Thermal Shock and Thermal Cycling: Choosing the Right Test

Two tests appear in almost every reliability plan, and they are frequently confused with each other. Thermal cycling and thermal shock both vary the temperature of an assembly, and they fail it by completely different mechanisms.

The Physical Difference

Thermal cycling changes the temperature slowly, typically a few degrees per minute, so that the whole assembly reaches a uniform temperature before the next transition. The result is a slow expansion and contraction with little gradient across the board.

Thermal shock moves the assembly between two baths at very different temperatures, so the surface changes temperature almost immediately while the interior lags. The result is a large temperature gradient through the material.

The distinction matters because the failures they produce are different, and passing one is not evidence for the other.

What Thermal Cycling Fails

Cycling stresses the interfaces where two materials with different expansion coefficients meet. The solder joint between a package and a board is the classic case, and the failure appears as a fatigue crack that grows from the outer edge of the joint inward.

The mechanism is fatigue, so the number of cycles matters more than the rate. A test with a wide temperature range and a slow ramp accumulates damage in the joint, and the characteristic life can be related to the service life through a model.

Plated through holes fail by the same mechanism, with the barrel cracking at the point of maximum strain, which is usually at the middle of the board thickness. Our coupon notes describe the structures used to measure this.

What Thermal Shock Fails

Shock produces a large instantaneous stress because of the gradient through the material. It fails things that are sensitive to a single high stress rather than to accumulated damage.

Delamination between layers, cracks in ceramic component bodies, cracks at the interface between a solder mask and copper, and barrel cracks in a hole with poor plating adhesion all appear in shock testing.

A single shock cycle can produce a failure, which makes the test a screening tool as much as a reliability test. It finds the parts that are already marginal rather than predicting the life of a good one.

Choosing Between Them

Where the product experiences slow ambient changes, such as outdoor equipment or a device that is switched on and off, thermal cycling is the closer model.

Where the product experiences a rapid change, such as a device plunged into cold water or exposed to a spray, thermal shock is the closer model. Where both occur, both tests belong in the plan.

Where the purpose is screening a production lot for workmanship defects, shock is the more efficient tool because it fails marginal parts quickly. Our quality notes describe how the screening result is handled.

Test Parameters That Matter

For cycling, the parameters are the high and low temperatures, the dwell at each, the ramp rate and the number of cycles. The dwell must be long enough for the whole assembly to reach the temperature, which a heavy fixture can prevent.

For shock, the parameters are the two bath temperatures, the transfer time and the dwell. The transfer time is the essence of the test, and a slow transfer makes it a cycling test with an extreme range.

The two bath temperatures are limited by the fluids, which impose a practical ceiling on the high temperature and a floor on the low. Air to air shock chambers avoid the fluids and produce a slightly gentler transition.

Monitoring the Test

The failure is detected by monitoring the resistance of daisy chained joints continuously. A crack that grows produces a measurable increase in resistance before it becomes an open, and continuous monitoring catches the intermediate stage.

The event threshold must be stated, because a test that reports only opens misses most of the life of the joint. The time of the first event is the result that the analysis uses.

Our component reliability notes describe how the joint life relates to the thermal excursion, which is what the data is used for.

Analysing the Failures

The failed joints should be sectioned to establish where the crack started and how it propagated. A crack that began at the interface and grew into the bulk indicates a wetting problem, while a crack that began in the bulk and grew outward indicates fatigue.

The position of the failures on the board is also information. Failures concentrated at the corners of a package or at the ends of a row indicate a differential expansion driven mechanism.

Our solder defects notes describe the appearance of each failure mode in a section.

Reporting and Interpretation

The report should include the profile measured on the sample, the monitoring method and threshold, the sample size and the treatment of any unit that failed for an unrelated reason.

A single failure at a low cycle count is not a life prediction. The result becomes useful when several units fail, so that a distribution can be fitted and a characteristic life estimated.

Where no failures occur, the result is a lower bound on the life under those conditions, and it should be reported as such rather than as a pass with unlimited margin.

Process Control and Verification

The measurements that matter are the repeatable ones: conductor width and spacing, annular ring, finished hole size, plating thickness and surface finish are all verifiable on a coupon that travels with the panel. Running a first article through the same checks as the production panel confirms that the two agree, and that comparison is the cheapest form of process control available at prototype stage.

Checks Before Release

The checks that matter are the ones performed on the product rather than on a sample kept for the purpose, because a coupon that travels with the panel is the only evidence about that panel. Where a requirement can be measured, it should be measured at the point of manufacture and recorded against the board or the lot it applies to.

A parameter that is set once and never re verified drifts, and the drift is usually discovered by a defect rather than by the record.

Verification and Records

The tooling, the material and the profile form one system, and a change to any of them should be assessed against the other two before it is released.

Points to Confirm at First Article

Where the process window is narrow, the measurement resolution has to be better than the window, or the data cannot distinguish a good part from a marginal one. A record that identifies the operator, the date and the settings is worth more than a record that identifies only the result.

FAQ

Is thermal shock more severe than cycling? It is more severe per cycle for gradient driven failures and less severe for fatigue driven ones, because the dwell is shorter.

How many cycles are enough? Enough that the failure distribution is visible or that the required life is exceeded with margin. A test stopped at the first failure has shown very little.

What does gopcb provide for these tests? We provide coupons with daisy chained joints for the components of interest, cycling and shock profiles measured on the sample, continuous resistance monitoring with a stated threshold, sectioning of the failures and a distribution analysis of the results.