RF Switches

Introduction

An RF switch routes a high-frequency signal from one path to another under electronic or mechanical control. In a radio system, switches select between antennas, swap a transceiver between transmit and receive, insert or bypass filters and attenuators, route signals through banks of components in test instruments, and steer power among the elements of a phased array. Wherever a signal must change paths without being torn down and rebuilt, an RF switch performs the task while preserving the integrity of the signal it carries.

Switching at radio and microwave frequencies is far more demanding than switching a direct-current circuit. The switch must present a controlled impedance, usually 50 ohms, in every state; it must pass the wanted signal with minimal loss while strongly blocking it in the off path; and it must avoid generating the spurious products that nonlinear behavior creates. These competing demands have produced several distinct switch technologies, each with a characteristic balance of speed, loss, isolation, linearity, and power handling. This article surveys those technologies, the common circuit configurations, the parameters that quantify performance, and the applications that drive switch selection.

Switch Technologies

Four families of devices dominate RF switching. Semiconductor switches, based on PIN diodes or field-effect transistors, offer fast, reliable, solid-state operation. Microelectromechanical (MEMS) switches occupy a middle ground between semiconductors and relays. Electromechanical relays provide the lowest loss and highest isolation at the cost of speed and life. Selecting among them is the central decision in any switch design.

PIN Diode Switches

A PIN diode has an intrinsic region between its P-type and N-type layers that, when forward biased, fills with charge and behaves as a low, current-controlled resistance, allowing RF signals to pass. When reverse biased or unbiased, the intrinsic region is depleted and the diode presents a small capacitance that blocks the signal. Crucially, at radio frequencies the diode does not rectify the signal; it acts as a bias-controlled resistor whose value the designer sets with direct current. PIN diodes handle high power, achieve high isolation and low insertion loss, switch in tens to hundreds of nanoseconds, and remain very linear. Their principal drawbacks are the continuous bias current required to hold the on state and the additional bias networks, chokes, and blocking capacitors needed around them.

FET and SOI Switches

Field-effect transistor switches use the channel of a transistor as a voltage-controlled resistance: a gate voltage turns the channel on to pass the signal or off to block it. Because they are controlled by voltage rather than current, FET switches draw essentially no static power, and they integrate readily into monolithic microwave integrated circuits alongside other functions. Gallium arsenide FETs have long served this role, but silicon-on-insulator (SOI) and silicon-on-sapphire technologies now dominate high-volume products. By building transistors on an insulating substrate, SOI suppresses the parasitic coupling that would otherwise spoil isolation and linearity, and it allows many transistors to be stacked in series so the switch can withstand large RF voltage swings. Modern SOI switches offer excellent linearity, very high integration, switching times of tens to a few hundred nanoseconds, and the multi-throw complexity that cellular front ends require, though their power handling and isolation generally trail those of PIN diodes and relays.

MEMS Switches

A radio-frequency MEMS switch is a tiny mechanical relay fabricated with semiconductor processes. An electrostatic force pulls a micromachined metal beam or membrane into contact, either touching a signal line directly or changing a capacitance, thereby opening or closing the RF path. Because the moving element is metal and the open gap is air, MEMS switches achieve very low insertion loss, very high isolation, and outstanding linearity that rivals or exceeds electromechanical relays, all in a part far smaller and lighter than a conventional relay. They switch in microseconds, faster than relays but slower than semiconductors, and consume negligible power. Their adoption has been limited by the high control voltages they often require, by reliability and packaging challenges, and by cost, so they appear chiefly in instrumentation and specialized systems rather than in mass-market handsets.

Electromechanical Relay Switches

An electromechanical RF relay moves a metal contact with an electromagnetic coil, making a physical metal-to-metal connection for the on path and a wide air gap for the off path. This direct metallic contact gives the lowest insertion loss, typically a few tenths of a decibel, and the highest isolation, often greater than 60 decibels, of any switch type, together with essentially perfect linearity because there is no semiconductor junction to distort the signal. Relays also handle large power. Their limitations are mechanical: switching takes milliseconds, the contacts wear and limit the part to a finite number of operations, and the coil consumes power and adds size and weight. Relays remain the choice for test equipment, high-power transmit paths, and any application where loss and linearity outweigh speed and life.

Switch Configurations

RF switches are described by the number of poles, the common terminals, and the number of throws, the selectable destinations, following the same naming convention used for ordinary electrical switches. The configuration determines how many paths the switch can route and how the part is used in a system.

Single-Throw and Double-Throw

The single-pole, single-throw (SPST) switch simply connects or disconnects one path, acting as an RF on/off element used to gate a signal or protect a sensitive input. The single-pole, double-throw (SPDT) switch selects between two destinations from one common port; it is the workhorse of transmit-receive switching, connecting an antenna alternately to a power amplifier and a low-noise receiver. SPDT switches are among the most widely used RF parts of all.

Multi-Throw Switches

Where more than two destinations must be selected, multi-throw switches such as SP3T, SP4T, SP6T, and SP8T connect one common port to one of several outputs. These route signals among multiple frequency bands, antennas, or filter paths, and they are central to multi-band cellular front ends and to the band-selection logic of broadband instruments. The transfer switch, a double-pole double-throw configuration, simultaneously swaps two signal paths and is used for redundancy switching, where a spare unit can be brought in to replace a failed one without disturbing the rest of the system. Switch matrices extend the idea further, connecting any of several inputs to any of several outputs for the flexible routing required in automated test systems.

Absorptive and Reflective Switches

Configurations are also distinguished by what happens to the off ports. A reflective switch leaves an unselected port as an open or short circuit, reflecting incident energy back toward the source. An absorptive, or terminated, switch instead connects each off port to an internal 50-ohm load, absorbing the energy and presenting a good match in every state. Absorptive switches protect upstream stages from reflections and maintain stable impedance, which matters when the switch feeds sensitive amplifiers or precision measurements, whereas reflective switches are simpler and slightly lower in loss.

Insertion Loss and Isolation

Two opposing specifications define the basic quality of any RF switch: how little it attenuates the signal it is meant to pass, and how strongly it blocks the signal it is meant to stop.

Insertion Loss

Insertion loss is the power lost in the on path, expressed in decibels, and a lower figure is better. In receive paths it directly degrades the system noise figure decibel for decibel, so a tenth of a decibel matters; in transmit paths it wastes amplifier power as heat. Insertion loss arises from the on-resistance of a semiconductor channel or forward-biased diode, the contact resistance of a relay, and the parasitics of the package and matching network. It generally rises with frequency, so datasheets specify it across the operating band. Relays and MEMS devices achieve the lowest values, semiconductor switches somewhat higher.

Isolation

Isolation measures how much the switch attenuates a signal in the off state, again in decibels, but here a higher number is better. Poor isolation lets unwanted signals leak through a supposedly disconnected path, allowing a transmitter to bleed into a receiver or one channel to contaminate another. Isolation is limited by the off-state capacitance of a semiconductor, which couples more signal as frequency rises, or by the residual capacitance across an open relay or MEMS gap, so isolation falls as frequency increases. Designers raise isolation by stacking series devices, adding shunt devices that short leaked signal to ground, or, in relays, widening the contact gap. There is an inherent tension between loss and isolation: techniques that strengthen one often weaken the other, and the art of switch design lies in balancing them for the application.

Linearity and Power Handling

An ideal switch would pass any signal level without distortion and survive any power, but real devices impose limits at both extremes. Linearity governs the largest signal a switch can carry cleanly, and power handling governs the largest signal it can survive.

Linearity

Linearity describes how faithfully the switch reproduces the signal as power rises. Semiconductor switches contain junctions and channels whose resistance varies slightly with the instantaneous signal voltage, and this variation generates harmonics and intermodulation products that can fall on top of wanted signals or violate spectral-emission limits. Linearity is quantified by the 1 dB compression point, the level at which the switch begins to compress the signal, and by the third-order intercept point, which characterizes intermodulation when several signals are present; higher values indicate better linearity. Relays and MEMS switches, having no semiconductor junction, are essentially perfectly linear. Among semiconductors, SOI switches built with stacked transistors achieve high linearity, which is one reason they have displaced earlier technologies in demanding cellular applications.

Power Handling

Power handling is the maximum signal power the switch can carry without damage or excessive distortion, specified separately for continuous average power and for brief peaks. Exceeding the rating risks burning out a semiconductor, welding relay contacts, or driving the device so far into compression that the signal is corrupted. Relays and PIN-diode switches handle the most power, making them suitable for transmit chains and high-power instrumentation, while FET-based switches generally handle less. Power handling falls at higher operating temperatures, so designers derate the device and provide adequate heat removal in high-power paths.

Switching Speed and Control

How quickly a switch changes state, and how it is commanded to do so, determine whether it suits a given task and how it is integrated into a system.

Switching Speed

Switching speed spans many orders of magnitude across the technologies. FET and PIN-diode switches change state in microseconds down to tens of nanoseconds, fast enough for time-division duplexing, beam steering, and rapid measurement sweeps. MEMS switches, being mechanical, take microseconds. Electromechanical relays are slowest at milliseconds, limited by the inertia of the moving contact. Speed specifications often distinguish the rise and fall of the RF envelope from the longer settling time required for the signal to stabilize fully, and applications that switch during live transmission care about both.

Control Interfaces and Bias

The control circuitry varies with technology. PIN-diode switches need bias networks that supply forward current to the on path and reverse voltage to the off paths, isolated from the RF by chokes and blocking capacitors. FET and SOI switches need only gate-control voltages and frequently embed a driver and decoder on the same chip, so a few logic lines, or a serial bus, select among many throws. MEMS switches require relatively high actuation voltages, often supplied by an integrated charge pump. Relays are driven by coil current, usually through a driver transistor. Many integrated switches present a simple parallel or serial digital interface that hides these details from the system designer.

Practical Design Considerations

Selecting and applying an RF switch involves more than reading the headline loss and isolation figures. Several secondary characteristics and layout practices decide whether the switch performs as intended in the finished product.

Impedance Matching and Return Loss

Each port of the switch should present close to 50 ohms in every state, characterized by return loss or voltage standing-wave ratio across the band. Poor matching reflects power, ripples the frequency response, and can interact with adjacent stages. Absorptive switches help by terminating off ports internally; otherwise the layout must provide controlled-impedance transmission lines and clean grounding right up to the switch terminals.

Settling, Video Feedthrough, and Transients

When a switch changes state it can briefly inject a control-related transient, sometimes called video feedthrough, onto the RF line as the bias or gate voltage moves. Sensitive systems must allow the switch to settle before sampling the signal, and designers may slow the control edges or add filtering to suppress these transients. Hot switching, changing state while significant RF power is present, stresses relay contacts and can degrade reliability, so many high-power designs sequence the controls to switch only when the power is removed.

Integration and Packaging

Package parasitics limit the usable frequency range and degrade isolation, so RF switches use low-inductance surface-mount or chip-scale packages, with waveguide or precision coaxial interfaces reserved for the highest frequencies. Highly integrated front-end modules combine switches with filters, amplifiers, and matching in a single package to save board space in handsets. The choice between a discrete switch and an integrated module trades flexibility and optimum performance against size and assembly cost.

Applications

RF switches appear throughout wireless and high-frequency systems wherever signals must be routed, selected, or sequenced.

Transmit-receive switching - SPDT switches connect a shared antenna alternately to the power amplifier and the low-noise receiver in time-division systems, the most common single use of RF switches.
Multi-band cellular front ends - Multi-throw SOI switches select among many frequency bands, antennas, and filter paths in smartphones, where high linearity and integration are essential.
Phased-array and beam-steering systems - Fast switches route signals and set phase paths among array elements in radar and satellite communication antennas.
Test and measurement instruments - Switch matrices and low-loss relay switches route signals among instrument ports and devices under test, where loss and repeatability dominate the requirements.
Filter and attenuator bank selection - Switches insert or bypass filters, attenuators, and amplifiers to reconfigure a signal chain across frequency or level.
Redundancy and protection switching - Transfer switches swap in spare units and gate sensitive inputs to protect receivers from overload.

Summary

RF switches route high-frequency signals between paths while preserving the controlled impedance, low loss, high isolation, and linearity that radio systems demand. Four technologies meet these needs in different ways: PIN-diode switches handle high power with fast, current-controlled operation; FET and silicon-on-insulator switches deliver low static power, high integration, and the multi-throw complexity of modern cellular front ends; MEMS switches combine near-relay performance with semiconductor-scale size; and electromechanical relays provide the lowest loss and highest isolation at the expense of speed and finite contact life.

Configurations from SPST through multi-throw and transfer arrangements, in reflective or absorptive forms, determine how many paths a switch routes and how it behaves at its idle ports. The key parameters, insertion loss, isolation, linearity, power handling, and switching speed, stand in tension with one another, so switch selection is a matter of matching the technology and configuration to the application. With attention to impedance matching, control and bias, settling and hot-switching behavior, and packaging, RF switches form an indispensable building block of wireless communication, radar, instrumentation, and every system that must steer signals at radio and microwave frequencies.