Smartwatch PCB: Fitting Radio, Display and Battery in a Case
A smartwatch is the most tightly packed consumer product in volume production. The case is a few centimetres across, the display occupies most of the front surface, a battery takes most of the internal volume and the radio still has to radiate. A smartwatch PCB is defined by what is left over after the mechanical parts have taken their space.
What Has to Fit on the Board
The board carries the application processor, the display interface, the radio with its front end, the sensors and their analog front ends, the power management and the battery connection. On many designs it also carries the antenna itself, because there is nowhere else to put one.
That combination puts a fast serial display link, a sensitive analog sensor path and a radio transmitter within a few millimetres of each other. The layout has to keep them apart while the mechanical design pushes them together, and the only tools available are ground plane discipline and careful placement.
Antenna Keep-Out in a Metal Case
The antenna keep-out requirement is the single largest constraint on the board. The region beneath and around the radiating element has to be free of copper on every layer, and the metal case of the watch is usually close enough to detune it further. In practice the antenna occupies one edge or one corner of the board, and everything else is arranged around that decision.
Because the case is metal, the design often has to route the radio energy through a defined aperture or use the case itself as part of the radiator, which changes both the tuning and the radiation pattern. Either way the antenna performance is verified on the assembled watch rather than on a bare board, and the surrounding circuitry has to be quiet enough not to desensitise the receiver.

Power Budget and Battery Life
The battery is the largest component in the product and the least negotiable. Everything else is designed around the capacity that fits, and the power budget is the number that decides which features can be enabled and for how long. A watch that displays continuously draws orders of magnitude more current than one that updates the screen once a minute, and that difference is what allows a week of use rather than a day.
On the board, that translates into a design that can shut down aggressively. Rails are switched individually so unused blocks can be isolated, the display supply is separate from the processor supply, and the radio front end is powered only for the fraction of time it transmits. Each of those decisions is a layout decision as much as a component decision, following the approach in ground routing and power trace planning.
Sensor Signal Integrity
The sensors are the reason the product exists, and their signals are small. A photoplethysmography front end measures a current of nanoamps superimposed on a large ambient background, and a motion sensor measures changes of a few milligees. Both are vulnerable to noise on their supply and to coupling from the display link.
The remedies are conventional but have to be applied without compromise on a board this small. Dedicated filtered supplies for the analog front ends, short and direct connections to the sensor pads, and a ground reference that is not shared with the display or the radio return. Where a sensor sits on the far side of the case, a flex tail carries its signals and needs the same care as the main board, as described in conformal coating and board protection.

Display Interface Routing
The display link is the fastest interface on the board, and it is also the one that has to travel the furthest in a mechanical sense, from the processor to the panel connector at the edge. It is routed as a controlled impedance differential or single ended group with matched lengths, on a layer with a continuous reference plane.
Because the display link radiates, it should be routed away from the antenna region and away from the sensor front ends. Keeping it on the opposite side of the board from the antenna is usually possible even in a watch, and it is worth the effort because a noisy display link reduces radio sensitivity in a way that is difficult to diagnose.
Timekeeping and Crystal Accuracy
A watch keeps time, which places a requirement on the clock source that a general consumer product does not have. The real time clock crystal has to hold accuracy across temperature, which means it needs to be a low frequency tuning fork type with good temperature behaviour, or the design has to provide temperature compensation.
The crystal is also one of the most sensitive components on the board. It should be placed close to its oscillator with short, symmetric connections, kept away from switching regulators and from thermal gradients, and its load capacitors should be connected to a quiet ground. A crystal that is a few degrees warmer than the rest of the board will drift, and in a device worn against the skin that is a real possibility.
Board Construction
Most smartwatch boards use a flex PCB or a rigid-flex construction, because the space is three dimensional and the display and the sensors are often not on the same plane. Rigid islands carry the dense components and flexible sections carry the connections to the display and the sensors, which also removes connectors and their height.
Layer count is high relative to the board area, typically eight to twelve layers, because every interface needs a reference plane and the supply rails are numerous. The construction therefore tends toward fine geometry, and the via design becomes a significant part of the layout, following the constraints in blind and buried via stack selection.
Component Height and Mechanical Tolerance
The board also defines the internal height of the product. The tallest component sets the gap between the board and the case back, and the display assembly sits in a recess on the other side. Any component that exceeds its planned height will press on the battery or prevent the case from closing.
That makes height a layout constraint rather than a manufacturing detail. A height map of the assembly, maintained alongside the layout, is the only reliable way to catch a violation before the parts are ordered, and it should be reviewed whenever a component is substituted.
FAQ
Can the antenna be printed on the watch board? It can, and it often is, but it needs copper-free space around it on every layer. Where the board area does not allow that, a chip antenna or a flex antenna mounted on the case is used instead.
How many layers does a smartwatch board need? Eight to twelve is typical. The count comes from the number of separate supplies, the number of high speed interfaces and the requirement for a continuous reference beneath each one.
Does a watch need a different surface finish? It should use one that survives the assembly and the storage of a high value product, and that is compatible with the fine pitch components used. A nickel gold finish is common for that reason.



