Don't-care conditions in K-maps: Visual walkthrough

Step 0 — What is even on the grid? (building the picture from zero)

A K-map (Karnaugh map, see Karnaugh Maps (K-maps)) is just a re-drawn truth table where the 16 input combinations become 16 boxes on a $4 \times 4$ grid. Rows are labelled by the pair $A B$ , columns by the pair $C D$ . The magic ordering is Gray order $00, 01, 11, 10$ — chosen so that neighbouring boxes differ in exactly one bit. That single-bit-neighbour property is the whole reason grouping works, and we'll lean on it hard.

Figure — Don't-care conditions in K-maps

Look at the picture: the mint boxes are the real $1$ s, the lavender boxes are the $X$ wildcards, the pale boxes are the locked $0$ s. Nothing is grouped yet. Our job: cover every mint box with as few, as large, groups as possible — borrowing lavender boxes only when it helps.

Step 1 — Read the axis labels: which boxes mean $A = 1$ ?

The red band in the figure is the whole $A = 1$ half. Notice what lives there: two real $1$ s (digits $8, 9$ ) and six $X$ boxes (digits $10$ – $15$ ). That mix is the opportunity.

Step 2 — Grow the biggest group: absorb six don't-cares to get $A$

WHY it's legal (and free): the six lavender boxes were $X$ . Calling them $1$ costs nothing, and it let a group of two real $1$ s balloon into a group of eight → one literal instead of several. This is exactly the parent's Rule 1: absorb an $X$ only when it enlarges a group that already covers a real $1$ . It does — so we do.

Step 3 — The remaining $1$ s: digits $5, 6, 7$ in the $B = 1$ row

The figure highlights digits $5, 6, 7$ (mint, still exposed) and points at the three lavender helpers $13, 14, 15$ directly below $6, 7$ (and $5$ ), which we're about to borrow.

Step 4 — First helper group: ${6, 7, 14, 15}$ gives $B C$

WHY: $B$ and $C$ are both pinned on ( $= 1$ ) throughout the group, while $A$ and $D$ each swing through both values and cancel. So this group of four costs just two literals, $B C$ , and it covers real $1$ s at $6, 7$ .

Step 5 — Second helper group: ${5, 7, 13, 15}$ gives $B D$

WHY this second group and not just one? Digit $5$ ( $= 0101$ ) is not inside the previous group ${6, 7, 14, 15}$ — it has $C = 0$ . So $5$ still needs covering. The group ${5, 7, 13, 15}$ mops it up, sharing the already-covered $7$ (overlap between groups is allowed and free), and it collapses to $B D$ . Notice we again borrowed $X$ s ( $13, 15$ ) purely to reach size-four → two literals instead of a clutter of size-two terms.

Step 6 — Assemble and read off the answer

Step 7 — The degenerate check: what if we had refused the don't-cares?

The figure contrasts the two outcomes side by side: left = don't-cares refused (many small groups, tangled expression), right = don't-cares embraced ( $A + B C + B D$ , three clean groups). Fewer terms means fewer gates — the payoff spelled out in Logic Gates & Hardware Cost.

The one-picture summary

One frame, three coloured groups laid over the finished map — red $A$ swallowing the whole bottom half, coral $B C$ down the $C$ -column pair, butter $B D$ down the $D$ -column pair — with the equation $F = A + B C + B D$ printed beside it. That is the entire derivation compressed to a glance.

Recall Feynman retelling — the whole walkthrough in plain words

We wanted a circuit that lights up when a decimal digit is $5$ or more. We drew 16 boxes, one per 4-bit pattern, arranged so neighbours differ by one bit. The digits we care about ( $5, 6, 7, 8, 9$ ) got a "yes" (1); digits $10$ – $15$ can never show up in decimal, so we wrote $X$ — a wildcard we may read as $0$ or $1$ , whichever is convenient. First we noticed the whole bottom half of the grid means " $A = 1$ ". It held two real yeses ( $8, 9$ ) plus six wildcards — so we called all six wildcards "yes", drew one giant group of eight, and everything except $A$ cancelled: that whole region is just $A$ . Then the leftover yeses ( $5, 6, 7$ ) each got a small $2 \times 2$ group, again borrowing nearby wildcards to reach size four: one gave $B C$ , the other gave $B D$ . Add the three: $F = A + B C + B D$ . If we'd chickened out and called every wildcard $0$ , the expression would've been longer and the circuit more expensive. The one rule that kept us honest: a wildcard may join a gang of yeses to make it bigger, but you never build a gang of only wildcards.

Connections

Karnaugh Maps (K-maps) — the grid and Gray ordering we read positions from.
Sum of Products (SOP) and Product of Sums (POS) — why each group becomes one OR-ed product term.
BCD - Binary Coded Decimal — source of the $10$ – $15$ don't-cares.
Boolean Simplification — the adjacency cancellation that drops variables.
Quine-McCluskey Method — the algorithmic version of this same grouping.
Logic Gates & Hardware Cost — why fewer terms = cheaper hardware.

Recall Quick self-test

Why did the eight-box group collapse to just $A$ ? ::: $A$ is the only variable constant across all eight boxes; $B, C, D$ each take both values and cancel. Which don't-cares did we use, and why? ::: $10$ – $15$ ; borrowing them let groups reach size four/eight, dropping variables. Why not a lone group around $10, 11$ ? ::: It would contain no real $1$ — an all- $X$ group is wasted hardware. Final expression? ::: $F = A + B C + B D$ .