Consider that you may have a card putting 15dBm (32mW) into a 2.2dBi antenna (on the AP), and thatyour card probably has an effective 2dBi gain antenna on it. Operating at 2.4GHz, at 1m, thereis 41dB of path loss between the two units. At 2m, you'll have another 6dB of path loss, for 47dB.So the signal leaves at 17dBm (from the AP), you lose 41dB due to LOS (Line of Sight) path lossat 1m, and the signal arrives at the antenna in your laptop at perhaps -24dBm, and is raisedto -22dBm by the antenna.This is quite close to the maximum signal level of 802.11g.
The IEEE standards state for 802.11b:
18.104.22.168 Receiver maximum input level
The receiver shall provide a maximum FER of 8×10–2 at a PSDU length of 1024 octets for a maximuminput level of –10 dBm measured at the antenna. This FER shall be speciﬁed for 11 Mbit/s CCKmodulation.
and for 802.11g:
19.1.2 Operational modes.
The changes to the base standard required to implement the ERP are summarized as follows:
1) The PHY uses the capabilities of Clause 18 with the following exceptions:
iii) The maximum input signal level (see 22.214.171.124) is –20 dBm.
1) The PHY uses the capabilities of Clause 17 with the following exceptions:
iv) The maximum input signal level (see 126.96.36.199) is -20 dBm.
Remember that, as above, your signal level is likely as high as -22dBm, as seen by the 'other'receiver.
Perhaps your AP puts up more than 15dBm transmit power. In any case, you're too close.
“quarter power” is likely down 6 dBm, and this probably allows the power level reaching theother receiver to be low enough that it doesn't saturate the front end, even with the attenuationpad(s) used.