Burn-In and Aging Tests for Assemblies

What an Aging Test Screens

Every population of manufactured assemblies contains a small number of units with a latent defect that will fail early. The distribution of failures over time follows a curve with three regions: a decreasing failure rate at the start, caused by these latent defects, a long flat middle where failures are random, and a rising rate at the end caused by wear-out. An aging or burn-in test is a deliberate attempt to move past the first region before the product is shipped, so that the units which would have failed in the customer’s hands fail in the factory instead.

Burn-In Versus Qualification

The two are often confused and they answer different questions. A qualification test establishes that a design and a process can meet a requirement, usually by testing a small number of units to destruction or to a defined life, and its purpose is evidence. A burn-in screen applies a short, mild stress to every unit to precipitate latent defects, and its purpose is to remove the weak units from a shipment. A qualification establishes what the product can do; a screen removes what it cannot. Where a burn-in screen is being used to demonstrate reliability, it is being used for the wrong purpose.

Temperature and Duration

The screen works because the time to failure of most failure mechanisms accelerates with temperature, so a short period at a raised temperature is equivalent to a much longer period at the operating temperature. The acceleration factor depends on the mechanism, and the well-known models differ significantly, which is why a screen is chosen empirically rather than calculated precisely. A common screen runs at the maximum rated ambient temperature, or slightly above, for a period long enough to precipitate the defects that matter, which in practice ranges from a few hours for a consumer product to several days for a product where a field failure is very expensive.

Power Cycling and Bias

Many mechanisms only appear when the unit is powered, and a surprising number appear only when it is cycled. A screen that holds the unit at temperature without power tests the materials, and a screen that powers it tests the electrical function, but neither tests the expansion and contraction that occurs when the unit heats and cools in use. Where the product will be cycled in service, the screen should include power cycling, and where a particular failure mode is suspected, such as a marginal solder joint or a connector contact, the cycling is what precipitates it. Bias also matters: a unit that is powered but not exercising its interfaces will not reveal a fault that only appears when a bus is active.

What It Cannot Find

Aging does not find design errors that are not stress related, and it does not find faults that require a specific stimulus the test does not apply. It also does not find a defect that appears only after a long period, since the whole point of the screen is to be short relative to the life. Where a product has a known wear-out mechanism, such as an electrolytic capacitor or a connector with a limited insertion life, the screen will not reveal it, and the answer for those is a design decision or an end-of-life calculation rather than a test.

Cost and Yield Effects

The cost of a screen is the chamber time, the fixtures, the monitoring equipment, the handling, the electrical test before and after, and the units that are scrapped. Against that it has to be set the cost of a field failure, which in automotive, medical or infrastructure products is much larger than the unit price. Where the screen is applied, the yield after the screen should be monitored: a drop in yield points to a process change, and a rising yield to a process that has become more consistent. The screen should also be reviewed periodically, because a screen designed for an older process may be masking a problem that a better process no longer has.

Planning the Test

Decide what the screen is for, which failure modes it is meant to precipitate, and what evidence will show that it worked. Choose the temperature and duration from those failure modes rather than from a catalogue, and include power cycling where the mechanism requires it. Define the acceptance criteria and the re-test after the screen, since a unit that fails after the screen has to be handled by a defined disposition rather than by an ad hoc decision. Finally, record the results, because the value of a screen is not only the units it removes but the trend it reveals.

Documenting the Screen

Whatever the screen consists of, it has to be written down with enough detail to be repeated: the chamber profile, the dwell at temperature, the number of power cycles, the bias conditions, the interfaces that were exercised and the test performed before and after. A screen that exists only as a habit is a screen that will drift, and a drift in the wrong direction either wastes capacity or stops catching the defects it was designed for. The record also makes the screen auditable, which matters where the product is supplied into a regulated market and the customer asks how the early failure rate is controlled. Where the screen is changed, the change should be treated as a process change and its effect on the post-screen yield measured, because that yield is the only feedback the screen provides about whether it is still doing its job.

FAQ

What is burn-in for? To precipitate latent defects in a short, stressed period so that early failures happen in the factory rather than in the field.

How long should it last? Long enough to precipitate the mechanisms that matter, from a few hours for consumer products to several days where a field failure is very expensive.

Is a burn-in the same as a qualification test? No. A qualification establishes what the product can do; a screen removes the units that would fail early.

Does it need power cycling? Where the failure modes are thermal or mechanical, yes, since a static temperature does not produce the expansion and contraction that cause them.

Can it replace design validation? No. It cannot find design errors that are not stress related or faults that need a stimulus the test does not apply.

Conclusion

An aging test is a screen against infant mortality, not a demonstration of reliability, and it should be designed around the failure modes it is meant to precipitate. Choose the temperature, the duration and the cycling from those mechanisms, define the disposition of failures, and track the yield trend. Test strategy sits alongside PCBA testing, the assembly processes it screens are described in PCB assembly, and the reliability discipline belongs to quality management. Screening is normally set up alongside prototype PCB assembly in 2026.