Solder Mask Sliver and Dam Width Control on Dense Boards
Solder mask is not a cosmetic layer. On a dense board it is a structural part of the assembly process, keeping solder where it belongs, protecting copper from the environment and defining the dams that stop a bridge between two adjacent pads. When the mask is squeezed into a narrow strip between features, that strip is called a sliver, and a sliver that lifts, cracks or washes away takes the process control with it. This article explains what makes a sliver fail, how dam width is set, and how to design dense areas so that the mask can do its job.
What a Sliver Is and Why It Fails
A sliver is a narrow isolated strip of solder mask, usually between two closely spaced openings, sometimes between an opening and the board edge. It has two failure modes. It can lift during development or during reflow, and it can crack along its length so that the two openings become effectively connected.
Both failures are driven by the same mechanics: the sliver is thin, it is bonded to a small area of substrate, and it sees the thermal expansion of the copper beneath it. A wide strip of mask can absorb that movement across its width, while a thin strip has nowhere to relieve the stress. The mask material itself is brittle once it is cured, so a sliver behaves like a small beam bonded at one edge and unsupported on the other.
Dam Width Fundamentals
The dam width is the width of the mask strip that remains between two adjacent openings, and it is set by the pad spacing, the opening size and the registration tolerance. For a given pad pitch, a larger opening leaves a narrower dam, so the opening should be no larger than the assembly process actually needs.
A wide opening is often specified out of habit, using the same mask expansion that was used on a coarse board. On a fine pitch layout that habit removes the dam entirely. The opening should be derived from the land pattern, the expected registration shift and the minimum dam width the fabricator can hold. Where the calculation says the dam will disappear, the pitch is the problem and the layout has to change rather than the process.
Registration and Its Effect on Dam Width
Solder mask registration is the alignment of the mask artwork with the copper pattern. It is never perfect, so the dam that appears on the finished board is the design dam minus the registration shift in the worst direction. A design that leaves exactly the minimum dam with perfect alignment will produce rejected boards as soon as the shift moves in the wrong direction.
The shift has two components. The lamination of the dry film or the exposure of the liquid mask introduces a registration error, and the copper pattern itself sits at a different position from the artwork because of the imaging and etching steps. Our tooling hole notes describe how the registration scheme is established across the panel.
Imaging and Development Limits
The resist being imaged has a resolution limit, and a feature below that limit will not reproduce consistently even if the artwork is perfect. Development removes unexposed material, and a narrow strip between two large openings is exposed to developer from both sides, so it can be undercut and washed away before the wider areas have cleared. That is why development time is a compromise: enough to clear a dense pattern, not so much that narrow features are lost.
The consequence is that the practical minimum sliver width is larger than the theoretical resolution. The figure should come from the fabricator and should be stated as a design rule with a margin, not as a target to be hit exactly.
Tenting Vias and the Consequence
Tenting a via covers the hole with solder mask to stop solder from wicking into it during reflow. The tent has to bridge the barrel, which means it needs a fully covered annular ring and a mask film thick enough to survive the process. A tent over a via on a dense layout competes directly with the surrounding dams for the same limited space.
Where the space does not allow a reliable tent, the alternatives are to plug the via, to leave it open with a mask dam around it, or to move it. Each has a cost, and the decision should be made at design time rather than discovered when the panel is imaged.
Copper Features That Create Slivers
Slivers are usually caused by copper, not by the mask artwork. A trace that passes between two pads, a thermal relief spoke that leaves a narrow gap, a via placed close to a pad edge and a copper pour that runs along a row of pads all squeeze the mask into a thin strip.
The fix is geometric. Moving the trace to the other side, adjusting the spoke width, shifting the via or opening the pour in that area all remove the sliver without changing the electrical function. Those changes are trivial at layout time and expensive later, and they rarely cost electrical performance because the sliver area was never carrying useful copper.
Design Rules for Dense Areas
A workable set of rules states the minimum mask dam width, the minimum sliver width, the allowed mask expansion for each pitch class and the maximum aspect of an opening. The rules should also state the minimum distance from a mask opening to a board edge and to a route path, because a sliver at the edge is often lost entirely during routing or profiling, leaving an exposed copper edge that later corrodes.
Two numbers matter most: the mask expansion and the minimum dam. Setting the expansion per pitch rather than globally is what allows a board with both coarse and fine areas to be built with one process.
Inspection and Acceptance Criteria
Inspection looks for missing dams, lifted slivers, cracks, and mask migration into an opening. A missing dam between two pads that will be wave soldered is a defect rather than a cosmetic finding, because it removes the barrier that prevents a bridge.
The acceptance criteria should be written in terms of function: does the remaining dam prevent a bridge, does the opening leave enough solderable area, and is the copper protected. Our land pattern notes explain how the pad geometry drives the mask opening, and our quality guide describes how a mask defect is classified.
Working With the Fabricator Early
The minimum dam and sliver widths are capability figures, and they change with the mask type, the copper thickness and the surface finish. Asking for them at the start of a layout review costs nothing, while finding out at the imaging step costs a panel and a schedule.
At gopcb those limits are recorded with the process capability for the design, and the fabrication notes that accompany an order state the mask type, the expansion and the minimum feature that the process will hold.
FAQ
What is a safe minimum dam width? It depends on the mask type and the copper thickness, but a common working figure is around the same order as the mask thickness, with more margin on a dense board. The fabricator capability figure should be used rather than a generic number.
Can a sliver be repaired? A small lifted sliver can sometimes be touched up, but a repair has to restore the function, not the appearance. On a production panel, rework of a mask feature is rarely justified.
Does a larger mask opening help assembly? It helps by exposing more solderable copper, but it also reduces the dam. The opening should be the smallest that still supports reliable paste release and inspection for the given pitch.
