A few weeks ago I saw an interesting Sonic Bomb Alarm Clock over at ThinkGeek. Being the ever aspiring young hacker, I decided that it wouldn’t be terribly difficult to implement something similar on my own.
Finding an alarm clock was easy enough. I snatched up an AC powered alarm clock for $5. Surprisingly the store offered a cheaper model that included AM/FM radio, but I opted for the basic clock for the sake of simplicity.
Deciding how to implement the “bed shaker” was decidedly more difficult. I uncovered two old-school Nintendo64 rumble paks. They’re a little bulky, but I decided they could do the job. What’s more, they have their own 2xAA power source, so I wouldn’t pull any power away from the clock. I probed the rumble pak’s board for a while and eventually found two solder joints that, when shorted, caused the rumble pak to, for lack of a better word, rumble.
At some point I shelved the idea of using the rumble paks in favor of a more professional approach. I decided tap into the power of the clock directly, hoping that a parallel circuit would be capable of driving enough amperage to power the clock and two small DC motors simultaneously.
I poured over the alarm clock for quite a while. I couldn’t find a timing device on the clock, so I made the assumption that the clock was using the alternating current as a timebase. I was discouraged because I know very little about AC circuits other than apparently it’s pretty dangerous. I sat down with the clock again today for another look. I was pretty confident that there would be DC power somewhere in the circuit to provide power for the IC chip. As soon as I sat down, I noticed the three diodes located near the incoming AC voltage. Sure enough, I had found a steady 12VDC power source.
My plan was to use a simple NPN transistor in series with the motor, and a capacitor to smooth out the voltage. I wagered that I could jumper a lead from the buzzer and siphon enough voltage to activate the base of the transistor. If all went according to plan, the vibrations would be synchronized with the sound of the alarm.
Of course nothing works out as originally planned. Even the tiny motor that I used for testing pulled too much current from the clock circuit, causing it to reset each time the alarm sounded. I tried using various resistors, ranging form 5k ohm to around 470 ohm to no effect (in retrospect I should have slapped a potentiometer in the circuit for testing). The power source could only output enough amperage to power either the clock or the motor, not both.
I was still concerned that my concept may be flawed, so I logged onto the ##electronics channel on the Freenode IRC server and asked the professionals there for some assistance. To my surprise, they informed me that my concept was indeed valid, and they confirmed my suspicions that the power supply simply couldn’t provide enough amperage to power both circuits simultaneously. The nice folks there gave me some good information about DC motors, how they actually require a relatively large power surge to get started, and recommended that I place a power diode between the motor and the transistor to prevent the fluctuating currents from the motor from destroying my transistor.
Overall, even though I wasn’t able to complete the project tonight, I’m still very pleased with my progress. I have several distinct options to proceed from this point:
- Return to my original concept using the rumble pak’s with their self-contained power supplies.
- Attempt to locate a power source that will provide 12VDC with more amperage.
- Wait for my Digital Alarm Clock Kit to arrive, which has a 12VDC relay already built-in.
I’ll try to get some pictures posted on the imagitronics project site of tonights festivities.
