Blind Via PCB Manufacturing: Stackup, Process and Reliability

A blind via is a hole that starts on an outer layer and stops at an inner layer instead of passing through the board. The idea is simple and the manufacturing is not: the hole has to be drilled to a controlled depth, plated without contaminating the layers it does not reach, and then buried under further lamination. That sequence is what makes high density interconnect possible, and it is also what makes it expensive.

What Makes a Via Blind

A through via passes from the top of the board to the bottom and can be drilled after all the layers are laminated together. A blind via connects an outer layer to one or more specific inner layers and stops there. A buried via connects two inner layers and never reaches the surface. Of the three, the blind via is the one that changes the process most, because the drilling and plating have to happen before the board is finished.

The benefit is escape routing. A through via blocks the layers it passes through, and on a dense design the blocked area becomes the limit on how many signals can leave a ball grid array. A blind via that stops at the layer it needs frees the space below, which is why the technique appears wherever a fine pitch device has to be fanned out.

Why a Design Reaches for Blind Vias

The first reason is routing density. The second is electrical. A shorter via has less capacitance and less inductance, and it does not carry a stub that extends past the connection point. At multi gigabit rates a through via stub resonates and degrades the channel, which is why fast serial links on thick boards either use blind vias or back drill the stubs away.

The third reason is mechanical and thermal. Blind vias free surface area for component placement and allow ground connections to be made where a through via would have created a thermal or assembly conflict. The trade is a stackup that requires sequential lamination and a cost that rises with each additional lamination cycle.

Stackup and Sequential Lamination

A board with blind vias is not laminated once. The inner core is imaged and laminated, vias are drilled and plated to the required depth, and then the outer layers are added in a second lamination. Repeat the pattern and the board becomes a three or four cycle build, each cycle adding cost, yield risk and lead time.

The stackup therefore has to be designed with the lamination cycles in mind. A symmetric arrangement keeps the board flat through repeated pressing, and the number of cycles should be the minimum that satisfies the routing. The structures and their relative costs are described in blind and buried via stack selection.

Drilling a Hole That Stops

Small blind vias are formed by laser drilling. The laser ablates the dielectric down to the copper of the target layer, and the copper itself stops the process, which is a natural depth control as long as the dielectric thickness is uniform. Laser formed vias are typically under a hundred and fifty micrometres in diameter and are used for microvia structures on the outer layers.

Larger blind vias may be mechanically drilled with a controlled depth, which is more demanding because the drill has to stop inside the board without touching the layer below. Depth control depends on the machine, on the flatness of the panel and on the thickness tolerance of the laminate, so the process window is narrow and the design should not assume a depth tolerance tighter than the fabricator can hold.

Aspect Ratio and Depth Control

Aspect ratio is the ratio of the hole depth to its diameter, and it governs how easily the hole can be plated. A hole that is too deep for its diameter cannot be plated uniformly, and the plating thins at the centre where the current density is lowest. Blind vias generally have a more favourable aspect ratio than through vias of the same board thickness, which is one reason the technology is used on thick boards.

The depth also has to land inside the target pad rather than at its edge. This means the pad on the target layer must be large enough to accept the depth variation, and the dielectric above it must be uniform. Both requirements are properties of the stackup rather than of the drilling machine, so they are settled before the board is ordered.

Filling the Via

After plating, a blind via that will be built over or used as a component pad must be filled. Copper filling plates metal into the hole until it is solid, which gives the best thermal and electrical performance and supports a pad that can be soldered or stacked on. Resin filling plugs the hole with a non conductive material and is then plated over, which is cheaper and adequate where the requirement is flatness rather than conduction.

Via filling is critical in HDI because the next lamination presses directly onto the filled hole. An incompletely filled via leaves a void that outgasses during lamination, forming a blister that may not appear until after assembly. The plating chemistry and the fill quality are therefore inspected on a sample, and the process is described in electroplating and via filling in HDI.

Design Rules That Keep It Manufacturable

Keep the number of lamination cycles low by using the smallest number of blind via levels that the routing needs. Give each blind via a target pad with enough margin for the depth tolerance and the layer to layer registration, and keep the via in the pad area where stacking permits rather than routing it out and back. Where a via is placed inside a pad, state the filling requirement on the drawing.

Check the aspect ratio for every via level against the fabricator capability rather than against a rule of thumb, and confirm the minimum annular ring on the layers the via reaches. The rules that govern the pad and ring dimensions are the same as for any other connection and are set out in PCB pad design standards.

Inspection and Reliability

A blind via cannot be inspected by looking into it. The accepted method is a microsection, in which a sample is cut and polished and the plating thickness, the depth and the connection to the target pad are measured under a microscope. X-ray confirms fill quality through a panel, and electrical test confirms continuity, but the barrel and the depth are verified by section.

Reliability testing follows the same logic as for any plated hole: thermal cycling to expose weak plating, and where the product requires it, a full qualification including reflow simulation. A blind via with a marginal plating thickness may pass the initial test and fail after a few hundred cycles, which is why the section data rather than the electrical data is what the reliability engineer looks at first.

FAQ

Does a blind via always need an extra lamination cycle? Yes. The via is drilled and plated before the outer layers are pressed on, so at least one additional cycle is needed for each level of blind via.

Can a blind via be mechanically drilled? Larger ones can, with depth control. Small vias are formed by laser because the laser stops naturally at the copper of the target layer.

How is a blind via inspected? By microsection on a sample. The section shows the plating thickness, the depth and whether the via reached the target pad properly.