The row-and-column approach is certainly the most convenient, but if you have a keypad with indiviual switches for you to connect, it is sometimes convenient to arrange them in a specific configuration. WIth an ASCII keyboard, just as an example, it is, with considerable thinking, possible to arrange the key matrix in such a way that the code returned when a key closure is sensed, reflects the ASCII code. With the advent of cheap microcontrollers, it became the more common practice to make the key matrix a simple row-and-column arrangement and determine the output value by means of a table-lookup scheme. The PC keyboard, for example, uses a muli-layered table lookup to encode, and subsequently to decode, the keyboard.

With a 4x4 keypad, using a micrcocontroller to encode the keypad, it is easiest to organize the switches such that you have four rows in which all the switches common to that row have corresponding terminals connected to a single "row" wire. The remaining contact on each switch is then assigned to a column, of which there will also be four members. The result is that you have eight wires, and whenever one of the row wires is connected to one of the column wires, it identifies which of the 16 switches is currently closed. When no switch is closed, there are no connections between rows and columns in the matrix.

RE