External antennae for notebooks and others...

Discussion in 'Wireless Internet' started by John Beeston, Jul 2, 2004.

  1. John Beeston

    dold Guest

    One of the other ones, but the same gang.
    They seemed to have good success. This is a hobby, not a need.
    The closest to need is a friend of mine who lives close enough to a
    commercial hotspot that he can associate to the hotspot, but never gets the
    login web page. He doesn't need much gain, and then he can give up his
    dialup access and phone line.
    I really am exploring what can be done with the mini-USB intact. If I want
    to pull off the ceramic antenna, I might as well start with a full sized
    USB that has an antenna, and follow one of David Taylor's designs.
     
    dold, Jul 5, 2004
    #21
    1. Advertisements

  2. John Beeston

    John Beeston Guest

    I, too, favour the idea of leaving the dongle as is,and simply fitting it to
    reflector... a bit like a bulb in a car headlight ... one wonders whether
    one of those big old shiny headlamps could be used?

    I see concerns about painting / illuminating the reflector. Presumably this
    is from the transmit side of the radio .. Is there an issue if the transmit
    side is enhanced far less than the receive side?

    I guess if there are two such reflector systems in use, the fact that both
    had improved reception would be of benefit as long as the noise levels were
    not increased too much. Presumably the focussing effect of the reflector
    would help reduce any noise not in the direct path?

    The various sales descriptions of these dongles refer to omni-directional
    antennae.... is this true for both transmit and receive? Is there any simple
    way of determining the signal strength pattern on transmit?

    John
     
    John Beeston, Jul 6, 2004
    #22
    1. Advertisements

  3. John Beeston

    Barry Jones Guest

    That sounds like a great idea. Easy to weatherproof, probably a tight
    focus, and you could probably mount it so that you can aim it like you aim
    a car headlight. Let me know how it works if you try it. My time will come
    next summer, when I put up a little hideaway in the back of the property.
    Supposedly the increased signal to noise ratio helps in the other direction
    (actually the receive function of the AP). I would do one side at a time,
    and see what you get.
    If you don't use an external amplifier, you don't have to worry about
    increased noise. Any increased noise using focussed antennae should be in
    proportion to the increased signal. In addition, you'll be blocking out
    noise from other sources.
    I've seen some sites listed in this group that have data on the signal
    strength patterns of various antennae. The omni's typically have a doughnut
    shaped signal strength pattern, oriented horizontally if the antenna is
    vertical. One's with higher dbi's presumably have a flatter doughnut.
    BTW, the above is not a good sig delimiter. It should be
    dash-dash-space-newline. If you can't change the AV's output, you can put a
    working delimiter above it, or above your name.

    Not really interested. We're in a text based world here.
     
    Barry Jones, Jul 6, 2004
    #23
  4. John Beeston

    dold Guest

    My plan is to put the USB dongle on a stick, and rotate it 360 degrees
    while observing NetStumbler against a reference WAP.

    In testing reflectors, I put the "D-Link" label away from the reflector.
    Maybe it would be better if it faced the reflector. Or was edge-on, or
    nose-on.

    Jeff provided a reference to http://www.component.tdk.com/2.4GHAnntena.pdf
    Presuming that the DLink is the 3mm CABPB0730A, there are three patterns.
    The third is apparently the flat orientation. I don't understand the
    distinction between the first two. I see that the graphic of the chip is
    oriented differently, but I don't know what it means in terms of the
    physical layout.
    It also appears that there is a frequency sweet spot that varies depending
    on which graph you are looking at. Just gain appears best for the higher
    channels. The pattern varies by frequnecy as well, not necessarily
    favoring the high channels. Depending on whether you have control of the
    WAP, that might be an important tuning aid.
     
    dold, Jul 6, 2004
    #24
  5. John Beeston

    dold Guest

    I looked at Fry's Electronics. The standard sized USB-802.11b adapters are
    still around $59. The DLink DWL122 is $19 after rebates. A refurb Netgear
    MA111(?) is $29. So the Mini-USB is a lot cheaper than the standard size.
    Ripping it open and soldering a pigtail directly to the board, ending in a
    radiator inside a can would be a very small and cheap version of the USB
    Cantenna http://www.nodomainname.co.uk/cantenna3/cantenna3.htm
     
    dold, Jul 7, 2004
    #25
  6. John Beeston

    John Beeston Guest

    But if a similar effect can be achieved by simply placing said device at the
    focus of a dish ... why get out the tools?

    If you are simply soldering a pigtail to another radiator... have you gained
    anything over relying on the built in radiator (ceramic or whatever)?

    Would the effectiveness of the new radiator, without the can, be better than
    the inbuilt antenna?

    John
     
    John Beeston, Jul 7, 2004
    #26
  7. John Beeston

    dold Guest

    That's a question best answered by empirical study.
    That's possible, even likely. You could attach a 3dBi rubber ducky in
    place of the -1dBi ceramic. (I don't know if it's -1dBi. That's what Jeff
    indicated, but I don't see a spec on the web page for the DWL-122.)

    Then you'd have something that looks like the $99 Linksys.
    http://www.linksys.com/products/product.asp?grid=33&scid=36&prid=612
     
    dold, Jul 7, 2004
    #27
  8. Who said that it works? The ceramic patch antenna sprays RF in almost
    all directions. Some of it illuminates the dish, but the rest goes
    off to places where you don't want it to go (spillover). It's not too
    bad for on receive, but it's hell on transmit where perhaps only 20%
    of the RF coming out of the patch antenna makes it to the dish. Also,
    dishes are at best 50% efficient, so we have more losses involved.
    There's also no guarantee that the dish will have anywhere near the
    theoretical gain defined by its aperature size (basically the diameter
    of the dish sets the theoretical maximum gain). Start reading here:
    http://www.w1ghz.org/antbook/contents.htm
    and be sure to at least read Chapter 4 at:
    http://www.w1ghz.org/antbook/chap4.pdf
    which covers the basics of dish illumination.
    I don't really understand the question. The basics are:
    1. Antennas do not generate any additional signal. They only
    redirect what is available.
    2. Antenna gain is simply redirecting that RF in one direction, at
    the expense of stealing RF that usually would go in another direction.
    3. Antenna gain, efficiency, bandwidth, and beamwidth can all be
    traded for each other. You don't gain in one of these, without losing
    in the others. (Kinda like: good, fast, cheap...pick two).
    Probably. The built in patch antenna is only slightly direction.
    Shoving it inside a can which forces all the RF to spew from the mouth
    of the can, instead of in all directions, is probably a net gain
    assuming that there are no disgusting reflections, vswr from the
    proximity of the metal can, or detuning of something on the radio. It
    also assumes that the surface of the can is low loss material (RF skin
    effect). Lots of assumptions, but a bit of careful construction
    should yield some gain.
     
    Jeff Liebermann, Jul 8, 2004
    #28
  9. Ahem. That's a question I can easily model using one of the NEC2
    tools. My favorite for this month is 4NEC2 version 5.39. Methinks
    TDK and others have NEC2 decks for their patch antennas (somewhere).
    I can use that to simulate the feed, and illuminate the can, dish,
    pizza platter, vegetable sieve, or garbage can of choice and produce a
    suitable model. My extensive experience with empirical (also known as
    cut-n-try or seat-of-the-pants) engineering is that it ALWAYS ends up
    worse than what theory and modelling predicts.
    The ceramic patch antennas vary from -4dBi to +2dBi depending mostly
    on physical size. The really tiny ones found on small USB radios tend
    to be the lower gains. The smaller R-SMA connector antennas are just
    1/2 wave coaxial antennas with a gain of about +1.5dBi (thanks to the
    use of cheezy coax). The slightly longer rubber antennas found on
    Linksys radios are only slightly better due to better coax.
    Incidentaly, these are measurements from my antique test equpiment
    pile and are subject to more errors than I care to admit. However,
    methinks they're close enough for comparison purposes.
    Yep. That antenna is better than the ceramic patch antenna. I
    haven't disected any of the latest USB radios yet, but I suspect that
    there are pads sufficient to mount an SMA or MMCX connector in place
    of the patch antenna, or simply attach a pair of 6.3cm wires to form a
    dipole and keep it simple.
     
    Jeff Liebermann, Jul 8, 2004
    #29
  10. John Beeston

    John S Guest

    Good discussion thread and info!
    Are you saying that a "dipole" works as a dipole without a balanced
    feed?
     
    John S, Jul 12, 2004
    #30
  11. Yep. The purpose of a balun (balanced to unbalanced transformer) is
    to prevent the coax cable from radiating and ruining the pattern. In
    the case of the dipole attached to the USB radio, there's no coax
    cable and no room for a balun. If you're thinking of attaching some
    small coax cable to the board, building a dipole at the end of the
    coax cable, and the perfect pattern is deemed desirable, you can get
    the same effect by simply wrapping the end of the coax shield with
    some lossy material. This will absorb any RF or VSWR that tries to
    use the coax as a re-radiator. I've used 1/4" magnetic recording tape
    for the purpose, but have never bothered to test how well it works (or
    if it works in the first place). Anyway, don't worry about balanced
    vs unbalanced.

    Incidentally, Proxim has the antenna measurements of their popular
    Gold/Silver Proxim/Orinoco/Agere/Avaya/Wavelan/Lucent card at:
    http://www.proxim.com/support/all/orinoco/technotes/antenna_patterns/classic_gold_pc_card.doc
    http://www.geocities.com/lincomatic/orinocoant.html
    The interesting part is the note that the reference point in the
    measurement of -26dB is equal to an antenna gain of 0dBi. Looking at
    the pattern, my guess is that the measured gain varies between 0dBi
    and -15dBi. (A simple dipole is +2.15dBi). And this is one of the
    better performing radios...
     
    Jeff Liebermann, Jul 12, 2004
    #31
    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.