Name the cause of four failures of domino gates, each taken from a SPICE run, from a numbered list of causes.
Hardware context: A chain of domino gates in a fast datapath, clocked by one precharge and evaluate clock. Find why a domino stage evaluated to the wrong value from its node voltages.
A block of domino logic fails in four ways. Each report below comes from an ngspice run on SKY130. Your job is to name the cause of each one before anyone changes the circuit.
This is a Debug problem. For each report, write the number of its cause from the cause bank. Some causes fit no report, and one number is used twice in the worked reports.
The check reads each output and reports the first one that is wrong.
Signals in the Timeline
| Signal | Width | Role |
|---|---|---|
item | 8 | input of each check |
Full specification
| Number | Cause |
|---|---|
| 0 | No fault: the gate behaved as it should. |
| 1 | An input was high during precharge in a footless gate. |
| 2 | The evaluate phase is shorter than the gate's evaluate delay. |
| 3 | Charge sharing: the dynamic node gave charge to inner nodes emptied in an earlier cycle. |
| 4 | An input fell during evaluate (the input is not monotonic). |
| 5 | No keeper: nothing restores the dynamic node after a disturbance. |
| Report | What the run showed | Cause |
|---|---|---|
| W (worked) | A two-input domino AND needs 60.8 ps from the clock edge to a rising output. A faster clock gives it an evaluate phase of 40 ps, and the next stage reads 0 when it should read 1. | 2 |
| X (worked) | A footed domino gate has its input high all through precharge. At the end of precharge its dynamic node is at 1.800 V, and the supply draws 86.6 nA. | 0 |
| Y (worked) | A three-input domino AND with a keeper. Its top input rises late in evaluate while the bottom input is 0. The dynamic node dips to 0.893 V and climbs back to 1.581 V. The output peaks at 0.145 V. | 0 |
| A | Stage 2 of a domino chain takes its input from stage 1's output through one more inverter. Stage 1 evaluates to 1. Stage 2's input ends at 0, but its dynamic node is emptied and its output reads 1.8 V. | ? |
| B | A three-input domino AND without a keeper. The previous cycle emptied the inner nodes of the stack. Now b = 1 and c = 0, and a rises late in evaluate. The dynamic node falls from 1.8 V to 0.795 V, with no path to ground. What pulled it down? | ? |
| C | The same gate as in report B. After the drop the dynamic node is still at 0.822 V at the end of evaluate, and the output stays wrong. Why does the node not come back up? | ? |
| D | A domino gate with no foot transistor. A static gate drives its input, and that input is 1 while the clock is 0. At the end of precharge the dynamic node is at 74 mV and the supply draws 180.5 µA. | ? |
Every gate has a precharge PMOS of 1 µm, an NMOS stack and foot of 2 µm, and an output inverter. A keeper, where there is one, is a PMOS of 0.42 µm by 1 µm with its gate at the output.
Examples
Example 1
Worked Report W- Input
report W
- Output
example_w = 2
The gate needs 60.8 ps to evaluate, and the clock gives it 40 ps. Nothing in the gate is wrong: the timing is. Cause 2.
Example 2
Worked Report X- Input
report X
- Output
example_x = 0
The foot transistor is off during precharge, so the high input cannot reach ground. The node reaches VDD and only leakage flows. No fault: 0.
Example 3
Worked Report Y- Input
report Y
- Output
example_y = 0
The dip is real, but the keeper restores the node and the output never comes near the switching point. No fault: 0.
Constraints and Contract
- Keep the module name answer and the names of its outputs.
- Every output is one number from 0 to 5.
Hints
Hint 1
Ask first whether the gate failed at all. A dip that recovers before the output flips is not a fault.
Hint 2
Reports B and C describe one event. One cause moved the node; another explains why it stayed moved.
Hint 3
For report A, follow the input of stage 2 through the extra inverter. Which way does it move during evaluate?
What the grader checks: The check reads seven outputs and compares each with the cause numbers set in JavaScript. Three of them are worked reports. Every number in the reports comes from a run of lesson G9.11 (data/G9-11-pins.mjs). Fixing the circuits is outside this problem.
7 checks on Icarus · browser samples / server checks. Run the examples first, then check the full regression. Full checks use the server judge. Downloads contain public samples only. A passing result establishes the tested cases, not a proof for all possible inputs.
See all 1 steps and the final capstone