It seems like a better option might be to fix this problem that the wlan uses the drive. You can set /proc/sys/vm/block_dump to 1 which will cause all disk accesss to be logged via klogd/syslogd; then you'll be able to figure out what caused the drive to spin back up (which file was accessed). Most likely it's some silly file that should be moved to tmpfs.