PCB Production Process Flow: Single-Sided, Double-Sided and Multilayer

A production process flow for a printed circuit board is a sequence of operations in which every step prepares the surface, the geometry or the copper for the step that follows. The sequence differs by board type, but the rules underneath it are shared, and knowing where the three flows diverge tells a designer which of their choices will be inexpensive to build and which will not.

The Three Board Types and What They Share

Boards are classified as single-sided, double-sided or multilayer for manufacturing purposes, and that classification describes how many conductive layers carry the interconnections. A single-sided board has conductors on one face only. A double-sided board has conductors on both faces, joined by plated holes. A multilayer board adds internal layers, which can be imaged at a finer resolution than the outer layers because they are never exposed to the mechanical damage that plating, etching and handling inflict on an outer surface.

All three begin the same way and end the same way. They begin with a sheet of copper-clad laminate cut down from a production panel, and they end with a tested board packed for shipment. Between those two points the flow splits and rejoins, and the number of splits is what sets cost and lead time.

Single-Sided Board Process

The single-sided flow is the shortest: cutting, drilling, image transfer, etching, solder mask and legend, surface finish, profiling, testing and packing. Because there is only one conductive layer, there is no plating step and no plating bus to add and later remove. The copper is patterned subtractively from the foil that the laminate arrived with, which means the thickness of the finished conductor is determined before the board reaches the imaging area.

The absence of plating has consequences. A single-sided board has no plated through holes, so every connection from one face to the other must be made by a component lead soldered on the conductor side. That constraint suits simple, low-density products and fits poorly with anything that needs a dense interconnect. Our notes on choosing the layer count describe where the break-even point between the options lies.

Double-Sided Board Process

The double-sided flow adds one operation before imaging: chemical deposition of a thin copper layer into the drilled holes, followed by electroplating that builds the thin deposit up to a thickness the circuit can rely on. Only then is the outer layer imaged and plated again to build the conductors, after which the unwanted copper is etched away.

Those two plating steps explain most of the cost difference between a single-sided and a double-sided board, and they also explain why hole wall quality matters so much. A hole whose wall is not completely covered by the electroless deposit cannot be repaired by the plating that follows, and the result is a joint that passes a continuity test and fails later. Our plating thickness notes describe what the deposit has to achieve.

Where the Multilayer Board Diverges

A multilayer board is not a double-sided board with extra copper. It is built from a set of thin double-sided cores that are imaged, etched and inspected as separate items, then stacked with insulating prepreg between them and bonded under heat and pressure into a single solid. After bonding, the outer layers are drilled, plated and patterned exactly as in the double-sided flow.

That structure is why the multilayer flow puts so much emphasis on the front end. Every inner layer is a finished circuit before the stack is bonded, so an error in an inner layer is not correctable at any later stage, and the whole panel is scrap. Design rules for inner layers are correspondingly tighter on registration and looser on surface finish, a distinction our notes on layer assignment set out in detail.

Drilling and Plating Sequence

Drilling happens once, before plating, and every hole in the stack is produced in that single operation. The drill must pass through copper features on the inner layers without tearing them, which is why entry and exit materials are used on top and bottom of the stack and why the drill wear is tracked by hit count rather than by time. A worn drill produces a hole that is oversize at the entry and barrel-shaped along its length.

Plating follows the drill, and the sequence of cleaning, conditioning, catalysing and depositing is what determines whether copper reaches the centre of the stack. A board with a high aspect ratio is harder to plate than a thin one because the chemistry has to reach further into a narrow hole, and the answer is more time and more agitation rather than more current. Our aspect ratio notes cover the limits.

Lamination and Registration

Lamination is the step that makes the multilayer board a single object. The imaged cores and the prepreg sheets are stacked in a fixture, with tooling pins or rivets holding the stack in alignment, and the whole assembly is pressed at a temperature and pressure profile that melts the resin, fills the gaps around the copper, and cures it. Copper features that are left with nothing beneath them can shift during this stage, because the resin flows.

Registration is the cumulative error of every layer relative to the drill. Each core is punched or drilled with its own tooling features, and the drill that opens the final through holes must land inside the annular ring of every layer it passes. Paid out as a budget, registration is what sets the minimum annular ring, and the annular ring is what sets the inner layer clearance. Our tooling and registration notes explain the mechanism.

Solder Mask, Legend and Surface Finish

Solder mask is applied after the outer layers are etched, over the whole panel, and opened at the pads that have to be soldered. The mask protects the conductors from oxidation, from solder bridging and from handling damage, and it also defines the only areas where solder is allowed to wet. Legend, the printed component outline and reference designator layer, is applied on top of the mask, usually by screen printing or by inkjet.

Surface finish comes last among the coating steps, because it must sit on bare copper. A finish is a compromise between shelf life, solderability, cost and flatness, and the choice interacts directly with the component pitch on the board. Our surface finish and etching process notes describe how the two steps constrain each other.

Test, Forming and Packing

Electrical test verifies that every net is continuous and that no two nets are shorted, and it is performed on the finished board rather than on the panel wherever the board geometry allows. Flying probe test needs no fixture and suits prototypes and small batches; a fixture with bed-of-nails access is faster for volume. The coupon that the test relies on is built into the panel border and is tested destructively to confirm plating thickness, dielectric spacing and surface finish.

Profiling separates the boards from the panel by routing, scoring or punching, and the choice depends on the board outline and the material. V-scoring leaves a shallow groove on both faces and is fast for rectangular boards; routing follows any contour but leaves a burr and a router path between boards. Packing completes the flow, and the requirement is that the boards arrive flat, dry and free of contamination, which in practice means vacuum-sealed with desiccant for thin or moisture-sensitive laminates.

FAQ

Why does the inner layer get imaged before lamination? Because after the stack is bonded, the inner layers are sealed inside the board and can no longer be modified. Working on thin cores before bonding also allows finer lines than working on a finished board, since a thin core can be handled flat in a clean area instead of on a panel that has already been drilled and plated.

Is a production process flow different for a prototype? The steps are the same, but the substitution of tooling for setup saves cost. A prototype panel is often made with a single drill programme and a single imaging set, and the same panel may be used for several different designs, which brings the cost down without changing the physics of any individual operation.

What does gopcb confirm before starting a panel? We confirm the layer stack-up, the material and its thickness, the copper weight, the surface finish, the minimum line and space, and the finished board thickness, because each of these choices changes the process flow and therefore the schedule. Confirming them before the panel is released is far cheaper than discovering a conflict after lamination.