Blind Via PCB Manufacturing: Structure, Process and Cost

A Connection That Stops Halfway

A blind via starts at the surface of the board and terminates at an inner layer without passing through the whole stack. It looks like an ordinary via from the outside, but underneath it the laminate is solid, which means the layers beneath are free for routing that a through via would have blocked. That single property is why blind vias, and the sequential lamination process that produces them, are the basis of high density interconnect construction.

The price of that freedom is process complexity. Blind vias require controlled depth drilling, or a lamination cycle in which the connection is made before the outer layers are added, and both routes demand tighter control than an ordinary multilayer build.

How the Three Via Types Differ

• Blind via: surface to an inner layer. The connection is visible from one side only.
• Buried via: inner layer to inner layer, invisible from both surfaces.
• Through via: top to bottom of the stack.

A design usually mixes them. Through vias carry power, ground and low speed signals that do not care about the routing space they occupy. Blind vias handle the dense escapes under a fine pitch package where the routing channel is the constraint.

Why Blind Vias Are Used

• Routing density. Removing a through via from a layer frees the channel for a signal that could not otherwise be routed there.
• Shorter signal paths. A shorter barrel means lower parasitic inductance and capacitance, which matters as the edge rate rises.
• Improved return paths. The unused portion of a through via acts as an unterminated stub; a blind via simply does not have it.
• Lower crosstalk and EMI. Fewer penetrations through the reference planes mean fewer discontinuities in the return path.
• Mechanical space. Compact products frequently cannot accommodate a stack of through vias beneath a fine pitch device.

These benefits are the reason HDI PCB construction exists at all. The trade is that the design freedom has to be paid for in process steps.

Design Requirements

• Stackup planning first. The layer arrangement determines which blind vias are possible, how many lamination cycles are needed and what the cost will be. It has to be agreed with the fabricator before the layout begins.
• Depth and tolerance. The hole must reach the target pad and stop. Too shallow leaves a weak connection; too deep punches through the plane below.
• Diameter and aspect ratio. Microvias are typically limited to roughly a 1:1 ratio of depth to diameter, because plating a deeper hole of that size is not practical. A 0.1 mm microvia through a 0.1 mm dielectric is the typical case.
• Controlled impedance. Where a high speed net transitions, the geometry around the via has to keep the impedance close to the trace value.
• Fill and cap. Any blind via used in a pad has to be filled and plated flat, or the paste deposit on that pad will be wrong.
• Thermal and stress evaluation. Stacked microvias concentrate mechanical stress, and the reliability of the stack depends on the material and the fill.

A DFM conversation belongs at the start of the design, not at the end. The stackup choices made in the first week determine whether the blind vias are manufacturable at an acceptable yield.

The Manufacturing Process

• Inner core fabrication. The cores that will sit in the middle of the stack are imaged and etched first.
• Drilling the blind vias. Laser drilling for microvias below about 150 micron, mechanical drilling with controlled depth for larger ones. Laser is preferred for the small sizes because it stops on the copper pad beneath the dielectric.
• Plating or filling the via. Either a plated barrel with a thin copper wall, or a full copper fill that makes the via a solid conductor. Resin filling with copper capping is also used where a conductive fill is not required.
• Sequential lamination. The drilled and plated core is bonded into a larger stack, and the process may be repeated to build up additional blind via levels. Each cycle adds a controlled amount of thickness and a corresponding amount of risk.
• Outer layer imaging and etching.
• Surface finish and final inspection.

The number of lamination cycles is the single largest driver of both cost and difficulty. A design that needs two blind via levels needs two cycles, and one built with a single cycle plus through vias is a fundamentally cheaper product. Deciding the via scheme before the layer count is fixed is therefore a cost decision as much as a technical one.

Quality Control

• Cross sectioning. A coupon is potted, ground and polished, and the blind via depth, the plating thickness and the fill quality are measured under magnification. This is the definitive check on whether the hole reached the right layer.
• X-ray inspection of the filled vias, which reveals voids in a copper fill and incomplete resin fill in a laminated stack.
• Thermal cycling where the reliability class requires it, since the interface between the fill and the barrel is where cracks begin.
• Electrical test for opens and shorts, which on a fine pitch HDI board often means flying probe or a dedicated fixture.
• Acceptance against IPC Class 2 or Class 3, depending on the product’s reliability requirement.

The coupon data and the cross sections are the evidence that the sequential lamination worked, and they should ship with the boards. Without them, a customer has no way to confirm that the inner layers of a finished board are what the drawing specified. That documentation is a normal part of quality management on an HDI program, and it is the reason the stackup should be treated as a controlled document.

Customisation Options

• Layer count: from four layers to twenty or more, with the blind via levels chosen to suit the routing requirement.
• Board thickness: typically 0.6 to 3.2 mm.
• Base material: standard FR-4, high Tg FR-4, Rogers laminates where the loss matters, and dedicated HDI materials.
• Finish: ENIG, ENEPIG, OSP and immersion silver, selected against the assembly process and the reliability class.
• Via fill: copper filled or resin filled with a copper cap, chosen according to the current, the thermal path and whether the via sits in a pad.

Cost

• Prototype, five to ten boards, six to eight layers: roughly 120 to 300 US dollars each.
• Small batch, 50 to 100 boards: about 35 to 80 each.
• Volume, above 1,000 boards: about 8 to 25 each.

The spread is set by the number of blind via levels, the lamination cycles, whether laser drilling is used, the via fill method, the laminate and the board area. A comparison between two quotations is only meaningful once the via scheme and the layer stack match, since a design with two blind levels and one with a single level plus through vias can differ by a factor of two on the same board outline. That comparison is easier when the specification is written down and priced through a formal custom PCB pricing request.

Lead Times

• Prototype: 7 to 10 working days.
• Fast HDI build: 5 to 7 days.
• Volume: 3 to 4 weeks.

The schedule is longer than an ordinary multilayer board because each lamination cycle is a separate pass through the press with its own queue. Protecting the date therefore starts with keeping the number of cycles to the minimum the routing actually requires.

Choosing a Manufacturer

Five capabilities matter: proven HDI and blind via experience, laser drilling and sequential lamination equipment in house, DFM support that can advise on the stackup rather than only accept it, the certification set the product requires, and stable delivery on small batches as well as volume. The engineering support is the one that is hardest to assess from a capability list and the one that most affects the outcome, because the stackup decision is made jointly. Where the assembled product is also part of the scope, keeping the fabrication and the PCB assembly with one supplier avoids a second interface at exactly the point where a via fill problem would show up as a solder defect. The wider set of process controls behind this type of build is described under PCB manufacturing.

FAQ

Are blind vias suitable for prototypes? Yes. Development boards in compact products routinely use them, and the prototype is where the stackup is proven.

What files are needed for a quotation? Gerbers, drill files, the stackup drawing, the impedance requirements and the finish specification.

How can the cost be reduced? Fewer lamination cycles, standard materials, fewer blind via levels and a reduced via count, in that order of impact.

What is the aspect ratio limit? Microvias are generally limited to about 1:1 depth to diameter. Deeper connections need a larger hole or a different construction.

Why does the via need to be filled? A via in a pad has to be filled and capped flat, otherwise the paste deposit on that pad is unreliable and the joint quality suffers.

Summary

A blind via connects the surface to an inner layer and leaves the layers below free, which is the mechanism behind dense HDI routing. Producing one requires controlled depth drilling or sequential lamination, and each additional blind via level adds a lamination cycle, a yield risk and a cost. Prototypes land between 120 and 300 dollars per board, small batches between 35 and 80, and volume between 8 and 25, with the position in that range decided by the via scheme rather than by the board outline. Decide the stackup early with the fabricator, keep the number of cycles to the minimum, fill and cap any via that sits in a pad, and verify with cross sections and X-ray rather than assuming the inner layers came out as drawn.