PCB Hole Wall Quality Criteria And How To Measure Them

A plated hole is the weakest structural element on most boards, and its quality is decided by a wall that cannot be seen without cutting the board open. Because the evidence is destructive, hole quality is judged through a combination of a fast indirect test and a slower direct measurement, and both need written criteria before they mean anything.

What Defines A Good Hole Wall

The requirements are simple to state. The copper foil of every inner layer must be exposed at the hole wall, the plating must be continuous and adherent over the full depth, and there must be no resin between the plating and the copper. The wall itself should be free of torn glass, pits, and voids that interrupt the barrel.

Everything else is a tolerance on those four statements. Plating thickness has a minimum, smear has a maximum, etchback has a range, and voiding has a limit. Because the four interact, a criterion written in isolation, such as a minimum thickness with no void limit, can be satisfied by a barrel that will still fail in service.

Backlight Grading And Its Limits

The backlight test, also called light of the hole, views a sectioned hole against a bright source and grades the fraction of the wall where copper is continuous. A dark line between the plating and the foil indicates resin or oxide. Published grading scales typically run from zero to ten, and the common production requirement is a grade corresponding to at least eighty percent coverage.

The method is fast and sensitive to smear, which makes it a good routine check. Its limitation is that it grades an appearance rather than a dimension. A hole can pass backlight and still be thin enough in the middle of the barrel to crack under thermal cycling, so the test complements a microsection rather than replacing it.

Microsection Criteria For Barrel Plating

The microsection gives the numbers. Copper thickness in the barrel is measured at the mid point of the board and at the surface, and the minimum is usually specified as a two-sided figure such as twenty to twenty-five micrometres for a standard class. Thickness at the surface is normally higher because the plating current is more accessible there.

The same section shows voids, cracks, and separations. Acceptance is usually written as no voids that penetrate the barrel, no separation between the plating and the foil, and no cracks in the copper after a thermal stress cycle. Each of these is judged on a defined number of holes taken from a defined location. Our notes on plating thickness explain how the measurement is taken.

Resin Smear And Nail Heading

Resin smear is measured directly on the section as a layer between the plating and the inner layer copper. A typical limit is that no smear is acceptable on more than a stated fraction of the interface, or that any measurable layer constitutes a failure depending on the acceptance class of the product.

Nail heading is a different phenomenon: the inner layer copper is deformed into the hole by the drill, so the copper appears thickened and pushed inward. A small amount is tolerated, while a pronounced heading means the drill parameters or the bit condition are wrong, and it usually accompanies a rough wall and other defects.

Etchback And Resin Recession Limits

Where etchback is required, the section is used to measure how far the copper protrudes into the hole. The target is normally a range, and the upper bound exists because excessive recession leaves an unsupported foil that can trap chemistry and cause plating voids or layer-to-layer shorts.

A specification that states only a minimum etchback encourages the process to run at the top of the range, which is where the failures come from. Writing the requirement as a window, for example ten to fifty micrometres, keeps the process in the middle where both the smear risk and the wicking risk are low.

Glass Fibre Protrusion And Pit Marks

Resin etches far more readily than glass, so the fibres next to a hole can stand proud of the wall. The plating must then cover an uneven surface, and over a protruding bundle the copper is thinner than elsewhere in the barrel. Pit marks, where a fibre has been pulled out leaving a cavity, are worse because they create a void by construction.

Acceptance criteria usually limit the size and the number of pits per hole and prohibit pits that extend into the wall beyond a stated depth. Both measures require a section, because neither is visible from outside. Where a material is known to be prone to fibre protrusion, the criteria should be agreed with the fabricator before the first build rather than after the first rejection.

Hole Wall Pull Away And Barrel Cracks

Pull away is a gap between the plating and the hole wall that develops during thermal stress, and it appears on the section after a stress cycle rather than before. It indicates poor adhesion, which in turn points at the desmear process or at contamination rather than at the plating bath. A criterion of no pull away after stress testing is demanding but common on high reliability products.

Barrel cracks appear in the same test and are usually a thickness problem, since a thin deposit cannot absorb the expansion mismatch between copper and laminate. The classic combination is a board that passes as fabricated and fails after three to five thermal cycles, which is why the stress test is part of the acceptance rather than an optional extra.

Sampling Plans And Sample Location

The sample location matters as much as the sample size. A section taken from the centre of the panel describes the nominal process, while a section from the edge describes the region with the most plating current and the highest thickness. Neither is representative on its own.

A practical plan takes sections from several panel positions and includes the smallest and the highest aspect ratio holes on the board, because those have the least margin. Our notes on hole types describe how the geometry differs, and board quality notes cover how the findings are recorded.

Writing Acceptance Criteria Into The Order

Acceptance criteria should name the test, the number of samples, the measurement location, and the limit for each property. They should also name the acceptance class that applies, since a product built to a general purpose class and one built to a high reliability class have different limits for the same defect.

Stating the class is what makes the criteria enforceable. A drawing that says only to remove all smear gives no basis for accepting a lot with a marginal result, and the argument then turns on opinion rather than on a document. Our notes on the hole copper measurements describe the values that typically appear in the report.

FAQ

Is backlight enough on its own? It detects smear well but says nothing about plating thickness or voiding. Pair it with a microsection for any product with a reliability requirement.

How many holes should be sectioned? Enough to cover the smallest and highest aspect ratio holes, taken from more than one panel position, at a defined frequency per lot.

What causes pull away after thermal stress? Poor adhesion between the plating and the wall, usually from incomplete desmear or from contamination introduced between the desmear and plating steps.