Procedures & Methods for PCBA Testing

Why must safety be the top priority when testing PCBAs? (must-read)

• Use current-limited bench supplies, soft-start rails, and GFCI/RCD protection when powering unverified boards. Stop immediately if there is smoke, odor, or sudden temperature rise.

• Enforce ESD controls: wrist straps, grounded benches (≤1 Ω), ionizers near fine-pitch devices, and humidity control (40–60%).

• Treat high-voltage tests (hipot, surge, leakage) as controlled operations requiring interlocks, two-hand switches, PPE, and trained operators.

• Burn-in or thermal stress must run inside guarded chambers with thermal cutoffs, fire detection, and remote cutoff for runaway failures.

Transition: When safety conditions are secured, testing becomes an optimization problem — selecting the right depth of coverage at the right stage.

---

What test strategy should you follow (the test pyramid)?

The PCBA test pyramid ensures broad coverage without wasting time on defective units:

1. Visual & AOI/X-ray — quick defect capture: solder bridges, lifted leads, missing parts.

2. Low-risk electrical checks — continuity, shorts/opens, rail pre-checks.

3. Structural tests — ICT, flying-probe, boundary-scan to capture net-level faults.

4. Functional Test (FCT) — behavioral verification under expected loads/stimuli.

5. Reliability screening — burn-in, cycling, stress tests to remove early-life failures.

6. Special tests — RF matching, EMC pre-checking, hipot for isolation-critical products.

Table — Typical Detection Capability by Test Stage

Test Stage	Finds	Misses	Best For
Visual / AOI	Missing parts, polarity, solder bridges	Hidden joints, electrical issues	All boards, early filtering
SPI	Paste volume & position	Component-level defects	SMT-heavy designs
X-ray	BGA/QFN voids, head-in-pillow	Electrical failures	BGA, power modules
ICT	Net-level faults, wrong passives	Firmware bugs, dynamic behavior	Medium–high volume
FCT	Functional behavior, sensors, comms	Marginal joints not triggered	Final product checks
Burn-in	Early-life failures	Mechanical defects	High-reliability products

Transition: With strategy defined, implement each method rigorously.

---

How do you run visual inspection and AOI correctly?

Purpose: Identify placement, soldering, and polarity defects before costly electrical tests.

Equipment: Lighted station, microscopes (10–40×), professional AOI with 2D/3D capability.

Procedure:

• Validate AOI rules using a golden board and ensure lighting profiles remain stable shift-to-shift.

• Confirm solder fillet geometry, coplanarity, lead wetting, and polarity marks.

• Capture AOI false calls and update libraries to reduce unnecessary operator load.

Pass/fail cues: Missing component, polarity reversed, significant void, gross bridging, lifted pad.

Transition: Even perfect placement cannot compensate for poor paste application.

---

How do you use SPI to stop printing defects from reaching placement?

Purpose: Detect insufficient/excessive paste volume — the root cause of 60%+ SMT defects.

Equipment: 3D SPI scanner with volume, height, area, and offset measurements.

Procedure:

• Characterize aperture target volumes and set SPC control limits (e.g., ±25%).

• Monitor trends such as slow drift caused by stencil wear or paste aging.

• Trigger automatic line stop on out-of-spec cluster events (≥5 related pads).

Pass/fail cues: Volume shortage, bridges predicted by excessive deposition, paste misalignment.

Transition: For BGA/QFN and dense power devices, X-ray is your only window into hidden joints.

---

How do you apply X-ray inspection for hidden-joint risks?

Purpose: View solder joints hidden beneath packages and detect structural failures.

Procedure:

• Capture orthogonal and oblique views for BGA ball integrity and void distribution.

• Compare void percentages to IPC-7095 guidelines or internal limits (often ≤20–35%).

• Run X-ray as a sampling plan or targeted inspection triggered by AOI flags.

Pass/fail cues: Missing balls, collapsed balls, >spec voiding, head-in-pillow.

Transition: After physical verification, move to electrical validation using ICT or flying-probe.

---

When should you use ICT versus flying-probe testing and how?

ICT and flying-probe complement each other:

Attribute	ICT	Flying-Probe
Fixture Cost	High	None
Throughput	Very high	Medium/Low
Flexibility	Low	Very high
Ideal For	Mass production	NPI, low/medium volume
Coverage	Excellent	Moderate–high

ICT Procedure:
Fixture self-test → isolation checks → net/continuity → passive values → diode/transistor → leakage tests → marked parametrics.

Flying-Probe Procedure:
Import netlist → probe continuity & shorts → measure passives → diode/transistor orientation → dynamic probing for selected circuits.

Pass/fail cues: Opens, shorts, wrong components, orientation errors, out-of-spec passive values.

Transition: Structural tests clear the board for a controlled, safe first power-up.

---

How to do a safe power-up / smoke test?

Purpose: Catch catastrophic shorts or faulty power rails without damaging the board.

Procedure:

• Set current limit (e.g., 0.1–0.3A for small boards, product-dependent).

• Ramp voltage slowly and monitor inrush.

• Observe the board via thermal camera within 3–5 seconds of startup.

• Increase current limits only after stable behavior is proven.

Pass/fail cues: Sudden overcurrent, localized heating (>10°C rise in 1–2s), unstable boot.

Transition: Safe units proceed to programming and calibration.

---

How should programming & calibration be performed?

Purpose: Apply firmware and store hardware-specific calibration constants.

Procedure:

• Verify connector sequencing, Vcc stability, reset states.

• Flash firmware and verify CRC/MD5 checks.

• Run calibration vectors (ADC offset, analog trims, Hall-sensor calibration, gain tuning).

• Bind firmware ID + calibration ID to PCBA serial for traceability.

Pass/fail cues: Programming timeouts, CRC mismatch, analog ranges out-of-spec.

Transition: Functional Test (FCT) verifies the product in its intended operational mode.

---

What makes an effective Functional Test (FCT)?

Purpose: Validate real behavior under realistic loads and environmental expectations.

Key design insights:

• Fixtures should simulate real connectors, mechanical tolerances, and thermal loads.

• Use controlled loads (electronic loads, resistor banks) and inject accurate signals.

• Include communication tests (UART, I²C, CAN, BLE, proprietary protocols).

Procedure:
Smoke check → normal power ramp → boot verification → exercise analog/digital interfaces → final serial write.

Pass/fail cues: Boot anomalies, comm failures, incorrect analog readings, protocol timing issues.

Transition: High-reliability assemblies require stress screening.

---

When and how do you run burn-in and reliability screening?

Purpose: Remove early-life failures and confirm thermal/mechanical robustness.

Procedure:

• Define stress profiles (e.g., 55°C/85°C cycling, load cycling, voltage margining).

• Instrument critical points (thermistors, current monitors).

• Capture failures and correlate with lot, operator, and assembly conditions.

Pass/fail cues: Any functional or parametric drift outside limits during stress.

Transition: Some markets require isolation or RF performance checks.

---

How do you perform HiPot, RF/EMC pre-checks and other special tests?

HiPot: Ramp voltage to test isolation paths; failure = leakage exceeding spec or arc events.
RF/EMC Pre-check: Validate antenna matching, emissions, and sensitivity before certification.
Surge/ESD Pre-check (optional): Evaluate robustness of power stages and interfaces.

Transition: All test steps must be integrated into one coherent flow.

---

What is the recommended end-to-end test sequencing?

Visual → SPI → AOI → X-ray (as needed) → ICT/Flying-Probe → Safe Power-Up → Programming & Calibration → FCT → Burn-in/Environmental Stress → Final Visual → Packaging.
Always verify fixtures with golden boards before each shift.

Transition: To ensure consistency, define quantitative pass/fail limits.

---

Which pass/fail gates and tolerances should you use (examples)?

Category	Typical Criteria
Power rails	±5% with load; ripple <50–100 mV (design-dependent)
Passive values	Resistors ±1–5%; capacitors presence/ESR per spec
DCIR / sag	Startup sag ≤1.0 V (product-dependent)
FCT vectors	100% pass on safety-critical I/O

Transition: When failures occur, structured triage speeds root-cause finding.

---

What are rapid triage steps for common failure signatures?

1. Same net failing across multiple units → probable PCB fab or stencil/paste issue.

2. Intermittent failures → mechanical/probing issues or marginal joints exposed by thermal cycling.

3. Programming timeouts → unstable Vcc, cable integrity, programmer firmware.

4. High inrush or smoke → shorted FETs, reverse polarity install, solder bridges under fine-pitch parts.

Transition: Fixtures and instruments must remain calibrated.

---

How often should you maintain fixtures and equipment?

• Run fixture self-test each shift; replace worn pogo pins.

• Calibrate meters, scopes, chamber sensors per schedule.

• Version-control all test code and maintain golden-board references.

Transition: Every test event must generate traceable data.

---

What data should you record for traceability?

Serial ID, operator, station ID, firmware version, measurements, pass/fail decision, SPI/AOI/X-ray images, IR captures, logs, timestamps.
Traceability is essential for root-cause analysis and regulatory compliance.

Transition: Finally, a few rules accelerate implementation.

---

What practical tips speed implementation and reduce waste?

• Gate programming and burn-in on AOI/ICT pass to avoid resource waste.

• Use SPC on SPI/AOI metrics to predict drift and prevent defects proactively.

• Maintain rapid escalation rules when SPC trends break control limits.

• Ensure MES linkage for lot, stencil, operator, and test history.

Transition: During NPI, test discipline prevents bad habits from entering mass production.

---

What should you do for first-article and NPI runs?

1. Run SPI & AOI on the first 3–10 boards; compare against golden board baseline.

2. Perform ICT/flying-probe for net integrity.

3. Use conservative smoke-test limits.

4. Flash firmware, run FCT, archive results.

5. Approve batch only after SPC stability and successful first-article validation.

---

Conclusion — one-line takeaway & three immediate actions

A disciplined test pyramid (AOI/SPI → ICT/Flying-Probe → controlled power-up → programming → FCT → burn-in) reduces defects, prevents expensive rework, and produces defensible traceable quality data.

Do these now:

1. Enforce fixture self-test + golden-board check at shift start.

2. Gate programming and burn-in on AOI/ICT pass.

3. Archive SPI/AOI/reflow and FCT logs linked to board serials.