How to set up a spare Linksys WRT54Gv2 as a wired access point network extender

Discussion in 'Linux Networking' started by Liam O'Connor, Mar 8, 2014.

  1. nope.
     
    DecadentLinuxUserNumeroUno, Mar 13, 2014
    #41
    1. Advertisements

  2. Which is to say that latency has exactly nothing to do with the cable
    used, with the twist rate, or the non-existant testing you claim.

    Latency is not a function of the cable.
     
    Floyd L. Davidson, Mar 13, 2014
    #42
    1. Advertisements

  3. If that is true, why can't you cite discussion of it by other
    sources? Why can't you even explain how it would happen?

    I'll repeat it again, there will not be any degradation of the
    signal or the performance due to the use of different cable
    pairs.
     
    Floyd L. Davidson, Mar 13, 2014
    #43

  4. Even though we fashion hand fabricated cat 7 runs at work for some mil
    stuff, it is not a recognized standard.

    We do it mostly for Tempest shielding requirements.

    The T i m m y T a r d will have to look that one up.
     
    DecadentLinuxUserNumeroUno, Mar 14, 2014
    #44
  5. It is a function of the test.

    Mustn't put badly fabbed cables into the system.

    Nice pass on backing up Bell.
     
    DecadentLinuxUserNumeroUno, Mar 14, 2014
    #45
  6. Do you know what "latency" means. A test does not have latency
    any more than a cable does.
    Poor digression on your part, but it does demonstrate a certain
    level of integrity, or lack of such.
     
    Floyd L. Davidson, Mar 14, 2014
    #46
  7. Is that a question?
    The test(s) utilize timing differences to make the aforementioned
    determinations.
    Where? You mentioned an asshole and described him like he was some
    fucking god.

    I mentioned that he was an asshole, and you didn't even retort. I
    guess I wouldn't want you in the same foxhole as I. I wont digress to
    relinquishment of trust. Digression, my ass.
    You sure you want to follow the troll into this realm, Floyd?

    Attacking my character... not too bright.

    You know I work on state of the art equipment, right?

    I help keep people like you safe.
     
    DecadentLinuxUserNumeroUno, Mar 14, 2014
    #47
  8. Liam O'Connor

    mechanic Guest

    Isn't there a famous scene in some Apollo film where a visitor to
    NASA asks a cleaner in a corridor: "And what do you do here", the
    response was "I'm sending a man to the Moon!"

    Eerily familiar.
     
    mechanic, Mar 14, 2014
    #48
  9. Timing differences is not latency and interface hardware does
    not make such tests.

    The signal *delay* through an interface is "latency", and it has
    virtually nothing (measurable) to do with the twist of the
    cable pairs. That is for the simple reason that variations in
    the rather huge by comparison latency of other parts of a transmission
    system would make such measurements impossible without significant
    expense.

    That is, yes the effect of different twist can be measured but the
    test setup will cost thousands of dollars per cable segment to do
    it. A million dollar network becomes a billion dollar network...

    Then it would be appropriate to purchase and use the test equipment
    required to assure cables meet specifications.
    It's a digression when you go off topic to something that has no
    significance, express yourself in an extremely crude and rude
    way for no purpose other than your own emotional satisfaction.

    My reference to Bell had logical significance to the discussion,
    yours was only emotional and only to you.
    Well, as you noted above, I "didn't even retort". I have no
    reason to follow you into that realm.
    That may be, but I certainly am not aware of whether you do or
    don't. What is very clear is that you do not have technical
    knowledge of the equipment we are discussing. That doesn't
    prevent you from using it to a purpose, or suggest that you
    don't have operational skills, but it limits your credibility
    in describing it's functionality.

    Using the equipment is one thing, designing or trouble shooting
    its function is something very different.

    As others have alluded, the janitor may not be a design
    engineer, but he will have exactly the same mission goal!
     
    Floyd L. Davidson, Mar 14, 2014
    #49
  10. Liam O'Connor

    mechanic Guest

    Yes there's nothing wrong with that as that kind of community spirit
    raises morale in the team, and makes clear what it is that gets them
    out of bed in the morning.

    As to this twisted pair thing, I've no idea what the specs are in
    particular cables but no-one seems to have mentioned the purpose of
    such a feature (if indeed it exists) - presumably to reduce
    crosstalk between pairs, but is that important in Ethernet
    applications?
     
    mechanic, Mar 14, 2014
    #50
  11. Yep. Maybe design engineers are more critical, and higher paid,
    but the janitor really does contribute to the mission too!
    Crosstalk between pairs is the main issue with higher speed
    networking! Almost every difference between a modern CAT5 or
    CAT6 cable and the typical twisted pair drop cable used for a
    telephone 100 years ago is specifically intended to increase the
    speed of data by reducing crosstalk.

    But... the network interface does not do standards compliance
    testing!

    A network interface today is dime-a-dozen mass produced drop in
    device. The test equipment required to do such testing is not
    inexpensive.

    It is probably also worth noting that such test equipment does
    not measure or even indicate what the twist rate is on a pair.
    It just measures crosstalk levels and indicates a pass/fail
    condition, but not a reason for either.

    If a given cable segment fails though, other tests are available
    which might (or not) indicate a reason. Time Domain
    Reflectometry testing can show "impedance bumps" at a given
    distance from the end of a cable. But again, it doesn't
    indicate why there is an anomaly. A cable might have been
    spliced, it might have a bad connection, it might be crimped, it
    might have a poorly fabricated connector, etc etc. Any of those
    will show up as a point where the impedance is out of
    specifications, and it requires a visual inspection to determine
    what physical the cause is.
     
    Floyd L. Davidson, Mar 14, 2014
    #51
  12. Liam O'Connor

    John Hasler Guest

    Think about two twisted pairs running close together for a long
    distance. The field is not confined between the wires in a pair: a bit
    leaks out into the surrounding space. If there is another pair close by
    the field couples a bit of the signal into it. If the pairs have
    exactly the same twist rate this coupled interference will always be in
    the same phase and so will add up over a long run and cause problems.
    If the twist rates are slightly different the interference will vary in
    phase by 180 degrees over a length proportional to the difference in
    twist rate and so will average to zero. The problem gets worse at
    higher frequency so Ethernet is more sensitive to it.

    This effect was observed (and dealt with) in long 600 ohm open wire
    lines more than 100 years ago.
     
    John Hasler, Mar 14, 2014
    #52
  13. Up to this point that is correct.
    The twist rate will not change the phase of the signal, either
    in the wire or induced into an adjacent wire, to any significant
    degree.
    While that change in phase rate will happen, it will not average to zero
    at any point in time at the cable termination. It will average to zero
    over a period of time, which is not useful
    That's not the significance though. Phase makes exactly zero
    difference in regard common mode rejection of induced noise. If
    the signals have a given phase relationship it will remain
    exactly the same as long as the *length* of each pair is the
    same.

    The problem with having two pairs with the same twist is that
    they can and will end up physically aligned in such a way that
    each tip, for example, will have the same proximity to the other
    tip over an extended distance that will not be on an average
    equal to the proximity to the other ring. Thus unequal voltages
    will be induced between the tip and ring of each pair.

    Because each pair has a differential signal, resulting in common
    mode rejection of induced voltages, the twist is there to ensure
    that exposure to both tip and ring is of equal intensity for any
    given external field. If a field, for example, extends along
    one inch of the pair, and the pair has 50 twists per inch, the
    exposure to the tip will be very close to exactly the same as to
    the ring. Without the twist the field would be closer to one
    wire than to the other, and the induced voltages would be
    different.

    If the source of a field is another pair though, having the same
    twist would allow the wires to have the same unequal exposure
    tip to tip over an extended distance because a twist in one pair
    would match the twist, and distance, to the other pair. By
    using a different twist in each pair there will be a varying
    exposure over a significant distance, and that will cancel out
    the unequal exposure over any short distance.

    It is purely the strength of the signal induced, which is a
    function of distance. Other variations could affect it, but the
    cable is of course manufactured to minimize the significance of
    any other variation. Thickness and composition of insulation
    are examples where uniformity is important.
     
    Floyd L. Davidson, Mar 14, 2014
    #53
  14. Yet I still get the feeling from each post you make that you are
    eerily retarded.

    Trust me, pal, my baseband gateways are the best wireless devices on
    the planet surface. Even the big players in Europe, asia and australia
    are buying our gear.

    That is just the gear that I am able to discuss.

    You lose... again.
     
    DecadentLinuxUserNumeroUno, Mar 15, 2014
    #54
  15. You are the apparently low integrity dork who brought Bell up.

    Talk about no significance. What did you steal to get where you are?
     
    DecadentLinuxUserNumeroUno, Mar 15, 2014
    #55
  16. I was referring to the troll who has been attacking my character.

    Nice attempt at twisting things though. Pretty obvious you knew what
    I meant.

    So, I guess now, you have reverted to playing? You can shove that
    stupid shit as well as your attacks on me right up your ass, pal.

    And you had the gall to cite me for being off topic.

    You want to attack someone, go find a bridge, and jump off. Attack
    yourself. Stop acting like a 4-F twit.
     
    DecadentLinuxUserNumeroUno, Mar 15, 2014
    #56

  17. It was mentioned. You just do not know how to read, much less read a
    thread.

    There was discussion about crosstalk at the terminations, where
    crosstalk occurs. What makes you think it would occur in the twisted
    segment?

    Learn to read instead of simpleton perusal. You won't look like such
    a simpleton that way.
     
    DecadentLinuxUserNumeroUno, Mar 15, 2014
    #57
  18. Crosstalk is not limited to "the terminations", nor for that
    matter is that even where most of it should normally occur. It
    is quite true that most often out of specification crosstalk
    will be due to faults at the connectors, but that is simply
    because no other part of the cable is usually worked on.

    With an in specs cable there is more crosstalk, generally, in
    about the first 10-20% of the length of the cable, at each end.
    That obviously includes the connectors, but it is far from
    restricted to it. The reason is simply that the RX pair has
    the lowest signal and the TX pair has the highest signal that
    they exhibit at any point along the cable.
    It would help your credibility if you'd drop this sort of
    comment. I virtually always reflects more on the person making
    such statements than on the one it is directed at. In your case
    that is universally true so far in this discussion.
     
    Floyd L. Davidson, Mar 15, 2014
    #58
  19. Liam O'Connor

    mechanic Guest

    Morning DLUNU, thanks to your minders for letting you out to play
    again today.

    Now if you were concerned about different twist rates for different
    pairs in a cable, the crosstalk you're worried about must occur in
    the 'twisted segment', no? If it occurred only at the terminations
    the twist rates of the various pairs could be the same with no
    penalty? See the argument?

    In telephone cables there were rules set out ages ago to try to
    minimise crosstalk by choosing pairs in appropriate groups in big
    transmission cables, I was just wondering if it was a real problem
    in Ethernet cables or even Ethernet patch cords? Seems unlikely but
    I wondered if there were any standards on this.
     
    mechanic, Mar 15, 2014
    #59
  20. You do not know the idiot. He has done this crap before. You'd be
    best to stay out of my conversations with others. Especially since you
    know nothing about his posting history.
     
    DecadentLinuxUserNumeroUno, Mar 15, 2014
    #60
    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.