PCB Trace Antenna versus Chip Antenna: Cost and RF Performance
Nearly every wireless product built in the last decade has had to answer the same question, usually late in the schedule: does the antenna live on the board as copper, or does it arrive as a component? A PCB trace antenna and a chip antenna can both deliver a working radio link, but they shift the cost, the board area, the risk and the responsibility for tuning in different directions, and the choice is easier to make before the layout exists than after.
Two Ways to Radiate
A PCB trace antenna is a pattern of copper on the board itself, shaped to resonate at the frequency of interest and fed by a controlled impedance line. Its dimensions are set by the wavelength in the laminate, so the pattern is shorter than a free space antenna would be, and its shape depends on how much space the layout can spare. Common forms include the inverted F, the meander, the monopole and the slot, and each of them trades bandwidth against area.
A chip antenna is a separate component, usually a ceramic or multilayer structure, that is placed on the board by the same pick and place machine that places the rest of the components and soldered in the same reflow pass. It occupies little board area but it is still a radio frequency device, which means it has a feed, a ground requirement, a keep-out area and a tuning network of its own.
What a PCB Trace Antenna Costs and What It Asks For
Because the radiator is copper, the bill of materials gains nothing and the assembly process gains no step. In volume that is a real advantage, and it is why the trace antenna is the default in cost sensitive consumer products. The price is paid in engineering. The pattern has to be designed for the specific board size and laminate, simulated, laid out with a clean reference plane beneath the feed, and then tuned on the assembled product rather than on a bare board.
Mechanical integration is the other constraint. The pattern must sit in a region where the ground plane is removed, which means the keep-out area consumes routing space on every layer, and the antenna must be kept away from metal, from the battery and from the display. If the enclosure changes, or if a metal component moves closer, the tuning can shift and the design cycle repeats.
What a Chip Antenna Buys You
A chip antenna arrives with a datasheet, a reference layout and a specified performance when that layout is followed. For a team without radio frequency experience, that is worth a great deal: the vendor has already solved the radiating structure, the matching network values are given, and the design task reduces to reproducing the recommended layout faithfully and leaving the keep-out area empty.
The trade is component cost, procurement, a small amount of assembly risk and a dependency on the vendor. The chip is small and its footprint is small, but the total area it needs is not necessarily smaller than a trace antenna, because the reference layout still includes the feed line, the matching components, the ground clearance and the specified ground plane geometry around it.
Size and the Real Footprint
Comparing the two by the size of the radiator is misleading. For a trace antenna the footprint is the pattern plus its keep-out plus the feed and matching network; for a chip antenna it is the component plus the reference layout area plus the matching network and its own clearance. On a compact board the chip often wins, but on a board with unused area at one end the trace antenna can be effectively free.
The enclosure decides which of those areas is actually available. Metal and moisture in the near field detune both types, and a hand or a finger in the wrong place changes the impedance of either. The design has to be judged in the product, not on the bench, and a design that works on an open board with a network analyser can fail in a plastic case with a battery two millimetres away.
Performance and What Actually Limits It
Neither type is inherently better. Antenna efficiency, bandwidth, return loss and radiation pattern depend on the ground plane, the surrounding clearance, the matching network and the losses in the laminate and the copper. A well designed PCB trace antenna on a board with a generous ground plane can match or beat a chip antenna in the same product, and a chip antenna placed against the vendor advice can be worse than either.
The most common failure is a ground plane that is too small. Both antenna types rely on the ground plane acting as the other half of the radiator, and a product whose ground occupies a few square millimetres will be limited no matter which antenna is chosen. Feed routing matters next: the line between the radio and the antenna must be kept at the specified impedance and kept short, and its geometry should be reproduced exactly from the reference design rather than approximated. The behaviour of the feed line and the wider routing rules that apply to radio frequency nets are covered in high frequency trace and data bus routing.
Matching, Tuning and the Enclosure
Both types need an RF matching network, and in both cases the network is tuned in the final product. The tuning is done with a vector network analyser or a spectrum analyser with a tracking generator, by trimming the feed point, adjusting the series and shunt elements, and in the case of a trace antenna sometimes by trimming the pattern itself. Tuning on the bare board and then closing the case is a common way to lose a working design.
The design should therefore leave a pi network or at least a couple of spare pads in the feed, with the components placed so that they can be changed without disturbing the feed geometry. Planning for the tuning step costs almost nothing at layout and saves weeks later. The wider set of layout rules that keep a mixed analogue and digital design quiet enough for a sensitive receiver is described in mixed signal PCB design guidelines.
Fabrication Tolerance and the Trace Antenna
A trace antenna is a copper dimension, so it inherits the tolerances of the etching process. On a narrow radiator trace an etch variation of a few tens of micrometres is a small fraction of the width, but on a fine pattern it can shift the resonant frequency measurably. Wider traces and generous spacing reduce the sensitivity, and the laminate properties matter as well because the dielectric constant sets the electrical length.
This is one reason to keep the antenna on the same layer and the same laminate class throughout a product family. Changing from one FR4 supplier to another, or changing the dielectric thickness, can move the resonant frequency enough to require retuning. Materials that hold their dimensions better through the press cycle keep the pattern closer to its design length, as described in PCB dimensional stability and expansion. Where the radio must be certified, that change triggers a repeat of the radio tests, which is a cost that a chip antenna with a fixed reference layout does not impose in the same way.
How to Choose
If the board has space at one end, the volume is significant, the schedule allows for tuning and the team can measure radio performance, the PCB trace antenna is usually the better economic answer. If the board is small, the schedule is short, the radio frequency experience is limited, or the product family needs a repeatable antenna across several board revisions, the chip antenna is usually the safer answer.
In either case the decision should be made with the ground plane, the enclosure and the certification requirement in view rather than by comparing the price of two components. The antenna is a system, and the component is only one part of it.
FAQ
Which antenna is cheaper? In component cost the PCB trace antenna is cheaper because it is copper rather than a purchased part. Once engineering and tuning time are included, the difference narrows and depends on the team.
Can a chip antenna be used without tuning? Not reliably. Following the reference layout gets close, but the enclosure, battery and ground plane still shift the match, and final tuning is normally required.
Does a trace antenna need a keep-out area? Yes. The ground plane must be removed under and around the radiator, and that area is unavailable for routing and for components.
