The goal of this project was to create a simple UART (Universal Asynchronous Receiver-Transmitter) device, which would be able to send digital messages over a serial interface. The standard technique of assigning boolean values to voltage was not used, instead a time-based method was used, i.e. a shorter pulse representing a logical “0”, and a longer pulse representing a logical “1”.

fig. 1 - click on image to enlarge

The mapping of logical values to time instead of voltage has a few unique benefits. Since the receiver’s clock is syncing once per bit instead of once per message, error in baud rate timing is nearly eliminated. The error for the whole message is the same as the error for each bit, instead of compounding for every bit in the message.

Message framing (start/stop bits, error detection) is necessary for voltage-based UART devices to function, yet it is not for time-based transmissions. The goal of this project is not to eliminate these structures however, but to potentially leave them up to the user to configure. For example, error detection would be just as important with a time-based UART, so a user with an especially noisy transmission line could configure the UARTs to include larger error detection information, while an application which prioritizes speed would use less.

fig. 2 - click on image to enlarge

A two wire interface was used (ground, data), as only one receiver and transmitter were assembled for testing. This setup would simply be doubled to allow both sides to send and receive. The top breadboard contains the sender, the second contains the receiver, and the third contains a display, which shows the last eight bits received. The transmission wire can be seen with white and black stripes, bridging the first and second boards on the left side.

fig. 3