Blind and Buried Vias in HDI PCB Manufacturing

When routing density rises far enough, a through hole becomes an unaffordable use of the board. It occupies a landing pad on every layer, blocks the routing channels around it on every layer, and prevents the layers above and below from being used for anything else. Blind and buried vias remove that cost by connecting only the layers that need to be connected, but they exchange a routing problem for a manufacturing problem, because the board can no longer be built in one lamination cycle and the holes can no longer be drilled from the outside.

Why Through Holes Are Not Enough

A through hole is drilled after all the layers are laminated, so its position is fixed relative to the finished board and the process that produces it is well understood. A blind via reaches an inner layer from the outside, and a buried via connects two inner layers without reaching either surface. Neither can be drilled through the finished stack, so the board has to be built in stages, with some holes formed and filled before the remaining layers are added. That sequence is the defining characteristic of an HDI board, and almost every manufacturing rule that follows is a consequence of it.

Structures and Stackup Notation

HDI stackups are described by the number of build-up layers on each side of a core, so a notation such as 1+4+1 describes one build-up layer on each side of a four-layer core. A 2+4+2 board adds a second build-up layer on each side and introduces second-order vias that land on a pad formed by the first build-up step rather than on the core. Stacked structures, where a second-order via sits directly on top of a first-order via, give the highest density and the tightest alignment requirement; staggered structures, where the vias are offset, are more forgiving to build. The choice between them belongs to the stackup decision, and the trade-offs are set out in this article on blind and buried via stack selection.

Build-Up Materials

The dielectric added in each build-up step is usually a resin-coated copper foil, consisting of a copper layer with an uncured resin film behind it. The copper is available in half-ounce and third-ounce weights, and the resin thickness is commonly between about 55 and 100 micrometres in five-micrometre steps. Because the layer is pure resin, it is well suited to laser drilling, and its dielectric constant is typically in the range of 3.2 to 3.8, with 3.6 a common value for design purposes. A laser-grade prepreg with a looser glass weave is also used, since the open structure removes more easily under the beam than a tightly woven fabric, but laser drilling through glass-reinforced prepreg remains more difficult than through resin-coated foil and is not available from every fabricator.

Sequential Lamination and the Order of Operations

The build sequence determines when each hole is formed and when the layers are bonded. A conventional flow laminates the core, drills and plates the buried vias, plugs them, forms the inner-layer circuitry, then laminates the build-up layer and drills the blind vias before the outer layers are patterned. A more recent flow moves the via filling later and changes the order of the plugging and the outer-layer processing, which shortens the sequence and reduces the number of times the panel passes through the plating line. In both cases the buried via must be filled before the next lamination, and the fill depth is the parameter that decides whether the result is acceptable. A fill that is too shallow leaves a depression into which the added resin sinks, creating a surface that may short to the outer layer; a fill that is too deep stands proud and deforms the inner layers under the press. A working window of roughly 70 to 100 percent of the via depth is typical, with the lower bound treated as a hard limit. The panel is also treated with a bonding process before and after the filling step, because the resin that covers the copper during plugging would otherwise prevent the necessary surface preparation on the layer beneath it, and because the thin treated layer can be damaged during handling and curing. Both omissions lead to the same failure, which is delamination under the press.

Drilling Sequence and Its Consequences

The order in which the mechanical and laser drilling steps are performed matters more than it appears. Drilling the laser vias first and the mechanical through holes afterwards leaves debris in the blind vias, and debris that cannot be fully removed produces a plating defect inside a hole that cannot be inspected from outside. The mechanical drill also exerts a lateral force on the panel, and a blind via located close to a through hole can be deformed by that force, which again shows up as a plating defect. For both reasons the mechanical drilling is normally completed before the laser step. The distance between a blind via and a nearby through hole should be checked in the layout, and where the two structures overlap, the design should be revised rather than the process adjusted. Where a blind via and a buried via are on the same net and overlap, the blind via can often be deleted without affecting the function; where they are on different nets, one of them has to move, and the hole edge to hole edge clearance should be kept at the fabricator’s stated minimum.

<img src="https://www.gopcba.com/wp-content/uploads/2026/08/2.png" alt="laser drilled microvia registration target on an HDI panel” />

Registration Targets and Panel Size

Blind vias are aligned to a target rather than to the panel edge, and the target is normally placed on the sub-outer layer so that the laser can be aimed at the pad that the via has to reach. Registration accuracy degrades with panel size, because the dimensional change of the laminate accumulates across the panel, so HDI panels are kept small; a limit of about 16 by 18 inches is common, and the border inside the panel edge has to be wide enough to carry the targets and the tooling features. Where several drilling operations are used, the tooling holes for one operation should be positioned so that they do not conflict with those of another, and where the design contains no through holes at all, additional plated holes have to be added at the panel edge purely to provide references for the imaging process.

Forming the Blind Via

Three methods are used to reach the target pad through the build-up dielectric. The first removes both the copper and the resin with a single ultraviolet laser pass. The second uses an ultraviolet laser to open the copper and a carbon dioxide laser to remove the resin beneath it. The third, known as the conformal mask method, images and etches a small opening in the copper at each via location and then removes the exposed resin with a carbon dioxide laser. The first two are accurate but slow, which limits their throughput; the third is fast and suited to volume production but is more exposed to imaging tolerance, which can produce an off-centre or damaged hole. The choice is therefore a trade between the registration the design can tolerate and the volume the product requires.

Annular Ring, Hole Size and Design Checks

Blind vias commonly range from 4 to 6 mil in diameter, a size at which the hole shape can be controlled well. When a blind via would exceed about 8 mil, mechanical drilling becomes the cheaper option and the design should be reconsidered, because the aspect ratio of a laser-drilled hole cannot be extended indefinitely without losing control of the hole shape. The annular ring is the parameter that most often forces a layout change, and its acceptable value depends on the process: a newer alignment scheme may require a minimum ring of 5 mil with 4.5 mil of pad remaining after the drill, while an older scheme may accept 4 mil and 3 mil respectively. When the ring cannot be achieved, the available levers are to move the trace, move the hole, reduce the via diameter, narrow the trace locally, or increase the clearance between the trace and the via pad. Every blind via should also be checked for a missing landing pad, for a diameter that differs from the others on the board, and for a position too close to the panel edge, since a single inconsistent via diameter can complicate the entire drilling program. The via structures that carry the highest current or the tightest pitch are normally filled and plated, and the requirements of that step are described in this article on electroplating and via filling in HDI.

Cost, Testability and Prototype Builds

Each additional build-up step multiplies the number of laminations, drilling operations and inspections, so the cost of an HDI board rises faster than its layer count suggests. For a first prototype it is worth asking whether the density genuinely requires the structure or whether a conventional multilayer board with a thinner core would serve, and the general requirements for prototype builds are covered in this article on multilayer PCB prototype requirements. Testability also has to be considered, because the internal vias cannot be probed after assembly; the design should provide surface test points for the nets that need to be verified, and the fabricator should be asked to provide cross-section data so that the fill and the plating can be confirmed on the first article.

FAQ

What is the difference between a blind via and a buried via? A blind via reaches an inner layer from one surface, while a buried via connects two inner layers and is invisible from both surfaces.

Why must a buried via be filled before the next lamination? Because the resin added on top will sink into an open hole or stand proud over an over-filled one, causing a depression or a deformation that affects the outer layers.

Why are HDI panels kept small? Because registration error grows with panel size as the laminate expands and contracts, and a smaller panel keeps the blind vias aligned with their targets.