Fault Models: Debug Playbook

Debug playbook

Debug Playbook for Fault Models focuses on stuck-at / transition / bridge coverage, untestable fault ratio. The goal is to convert metric movement into mechanism, owner, and release decision.

Debug aims to find the first incorrect assumption, not the loudest downstream symptom. Start with reproducibility and ownership.

Root-cause tree

ROOT-CAUSE TREE - Fault Models

stuck-at / transition / bridge coverage, untestable fault ratio regresses
        |
  setup changed?
    /        \
  yes         no
  |            |
constraint    silicon or
or ATPG       physical/test path
 /    \          |
SDC   model    chain/clock/power/diagnosis
diff  diff     isolate first failing signature

1. Freeze run tags for patterns, constraints, and tester setup.

2. Isolate first failing metric bucket and scenario.

3. Classify failure source: model, constraints, physical, or silicon.

4. Prove mechanism with one reduced replay or targeted run.

5. Apply smallest owner-controlled fix.

6. Re-run timing, power, and quality regression matrix.

Review memo template

STAFF DFT REVIEW MEMO - ATPG & Pattern Generation / Fault Models

1. Symptom
   - Watched metric: stuck-at / transition / bridge coverage, untestable fault ratio
   - Failing scenario: <mode/lot/corner/program>
   - Pattern class: <scan/transition/compressed/BIST/JTAG>
   - Tags: <constraints, patterns, tester program, netlist>

2. Mechanism hypothesis
   - Primary mechanism: Fault models define what manufacturing defects ATPG targets; model mix drives pattern count, runtime, and outgoing quality.
   - Competing hypothesis: <constraint issue, model issue, physical issue, silicon issue>
   - Missing evidence: <report, replay, diagnosis trace>

3. Proposed action
   - Minimal reversible change: <constraint fix, architecture tweak, pattern update>
   - Expected metric movement: <delta>
   - Regression risk: timing, power, quality, schedule

4. Signoff
   - Re-run artifact: fault coverage report, fault class breakdown, untestable list
   - Required owners: ATPG owner, DFT architect, quality owner
   - Final decision: release, waive, rollback, or escalate

DFT deep dive

ATPG quality comes from fault model choice plus legal constraints, not raw pattern volume alone.

Concept diagram

ATPG FLOW

fault model -> constraints -> generation -> simulation -> coverage closure -> signoff

Metric graph

COVERAGE GAP

target coverage
  ^
  |      o before closure
  |          o after fixes
  +---------------------> iteration

Reports and artifacts

• fault model coverage

• untestable class report

• constraint legality errors

• pattern signoff memo

Mini case study

Transition coverage stalled due to clock constraints mismatch; updated at-speed capture definitions recovered target.

Debug branches

• Classify untestable faults

• Diff ATPG constraints each run

• Pair coverage with pattern budget

Senior review question

Ask: what evidence proves this DFT decision is safe for production?

Key takeaways

• State metric, lot/corner context, and pattern tag with every claim.

• Treat timing, power, and quality as one signoff problem.

Common pitfalls

• Chasing coverage without legality checks.

• Ignoring test-power side effects of pattern changes.

• Debugging silicon without reproducible tags.

Principal DFT review addendum

Fault models define what manufacturing defects ATPG targets; model mix drives pattern count, runtime, and outgoing quality.

Metric: stuck-at / transition / bridge coverage, untestable fault ratio