Leak Testing For Sealed And Coated Assemblies

Where a product must survive water, dust or a cleaning process, the enclosure is expected to keep the environment out. The board inside contributes to that expectation: a coating, a potting compound or a sealed housing all form part of the barrier, and each of them can fail in ways that are difficult to see. Leak testing is the method used to prove that the barrier works before the product reaches a customer.

This article covers the methods used, what each of them can and cannot detect, and how the board design affects whether a test is meaningful.

The subject sits between several disciplines, which is why it is often handled late: the mechanical engineer owns the seal, the process engineer owns the test, and the layout engineer decides whether the board can be tested at all.

What Is Being Tested

A leak is a path for a fluid or a gas between two regions. For a sealed enclosure the concern is usually water ingress and the associated corrosion; for a coated board it is the integrity of the coating over the features it protects. The two cases need different tests, because a coating that is continuous over a conductor can still have pinholes elsewhere, and a housing that holds pressure can still admit moisture through a permeable gasket.

The requirement is normally expressed as an ingress protection rating, and the rating determines the test. Air pressure decay, water immersion, helium mass spectrometry and tracer gas methods each cover a different range of leak rates, and choosing the wrong method for the requirement produces either a test that passes everything or a test that fails good product.

Pressure And Vacuum Decay

The simplest method pressurises the sealed volume and measures how quickly the pressure falls. It is fast, it needs no special gas, and it can be integrated into a production line. Its limitation is sensitivity: a small leak produces a slow pressure change, and the measurement has to distinguish that change from temperature effects and from the compliance of the enclosure itself.

Making the method reliable requires controlling the temperature, allowing the enclosure to settle before measuring, and calibrating with a known leak. Where the volume is small, the decay from a given leak is faster and the method becomes more sensitive; where the volume includes a large sealed cavity, the same leak may be undetectable. The potting compound used inside the housing changes the effective volume and the thermal behaviour, which the setup has to account for.

Helium And Tracer Gas

Helium mass spectrometry detects leaks several orders of magnitude smaller than pressure decay, because the gas is light, inert and present in the atmosphere at a very low concentration. The sealed part is filled with helium or exposed to it, and the spectrometer measures the helium that escapes or enters. The method is the reference for high integrity seals.

Its costs are the equipment and the time. It also requires a vacuum chamber or a hood for the measurement, and the part has to be clean, because contamination can mask a leak or introduce a false signal. For a coated board, the helium method can detect a breach in a sealed overmould but it cannot tell where the breach is.

Temperature is the variable that most often makes a leak test untrustworthy. A sealed volume that is warmer than its surroundings will cool during the measurement, and the resulting pressure drop looks exactly like a leak. The standard remedy is to let the part reach thermal equilibrium before the measurement begins, and to run a reference measurement on a known good part alongside the production parts.

Testing A Coated Board

A conformal coating is not intended to be a hermetic barrier, so a leak test of a coated board tests something else: the continuity of the coating over the areas that matter. The practical methods are optical, with fluorescent inspection under ultraviolet, and electrical, by measuring insulation resistance or leakage current under bias after exposure to moisture.

The electrical methods are the ones that correlate with field behaviour, and the protective function of the coating is what they measure. A damp heat exposure followed by an insulation resistance measurement is a reasonable qualification test, and the same measurement on a production sample gives an indication of process control.

Test Points Gaskets And Sealing Surfaces

Any opening in a sealed enclosure is a potential leak, and a test point, a connector or a vent is an opening. A connector that is sealed to the housing with an O ring relies on the surface finish and the flatness of both parts, and the board design affects that: a connector whose mounting area is on a thin section of a warped board may not compress the seal evenly.

Where a test has to be performed through the enclosure, the port must be sealed afterwards to the same standard, and the sealing method should be specified. The manufacturing tolerances on the board outline and on the mounting features then become part of the sealing performance, not only part of the assembly fit.

Design Decisions That Help

Three design choices make leak testing easier. Keeping the sealing surfaces flat and free of interruptions from vias or test points avoids paths that the seal cannot follow. Avoiding sharp features under a coating improves its coverage and reduces the chance of a thin spot. And keeping the sealed volume predictable makes the pressure decay method usable.

It also helps to plan the test before the enclosure is designed. A test method that needs a fill port, an access point or a defined volume imposes constraints that are cheap to accommodate early and expensive to retrofit.

Interpreting A Failure

A failed leak test is a starting point rather than a conclusion. The leak rate that the test reports may be a real leak, a leak in the test setup, or a temperature effect that the setup did not compensate. The first step is to confirm the failure by repeating the test on the same part after allowing it to stabilise.

Locating the leak is the second step, and the methods differ. Immersion with a bubble test locates a gross leak; a mass spectrometer with a sniffing probe locates a small one; and a dye penetrant followed by a sectioning operation locates a leak inside a coating. The result then feeds back into the design, which is the point of the exercise.

Another consideration is the sequence of operations. A leak test performed before the coating is cured measures a different assembly from one performed afterwards, and the two are not interchangeable. The test should be placed at the point in the process where the barrier it verifies is complete, and that point should be documented rather than assumed.

FAQ

Can a leak test replace a moisture exposure test? No. A leak test measures a path; a moisture exposure test measures the effect of the environment on the assembly, including permeation through materials that are not leaks.

How often should production be leak tested? Every unit for a safety critical seal, and by sampling where the process is stable and the consequence of a leak is not severe. The sampling rate should be justified by the process capability rather than by habit.

Does potting remove the need for a sealed housing? It reduces the requirement, because the compound protects the conductors directly. It does not remove it where the assembly contains components that are sensitive to moisture themselves, such as sensors or connectors.