Slow wireless

Discussion in 'Tomato Firmware' started by alexo, Jul 10, 2008.

  1. alexo

    alexo Network Guru Member

    Hello, I have Tomato 1.19 installed on a WRT54GS.

    One computer is connected directly to the router, another is connected via 802.11g (has a LinkSys card). Both running Windows XP Pro.

    The wired connection gets normal speed.
    The wireless connection is very slow: DSL Reports speed tests indicate 5x to 10x slower, browsing is painful - web sites often time out.
    Windows shows signal strength as "excellent", reports 54Mbps.

    How can I find out what's wrong?

    Wife is upset, please help!

    Thank you,
  2. alexo

    alexo Network Guru Member

    More info:
    ISP is Rogers (cable)
    QOS is disabled

    Also getting sporadic ping timeouts
  3. LLigetfa

    LLigetfa LI Guru Member

    Check for RF interference. Tweak your RWIN.
  4. alexo

    alexo Network Guru Member

    Thank you for the reply LLigetfa,

    I don't have any 2.4GHz devices operating.
    Shouldn't RF interference lower the rate?
    How do I check?

    What should the RWIN be?

    Thank you,
  5. LLigetfa

    LLigetfa LI Guru Member

    You don't have to actually own the 2.4GHz devices to suffer interference from them. They could be owned by someone else.

    The RWIN value depends on your individual circumstance and MTU. Trial and error is the best way. I have mine set to 131040.
  6. alexo

    alexo Network Guru Member

    So the question remains: how do I find out if it is interference or something else, and what do I do if it is?

    Would a sub-optimal RWIN setting slow transfer rates 5x-10x and cause ping timeouts?
  7. LLigetfa

    LLigetfa LI Guru Member

    That very much depends on what is the source of interference. You can start by scanning for other WiFi with something like NetStumbler and if you find other strong signals, try to get 5 channels away from them. If the interference is from a hopper, you might be SOL.

    No to ping timeouts, yes to slowdowns. There was no mention of ping timeouts in your OP, only of web sites often timing out. Certain web sites timing out could also be a MTU issue.

    Do you get erratic ping times pinging your router? If so, that leans to my RF interference theory. What power did you set your AP to? Maybe you have it too high.
  8. alexo

    alexo Network Guru Member

    I'll check.

    There was in the 2nd post. You may have missed it.

    Actually, I was pinging an external site but you are right, I should ping the router.
    I'll check the AP as well.

    Thank you,
  9. alexo

    alexo Network Guru Member

    Router detects 2APs on channel 1 (strongest -92dBm), 5 APs on channel 6 (strongest -58dBm) and 2APs on channel 11 (strongest -89dBm). The client machine (in another room) sees a different picture but Windows does not provide details.

    I set the channel to 1.

    The transmit power of the router was set to 100mW (default is 42mW).
    Lowering it to 60mW raised the speed significantly (to 4+Mbps compared to slightly below 6Mbps wired). Lowering it further to the default 42mW slowed it again.

    Can anyone explain this?

  10. tomatofox

    tomatofox LI Guru Member

    Hi Alexo,
    i have same issues.
    The problem is, that the Tomato Firmware uses an old WLAN-Driver. So some WLAN-Clients have a bad perfomance. You can try the dd-wrt Firmware, i think there you will have better perfomance, because they use a newer wlan driver.
  11. alexo

    alexo Network Guru Member

    So you suggest that this is a Tomato issue?
    Interesting. What's the best way of bringing this to the attention of the author(s)?
  12. TexasFlood

    TexasFlood Network Guru Member

    Have you tried the "Tomato_1_19_ND" version with the newer driver? You'll find it on Jon's website.
  13. LLigetfa

    LLigetfa LI Guru Member

    I very much doubt it is a Tomato issue but you are certainly free to try dd-wrt. I use both and see no difference in WiFi speeds.
  14. mstombs

    mstombs Network Guru Member

    Yes this can be expected if the signal starts to distort at high power, the important signal to noise quality measure can be degraded.

    It is best to use better antennae than just boost the transmit power on the router - this improves the router ability to receive as well as transmit.

    You should also play with the orientation of the existing antennae - use netstumbler on the client for graphical display of the effect.
  15. LLigetfa

    LLigetfa LI Guru Member

    This description is not to be confused with the SNR normally referred to in WiFi. An analogy of the above is a cheap stereo in a quiet room cranked up to the max. The listening pleasure is less than ideal.

    The other SNR is when you have a stereo that is not cranked up but is in a very noisy room with many people shouting. THis is the SNR most referred to in WiFi.

    The proper terminology for the phenomenon the OP is experiencing is EVM (Error Vector Magnitude). Most cheap WiFi gear cannot transmit at higher power in 11g mode because of it. If you look at the technical specs, you often find that the radio firmware will reduce the power at the higher bit rates that require more complex modulation techniques. Only a few commercial grade radios can sustian the power at the higher bit rates and not raise EVM.

    In reality, the OP is probably suffering both EVM and SNR.
  16. TexasFlood

    TexasFlood Network Guru Member

    On my WRT54G v2.0, in my testing years ago, way before Tomato existed, I found 42mW was the best setting for me to avoid performance issues. So Tomato having the default set to 42 was perfect for me. This was honestly the result of my tests and had nothing to do with Douglas Adams or the Hitchhiker's Guide to the Galaxy. I think it's just further proof that 42 really is the answer to life, the universe, and everything, :-D .

    There used to be a few different scope images of what the signal looks like when cranking up the transmit power, here is the only one I can still find.

    So it seems there is going to be a point of diminishing returns with increasing the tnansmit power and this quite likley vairies somewhat with the specific hardware. Heck, it might even vary with the software. As my tests were done way back in the day with Sveasoft, I might have to try setting mine to 60 and see what results I get.
  17. alexo

    alexo Network Guru Member

    OK, this is getting interesting.

    I noticed that only the download speed is affected, the upload speed is the same as on the wired connection. Which lends credibility to the distortion theory.

    Now, what can I do to get the optimal speed?
    - Play with the antenna orientation (which one? there are 3: 2 on the router, 1 on the card)
    - Play with the output power
    - What else?

    Also, what tools I can run on the client (WinXP) to analyze the effects of the changes?

  1. This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
    By continuing to use this site, you are consenting to our use of cookies.
    Dismiss Notice