Is it possible that the OP wants to use an old-fashioned mechanical
alarm clock as the source? If so, that gives me an excuse for some
nostalgic recollections <g>.
About 20 years ago, chatting to a friend in the office about our
shared habit of early rising, he mentioned that he had to wake his
deaf teenage daughter each morning. She too had to leave early for
work but because of her handicap couldn't use normal alarms like a
clock or radio. That got me exploring some electronic possibilities:
Light - simulating the sun beaming down onto her face. A bright
spotlamp perhaps, or an infra-red lamp of the sort sometimes fitted in
bathrooms. Quickly dismissed as I didn't know enough to be sure of its
safety over prolonged exposure. I did try a few experiments with
flashing a normal domestic 100W spotlamp a couple of feet away but
decided that couldn't be reliably expected to wake her up, even if she
happened to be facing directly towards it.
Breeze - a pleasant stream of air from a bedside fan. Experiment
quickly showed it would need to be a largish fan to have any chance.
But impossible to devise a reliable way of ensuring it stayed in
position, short of strapping it to her head!
Electric Shock - fleetingly toyed with the idea of a (battery
operated!) wrist strap, delivering a 40 or 50V. Really was fleeting
<g>.
Vibration - finally settled on this, and built a gadget which she used
successfully for many months.
A small d.c motor with a lead weight attached eccentrically vibrated
sufficiently when placed under the pillow. Used an old, cheap
mechanical alarm clock with a modification to the striker mechanism so
that it repeatedly opened and closed an electric contact.
The clock was mounted on top of a home made wooden box, and the
ultra-simple electronics on a piece of stripboard. A jack plug and
socket were used to connect the circuit's output signal to the motor
unit via a pair of wires about a metre long. An l.e.d in parallel with
the socket provided a supplementary indication, useful for testing.
Improvised electrical contacts by epoxy gluing a strip of tin inside
the casing, isolated electrically by the glue, so that the hammer
could strike it without hindrance. While the contacts were closed, a
brief positive pulse was delivered to the base of a darlington pair
arranged as an emitter follower. If the plug was in place, then the
motor vibrated. In any case the l.e.d would light. With 3 x 1.5V
D-type batteries, after allowing for the two base-emitter drops this
delivered a little under 3V to the motor. Quiescent current was
negligible, so no additional ON/OFF switch was necessary in practice.
The existing mechanical one on the clock served the purpose. The
motor was mounted in a small plastic case. Finished unit:
http://www.terrypin.dial.pipex.com/Images/DeafAlarm.jpg
I was pleased to hear that it performed very well in practice. But
about six months later it got dropped from a shelf, disintegrating
beyond repair. At least, that's what my friend told me!