11 / 36 · Konzept
Test the cases that reveal mistakes
Choose cases that reveal reversed selection and sampling errors.
Lektionen sind frei lesbar. Schreibe dich ein, um deinen Fortschritt zu speichern.
Die Übersetzung ist noch nicht verfügbar. Die Originallektion wird angezeigt. (English)
Learning goals
- Choose inputs that distinguish correct and incorrect circuits
- Keep expected and observed values separate
Avoid tests that let mistakes pass
If a and b always match, a broken selector can appear correct. First vary sel with a=0 and b=1; then use a=1 and b=0 to check the opposite direction.
Calculate expected values before changing the input file. Rewriting expectations after seeing the result defeats the comparison. Keep the original and modified files distinguishable.
| Target | Input condition | Compare |
|---|---|---|
| Select direction | a != b; toggle sel | y |
| Unselected input | Hold selected input; toggle other | y remains stable |
| Storage | Change sel between edges | y versus sampled_y |
| All combinations | 8 combinations of a, b, sel | Truth table |
Start with the first mismatch
Record the first mismatching instant and its input combination. If y is wrong, inspect the selection expression and wiring. If only sampled_y is wrong, check initialization and clock edges first.
Covering eight input combinations does not verify every electrical property of a chip. The goal is to explain the 0/1 selection and storage behavior of this RTL example.
Selbst ausprobieren
y matches the truth table but sampled_y appears one edge off. What should you compare first?
Erklärung lesen
Compare the initial sampled_y, input transition times and first rising edge before changing the selection expression.