1. This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.

My ASUS RT-N16 wireless transmission speed useless

Discussion in 'Tomato Firmware' started by velis, Apr 22, 2012.

  1. velis

    velis Serious Server Member

    Pretty much what says in the thread title. I get -37 dBm, quality 62 while at the same time the connection is 72/72, 72 / 7 or 64 / 30 something or whatever. I have not seen my notebook of SGS2 connect, let alone transfer at anything higher.
    When I try to transfer a file, the transfer is at most 1MB/s or so no matter where I place the notebook. This transfer speed is achieved with 72/72 connection.

    I have read the transmission power thread where toastman was explaining stuff about this router a lot as well as multiple other threads on this forum looking for an answer, all to no avail.
    I have changed about 20 firmwares now, including tomato toastman, shibby, dd-wrt and asus firmwares. I rotated antennas and did whatnot, but none of these things really helps. The router behaves extremely consistently in its (non)performance.

    So ultimately I have a semi-great connection (would prefer 144), but transfer is completely useless.

    What am I doing wrong? Am I one of those few toastman mentioned that have uncorrectable connection issues with this router or am I just not setting parameter xxx to the right value?
  2. Gaius

    Gaius Networkin' Nut Member

    Most built-in notebook cards are the cheapest kind, which are locked at a certain transfer rate. Considering looking at your card's specs on the web and other people's experience with it. You might find that you have to upgrade your wireless card to something more robust. Your SGS2 is a phone and I'd never expect it to go over 65MBPS. Mobile phones just don't go up to 300MBPS or even 144MBPS yet.
  3. velis

    velis Serious Server Member

    This particular notebook is a Pavillion dv7 with dual 500GB drives and all the other extras which I have as replacement because my 8740w i currently in service. I don't think this is the cheapest notebook around. While I agree about SGS2, I'm not so sure about the notebook. Maybe I'll just wait to get the 8740 back before I actually return the router, but I really didn't expect such results. I upgraded to N16 so that I could stream 1080p which I can't do on my WRT54GL.

    Is there any way I can actually test my transfer speeds so that I could determine if it's the router or the other piece of HW's fault?
  4. Elfew

    Elfew Addicted to LI Member

    I have problems with transfer speed over wifi too... This problem is because wireless driver which tomato uses. You should try stock firmware and compare speeds...
  5. velis

    velis Serious Server Member

    Just reflashed ASUS I get a 5.5mbit connection with lots of ping drops. The ones that come through are all over the chart ranging from 1 to 100ms. It's been a while since I saw such horrid numbers. Took a while for file copy dialog to even show up :(
    Changing channel width to 40MHz helped a lot, but only restored file copy performance to 800KB/s speed. Seen better both with any of the tomatoes or dd-wrt.
  6. Elfew

    Elfew Addicted to LI Member

    Thats interesting, I cant help you... it solved my problem...

    However ASUS firmware doesnt have many things which I need so I prefer Tomato... (LAN spees is excellent but wifi is not good)
  7. velis

    velis Serious Server Member

    I have reached the best results so far using Shibby 092 K26 USB AIO
    Advanced wireless: frame burst: enable, Preamble: short. Aside from WPA2 AES, all other settings are on defaults. To my knowledge Shibby and Toastman builds are pretty much the same in this regard.

    I get DL 4 MB/s and UL 2 MB/s at 72/72 conection (-28dB 71 quality)

    This is more than I ever got from my WRT54GL, but it's not exactly what I was hoping for. I expected transfer speeds to reach at least 10 MB/s both ways.
  8. Riddlah

    Riddlah Networkin' Nut Member

    Currently using Toastman's build and I can get anywhere between 8mb/s and 12mb/s depending on number of users on wireless and usage. Average is usually around 10MB/s. This is transferring from my laptop to my NAS.

    What build of the Tomato firmware are you using? Also what wireless card is in your laptop? You may have to make a change to your wireless card settings to go above 144Mbps.
  9. Gaius

    Gaius Networkin' Nut Member

    Yeah I wish Tomato had a new driver.
  10. crhiles

    crhiles LI Guru Member

    I was getting similar speeds until I changed the channel width to 20mhz. I also set 802.11n preamble to GF-BRCM, WMM No ACK to enable. My speeds are now between 8 and 15 mb/s. Using the 40mhz channel width, on the status: overview page the interference level would show unacceptable.
    I'm running v1.28.7498 MIPSR2-Toastman-VLAN-RT K26 USB VPN-NOCAT.
  11. velis

    velis Serious Server Member

    @Riddlah: The driver on the notebook says Broadcom 4313 (VEN_14E4\DEV_4727\SUBSYS_1483103C&REV_01). Current build is Shibby 092 as stated in the post just above yours.
    @crhiles: I first matched your settings and then went into card advanced properties and matched all settings between router and computer that I could recognize. That mostly meant setting afterburner to yes and 802.11n preamble to green field. I did disable all the power saving and set all compatibility options to "compatible". Nothing had any effect on speed. I'm beginning to think the laptop can't connect at more than 72.
  12. Elfew

    Elfew Addicted to LI Member

    Yeah - short preamble and frame burst enable I have best results... 22Mb/s download, 26Mb/s upload...
    I wish if somebody take a time to boost wireless performance in Tomato
  13. velis

    velis Serious Server Member

    WTF?!? Got my 8740w back - it has a Intel(R) Centrino(R) Ultimate-N 6300 AGN. Plenty of trademarks, none of the performance.
    The damn crap connects in G mode only.
    I have tried to set the router to N only, though that's not really an option since I still have 2 G only devices, but that didn't seem to work (didn't kill the G devices) and the notebook again connected at G only.
    This is getting pretty ridiculous. In one week, I have tested one of the most popular (and expensive) phones and two pretty expensive notebooks and none of them produced any results to speak of. I think by now I should call this foul.

    Edit: I can pretty much confirm that wireless mode combo doesn't work in Shibby 092 for my router. The G devices work uninterrupted if mode is set to "N only". I have been playing with various settings for a while now and each "Save" caused the notebook to renegotiate, but always back to G, not once to N mode.
  14. Gaius

    Gaius Networkin' Nut Member

    Sorry but HP is a pretty terrible brand and they're known for putting crappy Intel and Broadcom wireless cards in their machines. Yes, they are garbage, even the 6300 in your Elitebook. Just a google search of your Broadcom wireless card brought up a lot of issues people have with their HP notebook, and that doesn't surprise me one bit. Next time, don't buy from companies like HP. Buy your notebook from companies that do custom builds.

    To remedy your issue, buy yourself a quality wireless USB adapter. That's what I did to fix a friend's crappy wireless on their Dell.
  15. Monk E. Boy

    Monk E. Boy Network Guru Member

    I hate to say it, but Intel wireless chipsets went straight to hell after their 802.11b chipsets. I like to think the hardware itself is halfway decent, but Intel just can't write software to save themselves. But it's still only halfway decent hardware because even Linux guys, writing their own software, haven't made it all that much better (it is, just not much better).

    I am a big believer in matching chipsets on the system to chipsets in the wireless card to both increase performance and (primarily) stability. Since the RT-N16 uses a Broadcom chipset, using a Broadcom WNIC should, in theory, be the best option (obv. if you buy a $4.99 Rosewill clearance special then all bets are off). The problem is that most Broadcom-based wireless cards are budget specials that aren't engineered properly, though the same can be said of most wireless cards these days.
  16. Gaius

    Gaius Networkin' Nut Member

    The major problem is in a lot of people's heads. They actually think that paying over $800 for a laptop will guarantee them perfect wireless performance and when they get a clue and buy a new wireless adapter or card, they tend to cheap out and get the cheapest crap they can find for less than $10 and they STILL get mad about bad wireless performance. ALL because they feel they shouldn't have to pay more for their "pricey" yet mediocre laptop to work. It's as if they have a mental disorder or something. They act like children and expect things to work without putting any work into finding out what will work and what won't.

    Whoever feels too entitled or too cheap, it's time to wake the hell up. You're not going to get good wireless from an HP laptop or similar brand and you're not going to get satisfaction for a measly $10 generic adapter from Ebay. You have to actually DO THE RESEARCH and NOT BE A TIGHTWAD.
  17. Planiwa

    Planiwa LI Guru Member

    So, in this conversation about 802.11N throughput rates on an RT-N16 running Tomato it has been stated:

    Enable Frameburst
    Use Short Preamble

    Often the Tomato GUI, the nvram command, and the wl command use different terminology.

    Thus, while all three agree on "frameburst", . . .
    .....  Term      Default  Other-value(s)
    GUI:   Preamble  Long     Short
    nvram: plcphdr   long     short
    wl:    plcphdr   long     auto    debug      (Long term: "PLCP Header)
    When Short Preamble is specified in the GUI, plcphdr is set to short in the nvram, and to auto in wl.
  18. velis

    velis Serious Server Member

    @Gauius: I think you're getting me wrong. Yes, I did kind of expect my 8740w to connect better, but I got the notebook way before I even started thinking about a N router. It just wasn't a requirement.
    What I'm bitching about is the actual state of affairs with the N standard. I have been spoiled by my WRT54GL, both its wired and wireless performance. Then I go and buy me a N router thinking stuff will be the same. Turns out that's reaching a bit too far.
    So in this thread I'm trying to determine if I just chose a bad router or what. I have a 60 day money back guarantee, but I don't want to use that unless I'm absolutely sure it's the router. So far quite a few posters confirmed that their routers of the same make can do good transfers and I haven't moved an inch. Perhaps I should get me a decent USB adapter to test things out, but by this point I fear that too won't get me anywhere.
  19. Planiwa

    Planiwa LI Guru Member

    Could someone please quantify the expected improvement from each of those two settings?
  20. Elfew

    Elfew Addicted to LI Member

    enable both - you will gain about 6Mb/s improvement (in my case)
  21. Planiwa

    Planiwa LI Guru Member

    Hm. 6Mb/s improvement over how much? 3Mb/s for each one? Or?

    On the box it says "300Mb/s data rate". If one setting improves the data rate by 3Mb/s, that's 1%.
    I wonder how one would measure a difference of 1% in the field?

    Of course if the measured data rate was much less, 3Mb/s would be a larger proportion. But surely one would not get a 3Mb/s improvement on 30Mb/s from one setting? So, what were the actual measurements, before and after?
  22. crhiles

    crhiles LI Guru Member

    Reading some other threads, I found Shibby talking about why you may only get a connection of 72Mbps. The thread is here: http://www.linksysinfo.org/index.php?threads/wnr3500lv2-shibby-092-wireless-signal-help.37357/ Quote Below:

    shibby20 post #5:
  23. Planiwa

    Planiwa LI Guru Member

    Aha. What is the difference between half of 300Mb/s and 144Mb/s? It is 6Mb/s.

    Perhaps that's the 6Mb/s to be (re-)gained with Frameburst and Short Preamble? :)

    Oh, wait, Shibby also mentions Afterburner.

    So, Afterburner, frameburst, and short preamble, to squeeze a percent or two out of the data-rate, and then 40MHz to double it. Except when 40Mhz makes it unstable and much worse than 20MHz.

    Got it. :)

    Oh, wait, wl has no setting called "afterburner". It does have ambinrate:
      wl -h abminrate
    get/set afterburner minimum rate threshold
      wl abminrate
    36 Mbps
  24. Monk E. Boy

    Monk E. Boy Network Guru Member

    300Mb/s is a signaling rate, not a transfer rate. Wireless signals are inherently less efficient than wired signals, which is why "54Mb" 802.11g connections only transferred at 20-25Mb/s, and 802.11b "11Mb" connections transferred at 2Mb/s (at best). Trying to quantify a change based on it's percentage of the signaling rate, especially when the signaling rate is an appreciable multiple of the transfer rate, is a recipe for frustration.

    Halve 144Mb/s (which is assuming you're in an environment without massive 2.4Ghz wireless interference, quite rare in urban environments - interference will further cut down your transfer rate) and then compare how big a jump 6Mb/s is... it should be around 8%.
  25. Planiwa

    Planiwa LI Guru Member

    So, who is advocating Afterburner?
    Are they aware that it bricks routers?

  26. Toastman

    Toastman Super Moderator Staff Member Member

    Afterburner bricks a particular persons router, and now we have another myth starting.

    Come on guys, get a sense of proportion!
  27. Planiwa

    Planiwa LI Guru Member

    The fact is that some people say "turn on afterburner", without giving any reason or any evidence.
    Other people follow this advice like sheep.

    When I ask for evidence, or quantified statements, or measurement procedures, there is silence.

    When I tell people about the relation between the GUI, the NVRAM, and the wl command, there are blank stares.

    Does anyone have a clue what changes actually happen to wl and to NVRAM when you turn on afterburner in the GUI? It seems that most people either don't care, or have no way of telling.

    [This is an invitation to anyone to show what does happen.]

    This state of affairs reduces everything to superstition for most people. At least, so it seems.

    Outrageous and misleading marketing claims are matched by equally outrageous and unsubstantiated words of advice.

    People often say: "do this, it works for me".

    People rarely say: "this means that and this is the reason and it relates to the other thing in this way and this is how it works in these circumstance but not in other circumstances.

    If there are myths, it is because of a lack of concept-related documentation, clarity, and transparency.

    When people report things such as "turning on afterburner has bricked my router twice", they should be encouraged to provide minimal replicable documentation. Such conversations should be encouraged.

    When people say: "turn on afterburner to get better throughput rates" they should be challenged to show their evidence, to give their reasons, and to state their measurement procedures.

    When people show their data, we may find that some of the observations were mistaken.

    But when they don't, all we can do is believe.

    Getting from "I'm a proud owner of blah blah and this works for me and it is the best and everyone else should do what I do" to data, critical thinking, evidence, reasons, . . .

    . . . that's how we can get past the myths.
  28. Elfew

    Elfew Addicted to LI Member

    OK, so I enabled frame burst and set - preamble - short.

    I got improvement in my speed - I have now about +6Mbps in download and +8Mbps in upload speed... stability is same - tested by web application.

    I am happy now! Thank you for great firmware!!!

Share This Page