Problem is solved... I was using the internal clock, but I didn't think that would stop me from getting a 2400 baudrate....but I found that if I disconnect the cable for the port and jiggle it against the header pins, I get bursts of garbage so I'm sure it would work with a crystal.