A board can pass an electrical test at room temperature and still develop an intermittent connection when heated or cooled. Thermal shock testing investigates that risk by exposing a defined specimen to rapid temperature changes and measuring its response. The useful result is a record of the specimen, exposure and failure criterion—not just a chamber setpoint and a cycle count.
Start by identifying what must survive: bare-board copper interconnects, a component package, or the solder attachments on an assembled board. For the broader cycling topic, see PCB thermal cycling testing. Here the focus is choosing and specifying a thermal shock test whose measurements can support a manufacturing or qualification decision.
Which thermal reliability testing method matches the failure risk?
A thermal stress test for PCB interconnects does not automatically qualify the components mounted on that board. Select the test vehicle and standard around the structure of concern.
| Object under investigation | Relevant method or specification | Evidence needed for the decision |
|---|---|---|
| Bare-board plated holes, vias and interconnections | IPC-TM-650 2.6.7.2C: Thermal Shock, Thermal Cycle and Continuity | Representative coupon construction, thermal history, specimen temperatures and resistance results |
| Component package | The component's qualification requirement; JEDEC temperature-cycling or thermal-shock method where specified | Manufacturer's report for the exact package/technology, with conditioning and test conditions |
| Surface-mount solder attachments | IPC-9701B thermal-cycling fatigue-life characterization | Board/package construction, assembly process, monitored failures and analysis under the stated conditions |
| Complete product operating through temperature changes | Product-specific environmental and functional test plan | Enclosure, mounting, powered state, operating limits and functional measurements |
IPC-9701B concerns fatigue-life characterization of surface-mount solder attachments. Its revised scope explicitly excludes qualification; a product's qualification decision needs its own agreed requirements. A bare-board coupon test and a solder-attachment characterization therefore cannot be exchanged merely because both use temperature cycles. IPC-9701B public scope, sections 1.0 and 1.4.
How is thermal shock different from thermal cycling?
The answer depends on the named method. IPC-TM-650 2.6.7.2C defines shock for unpopulated printed-board specimens using a surface-temperature change of at least 1°C/s, over at least the central 60% of both heating and cooling transitions. Its cycling definition is below that rate. This is a specimen measurement, not a universal definition based on a chamber's advertised °C/min rating. IPC-TM-650 2.6.7.2C, sections 3.1–3.2.
Three times must remain separate in a thermal shock test procedure:
- Transfer time: movement between zones, or switching the thermal environment.
- Medium recovery time: the air or liquid returning to its specified temperature after the change and loading disturbance.
- Specimen stabilization and dwell: the actual article reaching the required condition and remaining there for the prescribed exposure and measurements.
A fast transfer does not prove a fast specimen transition. Copper planes, board thickness, heatsinks, fixtures and the number of samples affect the loaded response. Place sensors where they demonstrate the required exposure, and retain both specimen and medium traces.
Air-to-air or liquid-to-liquid?
HILPCB offers integrated thermal-shock and thermal-cycling testing, material analysis and failure-analysis services, including air-to-air and liquid-to-liquid chamber options. Qualified capacity includes automated two-zone transfer, real-time monitoring and programmable profiles; the available equipment envelope is −80°C to +225°C with ±2°C chamber control, subject to the selected method, specimen and profile. Digital lot-data capture can support prototype through volume programs. Engineering services include CTE mapping, stress simulation, material selection, reflow/via support and microsection, X-ray, fractography and thermal-model analysis. Agree test limits and deliverables for each project; the equipment envelope is not a guarantee of specimen performance.
All test results are digitally logged and linked to manufacturing lots.
Air systems can move a basket between zones or switch conditioned airflow around stationary samples. Liquid systems immerse the specimen in different baths. Neither arrangement is inherently the right choice for every HDI, ceramic or metal-core board. Check the applicable method, heat transfer, fluid compatibility, coatings, seals and measurement access before selecting immersion.
ESPEC's air-to-air equipment documentation distinguishes specimen-temperature control and specimen-temperature trigger functions. These are useful examples of why product temperature and chamber control temperature need separate attention; they do not establish the equipment available for a particular HILPCB order. ESPEC air-to-air chamber documentation.
The profile illustration below shows alternating hot and cold exposures. It is a schematic, not a measured HILPCB test record or an acceptance profile. Its marked functional checks do not demonstrate continuous electrical monitoring during transitions.
What does a specified bare-board method require?
IPC-TM-650 2.6.7.2C, revision C dated March 2020, separates qualification from quality conformance. AABUS means agreed between user and supplier.
| Decision in that method | Qualification | Quality conformance |
|---|---|---|
| Temperature extremes and cycles | AABUS | 100 cycles; cold options −40°C, −55°C default, or −65°C; hot extreme follows section 6.5.1.1 |
| Drying before testing | At least 6 hours at 105–125°C | Depends on the normal assembly process and quality-conformance requirements |
| Prior assembly simulation | Six reflow simulations under method 2.6.27, unless an alternative is agreed | Same method requirement; record any agreed departure |
| Resistance reference | First cycle's peak resistance after stabilization at the high-temperature dwell | First cycle's peak resistance after stabilization at the high-temperature dwell |
Section 6.7.1 states a maximum resistance change of 5% unless otherwise specified. Summary tables and suggested drawing notes are abbreviated: the sample drawing uses a strict “less than” expression. Specify the agreed limit and equality treatment rather than copying fragments into a purchase order. IPC method, sections 6.1–6.7 and suggested drawing note.
The hot-end rule also requires an agreed basis for the material Tg. Its Tg-related term, reflow-related term and absolute ceiling must be considered together; the method's qualification branch does not simply inherit a fixed industry temperature range. Keep the selected conditions and any deviations in the signed test plan. The method's older temperature examples are historical references, not universal “commercial” or “aerospace” profiles.
For example, assume a project has already agreed a Tg of 180°C and an assembly-simulation peak of 260°C. The hot-end candidates described in section 6.5.1.1 are then 180 − 10 = 170°C, 260 − 25 = 235°C, and 210°C, giving 170°C as the lowest. This example is above the method's lower-bound consideration for the Tg term; it does not resolve that interpretation for lower-Tg materials. A different Tg measurement basis or assembly peak requires review of the selection.
How should resistance changes be evaluated?
Compare readings at the specified thermal state. Copper resistance changes with temperature, so comparing a hot test reading with a room-temperature baseline can confuse temperature response with damage.
Suppose a test plan uses a 5% maximum increase relative to the first cycle’s peak resistance after stabilization at the high-temperature dwell, which is 0.400 Ω. Later readings at the same specified hot condition give:
| Later resistance | Calculation | Change |
|---|---|---|
| 0.416 Ω | (0.416 − 0.400) / 0.400 × 100 | 4.0% |
| 0.424 Ω | (0.424 − 0.400) / 0.400 × 100 | 6.0% |
The first value is below that illustrative limit; the second exceeds it. Measurement uncertainty, fixture stability and the agreed decision rule still matter. Neither number identifies the physical location of a crack.
An end-of-dwell resistance reading and an event detector also answer different questions. A short open circuit during a transition can disappear before the next logged reading. If transient interruption is a failure criterion, specify event duration, resistance threshold, detector response and sampling coverage. “Electrical monitoring included” is insufficient without these details.
What can thermal shock analysis reveal—and miss?
Thermomechanical loading depends on the actual material combination, direction of expansion, temperature range and mechanical constraints. A single generic FR-4 CTE value cannot describe behavior both below and above Tg. Likewise, a high-Tg laminate or a ceramic substrate is not automatically the most reliable choice for every bonded structure.
For HDI boards, identify which blind, buried or stacked-via structures the coupon represents. For ceramic PCBs, include the copper/substrate and attachment interfaces of concern. A change in stackup, plating, fill or attachment process may invalidate a previous test comparison even when the board outline is unchanged.
| Observation | Useful follow-up | Limit of the observation |
|---|---|---|
| Resistance increases at a repeatable cycle or temperature | Confirm leads/fixture, localize the affected net, then examine the suspect interconnect | Resistance alone does not identify the crack mechanism |
| Intermittent continuity during transitions | Review event records alongside specimen temperatures and fixture motion | A final room-temperature continuity pass can miss the event |
| Suspected laminate or bonded-interface separation | Choose a suitable acoustic, sectional or other examination for the material and defect | Conventional X-ray is not a universal delamination detector |
| Solder-attachment failure on an assembly | Correlate package, board, alloy, process and mechanical constraint with physical analysis | A bare-board coupon result does not explain every assembly failure |
Preserve failed specimens and record the sequence of nondestructive and destructive analysis. A microsection samples a particular plane; its preparation and location affect what can be concluded. Report the evidence supporting the proposed cause, not just a photograph labelled “thermal failure.”
Can a thermal shock reliability test predict field life?
Thermal-shock testing can support reliability assessment only when its failure mechanism is relevant to the intended service conditions. Field-life prediction requires a justified acceleration model, material properties and test-to-use correlation; simple cycle-count multiplication is not sufficient.
IPC-9701B notes that solder fatigue damage in accelerated testing is not generally equivalent to operational damage; appropriate acceleration factors are needed. NIST describes Arrhenius acceleration in terms of a mechanism-dependent activation energy, with applications such as chemical reactions, diffusion and migration. It is not a default conversion for every mechanically driven fatigue failure. IPC-9701B, section 3.1, NIST Arrhenius model.
A Weibull fit describes the failure distribution in the data supplied to it. Before using it for a field prediction, retain the tested population, failures and surviving specimens, failure definitions, stress conditions, confidence bounds and the validated acceleration model. Combining several unrelated failure mechanisms into one curve can obscure the decision rather than improve it.
What should buyers request from thermal stress testing solutions?
A comparable quotation needs more than the cold and hot setpoints. Send HILPCB the board revision, stackup and material grades, coupon design, via structures, finish, assembly history and the reason for testing. Include the applicable standard and revision, qualification or lot-acceptance purpose, sample quantity, temperatures, cycles, conditioning, monitoring and failure criteria.
Ask the supplier to confirm:
- Which specimens represent the production construction, and which manufacturing lots they cover.
- The actual loaded test arrangement, sensor locations, thermal profile and stabilization criteria.
- The resistance reference, event-detection requirements and calibrated measurement capability.
- Individual sample results, interruptions, deviations, raw data and failed-sample retention.
- Test execution location, any external laboratory, analysis scope, schedule and price.
- Who can approve a deviation, retest, material change or lot disposition.
High-thermal PCB construction and SMT assembly can require different test vehicles and evidence. Confirm the quoted scope before manufacture: a general order should not be assumed to include liquid immersion, continuous transient monitoring, failure analysis or a field-life model for every production lot.
Frequently asked questions
What is the purpose of a thermal stress test?
It exposes a defined specimen to thermal loading to evaluate a particular failure risk. The method may concern bare-board interconnects, a component or an assembled product. Agree the object, exposure, measurements and acceptance criteria before interpreting the result.
What temperature range should a thermal shock test use?
Use the range required by the applicable method and product test plan. Material properties, assembly exposure and the purpose of the test matter. A temperature pair described as “commercial” or “aerospace” is not enough to establish the correct condition for a specific board.
Are 100 thermal shock cycles always enough?
No. The 100-cycle value in IPC-TM-650 2.6.7.2C belongs to its quality-conformance branch. Qualification conditions are agreed between user and supplier. Component qualification, solder-attachment characterization and complete-product testing have their own requirements and cannot be inferred from that bare-board cycle count.
Does faster transfer make a test more valid?
Not necessarily. Transfer time is only one parameter. Demonstrate the required specimen-temperature transition and dwell under the actual loading arrangement. A faster or more severe exposure can create a different failure mechanism from the one the program is meant to evaluate.
Can I replace air testing with liquid immersion?
Only if the governing method and approved test plan permit it. Assess fluid compatibility, absorption, coatings, sealing and electrical measurement arrangements. Similar bath and air setpoints do not establish equivalent specimen exposure.
Does a passing test guarantee reliable production boards?
It supports the specified decision for the represented construction, samples and conditions. It does not remove the need for manufacturing controls, assembly inspection, change control or product qualification. Keep the report linked to the actual revision and lot before using it to authorize release.