Our results were very successful. We were able to build a working fire alarm system much like the ones in large buildings.

Safety was a big concern of ours. We used no more than 5 volts in our system at a safe current ensuring that electrocution was not a concern. Additionally, there were no mechanical or other devices which could provide harm to the user. The system was designed with several fail-safes so that a zone disconnection would be noted by the user. The user should therefore be aware if the system is rendered useless and a building is without adequate fire detection. Additionally, in the event of an alarm, the user interface beeps until the alarm is acknowledged. This feature could be very useful in a busy dispatch room where only a visual notification would go unnoticed.

Interference with other projects was not really an our fire alarm system. There are no wireless circuits and the circuit was at a low DC voltage, the interference with other projects is probably less than the interference caused by the fluorescent room lights.

Along the same idea of safety and interference, an actual alarm could have been implemented without very much difficulty (via standard buzzers/sirens available), however, we chose not to for the sanity of those around us. An audible alarm was a concern presented to us by several TAs.

Our clocking code was found to be very accurate. Theoretically, the accuracy of our design is dependant on the accuracy of the microcontroller interrupts. In testing, the time was found to be accurate when compared with a clock.

The system was designed with usability as a top priority. The VT100 specification allowed for accurate text placement. In addition, features such as flashing that the system was in alarm, along with beeping the console, provided for increased visibility to the user during alarm states. The user interface uses numbers and the system is customizable by the user so that zone and system names are easily read and understood by the user.

One thing that was tried in the development stage was to use a 24VDC detector system using detectors obtained by Cornell's Environmental Safety (please see acknowledgements). Unfortunately, due to a lack of documentation, this was abandoned and the detector circuits were built from scratch using concepts similar to the commercially produced smoke detectors.