PLEASE HELP - Need advise (thats legal!)

Discussion in 'Wireless Internet' started by Kingz#1, Jun 25, 2004.

  1. Kingz#1

    Kingz#1 Guest

    I am trying to create a multi-point wireless network. The main server
    will be 10km away. I do have LOS. This sight will be having a
    OMNI-DIRECTIONAL antenna about 80ft above the ground (to clear and
    obstuctions). Once of the recieving ends (which is 10km away) will be
    using a 24dBi parabolic Andrew antenna to try and link to the
    OMNI-DIRECTIONAL antenna. Here is our delema.... We first considered
    using a 15.4dBi omni 70inch antenna (from fab-corp.com) with a 1w
    outdoor amp linked to a 200mw Long-Range 802.11b router. After finding
    out that this was ILLEGAL.... (I'm assuming that it is harmful to ones
    health). I would like to ask the experts in this group what is the
    MOST powerful antenna I can use without breaking the law or causing
    any kind of health concerns to prelonged exposure.

    BTW: Here is some more info on the setup.
    Both routers are the same (200mw Long Range 802.11b routers from
    fab-corp.com)
    24 dBi Andrew parabolic antenna (for the recieving end)
    both routers will be using 1ft low loss cable from antenna to router
    Also... from what I've read in this group... a 6dBi omni looks like
    the maximum legal limit for this amp. So my question is.... would the
    6dBi omni with the 1w amp be better than the 15.4dBi omni without the
    amp???? If there is any better combo... please let me know.


    Thanks!
     
    Kingz#1, Jun 25, 2004
    #1
    1. Advertisements

  2. Kingz#1

    David Taylor Guest

    will be 10km away. I do have LOS. This sight will be having a
    Why will one end have an omni directional antenna? I don't follow that,
    you should be using 2 directional antennas for a point to point link.
    I'm not in your country but I think you'll find that 2 24dBi parabolic
    antennas will do the link without any amplification anyway for that
    distance.

    David.
     
    David Taylor, Jun 25, 2004
    #2
    1. Advertisements

  3. Kingz#1

    Kingz#1 Guest

    David,

    I have to repeat myself 1 million times everytime I post........ I
    know 2 directional antennas would be more than enough.... for a point
    to point. But this is NOT a point to point connection. This is a point
    to MULTIPOINT connection....like I said earlier. I MUST (1000%
    certainty) have a omni antenna on one end.


    Thanks.... and please advise.
     
    Kingz#1, Jun 25, 2004
    #3
  4. Most people I know in the business stay away from amps unless it's the
    *only* way to get it to work. Amps amplify noise equally with signal,
    they don't always help. With 200mw radios, a 15dbi omni and a 24dbi
    parabolic, I can hardly imagine you having problems getting a 10km link
    to work unless there is a lot of 2.4GHz noise in the area already or you
    have problems aiming the omni (you do know high-gain omnis have a rather
    flat doughnut shaped radiation pattern, don't you?) Done a site survey yet?
     
    =?ISO-8859-1?Q?R=F4g=EAr?=, Jun 25, 2004
    #4
  5. Kingz#1

    Hans Vlems Guest

    Your combined specifications are difficult to meet: multipoint wireless + 10
    km distance + legal solution.
    I guess that you need to let one spec go to make this work. The legal aspect
    might be tied to health issues,
    though I'd rather think it is a matter of interference. Wireless signals do
    interfere and your strong signal
    will disturb all the weaker ones it finds in its path. Depending on your
    location, you might get away
    with strong signales, as long as nobody else suffers. I live in the
    Netherlands and for sure it would
    be cause for complaints.
    Signal amplification is the last thing I'd pursue. At the omnidirectional
    side, the antenna receives a lot
    of noise and that gets amplified as well. Amplifiers don't change the signal
    to noise ratio, they make
    it even worse.
    Basically the only advice here is to change the multipoint design to a
    collection of point-to-point
    connections. It is very likely that you don't want to hear that, but please
    remember that multipoint
    is equivalent to a a single point of failure. Not a good design, but YMMV,
    Hans
     
    Hans Vlems, Jun 26, 2004
    #5
  6. Kingz#1

    DS Guest

    (Kingz#1) wrote in @posting.google.com:
    The max TX power is 36 dBm. So base your calculation's on that. 200 mW is
    23 dBm. Assuming that the antenna is connected directly to the box, the
    greatest gain antenna would be 13 dBi. Now you have to account for cable
    loss. Cheap coax is high loss, low loss coax is expensive. The coax specs
    will give a loss per foot, or meter.

    If you axe the amp and go with the 15 db omni, and can manage only 2+ dB of
    coax loss, you'll be at max. This can be done by mounting the device in a
    nema box on the mast and using the PoE. I can only assume that you were
    looking at the one titled: Long Range 802.11b Multi-Client Bridge & POE
    Inj.

    Almost always better off not using an amp. Why add cost, another point of
    failure, and more complexity if you can get the same result's with a more
    simple approach.

    Also, while it's true that directional antenna's do focus the RF energy,
    there is still side lobes that vary from design to design. These lobes can
    be used to make successful connection's also as long as the system is
    planned that each subscriber units operates at similar power levels.

    Of couse, I was in at an ISP that had all type's of antenna's & coax's
    available to use to mix and match as necessary, so that made it quite a bit
    easier.......
     
    DS, Jun 26, 2004
    #6
  7. Kingz#1

    David Taylor Guest

    I have to repeat myself 1 million times everytime I post........ I
    Well, it wasn't clear to me, you only mentioned two locations!
    Why? If you have say 3 sites, use two access points in the middle site
    and use two DIRECTIONAL antennas, one pointing to each of the remote
    sites.

    How many other sites are there?

    David.
     
    David Taylor, Jun 26, 2004
    #7
  8. Kingz#1

    David Taylor Guest

    Basically the only advice here is to change the multipoint design to a
    Quite, that's two of us saying the same thing so it must be true! :)

    David.
     
    David Taylor, Jun 26, 2004
    #8
  9. 1rst problem is it looks like you will be using conventional, 802.11 for
    a multi-point system and you are going to get burnt by Hidden Node...
    Although if done correctly, all child sites will hear the parent the
    children will
    not hear eachother, and since 802.11 uses CSMA, they will talk at the same
    time..
    CSMA works by a client listening, and if clear it talks.. if it hears
    another system talking,
    the client backs off for longer and longer periods of time.. i.e.
    remember back in the
    Ethernet HUB days "collisions" This is how 802.11 works.. so by
    amplifiing the
    BSU and putting the SU's with directional antennae, the SU's will not
    hear eachother
    so when they all talk at the same time, from the perspective of the
    parent unit, the signals
    will interfere with eachother and cancel eachother out...

    Amplified systems and highly directional antennae are to be used in 2
    configurations..
    1) a point-to-point environment (reason why so many responses were
    confused in your post)
    2) multi-point system not 802.11... system with Polling... or clocking..
    i.e.
    polling like karlnet, Orinoco OR systems, and even the new crappy
    MP11 by Proxim
    clocking systems.. like the Proxim/WMUX Tsunami Multi-Point etc
    etc which use
    a GPS timed TDD system.

    Legalities: (your true question)
    The max EIGRP (effective power off the antenna) is truly 36dbm, but to
    produce this
    you have to play the db-1 rule to the output power of the radio as you
    keep increasing the
    antenna gain by 3dbi (double the effective power) Legally, you can't
    just take any system
    and do this... unless you want to certify your system radio, cable,
    antenna etc with the FCC.
    There are many already certified systems out there that you can use...
    i.e. if you use the karlnet or Orinoco OR systems, you can put them
    together with
    either a YDI (very noisy), or a Hyperlink AMP with a certain set of
    antennae and they
    are already certified solutions. These AMP's are available as inline
    remote at to be inserted
    right at the antenna side of the feed line with inline power.. that way
    it does not matter
    how much feed line you use... this is important so not to have to
    recertify a solution with a
    different cable length.



    Good luck!
     
    Fresnel Fadermargini, Jun 27, 2004
    #9
  10. Kingz#1

    Ron Bandes Guest

    Wired Ethernet uses CSMA/CD (Carrier Sense, Multiple Access with Collision
    DETECTION) and works as you described. 802.11 used CSMA/CA (CSMA with
    Collision AVOIDANCE) and doesn't suffer from collisions nearly as often as
    wired Ethernet.

    Ron Bandes, CCNP, CTT+, etc.
     
    Ron Bandes, Jun 27, 2004
    #10
  11. Kingz#1

    Hans Vlems Guest

    Saw your post too; obviously a case of great minds thinking alike, right ;-)
    I have no clue why the OP is so sure that an omnidirectional antenna is the
    only solution. OTOH he didn't really tell us what problem he's trying to
    solve.
    So unless the OP provides us with a little moer background this topic is
    closed.

    Hans
     
    Hans Vlems, Jun 27, 2004
    #11
  12. See Below

    Ron, I think you need to do a bit of research on Hidden Node... Hence the
    reason for polling/clocking MP systems.. I have been working in the space
    for 17 years!
    The Collision Avoidance part is where the Client listens first... but if a
    client can not hear
    another client oh well it talks... RTS/CTS was a weak attempt at solving
    this issue..
    It slowed down throughput and only worked for limited # of uers with packets
    within a
    certain size.

    -FF
    Expired CCIE, SMRE (Harris Senior Microwave RF Engineer)
    (since you list your certs :) )
     
    Fresnel Fadermargini, Jun 27, 2004
    #12
  13. Yep. Hidden nodes are a killer. Lots of token passing, round robin,
    polling, sychronous, and scanning type of solutions are available from
    various proprietary (non-802.11) vendors. While none of these do
    anything useful for small systems, large WISP (wireless ISP) systems
    can really benefit from an improved collision
    detection/avoidance/mitigation/bandwidth-allocation/whatever
    mechanism.

    Here's some graphs on the effects of collisions compared with a
    polling scheme.
    http://aqua.comptek.ru/test/HiddenNode/hidden_node_en.html

    Here's an open source attempt at a solution:
    http://frottle.sourceforge.net
    which uses QoS to allocate transmission bandwidth in order to reduce
    collisions.
    Yep. It makes no sense to send flow control packets for small data
    packet sizes. The overhead of the flow control packets is bad enough,
    especially since they're sent at 1Mbit/sec (so ever radio can decode
    them) in 802.11b. The radios have a CTS threshold setting. It just
    sends the data for data packets smaller than the threshold, and does
    the flow control exercise for larger packets. I wouldn't call it
    weak, but it certainly doesn't solve all the collision problems under
    all the circumstanced.
    Ummm... No certifications, memberships, or titles. I never had time
    to collect those and never needed them. Been doing radios in one form
    or other (service, design, marketing, management, politics) since
    about 1971. I added computahs to the mix in about 1981.
     
    Jeff Liebermann, Jun 27, 2004
    #13
  14. Kingz#1

    Dragon5126 Guest

    Your biggest problem isn't legal limits... it's attenuation. What is the
    length of transmission line that you will be using? Assuming a ground level
    location for your WAP, the signal to and from the antenna will be so
    degraded by feed line attenuation it will be un acceptable. Mounting the
    WAP in a weather proof enclosure directly at the antenna's feed point will
    help this issue. Run your network cable to the WAP at the antenna. This has
    worked well for me on numerous commercial installations I have performed. Be
    advised that every country has differing limitations on the ERP (effective
    radiated power). You need to check the regulations in your country ( in the
    United States the relevant document is FCC Part 15) and calculate the ERP of
    your installation, and adjust power accordingly (adjusting feed line length
    to attenuate the sufficient amount of RF is a very practical method).
    point of info: you do not need to use an omni antenna for multipoint
    access... 'stacked' directional antennas would be a better choice. The
    directional antennas should be chosen for beam widths (radiation patterns)
    and mounting so that they act as a single super high gain omni, that will
    increase the reception as well as transmission. There are many commercially
    made antennas that fit this description that would be preferable to a single
    high gain omni
     
    Dragon5126, Jul 3, 2004
    #14
    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.