What Is a Printed Circuit Board (PCB)? Layers, Vias, FR-4
A printed circuit board is the base that holds a device's parts. It is a board that does not conduct electricity, with thin copper wiring built on its surface and inside it. Chips, capacitors and other parts are soldered onto it and joined by that wiring. The invention replaced wiring parts by hand, one wire at a time, with mass-produced boards that are all alike.
What you'll learn
- What a printed circuit board does inside a device
- The board material (glass epoxy, FR-4) and how single-sided, double-sided and multilayer boards differ
- The holes called "vias" that join layers, and what the green film and white lettering are for
- How parts are attached: through holes, surface mount and reflow soldering
- How chip, package and board fit together, and what fast signals need from the wiring
Prerequisites: none. The case that holds a chip is explained in "What Is Semiconductor Packaging?"
Where you see it
Open up a broken remote control or game console, and you will find a green board. Black square parts and tiny grain-like parts sit on top of it. Many thin lines run across its surface. This board is called a printed circuit board.
In the industry, it is also called a printed wiring board. Its short name is PCB, from the first letters of printed circuit board. Almost every device that runs on electricity has one: phones, TVs, rice cookers, cars and more.
The chips on a smartphone's board are introduced in "What Chips Are Inside a Smartphone?" This article looks at the board itself, not the parts on it.
Board wiring replaced wiring by hand
Early electronic devices joined parts with wires, one wire at a time. Each joint was fixed with solder. Solder is a metal that melts at a low temperature. When it cools, it hardens and joins metal parts electrically. The work was done by hand, so it was slow, and wiring mistakes happened.
So people came up with a way to build the wiring onto a board first. A copper film is bonded to a board that does not conduct electricity (an insulating board). Chemicals then dissolve the copper that is not needed, leaving only the wiring pattern. The pattern is copied from a master, much like printing, which gave the board its "printed" name. Identical boards can be made again and again, with no wiring mistakes.
A well-known pioneer of this method is Paul Eisler, an engineer from Austria. In 1936 he moved from Vienna to Britain and built a trial radio using a printed circuit board. He filed his key patent in 1943. Yet the idea of bonding flat metal strips to an insulating board already appeared in a British patent filed in 1903. So who counts as the first inventor is disputed.
Board material: glass cloth and resin
The core of the board is usually a cloth woven from fine glass threads. This cloth is soaked in a resin that hardens with heat. The resin is a type called epoxy, so the board is called glass epoxy. The glass cloth makes the board strong. It also limits how much the board grows and shrinks with temperature.
The best-known grade of glass epoxy is FR-4. FR stands for "flame retardant," meaning hard to burn. Additives in the resin keep a fire from spreading, even if parts get hot. The grade name was set by NEMA, a US trade group of electrical manufacturers. Today, people often use FR-4 to mean board material in general.
A thin copper film is bonded to the board's surface. Boards sold for hobby projects are commonly 1.6 mm thick, with copper 35 µm thick (an example from Sunhayato's products). A µm (micrometer) is one thousandth of a millimeter. So 35 µm is less than half the thickness of a human hair (about 0.09 mm).
Single-sided, double-sided, multilayer: stacking wiring layers
A board with wiring on only one side is called a single-sided board. One with wiring on both sides is a double-sided board. As the number of parts grows, wires start to cross, and one side is no longer enough to route them. With two sides, a wire can hop to the other side where it would cross another.
Devices with even more wiring also build wiring layers inside the board. This is called a multilayer board. Thin boards and copper layers are stacked in turn, then pressed into one sheet with heat and pressure. The layer count goes from 2 (double-sided) to 4, 6, 8 and more.
Devices like smartphones pack many parts onto a small board, so they need many layers. A 2022 article in Dempa Shimbun, a Japanese electronics newspaper, describes the board maker Meiko. The company makes 10-layer boards for smartphone main boards. In the 2017 iPhone X, iFixit's teardown confirmed a main board made of two boards stacked and joined together.
Many inner layers are not thin lines but wide areas of copper. The 0 V line that serves as the voltage reference is called ground. Making the power and ground layers wide areas lets the board deliver power to parts steadily. Fast signal lines also need a copper area close by (explained later).
Vias: holes that join layers
When wiring sits on separate layers, it needs a path from an upper layer to a lower one. So a small hole is made in the board, and its inside is coated with copper by plating. Plating means growing a metal film on a surface in a chemical bath. This copper tube becomes the path between layers.
This hole is called a via. Vias that go all the way from top to bottom are drilled. Boards with very fine wiring also use lasers to make shallow holes that reach only the next layer. The smaller and shallower the holes, the more room is left on other layers for wiring.
The green film and the white lettering
A board looks green because its surface is covered with a thin resin film. This film is called solder mask (or solder resist). "Resist" here means something that keeps solder away. The film covers the copper, except for window-like openings where parts are soldered.
The film has three main jobs. First, it stops melted solder from flowing onto a nearby wire and joining the two. Second, it keeps air and moisture from corroding the copper. Third, it insulates the gaps between wires, so dust or moisture cannot cause current to leak.
There is no single answer to why green is so common. An explainer from Altium, which makes board design software, says green film is easy to form into fine shapes. It also says black film is hard to tell apart from the wiring, which makes inspection harder. Green, being easy to see and easy to make, came to be used as a kind of standard.
Now look at the white letters and lines on the surface. They are printed in ink: part names (R1, C3 and so on), outlines of where parts go, and marks for which way they face. This printing is called silkscreen (or "silk" for short). During assembly or repair, it shows at a glance what goes where.
Attaching parts: through holes or on the surface
The first method pushes a part's legs through holes in the board and solders them on the back. This is called through-hole mounting. The legs pass through the board, so the part holds up well to pulling and shaking. It is still used for large parts and for connectors that cables are plugged into.
The second method makes no holes. Parts are placed right on the board's surface and soldered there. This is called surface mounting. The legs do not go through holes, so parts can be smaller, and they can be placed on both sides of the board.
Where a part sits, there is a small square of copper. This square is called a pad, and it shows through a window in the green film. Almost all parts on today's smartphone boards are surface mounted. In 2024, Murata Manufacturing announced it had developed a capacitor measuring 0.16 mm × 0.08 mm. Capacitors are covered in "What Is a Capacitor?"
Reflow soldering: hundreds of parts at once
Surface-mount parts are usually soldered all at once, in three steps. First, a thin metal sheet with holes is laid on the board. Through it, a paste of fine solder powder is spread onto the pads only. Next, a machine picks up parts and places them on the paste, one after another. Last, the board passes through an oven. The heat melts the solder, which then cools and sets.
This method is called reflow soldering. Melted solder tries to pull itself together, the same way a water drop forms a ball. This pull draws small parts that were placed slightly off toward the middle of their pads.
Most solder today contains no lead. That is because, from July 2006, an EU law called the RoHS Directive restricted, in principle, lead in new electronics sold there. A common lead-free solder is tin with a little silver and copper, and it melts at about 217–220 °C. The oven heats a little above this. Parts are checked to make sure they can stand even higher heat. An industry standard (J-STD-020) tests them at 245–260 °C, depending on part size.
How chip, package and board fit together
The path from a chip to a device is easiest to picture as three nested levels. First, the silicon chip is wrapped in a case called a package. Then the package is soldered onto the printed circuit board. How chips are wrapped is explained in "What Is Semiconductor Packaging?"
In fact, the package can hold a small wiring board of its own. This is called a package substrate. It spreads out the chip's very fine electrodes (the points where current enters and leaves). The new spacing is wide enough to solder to the device's board. It stacks resin and copper layers just like a printed circuit board, but its wiring is much finer.
So "board" can mean two things: the small board right under the chip, or the device board that holds the parts. In this article, "printed circuit board" means the second one. The large central board of a device is also called the motherboard in a PC, or the main board (logic board) in a smartphone.
Paths for fast signals: keeping the wire width even
Boards in today's devices carry signals that switch between 0 and 1 very fast. As signals get faster, a wire stops acting like a plain wire. It must be treated as a path along which voltage changes travel like waves. To keep things simple, we skip the detailed theory here.
If the path suddenly gets "easier" or "harder" for the wave to travel along, part of the wave bounces back. When the bounced wave overlaps the original signal, the shape of the 0s and 1s breaks down and can be misread. This measure of how easily a wave travels is called characteristic impedance. It is set by the wire width, the distance to the copper area just below, and the board material.
So for fast signal lines, the width and layer thickness are worked out to keep the characteristic impedance constant. For example, USB 2.0 sends signals over a pair of wires. The pair's characteristic impedance is set at 90 Ω ±15% (a Texas Instruments explainer). The unit Ω (ohm) is the same unit used for resistance. The basic idea is explained in "AC vs DC Explained."
Common myths
Myth: Printed circuit boards are always green
In fact, the green is only the color of the surface film. There are also red, blue, black, white and other films. Hobby boards and PC parts that care about looks often use colors other than green. The color of the film does not greatly change the board's electrical behavior. The myth arises because most boards people see are green.
Myth: A board is just a plain sheet that holds the parts
In fact, the wire widths and the way layers are stacked decide whether a device works properly. The board's design also affects whether fast signals arrive intact, whether the supply voltage stays steady, and whether chip heat can escape. With the same parts, a poorly designed board means the device will not work properly. That is why board design, which decides how wires are routed, is a specialist job.
Myth: The lines you can see on the surface are all the wiring
In fact, on a multilayer board much of the wiring is on inner layers that cannot be seen from outside. What shows on the surface is only a small part. On a 10-layer board, such as a smartphone's main board, you can see only 2 layers: the top and the bottom.
By the numbers
| Number | What it is |
|---|---|
| 1943 | The year Eisler filed his key patent for the printed circuit board |
| 1.6 mm, 35 µm | Common board thickness and copper thickness for hobby boards (an example from Sunhayato's products) |
| 10 layers | An example layer count for a smartphone main board (Meiko; a 2022 Dempa Shimbun article) |
| 245–260 °C | Temperatures used to test whether parts can stand reflow heat (J-STD-020; for lead-free solder) |
| About 217–220 °C | The melting range of a lead-free tin-silver-copper solder (SAC305) |
| July 2006 | When the EU began restricting lead in new electronics sold there (RoHS Directive) |
| 90 Ω ±15% | The characteristic impedance set for the wire pair in USB 2.0 |
| Over 60% | Japan's Taiyo Holdings' global share of solder mask (a 2024 forecast by Fuji Chimera Research; EE Times Japan, March 2026) |
A research firm's forecast says one Japanese company makes more than half of the world's green-film material. A printed circuit board does not draw attention. Yet small advances in its materials and methods are what make phones so small and fast.
Learn more
- See the differences in a table: "BGA vs QFP vs QFN" (shapes of packages that go on a board) and "Single-ended vs differential signaling" (why signals are sent over a pair of wires)
- Terms for your next step: BGA, differential signaling, crosstalk (noise passing to a nearby wire), MSL (how easily a part is harmed by moisture), glass core substrate
- Try the numbers: PDN target impedance calculator (estimates how many capacitors a board's power wiring needs; somewhat advanced)
- Related reading: "What Is an Integrated Circuit?" (sorting out the words chip, die and package)
- As a career: work on board assembly and packages is covered in "Packaging Engineer"