PCB Drilling: Hole Types, Tolerances, and Cost Drivers

Every plated hole in a circuit board begins as a drilled hole. PCB drilling is the step that turns a laminated panel into something that can carry a signal from layer to layer, and the quality of those holes decides whether plating adheres evenly, whether impedance stays on target, and whether the finished assembly survives thermal cycling. On a dense multilayer stackup a single misdrilled hole can scrap the whole panel, so the process deserves more engineering attention than it usually receives.

What the Drilling Step Actually Does

Drilling is scheduled after lamination and before copper plating. The bit removes material to create the opening; the plating step that follows deposits copper on the barrel wall so the layers above and below the hole are electrically joined. That ordering matters, because the drill defines the geometry the plating has to cover. A rough barrel or a hole that breaks out of its pad cannot be rescued by better plating.

Two numbers dominate the discussion. The first is the aspect ratio, board thickness divided by drilled diameter. Standard FR-4 fabrication handles ratios up to about 10:1 comfortably, and controlled processes push toward 20:1 once the layer stackup has been kept balanced. The second is registration, the accuracy with which the hole lands inside its pad. Both are set by the drill program, the machine, and the entry and exit materials used on the stack.

Mechanical, Laser, and CNC Approaches

Most through holes are cut mechanically with a carbide bit. Mechanically drilled holes are economical at diameters above roughly 0.2 mm. Carbide is used because it holds an edge through thousands of hits, and the bit is normally retired on a fixed hit count rather than on visual inspection.

Laser drilling takes over below that range. A CO2 or UV laser can form microvias under 0.15 mm without the mechanical stress of a spinning bit, and it is the standard route for HDI builds where microvias are stacked between thin dielectric layers. Laser drilling costs more per hole because cycle time per hole is longer and the equipment is more expensive to operate.

CNC drilling sits on top of both. The machine is numerically controlled, so hole positions come from the drill file and are executed at high speed with minimal operator intervention. The practical benefit is repeatability across panels: tool wear, retraction height, and hit count are tracked automatically, and the same program produces the same result on panel ten thousand as on panel one.

Plasma etching is a niche fourth route. It is used mainly on PTFE and other fluoropolymer substrates, where mechanical cutting smears the soft dielectric and standard laser absorption is poor. Throughput is low, so it stays reserved for specialty RF work.

Hole Types and Where Each Is Used

A through-hole via spans the full board and connects the outer layers to any inner layer it passes. It is the cheapest and most common connection, and unless space is at a premium there is rarely a reason to avoid it.

A blind via starts on an outer layer and stops at an inner layer without passing through. It saves routing space at the cost of extra lamination and drilling cycles, because the hole must be formed after the layers it reaches are bonded. The blind and buried via selection rules set out where each type is worth the added process.

A buried via lives entirely between inner layers and is invisible after lamination. It consumes the most process steps and is reserved for dense HDI designs where routing channels on the outer layers are fully occupied.

Microvias are the smallest family, generally under 0.15 mm, and are almost always laser formed. They are used for escape routing from fine-pitch BGA packages and for stacked builds in portable electronics.

How Substrate Material Changes the Recipe

FR-4 is the baseline. Standard FR-4 drills predictably with carbide tooling, and most process windows are written around it. Higher-Tg grades are slightly more abrasive and shorten bit life.

Metal-core boards, usually aluminum or copper, are far harder. They need wear-resistant tooling, lower feed rates, and careful chip evacuation, because metal chips abrade the barrel wall if they are not cleared. Heat generated at the tip is also higher, so coolant flow and dwell time have to be managed.

PTFE laminates are soft and dimensionally unstable under heat. They tend to smear rather than cut, which leaves a residue that blocks plated copper from bonding to the barrel. Plasma desmear or a dedicated chemical etch is normally specified alongside drilling for these materials.

Ceramic-filled substrates are brittle and highly thermally conductive. Standard mechanical drilling chips them at the exit, so laser or ultrasonic processes are used instead.

Tolerances and the Standards Behind Them

Drill tolerance is the permitted difference between the diameter on the drawing and the diameter the machine produces. Under IPC-2221 the common allowance for finished holes of 0.3 mm and above is around plus or minus 0.05 mm, while laser-formed microvias are usually held to about 0.025 mm. Tightening beyond that is possible but is charged as a premium because it requires slower feeds, more frequent bit changes, and tighter incoming inspection.

Registration tolerance is a separate number and is often the one that actually fails. It covers positional error between the hole and the pad it must land on. When registration drifts, the hole breaks out of the pad, the annular ring on one side thins to nothing, and the connection becomes a reliability risk even if it passes continuity test. The underlying via design rules set the annular ring that absorbs this error.

Defects to Watch For

Breakout is the classic registration failure. It appears as a hole that has cut through the edge of its pad, leaving a crescent of copper instead of a full ring. The usual causes are drill program offsets, panel shift during stacking, or a bit deflecting as it enters.

Rough barrel walls show up after plating as uneven copper thickness. They reduce the current-carrying area of the barrel and create stress risers that crack during thermal cycling. Sharp tooling, adequate retraction, and a controlled hit count all reduce the risk.

Delamination is a heat problem. Excessive spindle speed, a dull bit, or too many hits in one location raise the local temperature until the resin between layers separates. The damage is usually hidden until cross-sectioning or reliability testing.

Hole position drift is a calibration issue. It traces back to a misloaded drill file, a worn collet, or a machine that has not been re-zeroed. Running a first-article cross-section on every new job is the cheapest way to catch it before the lot is complete.

Understanding Drilling Cost

Drilling cost is driven less by the number of holes alone than by how hard each hole is to make. A reasonable planning figure is between roughly one-tenth of a cent and five cents per hole, with the spread coming from four factors.

Diameter is the first. Small holes need slower feed rates, more precise machines, and bits that break more often, so each hit costs more. Hole count is the second, because machine time and tool wear scale with the total. Material is the third; aluminum, ceramic, and PTFE all cost more per hole than FR-4. Method is the fourth and the largest single multiplier, since laser drilling typically runs two to five times the cost of mechanical drilling for the same feature.

As a rough reference, a four-layer FR-4 board with about 5,000 plated through holes typically carries a drilling charge in the range of ten to twenty-five dollars, assuming standard tolerances and a uniform hole size.

Design Choices That Make Drilling Easier

Standardizing hole sizes is the highest-leverage change available to a designer. Every unique diameter adds a tool change, and tool changes dominate machine time on high-hole-count jobs. Collapsing a design from nine drill sizes to four can cut drilling time noticeably without touching the routing.

Avoiding microvias and blind vias where a through-hole via would fit is the second. Each blind or buried via adds lamination and drilling cycles, and those cycles cost far more than the routing space they save.

Keeping tolerances at standard values is the third. Specifying a tighter drill tolerance than the application needs is one of the most common ways to inflate a quote with no engineering benefit. Pad sizes should also be generous enough to preserve the annular ring if registration moves slightly, and stacked microvia counts should be kept low to avoid HDI complexity the design does not actually require.

Qualifying a Drilling Supplier

Not every fabrication house can drill every design, and the gap between capable and profitable suppliers is wider than most buyers expect. Four questions separate them quickly. What is the smallest via diameter and the highest aspect ratio the line holds in production, not in a sample? Which laser and CNC platforms are installed, and are they in-house or subcontracted? Is there documented experience with aluminum, ceramic, or PTFE? And is there internal cross-section and optical inspection capability to verify barrels before shipment?

A supplier that subcontracts drilling adds a handoff, and handoffs are where registration errors and lost panels accumulate. In-house drilling with first-article cross-sectioning is the pattern worth paying for on anything with fine geometry.

Where the Process Is Heading

Automated drill-path optimization is already reducing machine time by reordering hits to cut travel and spread tool wear. Stacked microvia processes are extending HDI density, and femtosecond laser sources promise cleaner holes in materials that currently smear. Machine-vision inspection is moving from sampling toward per-panel checking, which should catch barrel defects before plating rather than after. None of these change the fundamentals: the hole has to land where the drawing says, and the barrel has to be clean enough for copper to bond.

FAQ

What is the smallest hole a standard fabricator can drill? Mechanical drilling is reliable down to roughly 0.2 mm. Below that, laser drilling is used, and microvias in the 0.1 mm range are routine on HDI lines.

Does a tighter drill tolerance improve reliability? Only where the design actually needs it. Registration error and barrel quality usually matter more than absolute diameter, and over-tightening the tolerance mainly raises cost.

Can I mix through-hole via and microvia technologies on one board? Yes, and it is common in HDI. Each added via class brings its own lamination and drilling cycle, so keep the number of distinct via types as low as the routing allows.