I checked the circuit today in the labs. I used the lab power supply and got the same problem, but then with the capacitor attached it was fine. I couldn't check the wall wart because I didn't have the right plug, so ill check it tomorrow. I think the problem is noise (thanks Michael) and that the wall wart just has more than the lab supply and more than a single 100uf cap can filter out. I looked at the circuit supply's spectrum and there seems to be (apart from the high DC component) two spikes, one at ~6 and ~10 Khz, I'm thinking that because the pic works at high freq (8mhz) this might be the cause of the problem, but I don't really know. Ill try use a RC filter, maybe that will help.
Thaks a lot for the help guys