Physical Design · All levels
Signal Integrity and Variation Tricky Q&A
20+ senior Signal Integrity and Variation interview questions.
Q&A bank
Answer with mechanism, pitfall, regression check, and release judgment. At 10+ years, a correct definition is not enough.
Senior answer rubric
Start with the failing metric and analysis context.
Explain the physical or constraint mechanism.
Name the cheapest evidence-gathering experiment.
Choose a bounded fix and state what it can regress.
Close with signoff, waiver, or escalation criteria.
What is delta delay in SI context?
[INT][PD]
Q: What is delta delay in SI context?
A:
Change in victim net delay caused by coupling activity from nearby aggressors.
FOLLOW-UP TRAP: Defining it as base RC delay only.Why can SI reorder critical paths?
[INT][PD]
Q: Why can SI reorder critical paths?
A:
Aggressor patterns differ per path, so crosstalk shifts delay unevenly across candidates.
FOLLOW-UP TRAP: Assuming pre-SI top path stays top.How do you triage SI hotspots quickly?
[INT][PD]
Q: How do you triage SI hotspots quickly?
A:
Cluster by shared victims/aggressors and route window, then fix the highest-impact cluster first.
FOLLOW-UP TRAP: Treating each violation independently.What is the difference between noise glitch and delta delay?
[INT][PD]
Q: What is the difference between noise glitch and delta delay?
A:
Delta delay affects timing arrival; glitch concerns transient logic integrity on victims.
FOLLOW-UP TRAP: Using one metric for both.Why is pulse width important in glitch analysis?
[INT][PD]
Q: Why is pulse width important in glitch analysis?
A:
Victim cells filter short pulses; harmful glitches need both amplitude and width context.
FOLLOW-UP TRAP: Checking peak noise only.When is shielding preferred over spacing?
[INT][PD]
Q: When is shielding preferred over spacing?
A:
When space is limited and critical victims need deterministic coupling isolation.
FOLLOW-UP TRAP: Shielding every net by default.How does AOCV improve over flat OCV?
[INT][PD]
Q: How does AOCV improve over flat OCV?
A:
AOCV varies derates by path depth/distance, reducing excessive pessimism on some paths.
FOLLOW-UP TRAP: Calling it just another fixed derate.What does POCV add conceptually?
[INT][PD]
Q: What does POCV add conceptually?
A:
Statistical variation modeling with sigma-based assumptions for more realistic spread treatment.
FOLLOW-UP TRAP: Treating POCV as random runtime option.Why must derate policy be centralized?
[INT][PD]
Q: Why must derate policy be centralized?
A:
Inconsistent derates across teams break comparability and create integration disputes.
FOLLOW-UP TRAP: Letting each block pick its own margins.How do you validate a derate change?
[INT][PD]
Q: How do you validate a derate change?
A:
Run controlled before/after regressions across required modes and compare violation movement by group.
FOLLOW-UP TRAP: Changing derates without impact analysis.What route actions typically reduce SI delay harm?
[INT][PD]
Q: What route actions typically reduce SI delay harm?
A:
Spacing, shielding, layer reassignment, selective NDR, and aggressor control.
FOLLOW-UP TRAP: Only resizing endpoint cells.Why can SI fixes hurt congestion?
[INT][PD]
Q: Why can SI fixes hurt congestion?
A:
Extra spacing and shields consume tracks, potentially creating local route overflow.
FOLLOW-UP TRAP: Assuming SI fixes are free.How do you decide if SI is signoff-ready?
[INT][PD]
Q: How do you decide if SI is signoff-ready?
A:
Top victims meet delay/noise limits across required corners with stable trend and no unresolved hotspots.
FOLLOW-UP TRAP: One green run is enough.What is a common SI closure anti-pattern?
[INT][PD]
Q: What is a common SI closure anti-pattern?
A:
Disabling SI analysis late to recover schedule metrics.
FOLLOW-UP TRAP: Calling it temporary and never restoring.How should SI and extraction be coupled?
[INT][PD]
Q: How should SI and extraction be coupled?
A:
Use up-to-date parasitics and matching corner decks; stale SPEF makes SI conclusions unreliable.
FOLLOW-UP TRAP: Running SI on outdated extraction.How do mode changes affect SI analysis?
[INT][PD]
Q: How do mode changes affect SI analysis?
A:
Switching windows and aggressor activity can shift victim severity per mode.
FOLLOW-UP TRAP: Using functional-mode SI only.What belongs in an SI waiver?
[INT][PD]
Q: What belongs in an SI waiver?
A:
Victim scope, measured risk, mitigation plan, owner, and expiration criteria.
FOLLOW-UP TRAP: Permanent waiver with no owner.How do you explain SI risk to non-specialists?
[INT][PD]
Q: How do you explain SI risk to non-specialists?
A:
Show one victim path with aggressor coupling, resulting delay/noise impact, and concrete fix options.
FOLLOW-UP TRAP: Dumping raw report tables.Which cross-discipline handoff matters most for SI?
[INT][PD]
Q: Which cross-discipline handoff matters most for SI?
A:
Timing + routing + extraction alignment on decks, corners, and net naming.
FOLLOW-UP TRAP: Treating SI as routing-only problem.What is the interview-grade SI debug format?
[INT][PD]
Q: What is the interview-grade SI debug format?
A:
Symptom metric -> aggressor hypothesis -> targeted route experiment -> SI/timing regression proof.
FOLLOW-UP TRAP: Listing commands without rationale.How to drill this Q&A bank
Use each question as a two-minute mock. The target answer is not an essay; it is a structured closure response with a metric, mechanism, risk, and follow-up check.
Answer template
MECHANISM: what physical effect or tool step is involved
WHEN: where it appears in the flow
PITFALL: one wrong junior answer
CHECK: report/map/checklist item that proves the answerScoring
5/5: metric, mechanism, experiment, regression, and release decision named without prompting.
4/5: technically correct and names regression, but misses ownership or escalation criteria.
3/5: correct concept but no decision framework.
1/5: tool command or buzzword with no mechanism.
Principal-level review bar
Q&A bank pages in this course should be read like real closure review material. For a 10+ year PD engineer, the bar is not remembering terminology; it is making a release-quality decision under ambiguity.
What excellent looks like
Names the failing metric, corner/mode, database tag, and analysis switches before proposing a fix.
Separates data, constraint, physical, tool, and methodology root causes instead of treating all failures as optimization problems.
Chooses experiments by information gain and reversibility, not by habit.
States regression blast radius across timing, route, power, PV, DFT, package, and tapeout manifest.
Turns recurring failures into methodology guardrails, dashboards, or checklist items.
Closure note template
STAFF / PRINCIPAL CLOSURE NOTE
Context:
stage: <pre-CTS | post-CTS | post-route | post-fill | signoff>
tag: <database / netlist / SDC / library stack>
failing metric: <exact report line>
affected scope: <block / hierarchy / path group / power domain / region>
Hypotheses:
H1: <most likely physical or constraint mechanism>
H2: <competing explanation>
H3: <methodology or input-data issue>
Decision:
next experiment: <cheap check that can falsify H1>
fix candidate: <minimal reversible change>
rollback trigger: <metric that says the fix is wrong>
regression set: <timing / route / power / PV / DFT / package>
escalation owner: <team or reviewer>Tradeoffs a senior engineer must discuss
Technical tradeoff
Interview performance is structured closure thinking under time pressure. Explain not only the preferred fix, but what margin or schedule you are spending to get it.
Cross-team tradeoff
What must RTL, synthesis, CAD, STA, DFT, package, IP, or foundry agree to before this decision is final?
Which artifact becomes the source of truth after the decision: report, waiver, manifest, ECO script, or methodology deck?
What is the cost of being wrong: one rerun, ECO churn, mask risk, performance loss, or silicon escape?
Leadership communication
"The current blocker is <metric> in <corner/mode/stage>. The leading cause is <mechanism>. I recommend <fix> because it is bounded and reversible. The regression surface is <domains>. If it fails, we escalate to <owner> with <evidence>."Key takeaways
Always connect the concept back to a measurable signoff artifact.
A fix is not complete until you can name the regression checks.
Common pitfalls
Optimizing by habit instead of reading the current report.
Forgetting that a local fix can regress timing, routing, power, or PV elsewhere.