On Sun, 11 Jul 2004 21:28:06 -0700, Hamish McStill <(E-Mail Removed)>
wrote:
>On Thu, 08 Jul 2004 00:37:44 -0700, Jeff Liebermann
><(E-Mail Removed)> wrote:
>>For a PCMCIA card, you want a big fat antenna that protrudes. That's
>>because the thin, low profile, slightly protuding type of antenna are
>>invariably half wave dipole traces on G10 circuit board material. The
>>dipole is "folded" which reduces its gain. G10 is also very lossy at
>>2.4Ghz. Basically, the circuit board antennas are cheap and suck.
>>
>>For an example of one of these antennas, see photo at:
>> http://jeffl.ihwy.com/linksys/wpc11/ls-pcmcia-0.jpg (120K)
>>The two hook shaped traces are the two elements of the dipole.
>Is it possible to attach an external antenna to the WPC11 without
>destroying it?
With my web site of "LearnByDestroying", you ask such a question?
Enlightenment comes only after much suffering and the sacrifice of
hardware.
However, I'm working on the method of doing this without rampant
destruction right now. It's the same technique that I use to couple
an external antenna to a GPS (1575Mhz) that has no external antenna
connector. Basically, it's a full wave loop antenna, wrapped around
the PCMCIA card antenna, coupled to a coax cable, when ends up
connected to an external antenna.
It works well with a GPS because the external antenna has +25dB of
gain making the antenna amplifier the prime determination of
sensitivity. A little loss at the coupling isn't going to make much
difference in a GPS. Unfortunately, I don't have this advantage with
a 2.4GHz wireless card. I have to couple efficiently. Resonant
coupling is tricky because the resultant tuning would be rather
mechanically critical.
So far, the best I've done is turn the meandering line antenna (MLA)
or PIFA (patched inverse "F" antenna) into part of a big coupling
loop. It's a flat 3mm wide, 12.5cm long, flat copper strip wrapped
around the PCMCIA card. It's amazing and pure luck that the length of
a full wave loop fits perfectly around a PCMCIA card. I've been able
to couple +4dBm out to my spectrum analyzer, which is about -8dB
coupling loss (ignoring about -2dB loss in the coax cable and
connectors). Not very good, but better than no antenna at all.
Most of the loss is in the impedance mismatch between the very high
loop impedance and the 50ohm coax impedance. I'm playing with various
loop configurations and matching schemes right now. If I can get -3dB
coupling loss, I'll consider it done. Incidentally, I decided that I
couldn't use a patch antenna for coupling because over half of the RF
from the card will radiate away from the patch an not be coupled to
the coax. The loop goes all the way around the card and should grab
all the RF.
In theory, the best will be a 12.5cm circumference loop, in the form
of a solid flat wire ring, with one point designated as the ground
point (where the coax shield is attached), and some point around the
circumference as the 50ohm tap, with a tuning capacitor in series from
the tap to the coax center conductor (to tune out the inductance of
the exposed center conductor). This is a gamma match which is
disgustingly narrow band and seriously critical in tuning. I've had
the loss down to -2dB but found the construction to be so critical
that simply moving anything will detune the gamma match. Obviously,
this isn't gonna work at a coffee shop.
I've given up for the day. Photos, when I have something worth
photographing. If you wanna tinker, just take a 12.5cm long wire and
solder the ends to a piece of 50ohm coax. Make sure that the 12.5cm
length includes the exposed part of the coax cable. Bend the loop
around the antenna part of the PCMCIA card. Position the coax at the
edge of the card, not the top or bottom to prevent coupling directly
to the coax. It won't be fabulous coupling, but it will be better
than no external antenna.
--
Jeff Liebermann
(E-Mail Removed)
150 Felker St #D
http://www.LearnByDestroying.com
Santa Cruz CA 95060 AE6KS 831-336-2558