Reliability Test Planning
Quality & Inspection Standards
PRODUCTS CENTER
Get Custom Quote
PRODUCTS CENTER
Reliability Test Planning – A Risk‑Based Approach for PCB Qualification
A test plan gap – thermal shock only, no vibration
An automotive supplier developed a comprehensive reliability test plan for a new ADAS controller. It included 1000 thermal cycles (-40°C to +125°C), humidity testing, and HALT. However, the plan omitted random vibration. After launch, field failures occurred at 18 months due to solder joint cracks under under‑hood vibration. The vibration spectrum had never been characterized. The lesson: a reliability test plan must be driven by the actual field environment, not a generic checklist. We now require a formal environmental profile (temperature, humidity, vibration, shock) before test plan development.
The Four Pillars of a Reliability Test Plan
A robust reliability test plan must answer four questions before any test is selected:
- What can fail? – Failure modes and mechanisms based on design and materials.
- Under what conditions? – Field environmental profile (temperature, humidity, vibration, shock, power cycling, etc.).
- How much margin is needed? – Acceleration factors and test durations must correlate to field life.
- What constitutes a pass? – Quantitative acceptance criteria (not just "no visible damage").
Test plans that skip these four steps are often under‑ or over‑tested, wasting time and missing critical failure modes.
Risk‑Based Test Tiering – Class 2 vs Class 3 vs Custom
Not all products require the same test intensity. We use a three‑tier risk assessment to size the test plan appropriately.
| Tier | Criteria | Test Plan (Minimum) | Sample Size |
|---|---|---|---|
| 1 (Low risk) | Consumer, office, short life (<3 years), benign environment | Thermal cycling (100 cycles, -20/+70°C) + visual inspection + microsection (5 samples) |
5 boards |
| 2 (Medium risk) | Industrial, telecom, medical (non‑critical), 5‑10 year life, moderate vibration | Thermal shock (500 cycles, -40/+85°C) + random vibration (3 g RMS, 200h) + humidity (85/85, 500h) + microsection (5 samples) |
10 boards |
| 3 (High risk) | Automotive under‑hood, aerospace, military, ≥10 year life, extreme environment | Thermal shock (1000 cycles, -40/+125°C) + combined vibration‑temperature + mechanical shock (50 g, 6 ms) + HALT (to find margin) + microsection + dye & pry (10 samples) |
15‑30 boards |
Internal rule: For any product with BGA pitch ≤0.5 mm or layer count ≥12, we default to Tier 2 or 3 regardless of the stated application. HDI structures are inherently more sensitive to stress.
Test Selection Matrix – Matching Stress Type to Failure Mechanism
Each test in the plan should target a specific failure mechanism. The matrix below links typical PCB failure modes to the appropriate IPC‑TM‑650 or industry test methods.
| Failure Mechanism | Primary Test Method | Accelerated Condition | Key Measurement |
|---|---|---|---|
| PTH barrel crack (thermal fatigue) | Thermal shock (2.6.7) | -40/+125°C, 1000 cycles | Microsection crack length, resistance change |
| CAF / ECM (electrochemical migration) | SIR test (2.6.3.7) | 85°C/85% RH, 50V, 1000h | Insulation resistance >10¹⁰ Ω |
| Solder joint fatigue (vibration) | Random vibration (IEC 60068‑2‑64) | 5‑500 Hz, 3 g RMS, 200h | Daisy‑chain resistance, post‑test microsection |
| Delamination / blister (moisture + heat) | Thermal stress (2.6.8) | 288°C solder float, 20s | No delamination or blistering |
| Solderability loss (aging) | Steam aging + edge dip (2.6.1, 2.4.14) | 8h steam aging, 260°C solder, 5‑10s | ≥95% wetting |
Sample Size and Statistical Confidence – Success‑Run Theorem
Many test plans specify sample sizes arbitrarily. We use the success‑run theorem to size samples for zero‑failure testing:
n = ln(1 − C) / ln(R)
where C = confidence level, R = reliability (probability of success).
For example, to demonstrate 90% reliability with 90% confidence (common for Class 3), n = ln(1‑0.9)/ln(0.9) ≈ 22 samples. For 95/95, n = 59. However, destructive testing (microsection) often uses smaller samples (n=5‑10) due to cost, accepting lower confidence.
| Confidence / Reliability | Required Samples (Zero Failures) | Typical Application |
|---|---|---|
| 90% / 90% | 22 | Industrial, telecom |
| 95% / 95% | 59 | Automotive (non‑safety), medical |
| 99% / 99% | 459 | Aerospace, ASIL‑D (impractical; use prior data) |
| 80% / 90% (destructive tests) | 10 | Microsection, cross‑section only |
For high‑confidence requirements, we combine test data across multiple lots to increase effective sample size.
Test Sequence – The Order Matters
Tests should be ordered to avoid pre‑damaging samples before they experience the primary stress. Our recommended sequence:
- Baseline measurements – Electrical test, impedance, visual inspection, microsection (on separate coupons).
- Environmental stress (non‑destructive) – Thermal cycling, humidity, vibration (without disassembly).
- In‑process checks – Periodic resistance monitoring during stress (e.g., daisy‑chain).
- Post‑stress electrical test and visual inspection – Detect functional failures.
- Destructive analysis – Microsection, dye & pry, cross‑section on a subset (5‑10 samples).
- Remaining life assessment – Continue stress on survivors to find margin (HALT approach).
We have seen test plans where microsection was performed before thermal stress – this destroys evidence of thermal‑induced cracks. Always use separate coupons for pre‑ and post‑stress analysis.
Supplier Audit Checklist – Reliability Test Plan
- Does the supplier have a documented environmental profile for your product's field use?
- Do they use success‑run statistics to size test samples, or do they guess?
- Are test reports traceable to specific lot numbers and date codes?
- Do they perform in‑situ monitoring (daisy‑chain) during stress tests, or only post‑test?
- What is their process for handling test failures – root cause analysis and corrective action?
Related Engineering Resources
Frequently Asked Questions
A> It depends on the daily temperature variation and acceleration factor. A typical automotive under‑hood profile uses 1000 cycles (-40/+125°C) to represent 10 years. Use the Coffin‑Manson model to calculate cycles: N_field = N_test × (ΔT_test / ΔT_field)^1.9.
A> Yes for realistic simulation, but it's more expensive. If budget allows, run combined tests. If not, run sequential tests but with reduced durations (e.g., 200h vibration + 500 thermal cycles, not 500+500).
A> Only if the tests are non‑destructive and sequential (e.g., thermal cycling → humidity → vibration). Destructive tests (microsection) require separate samples. Plan sample allocation before test start.
A> Qualification tests verify that the product meets specifications (pass/fail). HALT pushes the product to failure to find design margins and weaknesses. Both are valuable – qualification for compliance, HALT for robustness.
A> Annually, or whenever the field environment changes (e.g., new installation location, different customer usage profile). Also after any major field failure that indicates a gap in the test plan.
Reference standards: IPC‑6012 (Qualification), IEC 60068 series (Environmental testing), JESD22 (Reliability). Statistical methods: Success‑run theorem, Coffin‑Manson acceleration model.
Developing a reliability test plan for your product?
We help define test levels, sample sizes, acceleration factors, and acceptance criteria based on your field environment.
Request a Test Plan Consultation →Get Custom Quote
PRODUCTS CENTER
World-class PCB & PCBA manufacturer focused on high-frequency and high-reliability boards for defense, aerospace, automotive and industrial electronics.
You can unsubscribe at any time with a single click.
Copyright © 2025 UltroNiu Electronics Group All Rights reserved.