ADSL speed drop, what to look for?

Discussion in 'Broadband' started by cl, Oct 30, 2014.

  1. cl

    Woody Guest


    Truth be known its probably 'Merkin speak.' They have very
    different ways of measuring things.

    For example, in the UK we usually talk of peak levels when
    it comes to measuring audio and particulary transmitter
    modulation, whereas the Yanks record average levels - and
    believe me that can cause some wonderous confusion!
     
    Woody, Oct 31, 2014
    #21
    1. Advertisements

  2. One thing that makes it so that it can't be that simple is that there
    isn't just one SNR. The SNR of a phone line (that was never designed to
    work at such high frequencies) varies massively with frequency and it's
    not as simple as just being worse at higher frequencies either, it'll
    also be worse at frequencies where there are sources of interference
    such as radio stations or nearby equipment such as switch mode power
    supplies.

    ADSL divides the available frequency spectrum into blocks of 4.3125kHz
    and assesses the SNR of each one separately to see how much data that
    section can carry reliably.

    Using my phone line as an example the downstream SNR is mostly between
    45dB and 55dB from about 165kHz (lowest frequency used for downstream)
    and about 520kHz, then is falls down to unusable (somewhere around 3dB)
    at about 1070kHz. But there are all sorts of wiggles and notches in the
    graph of SNR against frequency. I can see big notches at or near the
    frequencies of strong radio signals 198KHz, 558kHz, 693kHz, 909kHz (wide
    notch, very strong signal here) etc.

    See:
    http://www.bdg2.dsl.pipex.com/GDL/SNRpertone-2014-11-01-0025.png

    (The X axis is frequency divided by 4.3125kHz).

    This one shows how much data is being sent in each 4.3125kHz "band":
    http://www.bdg2.dsl.pipex.com/GDL/Bitloading-2014-11-01-0025.png

    which as you can see is closely related.
    While the actual SNR is a whole lot of different numbers the modem tries
    to spread the SNR margin evenly over all the frequencies so they're all
    more or less working XdB away from their theoretical limit. Where XdB is
    the SNR margin.
     
    Brian Gregory, Nov 1, 2014
    #22
    1. Advertisements

  3. True, but wherever you are, I think it's still proper to indicate
    power in units of power, and voltage in units of voltage, and not to
    mix them up. Declaring a power level in units of voltage seems a bit
    like saying "Number of apples = x oranges", yet that's what they do.

    Rod.
     
    Roderick Stewart, Nov 1, 2014
    #23
  4. cl

    NY Guest

    Both routers that I have (TP Link TD-W8970 and Netgear DG834PN) just quote
    sync speed (Kbps), SNR Margin (dB) and Line Attenuation (dB). No mention of
    power - whether measured in mW or mV :)

    Being able to see a spectrum of noise and power as a function of frequency
    (tone number) would be useful for diagnosing line problems. It's a shame
    more routers don't have this capability.
     
    NY, Nov 1, 2014
    #24
  5. Note: I said "cable modems". Those are ADSL modem/routers.

    Rod.
     
    Roderick Stewart, Nov 1, 2014
    #25
  6. cl

    NY Guest

    Sorry. Didn't see where you said "cable modems".
     
    NY, Nov 1, 2014
    #26
  7. This is possibly due to an old standard of reporting power levels in dB
    relative to, if I remember correctly, 1mV into 600 ohms.

    But cable isn't 600 ohms.

    Anyway I imagine it's just shorthand for dB relative to some standard
    level the full name of which would be too long to use in the status
    screen of a modem.

    Similar to the way dBA can be used as a shorthand when A weighted sound
    levels being quoted.
     
    Brian Gregory, Nov 1, 2014
    #27
    1. Advertisements

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.