Just another post to say I agree with the switch bounce problem.

I did a experiment years ago with a transistor driving a mechanical relay (of course, don't forget the anti-parallel diode to protect the transistor). I sampled the voltage across the relay with a Digital Scope, and found how long the bounce can be, some milliseconds going back and forth between 0V and +5V, like a 'spring'. It is a nice experiment to try and will open many eyes, but you need a DSO.

Regards,
Txip.