A hazard-cover term is a product term (or, in POS, a sum factor) you add on purpose even though the function doesn't need it. Its job is timing, not logic: it holds the output steady while another input hands the output over from one term to another.
For a static 1 hazard in SOP form, take the two terms involved in the hand-over, say and . Their consensus is : drop the variable that appears both plain and complemented, and AND what's left. By the consensus theorem, = , so the function is unchanged. But while Y = Z = 1, the term stays 1 no matter what X does, so the output can't dip.
On a karnaugh map, the cover term is simply the group covering the two adjacent 1s that sat in different groups.
For a static 0 hazard in POS form, the cover is the dual: the consensus factor.
This is one of the few places where the minimal expression isn't the best circuit. The cover term costs a gate, and a minimizer would remove it, so mark it as intentional.
| A\BC | 00 | 01 | 11 | 10 |
|---|---|---|---|---|
| 0 | 0m0 | 1m1 | 0m3 | 0m2 |
| 1 | 0m4 | 1m5 | 1m7 | 1m6 |