PCB Thermal Cycling: What the Test Proves

What Thermal Cycling Actually Tests

A thermal cycling qualification exposes a board or an assembly to repeated excursions between a cold and a hot extreme, with defined ramp rates and dwell times, and then inspects it for damage. What it is really measuring is the difference in expansion between materials that are bonded together. Copper expands at roughly seventeen parts per million per degree Celsius, the resin and glass laminate at a lower and anisotropic rate in the plane and much higher through the thickness, and a solder joint has its own coefficient. Every temperature excursion strains the interfaces, and the cycles accumulate fatigue until something cracks. A thermal cycling test does not measure strength; it accelerates a failure mechanism so that a design or process weakness becomes visible in weeks instead of years.

The Failure Mechanisms It Exposes

The classic failure is a barrel crack in a plated through hole. The barrel is copper, constrained in the through thickness direction by the laminate around it, and the laminate expands far more than the copper in that direction. The barrel is therefore stretched on every hot excursion, and the fatigue grows at the point where the plating is thinnest, usually the middle of the hole and often at a corner where the barrel meets an inner layer pad. A second mechanism is the corner crack where the barrel meets the surface pad, driven by the same mismatch. A third is delamination between the resin and the glass, usually initiated by moisture absorbed before assembly and expanding rapidly at reflow temperature. A fourth is the cracking of plated holes in a thick board, where a partially cracked barrel passes electrical test and fails after cycling. Fifth, in the assembly rather than the board, is fatigue of a solder joint on a component with a large CTE mismatch, such as a ceramic package on an FR-4 board.

Test Conditions and What They Mean

A cycle is defined by two temperature extremes, the dwell at each extreme, the ramp rate and the number of cycles. A common board level test uses minus forty and plus one hundred and twenty five degrees Celsius with dwells long enough for the sample to soak. A longer dwell is more severe than a faster ramp for barrel fatigue, because the failure depends on the time spent at temperature and on the full expansion being reached. The number of cycles in the qualification is chosen to represent the service life with an acceleration factor. The result is reported as the number of cycles to failure or as a pass at a specified number, and it must be read together with the inspection method, because a barrel crack at twenty percent of the wall may be found by cross section and missed by an electrical test, or vice versa.

How Design Affects Cycle Life

Cycle life is largely set at design time. Plating thickness is the strongest single factor, because the fatigue life of the barrel rises steeply with copper thickness, which is why high reliability classes specify a higher minimum. Aspect ratio matters because a deep, narrow hole has a thinner barrel in the middle and a longer constrained length. The hole size relative to the pad, and the presence of a large inner layer pad attached to the barrel, change the local stiffness and shift the failure location. Laminate choice matters, and a high glass transition material with a lower through thickness expansion reduces the strain per cycle; where the mismatch is severe, a lower expansion laminate or a different construction is needed. In the assembly, the compliant lead form, the solder volume, the thermal mass and the placement of heavy components all affect joint life.

Acceptance and Inspection

The acceptance criteria come from the applicable standard and class, and they are expressed in terms of the number of cycles and the post test condition: no crack through the full thickness of the barrel, a limited amount of corner cracking depending on class, and no delamination beyond a defined percentage of the interface. Inspection is normally a cross section of a sample, sometimes supported by resistance monitoring during the test, which detects an increase in the barrel resistance as a crack propagates. Monitoring during the test is more informative than a final inspection alone, because it identifies the cycle at which the damage began and therefore how much margin exists.

Designing for Cycle Life in Practice

Specify the hole wall copper above the minimum where the application will cycle, and confirm it with cross sections at the thinnest point. Keep the aspect ratio inside the comfortable range, or use blind, buried or microvia connections to shorten the barrel. Avoid stacking many high expansion interfaces in one area and be careful with very thick boards. Where the application cycles hard, consider a laminate with a lower expansion and a higher glass transition temperature, and treat the assembly as a joint design problem rather than only a board problem. Finally, run the test on a coupon that shares the panel and the process with the product, because the process is what determines the plating thickness and the laminate quality that the test is actually measuring.

FAQ

What does a thermal cycling test prove? It exposes the fatigue mechanisms caused by mismatched expansion, mainly plated barrel cracking and solder joint fatigue, and measures how many cycles a design and process can survive.

What is the most important factor in barrel life? The thickness of the plated copper, followed by the aspect ratio and the laminate expansion. Higher plating thickness increases fatigue life steeply.

Can a cracked barrel pass electrical test? Yes. A partial crack may leave continuity intact until it propagates, which is why resistance monitoring during the test and cross section afterwards are both used.

Why do some boards delaminate after reflow? Moisture absorbed into the laminate expands rapidly at reflow temperature. Baking before assembly and controlling storage remove most of the risk.

How many cycles should a board survive? The number is set by the service life and the acceleration factor of the chosen test conditions, and it is normally specified with the class and the acceptance criteria for post test inspection.

Conclusion

Thermal cycling is a fatigue test of the interfaces in a board, and the result is decided by plating thickness, aspect ratio, laminate expansion and the assembly design long before the test is run. Specify the hole wall copper above the minimum for cycling applications, keep the barrels short and well plated, choose the laminate against the expected temperature range, and confirm the result with a coupon cross section rather than a single electrical pass. The plating and class requirements are part of PCB capabilities, the stackup decisions belong in PCB design and layout, and the process that produces the barrels is described in PCB manufacturing. A prototype PCB assembly run before a cycling qualification makes the failure mechanisms visible while they are still cheap to fix in 2026.