Connect with us

GSM modem antenna termination - RF query

Discussion in 'Electronic Design' started by rob, Jul 11, 2006.

Scroll to continue with content
  1. rob

    rob Guest

    We've got some Intercel GSM / GPRS modems in one of our products.

    The modems have a couple of solder pads for connecting to the antenna via
    coax. We've been playing around with different termination routes and seeing
    different power consumptions and were wondering how this may relate to the
    matching of the antenna connection. We are using a good quality Belden
    RG-178 cable.

    We are terminating as recommended in the manuf. data sheets and it all works
    ok however we find the variation in power consumption, estimated at ~ 50%

    Our layout is:

    modem |- - - - - - - - - - - - - - -|------===============
    |------------------------ |====/

    | coax 12 -->80mm | SMA | 30mm stubby antenna

    Is varying the interconnecting coax length between say 12mm and 80mm likely
    to significantly change antenna matching - radiated pwr / rcv signal??

    Any advice for some RF dummies is welcome!

  2. PeteS

    PeteS Guest

    Rob wrote:
    Is varying the interconnecting coax length between say 12mm and 80mm
    to significantly change antenna matching - radiated pwr / rcv signal??

    Certainly, if the RF track from the modem is not impedance controlled.
    If it's not impedance controlled, then you'll set up a nice big
    mismatch, and the effective gain/loss will depend on the number of 1/4
    wavelengths from the coax termination point (plus some unknown offset
    for the actual electrical length of your tracks).

    In addition, RG178 is rather lossy, but the distance you are talking
    about would not give you the ratio you are seeing.

    The most probable cause is mismatch on the tracks and the coax
    connection, imo.


  3. rick H

    rick H Guest

    It may be nothing to do with the load; are you testing your GSM modem and
    antenna into a test-set or on-air? If you're not using a test-set, in
    which the transmitted power is under your control, then the GSM module's
    transmit power is under the control of the network, and you will indeed
    see the current consumption go up and down depending on what power level
    the network has instructed your GSM module to transmit at.
  4. I'm not sure how it involves your question, but here is a comment:

    This sort of antenna (center conductor extended out of coax) does not
    limit the external RF current to only the exposed center conductor.
    There will also be external RF currents along the outside of the coax
    shield (as if it were part of a dipole fed from the point where the
    stub exits) as the internal coax shield current escapes and makes its
    way to the outside of the shield. This makes the length and shape of
    the coax (and the rest of the modem and its interconnecting wiring)
    part of the radiating structure.

    You could almost eliminate this effect if you added a half wave
    diameter disk to he SMA connector or a few 1/4 wave radials.

    Or, if you can find a ferrite that is effective at this frequency, add
    a bead 1/4 wave back from the SMA connector to bounce the external
    shield current back to the connector at about a resonant length, so
    that this current uses only an intended part of the coax to radiate.
  5. PeteS

    PeteS Guest

    The amusing thing is, the better the signal, the lower the transmit

    In place, do a series of AT+CSQ commands (signal strength at the
    receiver). I say do a series because the cell system signal strengths
    can easily vary by 20dB or more (I see it regularly) so you need to
    capture a representative set of data for each configuration.

    Note that CSQ is related to receiver signal strength, not transmitted


  6. Chris Jones

    Chris Jones Guest

    Do you have a network analyser? If you are making your own antenna then you
    should check it on a network analyser. Perhaps you can borrow one or visit
    someone who has one. Note that things near the antenna will affect the
    matching. It may not be easy to make an antenna that is matched well in
    both the 900 and 1800 bands (or US equivalent if that's where you are).
    Perhaps you would be better off buying the antenna but I don't know what
    they cost in small quantity.

  7. rob

    rob Guest

    Thanks Chris, we are using an off the shelf antenna. Unfortunately we don't
    have access to any radiocomms test gear (or the expertise to use it).
  8. rob

    rob Guest

    Thanks for the comments guys, we do check the rcv signal strength. As you
    said it does move about quite a bit. We test the units on air as we don't
    have any RF test gear. We think/know the modem self regulate its output
    power depending on the efficacy of the antenna termination. I'm not sure if
    this is related to or the same thing as setting o/p power in response to
    feedback from the tower.
  9. rick H

    rick H Guest

    Hi, Rob.

    GSM handsets/modules aren't free to self-regulate their own output
    power - they always set the transmit power to the level dictated by the
    network. The network can look at the signal received from your
    transmitter and decide whether it needs more power or can accept less
    power from you.
    The "self-regulation" you talk about is probably the module's
    power-control loop, which is a circuit on the module to regulate the power
    to the level required by the network, but this is just a feedback loop,
    and has nothing to do with an autonomous decision made by the module as
    to what power to transmit at.

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.
Electronics Point Logo
Continue to site
Quote of the day