Samuel Buckstein edited COMMON.tex  about 10 years ago

Commit id: 4a2a880abbf1f486eb8d4d23b14533d10476ef3b

deletions | additions      

       

It was at this time that it was discovered that the holes cut for the light sensors were misaligned with the position of the candlelights in the tray, and as a result the photodiodes were too far from the candlelights to detect a strong enough signal for the PIC to interpret.   Furthermore, the IR detection system was interfering with the signal being sent to the PIC to detect flickering by saturating the photodiode photodiodes  with infrared light. It was decided to dismount the sensor array, dismount all the circuitry on the sensor array, cut new holes for the photodiodes so that they would be located directly above the candlelights, and also separate the IR emitters and recievers from the yellow photodiodes. It was also discovered that the inputs to the multiplexers had been connected in the opposite order in error, and more serious, the PIC was unable to recieve any input except from the first well. The muxes were tested on a protoboard and found to work fine. In between dismounting the circuitry from the sensor array, cutting new holes and remounting the boards, the mux boards were inspected, cleaned and repaired where necessary.   This also provided an opportunity to rethink the wiring of the sensor array so that the intereference issues discovered because of unequal grounds could be solved. This was achieved by using metal tape to restrain wires in neat paths and by keeping the grounds of the IR and yellow light detection systems separate.   The revised sensor array solved the issue of unequal grounds, but even at distance, the IR light was being detected by the photodiodes used for flickering detection.   It was then decided to remove the IR presence detection system entirely, and instead use the photodiode to detect both presence and flickering.   This was accomplished by again removing the sensor array and mounting a series of super-bright LEDs beneath the tray, one LED per well. The LEDs provided a gradient of light that allowed the PIC to distinguish whether a candlelight was present or not beneath the light sensor by comparing the voltage seen by the PIC when no candlelight obstructed the LED from the light sensor against the lower voltage seen when a candlelight obstructed the light path. The LEDs were mounted on the switch activation frame.  This also provided an opportunity to rethink the wiring of the sensor array so that the intereference issues discovered because of unequal grounds could be solved. This was achieved by using metal tape to restrain wires in neat paths and by keeping the grounds of the IR and yellow light detection systems separate.