Fillet Heel Cracks: Where Solder Joints Start to Fail

The heel of a fillet is the region where the solder rises from the pad onto the lead, and it is the point of highest stress in a through hole solder joint. A heel crack begins there and propagates through the fillet or along the interface, eventually opening the connection.

Because the crack starts at a location that is hidden by the component body or by the fillet itself, it is rarely seen at inspection. It appears as an intermittent open after thermal cycling, or as a resistance rise in a monitored test.

What a Heel Crack Is

The failure is a crack that initiates at the outer surface of the fillet where it curves up toward the lead. In a microsection it appears as a line running from that surface inward, sometimes through the bulk alloy and sometimes along the boundary between the alloy and the lead.

The direction of the crack follows the stress. A joint that is loaded by differential expansion of the board and the lead usually cracks across the fillet, while a joint loaded by bending of the lead tends to crack along the interface.

Where the Heel Sits in a Joint

A complete fillet is concave, rising from the pad to a defined wetting angle on the lead. The heel is the transition between the flat portion on the pad and the rising portion on the lead, and it is where the section thickness changes fastest.

Fillet geometry is specified by the applicable standard in terms of wetting angle and coverage. A fillet that is too small concentrates stress in a thin section, while one that is excessively large adds material that increases the load on the lead. Both extremes raise the risk at the heel.

Thermal Cycling and Stress Concentration

During thermal cycling the board expands in plane while the lead moves with the component, and the difference is accommodated by the joint. The heel has to absorb that movement in a small volume of alloy, which is why it is the first place to crack.

The severity depends on the temperature range, the number of cycles and the difference in expansion between the materials. A cycle from minus 40 to 125 C repeated a thousand times is far more damaging than a single excursion to the same peak, because the damage accumulates. The rate at which it accumulates depends on the alloy, the joint geometry and the stiffness of the lead, so two joints on the same board can have very different lives.

Lead Forming and Component Stress

A lead that is pre loaded when the component is inserted stores elastic energy that pushes on the joint from the moment it solidifies. The heel then carries a permanent load in addition to the cyclic one, and the time to failure is shortened.

Lead forming also controls where the joint can move. A lead with an inadequate stress relief bend transmits assembly and handling loads directly into the fillet, while one with a generous bend allows movement to be absorbed away from the heel. This is a design decision that no amount of process control can correct later. Insertion tooling that seats a lead without bending it, and a seating height that leaves the lead free to move, both reduce the pre load the joint carries.

Pad and Mask Geometry Effects

The pad decides how much solder is available for the fillet and how far the fillet can spread. A pad that is barely larger than the hole leaves a small fillet, and a pad connected to a heavy plane draws heat away so that the fillet forms differently on the two sides of the joint.

Mask encroachment is a related factor. A solder mask that overlaps the pad restricts where solder can wet, and the restriction usually appears on one side only, producing an asymmetric fillet with a thinner heel on the constrained side. Mask registration should therefore be checked whenever heel cracks appear on a particular side of a component. Symmetry of pad size between the two sides of a hole matters for the same reason.

Alloy and Microstructure

Lead free alloys are stiffer and less ductile than the tin lead alloys they replaced, so a given amount of differential movement produces a higher stress at the heel. The same change also raises the temperature at which the joint is at its strongest relative to the cycle range.

Microstructure matters within the alloy. Intermetallic layers that are too thick are brittle and provide a preferred crack path along the interface, while a coarse grain structure offers less resistance to crack initiation. Both are controlled by the thermal history of the assembly, as the work on bend testing and on intermetallic growth shows.

Detection in Section and by Test

Microsectioning is the direct method. A section through the joint shows the crack and its path, and it allows a judgement about whether the failure started at the surface or at the interface. Sections should be taken after a thermal excursion, because a joint that has not been cycled will often show no crack at all.

Thermal cycling with resistance monitoring is the traditional test, and it produces a time to failure that can be compared between designs. The microsection preparation after the test must not smear the soft alloy, since a preparation artefact at the heel looks very much like the defect being investigated.

Design Rules That Reduce Risk

Reducing the stress at the heel begins with geometry. Provide an adequate stress relief bend in the lead, avoid placing a stiff component directly beside a stiff board feature, and keep the pad large enough for a well formed fillet without adding unnecessary solder volume.

Material selection is the second lever. A laminate with a lower expansion in the z direction moves less, and a lead material with a closer expansion match to the board reduces the differential that the joint has to absorb. Both changes are cheap at design time and impossible to apply afterwards.

Process Controls and Records

The process contributes through fillet shape and through thermal history. A profile that gives the correct wetting angle and a consistent fillet on both sides of the lead produces a joint that resists cracking, while one that leaves an asymmetric fillet starts the joint at a disadvantage.

Records should capture the profile, the alloy, the pad and mask dimensions and the results of any thermal cycling test. Read together, they show whether a heel crack population is a design issue or a process drift, which is the difference between a change of drawings and a change of settings.

FAQ

What causes a heel crack in a solder joint? Stress concentration at the point where the fillet rises onto the lead, driven by differential expansion during thermal cycling and aggravated by pre loaded leads or an asymmetric fillet.

Can a heel crack be seen without sectioning? Rarely. The crack begins at a surface that is usually hidden by the component body or by the fillet curve, so it is normally found by thermal cycling with monitoring or by microsection.

Does a larger fillet prevent heel cracks? Not by itself. A larger fillet can increase the load on the lead and the pad, and the aim is a correctly formed fillet with the right wetting angle rather than simply more solder.

}