Biased HAST Testing: Humidity Bias Reliability for PCBAs

Biased HAST puts a powered assembly into a saturated steam environment and holds it there, typically at 130 C and 85 percent relative humidity, until an insulation failure appears or the test is complete. It compresses months of damp heat exposure into a few days, which is why it is used to qualify coatings and cleanliness rather than to describe field life.

The word biased is the important part. Applying a bias voltage across adjacent conductors turns a passive moisture test into an electrochemical one, because the field drives ion migration and makes failure mechanisms grow that a powered-off test would never reveal.

What Biased HAST Measures

The test looks for loss of insulation between conductors that should stay isolated. That includes electrochemical migration across the surface, conductive anodic filament growth inside the laminate, and leakage through a coating that has absorbed moisture or lost adhesion to the board beneath it.

It is not a general purpose reliability test. A design with no fine spacing, no coating and no ionic risk gains little from it. The test earns its cost on assemblies with tight conductor spacing, high impedance nets or a conformal coating, where small changes in process can create large changes in leakage.

Unbiased and Biased Test Conditions

The two variants answer different questions. The unbiased test, described in the JESD22-A118 method, measures the moisture resistance of the package or coating without the influence of an electric field. The biased test, following JESD22-A110, applies a voltage to the specimen during the exposure and expects the failure to arrive earlier.

Typical biased conditions are 130 C, 85 percent relative humidity, a saturated vapour pressure near 2.7 bar, and a duration of 96 to 100 hours, though 110 C and 85 percent is also used where the assembly cannot tolerate the higher temperature. The chosen condition should be stated with the duration and the bias, because a result quoted without them cannot be compared with anything.

Chamber Conditions and Acceleration

HAST accelerates the conventional 85 C and 85 percent humidity test by raising both temperature and pressure. The relationship is not linear, so a result at 130 C cannot simply be scaled by a constant to predict years at 40 C. The test ranks designs and detects process escapes; it does not by itself produce a lifetime number.

Because the acceleration is aggressive, small chamber errors matter. A door seal that lets the chamber run below saturation, a thermocouple reading the air rather than the specimen, or a humidity sensor that has drifted will all change the effective stress, and the failure that results will be attributed to the product rather than to the chamber.

Test Board and Bias Network Design

The coupon has to represent the production spacing and the production coating thickness, or the result describes something else. Comb patterns with the same conductor width and gap as the real product, built on the same laminate and coated in the same pass, are the only configuration that gives usable evidence.

The bias network needs as much care. Voltage is applied between adjacent fingers through series resistors that limit current when a short forms, and the resistors are chosen so that a failure changes the measured voltage without burning the coupon. Leaving the current unbounded destroys the evidence that the failure created.

Ionic Contamination and Coating Interactions

Residues left by flux, plating and handling dissolve in the condensed water and form the electrolyte that makes migration possible. This is why the test is so often run alongside a cleanliness measurement, since the two together show whether the assembly starts clean and whether it stays isolated under stress.

A coating changes the picture rather than removing it. A well adhered film blocks the electrolyte, but a coating that has poor adhesion at the conductor edges allows water to creep underneath and then traps it against the surface, which is often worse than no coating at all. The coating adhesion test and the ionic contamination verification are the two checks that predict how a coated assembly will behave in the chamber.

Failure Modes the Test Exposes

Surface electrochemical migration is the most common result. Metal dissolves at the anode, moves through the water film and deposits as a dendrite at the cathode, and the dendrite grows until the gap closes. Dendrites are usually found by microscopic examination after the failure, near the conductor edges where the field is highest.

<img src="https://www.gopcba.com/wp-content/uploads/2026/08/44.jpg" alt="Comb pattern test coupon wired for insulation resistance monitoring” />

Conductive anodic filament growth is the second mechanism, and it happens inside the laminate along the glass fibre bundles rather than on the surface. It is invisible from the outside and cannot be prevented by cleaning, because the path lies beneath the resin. Its behaviour under bias is covered in the work on conductive anodic filament formation.

In-Situ Monitoring and End Point Testing

Monitoring insulation resistance during the test is far more informative than measuring it only at the end. An in-situ record shows when the leakage started to rise, whether it recovered when the bias was removed, and whether one coupon failed early while the others stayed stable, all of which point to a cause rather than merely to a failure.

An end point measurement alone can be misleading because a dendrite may be destroyed as it burns open, leaving a coupon that passes the final check but contains the evidence of a mechanism that will return. Where in-situ monitoring is not available, the coupons should be examined microscopically even if they pass, and the surface insulation resistance method used as the measurement basis so the numbers are comparable between lots.

Sample Size, Handling and Control Lots

A handful of coupons proves very little, because the mechanisms are statistical. A control lot built with a known good process should travel with every test run, so that a failure can be attributed to the test article rather than to a chamber that has gone out of specification.

Handling between coating and test is part of the experiment. Bare fingers on a coated surface, a coupon stored in a cardboard box, or condensation formed during transfer all add contamination that the process did not produce. Coupons should be handled with gloves, stored in a dry environment and transferred to the chamber in a sealed carrier.

Interpreting Results and Failure Analysis

The first question after a failure is whether the leakage is on the surface or below it. Optical inspection finds dendrites, while a microsection is needed to show filament growth inside the laminate or a delaminated coating interface. The two conclusions lead to completely different corrective actions.

The second question is whether the failure is marginal or systematic. One coupon that failed at 60 hours out of twenty that completed 100 hours suggests a localised escape, whereas a cluster of failures at the same time suggests a process shift. Recording the time to failure for each coupon, rather than a simple pass or fail, is what makes that judgement possible.

FAQ

What is the difference between biased HAST and unbiased HAST? Both use the same temperature and humidity, but the biased version applies a voltage across the conductors during the exposure. Bias accelerates electrochemical failure mechanisms such as migration and filament growth, which the unbiased test cannot stimulate.

Which conditions are used for biased HAST? A common condition is 130 C, 85 percent relative humidity for 96 to 100 hours, following JESD22-A110. Where the assembly cannot tolerate that temperature, 110 C and 85 percent is used with a longer duration, and the condition must always be reported with the result.

Can biased HAST predict field lifetime? Not directly. It is an accelerated comparison test that ranks designs and detects process escapes. Converting the result into a field lifetime requires a validated acceleration model and a much broader qualification programme.