why i asked does anyone know how to make it work?

i bought "the best" (supposidly) wireless card etc... still didnt work for me correctly. i have used not just laptops i have used a vast amount of different cards, computers, manufactures and all of them have the same result. MIKROTIK access point mode has slow throuput.

that is useing the same internet, same routers, same computers, same switches. the only difference is the access points.

I have used either speakeasy speed test or speedtest.net for my testing. they both have the same results. my dlink is over 3 times faster then the mikrotik.

i want to get away from the box manyfactures. so if anyone can actulay help please do. this is not an arguement. this is a problem that alot of people are having. I KNOW that mikrotik to mikrotik works fast. been there done that. this is a indoor access point.

+1
+1

+1

maara · Tue Dec 06, 2011 2:24 pm

i dont think that anyone is really reading anything. first and formost. ACCESS POINT MODE INDOORS!!

that is why i posted on here. i have yet seen anyone post true access point configureation that said it works great.

i admit that i dont know everything and that is why i asked does anyone know how to make it work?

i bought "the best" (supposidly) wireless card etc... still didnt work for me correctly. i have used not just laptops i have used a vast amount of different cards, computers, manufactures and all of them have the same result. MIKROTIK access point mode has slow throuput.

that is useing the same internet, same routers, same computers, same switches. the only difference is the access points.

I have used either speakeasy speed test or speedtest.net for my testing. they both have the same results. my dlink is over 3 times faster then the mikrotik.

i want to get away from the box manyfactures. so if anyone can actulay help please do. this is not an arguement. this is a problem that alot of people are having. I KNOW that mikrotik to mikrotik works fast. been there done that. this is a indoor access point.

+1

TrollMan · Tue Dec 06, 2011 3:22 pm

Mikrotik is now moving into the office space and home space with its line of routers with built-in wifi. To be 1/3 of the speed of the cheepest routers will be a key issue to solve if they wish to stay in that product area. Now that NV2 is fast, have a look at AP client speed with windows, macs, smartphones connected!

TKITFrank · Tue Dec 06, 2011 3:29 pm

I have to disagree, See my post and this is a (my) home router!

http://forum.mikrotik.com/viewtopic.php?f=7&t=56161

maara · Tue Dec 06, 2011 3:33 pm

I have to disagree, See my post and this is a (my) home router!
http://forum.mikrotik.com/viewtopic.php?f=7&t=56161

I'll give it a try tonight but the 802.11n sucks on the mikrotik anyway...

(at least at this moment...)

maara · Tue Dec 06, 2011 3:34 pm

I have to disagree, See my post and this is a (my) home router!
http://forum.mikrotik.com/viewtopic.php?f=7&t=56161

is there any configuration difference between your 5ghz and 2,4ghz cards in your mikrotik?

TrollMan · Tue Dec 06, 2011 3:50 pm

I have to disagree, See my post and this is a (my) home router!
http://forum.mikrotik.com/viewtopic.php?f=7&t=56161

Your test is with 10 streams? At least it looks like it. The issue I have, and I think most people have is with one stream, The connection speed for me seems not to matter that much, I can still not get more than 7,35 MB/sec in TCP in a single stream, and using my Netgear 3700 I can get about 22 MB/s in TCP mode. I have tried same settings as you and it does not help. Same result in 2.4 as in 5GHZ for me. But running multiple streams(10-100) I can get good results.

TKITFrank · Tue Dec 06, 2011 4:33 pm

I have to disagree, See my post and this is a (my) home router!
http://forum.mikrotik.com/viewtopic.php?f=7&t=56161
is there any configuration difference between your 5ghz and 2,4ghz cards in your mikrotik?

No there are non, They are configured the same.

Your test is with 10 streams? At least it looks like it. The issue I have, and I think most people have is with one stream, The connection speed for me seems not to matter that much, I can still not get more than 7,35 MB/sec in TCP in a single stream, and using my Netgear 3700 I can get about 22 MB/s in TCP mode. I have tried same settings as you and it does not help. Same result in 2.4 as in 5GHZ for me. But running multiple streams(10-100) I can get good results.

Correct I use 10 streams. But I can get around 40-50Mbit/s using Btest.exe one stream TCP to/from the router (And then the CPU of the router is the limit). I have no fresh figures on this test with only one stream. I could get it if I find the time to assemble the test machines again.
BTW, Do you refer to MB/s as MegaBytes or Megabits? From how I read it, It is MegaBytes and then using only 300Mbit N that is quite good figures for laptop communication.

In my case 2,4Ghz is quite troublesome since the area has a lot of interference. And from tests at my work where we have schools running both 2,4Ghz and 5Ghz the 5Ghz band can use the bandwidth much better.

TrollMan · Tue Dec 06, 2011 5:47 pm

I have to disagree, See my post and this is a (my) home router!
http://forum.mikrotik.com/viewtopic.php?f=7&t=56161
is there any configuration difference between your 5ghz and 2,4ghz cards in your mikrotik?
No there are non, They are configured the same.

Your test is with 10 streams? At least it looks like it. The issue I have, and I think most people have is with one stream, The connection speed for me seems not to matter that much, I can still not get more than 7,35 MB/sec in TCP in a single stream, and using my Netgear 3700 I can get about 22 MB/s in TCP mode. I have tried same settings as you and it does not help. Same result in 2.4 as in 5GHZ for me. But running multiple streams(10-100) I can get good results.
Correct I use 10 streams. But I can get around 40-50Mbit/s using Btest.exe one stream TCP to/from the router (And then the CPU of the router is the limit). I have no fresh figures on this test with only one stream. I could get it if I find the time to assemble the test machines again.
BTW, Do you refer to MB/s as MegaBytes or Megabits? From how I read it, It is MegaBytes and then using only 300Mbit N that is quite good figures for laptop communication.

In my case 2,4Ghz is quite troublesome since the area has a lot of interference. And from tests at my work where we have schools running both 2,4Ghz and 5Ghz the 5Ghz band can use the bandwidth much better.

MB/s is megabyte per second, I think that 7.xMB/s, about 60-65mbit/s is way too low for one stream, and thats about 1/3 of consumer equipment on the market. I too use the RB800@1066 and a 52nm dual chain card. I think that I can not get more by tweeking settings/cards/antennas than this, I think routeros needs to be tweeked just like its been for NV2. Btw, using one chain(link at 150mbit) gives the same result as 2 chains(link at 300), about 60mbit/s. My conclusion is that routeros has either latency added to each stream, or some other issue.

TKITFrank · Tue Dec 06, 2011 6:32 pm

MB/s is megabyte per second, I think that 7.xMB/s, about 60-65mbit/s is way too low for one stream, and thats about 1/3 of consumer equipment on the market. I too use the RB800@1066 and a 52nm dual chain card. I think that I can not get more by tweeking settings/cards/antennas than this, I think routeros needs to be tweeked just like its been for NV2. Btw, using one chain(link at 150mbit) gives the same result as 2 chains(link at 300), about 60mbit/s. My conclusion is that routeros has either latency added to each stream, or some other issue.

Well I use both Cisco and HP at work. I have only done tests on the HP but they do not deliver more than 70-90Mbit using 5Ghz (N) 300Mbit.
What type of transfer are you using to get the 22MB/s speeds? FTP or iperf? And does it deliver this when it is in 300Mbit mode as well?

Anyhow there are always space for improvement in all products so if they do so I am also happy:)

TrollMan · Tue Dec 06, 2011 9:42 pm

MB/s is megabyte per second, I think that 7.xMB/s, about 60-65mbit/s is way too low for one stream, and thats about 1/3 of consumer equipment on the market. I too use the RB800@1066 and a 52nm dual chain card. I think that I can not get more by tweeking settings/cards/antennas than this, I think routeros needs to be tweeked just like its been for NV2. Btw, using one chain(link at 150mbit) gives the same result as 2 chains(link at 300), about 60mbit/s. My conclusion is that routeros has either latency added to each stream, or some other issue.
Well I use both Cisco and HP at work. I have only done tests on the HP but they do not deliver more than 70-90Mbit using 5Ghz (N) 300Mbit.
What type of transfer are you using to get the 22MB/s speeds? FTP or iperf? And does it deliver this when it is in 300Mbit mode as well?

Anyhow there are always space for improvement in all products so if they do so I am also happy:)

I used windows file copy from a WHS server that wire copy gives about 90 MB/s at the lowest. For me, its the best test since it gives real world figures. Its 176 mbit/s that I do get with the netgear 3700 kit and a laptop with a 6300 intel wifi chip. I've tried atheros, intel 4685( I think), 6100, 6200 and 6300. Roteros I do seem to get the best with this setting:

0  R name="wlan1" mtu=1500 mac-address=00:0C:42:67:93:F4 arp=enabled interface-type=Atheros 11N mode=ap-bridge ssid="R11" frequency=2412 
      band=2ghz-onlyn channel-width=20/40mhz-ht-above scan-list=default wireless-protocol=802.11 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

set default authentication-types=wpa2-psk group-ciphers=aes-ccm group-key-update=5m interim-update=0s management-protection=disabled \
    management-protection-key="" mode=dynamic-keys name=default radius-eap-accounting=no radius-mac-accounting=no \
    radius-mac-authentication=no radius-mac-caching=disabled radius-mac-format=XX:XX:XX:XX:XX:XX radius-mac-mode=as-username \
    static-algo-0=none static-algo-1=none static-algo-2=none static-algo-3=none static-key-0="" static-key-1="" static-key-2="" \
    static-key-3="" static-sta-private-algo=none static-sta-private-key="" static-transmit-key=key-0 supplicant-identity=MikroTik \
    tls-certificate=none tls-mode=no-certificates unicast-ciphers=aes-ccm wpa-pre-shared-key="" wpa2-pre-shared-key=123123123123

/interface wireless
set 0 adaptive-noise-immunity=ap-and-client-mode allow-sharedkey=no antenna-gain=0 area="" arp=enabled band=2ghz-onlyn basic-rates-a/g=\
    6Mbps basic-rates-b=1Mbps bridge-mode=enabled channel-width=20/40mhz-ht-above 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 distance=indoors frame-lifetime=0 frequency=2412 frequency-mode=\
    manual-txpower frequency-offset=0 hide-ssid=no ht-ampdu-priorities=0 ht-amsdu-limit=8192 ht-amsdu-threshold=8192 ht-basic-mcs=\
    mcs-0,mcs-1,mcs-2,mcs-3,mcs-4,mcs-5,mcs-6,mcs-7 ht-guard-interval=any ht-rxchains=0,1 ht-supported-mcs="mcs-0,mcs-1,mcs-2,mcs-3,mcs-4,\
    mcs-5,mcs-6,mcs-7,mcs-8,mcs-9,mcs-10,mcs-11,mcs-12,mcs-13,mcs-14,mcs-15,mcs-16,mcs-17,mcs-18,mcs-19,mcs-20,mcs-21,mcs-22,mcs-23" \
    ht-txchains=0,1 hw-fragmentation-threshold=disabled hw-protection-mode=cts-to-self hw-protection-threshold=2340 hw-retries=7 l2mtu=\
    2290 mac-address=00:0C:42:66:93:F4 max-station-count=128 mode=ap-bridge mtu=1500 name=wlan1 noise-floor-threshold=default \
    nv2-cell-radius=30 nv2-noise-floor-offset=default nv2-preshared-key="" nv2-qos=default nv2-queue-count=2 nv2-security=disabled \
    on-fail-retry-time=100ms periodic-calibration=default periodic-calibration-interval=60 preamble-mode=both proprietary-extensions=\
    post-2.9.25 radio-name=000C426693F4 rate-selection=advanced rate-set=default scan-list=default security-profile=default ssid=R11 \
    station-bridge-clone-mac=00:00:00:00:00:00 supported-rates-a/g=6Mbps,9Mbps,12Mbps,18Mbps,24Mbps,36Mbps,48Mbps,54Mbps \
    supported-rates-b=1Mbps,2Mbps,5.5Mbps,11Mbps tdma-period-size=2 tx-power-mode=default update-stats-interval=disabled wds-cost-range=\
    50-150 wds-default-bridge=none wds-default-cost=100 wds-ignore-ssid=no wds-mode=disabled wireless-protocol=802.11 wmm-support=enabled
add area="" arp=enabled bridge-mode=enabled default-ap-tx-limit=0 default-authentication=yes default-client-tx-limit=0 \
    default-forwarding=yes disable-running-check=no disabled=no hide-ssid=no l2mtu=2290 mac-address=02:0C:42:66:93:F4 master-interface=\
    wlan1 max-station-count=6 mtu=1500 name=R11_Guest proprietary-extensions=post-2.9.25 security-profile=guest ssid=R11_Guest \
    update-stats-interval=disabled wds-cost-range=0 wds-default-bridge=none wds-default-cost=0 wds-ignore-ssid=no wds-mode=disabled \
    wmm-support=enabled
/interface wireless manual-tx-power-table
set wlan1 manual-tx-powers="1Mbps:17,2Mbps:17,5.5Mbps:17,11Mbps:17,6Mbps:17,9Mbps:17,12Mbps:17,18Mbps:17,24Mbps:17,36Mbps:17,48Mbps:17,54M\
    bps:17,HT20-0:17,HT20-1:17,HT20-2:17,HT20-3:17,HT20-4:17,HT20-5:17,HT20-6:17,HT20-7:17,HT40-0:17,HT40-1:17,HT40-2:17,HT40-3:17,HT40-4:\
    17,HT40-5:17,HT40-6:17,HT40-7:17"
set R11_Guest
/interface wireless nstreme
set wlan1 disable-csma=no enable-nstreme=no enable-polling=yes framer-limit=3200 framer-policy=none
set "(unknown)"
/interface wireless align
set active-mode=yes audio-max=-20 audio-min=-100 audio-monitor=00:00:00:00:00:00 filter-mac=00:00:00:00:00:00 frame-size=300 \
    frames-per-second=25 receive-all=no ssid-all=no
/interface wireless sniffer
set channel-time=200ms file-limit=10 file-name="" memory-limit=10 multiple-channels=no only-headers=no receive-errors=no \
    streaming-enabled=no streaming-max-rate=0 streaming-server=0.0.0.0
/interface wireless snooper
set channel-time=200ms multiple-channels=yes receive-errors=no

I have tried more or less all settings in both 2.4ghz and 5ghz, and really there is no way to get current routeros to perform close to the netgear 3700 router. I do understand that I have about max of what I currently can expect from MT.

TKITFrank · Tue Dec 06, 2011 10:52 pm

Hi,

Seems fine

Saw a couple of things that can be adjusted...

preamble-mode=both
>Try short, It should choose that by default but in many cases it does not, This made a big difference for me.

hw-protection-threshold=2340
>Verify this against the settings of your wifi card. I use 2347.

security-profile=default
>Default should be no encryption an also the best, But using AES should not do any difference in performance due to Hardware acceleration.

Also do a wireless scan and see how your interference is. At my location 2,4Ghz is useless. I can see 10-15 2,4hz N AP's.
5Ghz is free from interference at the moment. And from what I have seen 5Ghz works better also even if I take the interference out of the equation.

The "Netgear 3700" Seems to be using 4 spiral stream's and MikroTik uses 2 so I am not so surprised that it preforms better. It would be much more accurate if you compare it with an equal product

It would however be nice to see Mikrotik develop a unit with those specs

TrollMan · Wed Dec 07, 2011 10:32 am

The settings I use, I've done a lot of trials to find them to work the best for me

Security is a must so running without it is not an option !

When I run 5GHz I do get results that are about the same, BUT I have not compared them or tested on the netgear. Here, the netgear could perform even better. The reason that I do not run 5GHz at home is because when not testing, but using wifi I get lower signal strength and also lower link speed compared to 2.4 GHz from where I usually sit and surf.

The netgear has a AR9220 for 2.4 GHz and AR9223 for 5Ghz, now thats the same as the MT 52nm card when running 2.4GHz. So in my point of view the RB has a much better spec, Better antennas, better CPU and should at LEAST have comparable result.

DogHead · Thu Feb 02, 2012 6:49 pm

We have been testing with WDS and MIMO, specifically to create Mesh networks.

What we have found is that if we set up with two RB433UAH + RB52Hn routers using ROS v5.12 using AP mode on both with dynamic mesh wds and mesh bridge interfaces with only AES CCM WPA2 security we can get links to associate and pass traffic at a maximum of 30Mbps TCP.

However, if we put one of the routers into Station WDS mode, we can get 70+Mbps TCP.

The net effect is that we cannot build a HWMP mesh using MIMO that can perform above 30Mbps.

We have tried building two radio systems with all sorts of combinations of filtering, bridging, etc using one radio in Station WDS and one in AP, but have yet to come up with a bridged solutions that works.

Is this a bug? Is it related to 802.11 specs? I remember seeing someone say that there was a spec conflict that limited WDS speeds, but would like to understand what exactly it is and how to work around it.

brandonrossl · Thu Feb 02, 2012 10:46 pm

I thought WDS only uses a max of WEP encryption?

DogHead · Sat Feb 04, 2012 5:17 am

Yes, after testing, you are correct that WDS will only function with WEP, which means that it cannot bridge using WDS at MIMO rates.

That means that there is no benefit to use MIMO with HWMP mesh.

Is there any way this can be fixed?

Real competitive disadvantage when people ask for mesh. Yes, we never recommend it, but people want it anyway.

jandafields · Sat Apr 07, 2012 8:02 pm

So, after reading everyone elses conclusions here, this is what I'm seeing as the bottom line regarding Mikrotik 802.11n:

For point-to-point and backhauls, Mikrotik is fine.

For an indoor Access Point, do not use Mikrotik. You get good results with 10 TCP streams on Mikrotik... but most people need good results on a single stream. So, use Linksys or D-Link or whatever your favorite Walmart brand is to get good results even on a single stream.

Please correct me if I'm wrong.

czolo · Sat Apr 07, 2012 8:43 pm

So, after reading everyone elses conclusions here, this is what I'm seeing as the bottom line regarding Mikrotik 802.11n:

For point-to-point and backhauls, Mikrotik is fine.

For an indoor Access Point, do not use Mikrotik. You get good results with 10 TCP streams on Mikrotik... but most people need good results on a single stream. So, use Linksys or D-Link or whatever your favorite Walmart brand is to get good results even on a single stream.

Please correct me if I'm wrong.

OK
You are wrong

Bad results in TCP sugests problems with errors on wifi. All hardware will have bad results if there will be errors in transmition. So MT with good antenna will be better than Linksys or any other vendor hardware with bad antenna. It's not magic - it's physic and 802.11 specification.

jandafields · Sat Apr 07, 2012 11:51 pm

Agreed, but these errors seem to only occur with mikrotik and not with other routers used in the same area with the same laptops... Again to my original statement of mikrotik bad Walmart good.

czolo · Sun Apr 08, 2012 12:21 am

Agreed, but these errors seem to only occur with mikrotik and not with other routers used in the same area with the same laptops... Again to my original statement of mikrotik bad Walmart good.

The same antenna system too? I don't think so.

jandafields · Sun Apr 08, 2012 1:19 am

Agreed, but these errors seem to only occur with mikrotik and not with other routers used in the same area with the same laptops... Again to my original statement of mikrotik bad Walmart good.
The same antenna system too? I don't think so.

Well, they are the "same" in that they both use their basic built-in antenna (think of the Mikrotik 751 for example). The consumer routers have built in hidden antennas as well. Are you saying that Mikrotik can only compete with the performance of Linksys if you buy a better antenna for the Mikrotik?

I'm not trying to prove Mikrotik bad if it isn't. I would love to use and recommend Mikrotik exclusively, but it's hard to do so when Linksys and D-Link work so much better in real world application.

czolo · Sun Apr 08, 2012 1:41 am

Well, they are the "same" in that they both use their basic built-in antenna (think of the Mikrotik 751 for example). The consumer routers have built in hidden antennas as well. Are you saying that Mikrotik can only compete with the performance of Linksys if you buy a better antenna for the Mikrotik?

I'm not trying to prove Mikrotik bad if it isn't. I would love to use and recommend Mikrotik exclusively, but it's hard to do so when Linksys and D-Link work so much better in real world application.

Internal antenna in RB751 is ... small ... it fits in the box ... and it's all good I can say about it. Remamber that I'm not saying it's bad

(but I'm obviously suggesting that) So buy better antenna and test it again.

TrollMan · Sun Apr 08, 2012 9:55 am

So, after reading everyone elses conclusions here, this is what I'm seeing as the bottom line regarding Mikrotik 802.11n:

For point-to-point and backhauls, Mikrotik is fine.

For an indoor Access Point, do not use Mikrotik. You get good results with 10 TCP streams on Mikrotik... but most people need good results on a single stream. So, use Linksys or D-Link or whatever your favorite Walmart brand is to get good results even on a single stream.

Please correct me if I'm wrong.
OK
You are wrong
Bad results in TCP sugests problems with errors on wifi. All hardware will have bad results if there will be errors in transmition. So MT with good antenna will be better than Linksys or any other vendor hardware with bad antenna. It's not magic - it's physic and 802.11 specification.

could you verify your statement? I see no one pulling good throughput on single stream in consumer ap mode. What setup do you have?

I can for my self verify that an rb800 with 52n card with antennas that cost about 50% of The complete linksys box can get above 30% through output of the linksys. For me it's all about the routeros not handling this setup and not about physics of 802.11n

czolo · Sun Apr 08, 2012 8:14 pm

could you verify your statement? I see no one pulling good throughput on single stream in consumer ap mode. What setup do you have?

I can for my self verify that an rb800 with 52n card with antennas that cost about 50% of The complete linksys box can get above 30% through output of the linksys. For me it's all about the routeros not handling this setup and not about physics of 802.11n

OK. Please do some test:
- test your link with UDP simplex, than test with TCP duplex. Check BER in registration.

Some theory:
TCP througput = (MSS/RTT)*(1/sqrt(pl)

MSS - Maxiumum_Segment_Size
RTT - Round_Trip_Time
pl - Packet_Loss

As you can see from this - if there are errors on the link, one session througput must be low.

If you get only 30% speed of Linksys - you have to made bad config.

I have MT AP's in my office and in my home. Both are with external antennas (Atom + pci card). In one session with N standard I can get 45Mb TCP throughput in one session.

TrollMan · Sun Apr 08, 2012 8:44 pm

Hmm, 45megabit??? That's really low. I get about 75 Mb from mb and 200+ from my linksys.... I have about 20 hours of fiddling with this so I'm confident that it's not related to antennas or spectrums used.

czolo · Sun Apr 08, 2012 8:55 pm

Hmm, 45megabit??? That's really low. I get about 75 Mb from mb and 200+ from my linksys.... I have about 20 hours of fiddling with this so I'm confident that it's not related to antennas or spectrums used.

In my network works 3 computer's. All is working in 20Mhz channel, so I think that 45Mb is not bad result in ONE SESSION TCP.

TrollMan · Sun Apr 08, 2012 10:31 pm

Hmm, 45megabit??? That's really low. I get about 75 Mb from mb and 200+ from my linksys.... I have about 20 hours of fiddling with this so I'm confident that it's not related to antennas or spectrums used.
In my network works 3 computer's. All is working in 20Mhz channel, so I think that 45Mb is not bad result in ONE SESSION TCP.

Anything less than 50% is bad if you have moderate free channel. Running 40mhz channel is greater of a challenge to get high percentage since its less lightly to be free in that channel span. Your at 30% and thats crap in my books. Thing is, my linksys 3700 has cheap integrated antennas, 2 for 2,4 ghz and 2 for 5 ghz and mb has expensive eternal antennas. Doing the math here should be easy, not a SINGLE person here has reported TCP single session even close to what cheap consumer products deliver. No matter mb settings its has not been done. For sure, multiple streams 10+ its been shown, and in nv2 or similar backhaul its been done. But not in consumer AP mode.

czolo · Sun Apr 08, 2012 10:45 pm

Hmm, 45megabit??? That's really low. I get about 75 Mb from mb and 200+ from my linksys.... I have about 20 hours of fiddling with this so I'm confident that it's not related to antennas or spectrums used.
In my network works 3 computer's. All is working in 20Mhz channel, so I think that 45Mb is not bad result in ONE SESSION TCP.
Anything less than 50% is bad if you have moderate free channel. Running 40mhz channel is greater of a challenge to get high percentage since its less lightly to be free in that channel span. Your at 30% and thats crap in my books. Thing is, my linksys 3700 has cheap integrated antennas, 2 for 2,4 ghz and 2 for 5 ghz and mb has expensive eternal antennas. Doing the math here should be easy, not a SINGLE person here has reported TCP single session even close to what cheap consumer products deliver. No matter mb settings its has not been done. For sure, multiple streams 10+ its been shown, and in nv2 or similar backhaul its been done. But not in consumer AP mode.

I don't belive your result. What exacly model of Linksys do you use? Can you give us a weblink?
Show me your TCP 200Mb+ result. And let me be first person here showing you real power of MT software in ONE SESSION TCP:

TCP one session.JPG

TrollMan · Sun Apr 08, 2012 11:24 pm

I don't belive your result. What exacly model of Linksys do you use? Can you give us a weblink?
Show me your TCP 200Mb+ result. And let me be first person here showing you real power of MT software in ONE SESSION TCP:
TCP one session.JPG

I've moved my linksys to another net, running my wireless IP security cams, thus I can't prove it. Its a linksys 3700 http://laptoprepair.ca/news/15559.html

your image, TCP-test3.jpg is running 20 streams not one. Thus I cant see it as a proof of anything. What setup do you have, looks like your running a 3*3 mimo setup(ive tried this too with AR9380 card and had dropouts every minute). For me it did not work well. And from what your showning is 109mbit in one stream @ 3*3 (450 Mb connectionlink) and 171 @ 3*3 with 20 streams. How does that prove a point? Your max at 171 is still only 38% max and your one session is at 24%. Way less than 50%

czolo · Sun Apr 08, 2012 11:42 pm

I don't belive your result. What exacly model of Linksys do you use? Can you give us a weblink?
Show me your TCP 200Mb+ result. And let me be first person here showing you real power of MT software in ONE SESSION TCP:
TCP one session.JPG
I've moved my linksys to another net, running my wireless IP security cams, thus I can't prove it. Its a linksys 3700 http://laptoprepair.ca/news/15559.html

your image, TCP-test3.jpg is running 20 streams not one. Thus I cant see it as a proof of anything. What setup do you have, looks like your running a 3*3 mimo setup(ive tried this too with AR9380 card and had dropouts every minute). For me it did not work well. And from what your showning is 109mbit in one stream @ 3*3 (450 Mb connectionlink) and 171 @ 3*3 with 20 streams. How does that prove a point? Your max at 171 is still only 38% max and your one session is at 24%. Way less than 50%

You shows a web-link to Netgear and you are writing about Linksys. Be ohnest and see that my test is FULL DUPLEX. So it's 110Mb in both directions, so it's 220Mb all throughput. 220Mb from 450Mb in N standard in ONE SESSION TCP! Read 802.11 standard and than count percentage of frequency real throughput usage. This RB can't handle more (RB800), that's why result is only like that. It's late, but maybe tomorrow I can make some other tests. This link can handle 170Mb in 20 sessions and 165Mb in one session, so maybe it's interesting. You have bad results with MT, perhaps because of: bad antenna systems, noisy environment, bad configuration...
You are comparing unknown soho router, that CAN'T handle even 100Mb TCP simplex one session test to RB751, which internal antenna is ... hmm let's say it's a crap. I'm showing you real potential power of MikroTik software with good antennas.

jandafields · Mon Apr 09, 2012 12:49 am

This thread, multiple pages long, is full of examples from multiple different people that all say they get worse performance from Mikrotik than Linksys. Are you saying that everyone except you has a bad config? That is highly unlikely.

I think my conclusion stands solid: For PtP or Backhaul, use Mikrotik. For home/office AP, use Linksys.

TrollMan · Mon Apr 09, 2012 1:12 am

czolo, your correct, it was a netgear device and not linksys. not really an unknown soho router. It has AR9223 chip for 2,4 ghz and AR9220 for 5ghz and a AR 7161 680 MHz cpu with 64MB ram. It also has cheap antennas INTEGRATED into the device.
My RB800 running at 1000MHz with 52nm card and (http://landashop.com/catalog/dipole-ant ... -1723.html) antennas. Now even if there where issues with noise I would think that the "power" of microtik would easly beat my cheapo soho router. But as you can read everyone has issues with RB running in soho environments. I've never tested the 751 nor will I ever but I do expect my setup do push more than that.

czolo · Mon Apr 09, 2012 2:23 pm

This thread, multiple pages long, is full of examples from multiple different people that all say they get worse performance from Mikrotik than Linksys. Are you saying that everyone except you has a bad config? That is highly unlikely.

I think my conclusion stands solid: For PtP or Backhaul, use Mikrotik. For home/office AP, use Linksys.

Mikrotik really can be used in soho environments (homes/offices). But comparaing non dedicated to this environments hardware to linksys/netgear or any other brand indoor wifi router is bad idea. This is like complaing for problem with parking a truck in hypermarket parking place. If you redesign normal MT system, add proper antennas and configure it properly it works the same or even better than others soho APs.

AIVAS · Mon Apr 09, 2012 3:31 pm

I notice alot of people in this thread seem to have intel cards as the clients for their home Mikrotik APs. I have done an install at a school I worked for where alot had atheros cards, but the others were intel. The intel cards also suffered poor bandwidth, usally half that of the atheros cards.

Much to my surprise I stumbled across an intel document while googling that told me intel cards only support 40mhz on certain channels, depending on whether you're bonding the upper frequency or lower frequency (yes, the 802.11 spec refers to 40mhz as a bonded 20mhz, whatever.. ).

For the interest of others, i've included the list of channels below, that was complete for my region. At least, configuring the correct channel on the ap for upper or lower bonding allowed the intel cards to obtain a sync of 270-300. Also, Intel (as previously mentioned in this thread) do not handle any bonding on 2.4ghz.

Upper bond channels
5180,5220,5260,5300,5500,5520,5580,5620,5660,5745,5785
Lower bond channels
5200,5240,5280,5320,5520,5560,5600,5640,5680,5765,5805

I have also found on some clients that enabling all ampdu priorities and enabling both chains for tx and rx on the AP really helped the bandwidth.

This was on approx 26 RB600A units equipped with 2xR52N units, one for 5ghz and one for 2ghz.

Clients had a mix of the following
Generic 9220 aheros minipci-e
Intel 4965agn
Intel wifi link 5300agn
Intel wifi link 6300agn

Hope this helps the intel users at least.

czolo · Mon Apr 09, 2012 4:52 pm

I notice alot of people in this thread seem to have intel cards as the clients for their home Mikrotik APs. I have done an install at a school I worked for where alot had atheros cards, but the others were intel. The intel cards also suffered poor bandwidth, usally half that of the atheros cards.

Much to my surprise I stumbled across an intel document while googling that told me intel cards only support 40mhz on certain channels, depending on whether you're bonding the upper frequency or lower frequency (yes, the 802.11 spec refers to 40mhz as a bonded 20mhz, whatever.. ).

For the interest of others, i've included the list of channels below, that was complete for my region. At least, configuring the correct channel on the ap for upper or lower bonding allowed the intel cards to obtain a sync of 270-300. Also, Intel (as previously mentioned in this thread) do not handle any bonding on 2.4ghz.

Upper bond channels
5180,5220,5260,5300,5500,5520,5580,5620,5660,5745,5785
Lower bond channels
5200,5240,5280,5320,5520,5560,5600,5640,5680,5765,5805

I have also found on some clients that enabling all ampdu priorities and enabling both chains for tx and rx on the AP really helped the bandwidth.

This was on approx 26 RB600A units equipped with 2xR52N units, one for 5ghz and one for 2ghz.

Clients had a mix of the following
Generic 9220 aheros minipci-e
Intel 4965agn
Intel wifi link 5300agn
Intel wifi link 6300agn

Hope this helps the intel users at least.

Really interesting and important informatin THX

TrollMan · Mon Apr 09, 2012 10:29 pm

Intel does support 40mhz channel bonding in 2.4ghz but you have to enable it in the driver. Your correct that intel is picky on upper/lower since it wont let you get outside band and thus have to stick with frequencies that make sure the complete 40 is within band.

Intel also has power management that can spook and get low speeds, PSP. To set PSP to CAM( fixes most intel wifi speed issues with some AP) you need the admin tool since the setting is not possible to change without the tool.

0ldman · Tue Apr 10, 2012 4:09 pm

Intel has had issues with their wifi chipsets since the 2100. I had hoped they had gotten them straightened out.

Guess not.

I still try to keep one around so I have a baseline with a crappy card for when some of the hotel customers are having issues.

TrollMan · Thu Apr 12, 2012 9:53 am

czolo, what antennas do you recommend for soho environment?

czolo · Thu Apr 12, 2012 9:03 pm

czolo, what antennas do you recommend for soho environment?

There is no simple answer. For what frequency, will be there just A/B/G or N with 2x2 or 3x3?
In office better results you can get with small omni antenna like this:

than with full profi outdoor antenna like this:

Important thing is also tx power. In MT software - Default tx power in office is usual to strong value.

TrollMan · Fri Apr 13, 2012 6:18 pm

LBW_5DB.jpg

For 2:2 or 3:3, AP in 2,4 ghz mode. AS I can see the antennas looks like they would have too much gain to them, and also recieving range too large. I have smaller now and I think I will keep them since I live in a wooden house and have full coverage indoors. Thus I dont think that for me a change of antennas would matter at all. Thus my conclution is that my soho router still kicks RB. My thought would be that RB would handle noise, interference better and at least have same thought put in ALL conditions.

I think that RB has a lot of work to do to get the n stack up to par with soho routers, but I also think that the other modes are really good. BUT since MT is moving in to soho market this is KEY for success here.

WirelessRudy · Sat Apr 14, 2012 12:35 pm

My experiances with buld soho routers are that they do fine as long as you only have one or maybe two devices connected to them and only one unit is doing a lot of traffic.
The moment more devices want to communicate, special when some demand prolonged data streams (heavy downloads, streaming video) most of the cheap Soho routers soon start to fail. They usually combine high output radios (so nice high signal strentghts) with simple OS and cheap cpu's. Can't be else if you'd consider nowadays you buy a full triple antenna (3-chain) 802.11n router for les than 50 euros... they have to save somewhere! (And the power is what the buyer sees first, it has to be high, but the cpu power buyer will only find out in hars conditions...)
MT is doing much better when it comes to serving more devices in more difficult conditions.

But in the end, if you are really looking for the ultimate in wifi you'd better look in the carrier grade range.
An indoor AP might cost you 4 or 6 vold of a TP-link but where a TP-link, or even a MT router, would have be trown into the bin because it can't cope some stuff of for instance Ruckus is just waiting to get even more units on.....
I have been playing with some of their stuff and once you have done that you know where the real gain is to be made in wireless internet: managing and using RF antenna/radio technology in an intelligent way, not dumm sheer power what seems to be the fasion in soho grade stuff.....

TrollMan · Sun Apr 22, 2012 7:59 pm

With 5.15 there has been great improvements, a quick test gives me 152 mbit throughtput TCP one way one session. Now thats at least DOUBLE compared to 5.14 results! Now, some more improvements and ROS will be on par with generic wifi routers.

moffeloffen · Tue Jun 05, 2012 12:00 pm

Hi!

I have the same problem. i got a cheap TP-link ap N router, and it easyli pulls 9-11MB /sec.
But when i use mikrotik equipment it is like 2-4MB /sec.

I have tried different wireless card like SR71a, SR71-12 and mikrotiks own Rh52hn.
Even tried different board like x86 ones, rb433 rb411ah and rb800. I know for sure mikrotik to mikrotik PP links works.
Tried different laptop cards, but it works nice whit a another router, soo it not the card i belive.
I have tried most settings and cant find a solution for it.

moffeloffen · Tue Jun 12, 2012 10:13 am

anyone that have a solution to this low speeds?

Tue Jun 12, 2012 3:21 pm

anyone that have a solution to this low speeds?

please provide us support output file from your configuration - send that file to support@mikrotik.com
Also what TP-link router you used and how close you were when you have done that testing and how the testing was done?

voona · Sat Jul 07, 2012 2:33 pm

Just to add to this, im having the same issue with N in 'AP' mode getting between 5-6MB/sec. Tried every setting possible, did full wireless survey and still getting the same result.

Purched a new Cisco 1140 2.4 G AP, exactly the same settings WHAM 22MB/sec soild. As other people have stated Mikrotik is awesome for P2P but in AP mode its terrible

craigmitchell · Sat Aug 18, 2012 6:49 pm

Hi guys, just to give my take on this.

The results below come from 3 tests using speedtest.net and taking average.

Laptop: Lenovo T420s with Intel Centrino Advanced-N 6205
RouterBoard: RB751G with country=finland (17dBm), bandwidth=20/40 HT above

Laptop approx 50cm from RB751 (signal level -29dBm)

Laptop: 802.11n Channel Width for band 2.4 set to 20MHz only

mode reported line speed DL/UL mbps
------------------------------------------------------
802.11B 11Mbps 5.6 / 5
802.11G 54Mbps 23 / 10 (my max uplink)
802.11N 65Mbps 37 / 10 (my max uplink)
802.11B/G/N 65Mbps 37 / 10 (my max uplink)

Now changing the laptop channel width to auto (assume = 20/40MHz) for 2.4 GHz band

mode reported line speed DL/UL mbps
------------------------------------------------------
802.11N 150Mbps 89 / 10 (my max uplink)
802.11B/G/N 150Mbps 89 / 10 (my max uplink)

When I transfer a file from LAN server, I get a file transfer over WLAN of 9.91 MB/s (NetPerSec.exe reports average of 88.3 Mbps for the transfer).

When I try from a further distance of say 10m (+wall), I have signal level of -80dBm

mode reported line speed DL/UL mbps
------------------------------------------------------
802.11B/G/N 90-150Mbps 69-72 / 10 (my max uplink)

The settings on the RB751G from 'system reset-configuration' is just adding WPA2 key, country=finland, ie not using TKIP.

RB751_file_transfer_WLAN.png

hkr · Sat Oct 20, 2012 12:23 pm

Hi All,

I have been using an RB600A as a home (indoor) router, equipped with the R52n (AR922X Wireless Network Adapter (rev: 1) wireless card for several years now.

When I bought the stuff, I spent a couple of days trying to find an acceptable settings. Then I ended up - like others - wiring up a "cheap AP" next to the RB600A. Very disappointing, I have been hardly unable to exceed about 6 Mbps down and 2 Mbps up measured to speedtest.net. However, all the cheap APs I have tested let me measure 20+ Mbps on any client device (laptops, phones, tablets etc), which is fine given that my internet connection was 25 Mbps.

Now that my main internet-connection is upgraded to 120 Mbps, I thought I would look back at Mikrotik and see if there is any sample config better suited for SOHO WLAN. Is there any?

Mon Oct 22, 2012 3:21 pm

use the latest RouterOS version, make sure you have distance setting to indoors and rate-selection to advanced.

Jorbu · Tue Feb 05, 2013 4:18 am

Noticed the recently released 6.0rc8 had "wireless - fixed AES encryption speed issues (upgrade suggested)". I decided to upgrade a couple of access points and test. Before the upgrade I was running v5.23, my max 1 session tcp speed was between 18 - 21mbps (one way). After the upgrade, I'm now able to hit 50mbps down, 57mbps up from my macbook laptop. All tests were performed using btest to a router behind the WAP. I'm running 1 chain, 20MHz channels on a RB711ua-2HnD board. I will test two chains soon and upload my findings.
Have you tested the new release? Were your results positive?

Who is online