Results:
Our final product ended up working with our original design. The part that caused the most problems was the photosensors. Initially, a photoresistor was used which just varies the resistance depending on the intensity of the light that is shined on it. It turned out that this method leads to a lot of noise which causes random interrupts to trigger. The best solution was to use a phototransistor, which acts much more like a diode. Instead of varying the resistance, the connection is either open or shorted causing a more distinct ‘0’ or ‘1’ state. This reduced the noise and caused the interrupt to occur consistently.
Another issue we had was the embedded safety mechanisms of the hard drive. The safety mechanism prevents the hard drive from spinning when there is a signal being sent to the actuator coils. This would not allow us to use the actuator to create the alarm sounds and the ticker. We found out that signals being sent to the actuator are only checked when the hard drive boots up, so to get around this problem, we leave the actuator circuit unplugged and then manually push the circuit back in to create a connection. We would then proceed to turn the hard drive on. Once it is started, it is safe to let go and leave the circuit unplugged again.
Conclusion:
We were able to implement our design with our components. The timing of the software and the interactions with the hardware allowed the LED lights to properly display lines corresponding to hands of an analog clock. If we were to do the project again, it would be much easier to use a hard drive with multiple platters. The one we used only had one platter, so there was not much space inside of the hard drive. With multiple platters, we could put the LEDs behind the disk without cutting the back and having to leave support beams. One of the other disks could also be used for reference point sensing. A hard drive without safety mechanisms would have also made the project easier so that we wouldn’t have to implement a manual override.