Anatomy of a Printed Circuit Board: Pads, Vias and Layer Types

A printed circuit board looks like a single object, but it is a stack of different materials and a set of features that each do one job. Knowing what those features are called and what they are for makes it possible to read a fabrication drawing, to discuss a change with a fabricator, and to understand why a design that looks simple to draw is difficult to build. This is the vocabulary, from the pad outward.

The Features on the Board

The pad is a defined area of exposed metal where a component lead or a termination is soldered. Its size is set by the component and by the assembly process, and it is the only part of the copper that has to remain free of solder mask.

The via is a plated hole that connects copper on different layers. A via that runs through the whole board is the cheapest form; a blind or buried via connects only some layers and costs more because the board has to be built in stages. A via is not intended to hold a component, so it can be smaller than a pad, but it still consumes routing area on every layer it passes through.

The mounting hole fixes the board into the enclosure or the chassis. It may be plated and connected to a net, usually ground, or it may be unplated with a keep-out region around it.

The trace is the copper path that carries a signal or a current between pads. Its width carries the current and its geometry carries the impedance, so a trace is a designed object rather than a drawn line.

The connector joins this board to another board or to a cable, and the component is the part that performs the function. Both are placed on the board, but only their footprints are part of the board itself.

The copper fill is a large area of copper, usually connected to ground, that fills the space between traces. It lowers the impedance of the ground network, spreads heat, and reduces the amount of etchant that has to be removed, which makes the etching more uniform.

The board outline, sometimes called the electrical boundary, is the contour that defines the finished size. No component and no copper feature may extend beyond it, and the fabricator uses it to route the panel.

Single, Double and Multilayer Construction

A single-sided board has copper on one face only. The components sit on the bare side and the copper side carries the routing and the solder joints. The construction is cheap, but the routing is limited to one layer, so it survives only in low-density products.

A double-sided board has copper on both faces. The two sides are joined by plated through holes, so a trace can cross under a component on the other side. Almost all general-purpose electronics are built this way, and it is the most cost-effective construction for moderate complexity.

A multilayer PCB adds inner layers between the two outer ones. The inner layers carry signal routing, power planes and ground planes, and they are connected to each other and to the surface by vias. The layers are bonded under heat and pressure with prepreg and core material between them, and each layer pair is etched separately before the stack is pressed. Layer count is the main cost driver, and it is chosen by the routing density, the number of power rails and the impedance control requirements rather than by preference.

<img src="https://www.gopcba.com/wp-content/uploads/2026/05/3D在线AOI-1.jpg" alt="multilayer PCB cross section with inner layers” />

The Working Layers in a Design

The layers a designer works with are not the same as the physical layers of the finished board. They are the views the design is built from, and each one becomes a different piece of data for the factory.

The signal layers hold the traces. On a simple design the outer two layers are enough; on a dense one there are internal signal layers as well, and their number is what the designer negotiates with the fabricator. Top and bottom also carry the components and the copper fill, so they are usually the layers where routing and placement compete for space.

The solder mask layer, called by its material name on the top and bottom, is the coating that keeps solder away from the areas that should not be wetted. It defines the openings around the pads, and its expansion value is a fabrication parameter, not a design preference.

The paste layer is the stencil data. It has the same shape family as the solder mask but a different purpose, and the difference matters: the solder mask opening goes to the board shop and the paste opening goes to the stencil supplier. A design that treats them as one shape produces a stencil aperture that is larger than the pad.

The silkscreen layer carries the reference designators, the polarity marks, the connector labels and the company identification. It is printed on the finished board, so it has to stay clear of pads and of any area that will be soldered, and the minimum line width is a fabrication capability rather than a design choice.

The internal plane layers are solid copper assigned to a net, usually power or ground. They give a low-impedance supply and a continuous reference for the traces above them, and they are the reason a four-layer board behaves very differently from a two-layer board with the same routing.

Finally there are the mechanical and documentation layers: the board outline, the drill guide that positions the holes, the dimensions layer, and the assembly drawing. These layers are not manufactured, but the release is incomplete without them, and a missing outline is one of the most common reasons for a data package to be queried.

Why the Vocabulary Matters in Production

The features above become decisions with cost attached as soon as the design enters the factory. A pad that is too small for the intended part causes placement defects and rework. A via that is placed inside a pad on a fine pitch device makes the paste deposit uneven and produces voiding. A copper fill that comes too close to a trace without clearance creates a short after etching. A silkscreen paragraph printed over a pad is a solderability defect.

None of those problems are electrical in origin, which is why they are missed in a schematic review and caught in a data review instead. Our PCB manufacturing checks look at the features as the process sees them, and the results are only useful when the designer and the fabricator are using the same words for the same things.

Choosing a Construction

The practical rule is to use the lowest layer count that meets the electrical requirement, then to add layers only where they buy something specific. Two extra layers are cheaper than a redesign that fails an EMC test, but four layers added for convenience are a permanent cost on every board built.

Where the density or the impedance requirement forces a multilayer stack, the layer sequence matters as much as the count. A stack that places a ground plane directly under the top signal layer gives controlled impedance on that layer with no extra work, and it makes the return path continuous for everything routed there. A stack that separates the planes with thick dielectric loses that benefit. Our multilayer fabrication group reviews the sequence against the manufacturing capability while the stack is still being defined, which is the point where a change is free.

For a project that is still choosing, the sensible sequence is to define the board outline and the mounting scheme, choose the construction and the layer sequence, then place the parts and route. Our design and layout team can take that work at any stage, and the layer decisions are the first thing reviewed.

FAQ

What is the difference between a pad and a via? A pad is a solderable area for a component; a via connects layers internally. A via may be covered by solder mask, while a pad must be exposed.

Why does a copper fill matter? It lowers the impedance of the ground network, helps spread heat, and makes the etching more even across the panel.

How many layers should a design use? As many as the routing density and the number of supply rails require, and no more. Each additional layer increases the cost of every board built.