Solder Joint Pull Testing and Failure Mode Analysis

A pull test measures the force needed to separate a solder joint, and its real value is not the number but the failure mode that accompanies it. Two joints can break at the same force, one cleanly through the solder and the other by lifting the pad off the laminate, and only the second one tells you that the board rather than the process is the weak point.

The test is therefore read in two stages. The force is compared against the requirement for the lead diameter or pad size, and the fracture surface is classified so that the result points at a cause. Without the second stage, a passing number can hide a pad crater that will appear as a field failure later.

What a Pull Test Measures

A pull test applies a tensile load to a joint, usually through a hook on a lead or a grip on a wire, and records the maximum force and the extension at which it occurs. For through-hole joints, the load path runs from the lead through the solder fillet into the plated barrel and the pad.

Every element in that path can be the weakest link. The measurement therefore describes the assembly rather than the alloy, which is why the same solder can pass on one board design and fail on another with a thinner annular ring or a smaller pad.

Test Methods and Grip Geometry

Grip geometry changes the result. A hook that bears on the lead close to the board creates a different stress distribution from one that pulls from further away, and a 45 degree pull loads the fillet in shear as well as in tension, which is often closer to service conditions.

Pull rate matters for the same reason: a fast pull measures a more brittle response. Most procedures settle on a steady rate rather than a shock load, with the joint pre-loaded lightly so that slack in the fixture is not recorded as extension. Consistency between operators matters more than the exact value chosen.

Force Limits and Acceptance Values

Acceptance values are set by the applicable standard and scaled by lead diameter, pad size and alloy, so a single number applied to every joint on a board is not meaningful. For a typical 0.8 mm through-hole lead, minimum values in the region of 10 to 20 N are common in published requirement tables.

Where a requirement does not exist for the joint in question, the process owner should establish one from the distribution of good boards rather than from a single sample. A limit set at three standard deviations below the mean of a proven process is defensible in a way that a copied number is not.

Failure Mode Classification

Failure modes are grouped into a small number of families: solder fracture, interfacial fracture between the solder and the intermetallic layer, pad lift or cratering in the laminate, and lead or wire fracture away from the joint. Each family has a different cause and a different corrective action.

Solder fracture with a ductile, torn appearance usually indicates a joint that was starved of solder or cooled too quickly, while a clean interfacial break points to a contaminated or over-aged surface. Pad lift points at the laminate and the thermal profile rather than at the alloy, and the distinction is what makes the classification worth the effort.

Intermetallic Layer and Aging

The intermetallic layer at the solder to copper interface grows with time and temperature, and it is a brittle phase. A layer of 1 to 2 µm behaves well, while a layer grown to 4 µm or more by extended high temperature exposure becomes the fracture path and the joint loses strength.

Where the pull result falls after thermal cycling, the microsection analysis of a tested joint shows whether the intermetallic layer or a crack initiated elsewhere is responsible. Comparing the thermal shock and thermal cycling exposures is also worth doing, because the two drive different failure mechanisms.

Pad Cratering and Laminate Strength

Cratering is a laminate failure that starts under the pad and propagates into the resin, and it is often caused by the pull test itself when the joint is stronger than the board. A crater that becomes visible during testing is evidence about the design, but it also destroys the sample.

Signs of incipient cratering include a pad that lifts with a cone of resin attached and a crack that runs from the pad edge into the weave. The pad cratering pattern is easiest to see in a polished section, and it should be recorded with the pull result rather than separately.

Sample Selection and Board Support

Samples should come from the boards that were exposed to the process being evaluated, not from untested stock, and they should represent the smallest pad and the heaviest thermal mass on the panel. Testing only convenient joints produces an optimistic picture that the design does not deserve.

Board support is part of the set-up. A board that is free to flex records a lower force, because the panel bends before the joint reaches its limit, and the deformation is easily mistaken for a ductile joint. The same support fixture should be used for every test so that results remain comparable over time.

Reading the Results Against the Process

A single low result is an outlier until it repeats. Plotting the results per shift and per joint type shows whether the process has shifted or whether one sample was damaged before testing, and the plot is far more useful than a folder of individual readings.

Where the distribution widens rather than moves, the cause is usually variation in the process rather than a change in the material. Solder temperature, flux activity and the time the board spent in the wave are the usual suspects, and the solderability test result on incoming material should be checked at the same time.

Records and Routine Monitoring

Records should store the force, the failure mode, the pull rate, the grip type and the joint identification. Those fields turn a destructive test into a process control tool rather than an annual audit. Those five fields allow two engineers to compare trends years apart, which is the practical test of whether a record is worth keeping.

Routine monitoring frequency follows the product: once per shift for a general assembly, once per batch or once per panel for a safety related product. Whatever the frequency, the same fixture, the same rate and the same failure mode definitions should be used, because a change in any of the three breaks the comparison.

FAQ

What does a solder joint pull test actually prove? It proves the strength of the weakest element in the joint path, which may be the solder, the interface or the laminate, and the failure mode identifies which one that was.

Why does the same joint give different forces on different fixtures? Grip geometry, pull rate and board support all change the load path, so a fixture change is a process change and the results before and after it are not comparable.

Should a pad crater found during testing fail the board? Yes. A crater indicates that the laminate failed before the joint did, which is a design or process issue rather than a measurement artefact.