Source

Adapted from my own handwritten and hand-typed notes from Year 1 to Year 3 of my EEE degree

Meaning

Sometimes a function is only specified for some input combinations, either because those inputs can never happen or because the output does not matter for them. These rows are don't-care conditions, written as x, X or d.

Notation:

means for rows 1, 2, 5, is undefined for rows 4 and 6, and elsewhere.

Using Don't-Cares on a K-Map

Enter an x in each don't-care square. When grouping, each x may be treated as 1 or 0, whichever helps:

  • take x = 1 only when it makes a group larger (fewer literals)
  • take x = 0 otherwise and simply leave it out of every group
  • never make a group that contains only don't-cares

Rule of thumb: don't care about the don't-cares first. Group the real 1s, then use x squares only to enlarge those groups.

Worked Example

:

00011110
001xx
11001
  • with the x at gives (x at 110 used as 1)
  • with gives
  • the x at 100 is not needed, so it is taken as 0

4-variable example:

Solving a Word Problem

Example: and are 2-bit numbers and when they are equal. The 16-row truth table has only for rows 0000, 0101, 1010, 1111. None of these 1s are adjacent on the K-map, and the result is

where is XNOR. This is the same idea used in magnitude comparators.

Other Minimization Methods

For more than 4 or 5 variables a K-map becomes hard to use. Alternatives are the Variable-Entered Map (VEM) and the tabular Quine-McCluskey method, which follow the same prime implicant idea as Karnaugh Map Minimization.