Out of 10 people I know who tried them (myself included) only one person has had luck with those Linksys WAP11s. The really scary part is, everyone I know had different problems with them.
Your problem sounds familiar though, nonexistent sustained speeds.
I just re-read your October update; you upped the voltage to send power over some unused wires of the ethernet cable. Um. Forgive me if I'm utterly wrongheaded, but, that cable resistance would lower the amperage, not the voltage, wouldn't it? Also, for a run that long, wouldn't AC be more effective than DC?
If it's underpowered, it would be plausible to get decent short bursts but nothing sustained.
You probably already knew about Robert X. Cringely's efforts to get a long-distance wireless link going, but just in case...
AC would be bad in a POE situation. imagine the interference caused by the 60 hz power a mere milimeter from the data lines. plus then you'd need step-up and step-down transformers since you'd need to shoot the voltage up to several thousand in order to overcome the limited amperage-handling ability of the very very thin cables.
I was thinking of extension cords, not POE.
If you want to check to see if power is a problem, you can use a multimeter to check voltage & amperage.
If you want to know if it makes a difference, you could put a few batteries in series - that _might_ work for long enough to test the bandwidth for a few minutes.
If you need thousands of volts to get enough AC current up, wouldn't the amount of DC be enough to either burn the cable or the electronics at the ends?
Um. Look, this isn't my area of expertise, so I'd keep it simple, because it's harder to mess up.