The parasitic elements of a Yagi-Uda array do not need to be grounded (earthed) in the middle. The boom need not even be conductive. At the frequency 802.11G operates, a yagi could even be etched on a piece of single side printed circuit board.  You might consider an etched Log Periodic Dipole Array (LPDA) since gain is your objective and not necessarily F/B.

It would indeed be possible to couple to your WLAN card with an external dipole, but I think I would try a full wave loop.  At 2.43 GHz, you would need a loop 39.29mm in diameter.  This also could easily be etched on a piece of single sided Printed Circuit Board.  it could be cable tied, velcro'd glued or taped to your Belkin 802.11 card.  To find the best place, try using a spectrum analyzer hooked to the output of the loop.

Good luck and please post how your experiment works out.

I remain,

The Old Soldering Gunslinger

Thanks for your thoughts and ideas.

The idea of measuring the Belkin’s output in order to establish the optimum coupling is obvious from reciprocity but apparently not obvious when looking at the problem from my kitchen! I might try that if I get a few minutes spare at work.

Trying to couple energy off your card and adding a passive antenna probably won't get you much. You're coupling loss would be too much to overcome any gain increase. You could try to improve your antennas radiation pattern, but that would be awkward.

Can you tell if you just need to transmit more signal, or is your received signal also too weak.

If you only need to transmit more energy, then making an antenna/amplifier/antenna setup would boost your transmit power (possibly illegally) and you could transmit further. Coupling to your present antenna at -20 dB is reasonable, then add 30-40 dB gain amplifier and another antenna separated from the notebook antenna (to prevent oscillation). Look for mini-circuits amps, they're cheap and battery operated.

I think a passive system ought to work, but I am having no success to date. A corner reflector should give a reasonable gain so I made one out of aluminium foil , stiffened by cardboard from a cereal box. Good experimenters tools! I put this around the lump sticking out of the side of the laptop case (the plug in card has a noticeable lump on the end). Unfortunately the Belkin software doesn’t update the signal strength in real time so tuning for optimum signal strength is all but impossible. There is a nearby network to receive but the signal is so weak I can only receive it when in my loft! Not an ideal tuning environment. The Net Stumbler software also fails to update the signal value correctly which is also a pity.

I think the antenna is a diversity pcb antenna, meaning there are two polarisations and the system picks the one with the biggest signal.

I think a passive system ought to work, but I am having no success to date. A corner reflector should give a reasonable gain so I made one out of aluminium foil , stiffened by cardboard from a cereal box. Good experimenters tools! I put this around the lump sticking out of the side of the laptop case (the plug in card has a noticeable lump on the end). Unfortunately the Belkin software doesn’t update the signal strength in real time so tuning for optimum signal strength is all but impossible. There is a nearby network to receive but the signal is so weak I can only receive it when in my loft! Not an ideal tuning environment. The Net Stumbler software also fails to update the signal value correctly which is also a pity.

I think the antenna is a diversity pcb antenna, meaning there are two polarisations and the system picks the one with the biggest signal.

I am not sure if the Soldering Gunslinger’s reciprocal antenna tuning idea is workable because it is not clear if the antenna on the laptop is transmitting when searching for a network. Certainly the power level being emitted is low (on average) as seen on my TriField meter, or rather as nothing is seen on it. Probably the transmit burst, if any, is a very short burst at best, and therefore not detectable on a long integration time power detector.

This is a pretty cool project.  If you want to invest an a wireless card where you can hook an external antenna up, then I would try putting a powered RF pre-amp at your frequency in beween an antenna and the card.  Remember that you are going to amplify not just the signal but the whole noise-floor as well, so you might need some filters in there.

An amplifier would mean that I could listen but not talk, which is less than ideal. However I could borrow a microwave amplifier from work to get some sort of signal to play with in terms of coupling it into the card.

may I add a comment that increasing the wifi signal level is illegal in most countries,as this interferes with other users in this band, and would reccomend a change in antenna instead.

Good point Johnor,
I believe that there is a rule that each time you increase antenna gain 6 dB, it requires 3 dB reduction in transmit power, that way you can have extended range with narrow field of view from your antennas. I'm certain there would be an upper limit for Effective Radiated Power (ERP) = Power x Antenna Gain also.
Check FCC guidlines for the details.
kch