Laptop PCB Design: Layer Count, Space and Thermal Limits

A laptop main board is one of the densest products in consumer electronics. It carries a processor with a thousand or more connections, a memory interface running at multi-gigabit rates, several high-speed serial links to the display and the storage, a power delivery network that supplies tens of amperes at under a volt, and the connectors for everything that plugs into the machine. All of it fits into a board a few millimetres thick inside an enclosure with no room for a heatsink on most of it.

This article covers the structure of a laptop PCB, the constraints that shape it, and the design decisions that follow from them.

What Is Inside a Laptop Board

The functional blocks are the processor and its memory, the power delivery network, the embedded controller, the display and storage interfaces, the audio and USB subsystems, and the connectors. Because the board is a single assembly that everything plugs into, the connector placement is fixed by the mechanical design of the chassis before the layout begins.

Around the main board sit several smaller boards and flexes: the power button board, the touchpad, the keyboard, the antenna modules, the display cable and the battery connector. Each of those is an interface with its own mechanical constraints, and each is a place where a late change can force a redesign of the main board rather than of the small one.

Layer Count and Build-Up

A laptop board typically uses eight to twelve layers, often with HDI features in the region around the processor and the memory. The processor fanout is the constraint that sets the count, because a fine-pitch ball grid array with a thousand balls cannot be escaped on a conventional stack without an impractical number of layers.

The build-up structure matters as much as the total count. Where the outer layers carry the fine-pitch fanout, they need thin dielectric and small vias, while the inner layers can be conventional. That mixed construction is what allows the layer count to stay manageable. The HDI technology and microvia rules describe the structures and the design rules that apply in those regions.

Space, Connectors and Flexes

The board outline is dictated by the chassis, and so is the height available on each side. That height limit constrains the components that can be used: a connector is often selected for its profile rather than for its electrical performance, and a capacitor may have to be a low-profile part even when a taller one would be cheaper.

Where a function cannot be placed on the main board, it moves to a small secondary board connected by a flex. That is how the display, the touchpad and the power button are usually handled. Each of those connections is an interface that has to be planned, and the flex routing has to respect the bend radius and the shielding requirements of the signal it carries.

Power Delivery

The processor rail runs at a voltage below one volt and a current in the tens of amperes, which makes the power delivery network a layout problem rather than a component problem. The voltage drop across the copper between the regulator and the processor has to be a small fraction of the supply voltage, which means a lot of copper in a small area.

Decoupling has to be spread across the die footprint, and each capacitor has to be connected with the shortest possible path to the plane pair. The trace width and current calculation gives the copper geometry for the current, and the plane pair beneath the processor provides the high-frequency energy that the discrete capacitors cannot deliver quickly enough.

Thermal Budget

The processor is the dominant heat source, and the board is the spreading element between it and whatever cooling the chassis provides. In a thin laptop that is usually a heat pipe connected to a fan, with the board acting as a path rather than as the final heatsink.

Thermal vias carry heat from the processor pads into the internal planes, which then spread it laterally. The number and placement of those vias, and the copper area around them, set how efficiently the heat reaches the heat pipe. Components that cannot tolerate the local temperature, such as the battery and the display cable, have to be kept away from the hot region, which is a layout constraint before it is a thermal calculation.

High-Speed Interfaces

The display link, the storage link, the memory bus and the USB ports all run fast enough to require controlled impedance and length matching. Differential pairs have to be routed with consistent spacing and length, referenced to a continuous plane, and kept away from the power delivery network and the switching regulators.

The rules for these interfaces, including the via structure at the package and the connector transitions, are described in the high-speed interface layout notes. The most common failure in a laptop layout is a differential pair that is routed correctly in the middle and poorly at the ends, where it has to escape the package and reach the connector.

Signal integrity in this environment is largely a question of the return path. A pair that crosses a plane split, or that runs for a distance without an adjacent reference, radiates and couples to its neighbours, and the resulting impairment appears as jitter in the eye. Keeping the reference plane continuous under every high-speed net, and stitching it wherever the layout forces a change of layer, does more for the link margin than any amount of tuning at the transmitter.

Manufacturability

A board at this density has to be designed with the fabricator’s capabilities in hand. The minimum via diameter, the minimum capture pad, the registration tolerance and the achievable aspect ratio all determine whether the design will yield, and they differ between suppliers. Designing to a capability that the chosen supplier does not have produces a board that costs far more than budgeted, and the delay usually arrives after the tooling has been cut. Where the design pushes the process, a quick capability review before the layout is frozen is the cheapest risk reduction available, because it converts an assumption into a confirmed limit.

The assembly side matters equally. A board with components on both sides and a large processor needs a reflow profile that does not damage the parts on the second side, and the stencil design and the placement sequence have to be planned before the layout is finalised. Involving the assembly house during the layout is normal practice for this class of product.

FAQ

Why do laptop boards use HDI? Because the processor and memory packages cannot be escaped on a conventional stack within the available board area. HDI allows vias inside pads and routing on more of the outer layer, which is what makes the fanout possible at the layer count the chassis allows.

Can a laptop board be repaired? Sometimes. Connectors, capacitors and some power components can be replaced, but the processor and memory are usually soldered and the board is replaced rather than repaired. Designing for repair means keeping the replaceable parts accessible, which is a layout decision.

How much copper is needed for the processor rail? Enough that the resistive drop between the regulator and the processor is a small fraction of the supply voltage at the maximum current. Since the voltage is under one volt, that typically means multiple large copper areas on several layers connected by many vias.