Printer Friendly

Switching to suit the application.

The characteristics of RF and microwave switches are described by typical or guaranteed performance usually plotted as a function of frequency. Insertion loss, return loss, and isolation make up one set of parameters often included in the switch specifications. These are scalar dB power quantities, and typical measurements may be shown relative to the guaranteed limits.

Alternatively, some datasheets state voltage standing wave ratio (VSWR), isolation, and insertion loss. These simply may be listed as maximum values within a frequency range rather than presented graphically.

Several types of devices are used as RF/microwave switches including FE'Ts, subminiature relays, MEMS switches, PIN diodes, and coaxial relays. What makes one technology better than another for a given application?

To help distinguish one from another, Agilent Technologies' Cheah Kai-Nian, product marketing engineer, listed 12 parameters and ranked them for coaxial relays, PIN diode switches, and FET switches. The radar diagram in Figure 1 was constructed from this data. For all 12 axes, performance increases going away from the center.

Clearly, no one technology excels in every respect. The specific application governs the best choice among the alternatives. For example, a coaxial relay can have very high-frequency operation with excellent insertion loss and power handling. A PIN diode switch is faster and has infinite life but can only handle relatively low power and exhibits a larger insertion loss. A FET-based switch has even higher insertion loss but lower video leakage.

Subminiature electromechanical relays and reed relays have the advantage of small size but are limited to lower frequencies.

FET and PIN diode switches are damaged by overvoltage caused either by ESD events or by the signal itself. Coaxial relays are sensitive to vibration but relatively unaffected by ESD.

Relationships At RF and microwave frequencies, S-parameters often are used to describe performance because S-parameter measurement does not involve shorts or opens both difficult to accurately achieve at high frequencies. Equations 1 through 4 relate S-parameters to other terms with which they are closely associated.


where: IL= insertion loss in dB


where: RID n 7--input return loss in dB


where: RLout = output return loss in dB


where: Irev = reverse isolation in dB

In each equation, only the magnitude of the relevant S-parameter is involved because the related loss quantities are powers. S-parameters are ratios of voltages, and power is related to the square of voltage. That is the reason the equations have a factor of 20 in the exponent rather than 10. S-parameters have complex values and can be presented on a Smith chart as is commonly done for amplifier gain and the impedance or admittance loci of coupling networks. Typically, manufacturers do not provide complex S-parameters for switch modules.

Insertion Loss

Using Agilent Technologies' 18-GHz FET-based Model U9397C Switch as an example, Figure 2 shows the insertion loss data copied from the datasheet along with the derived values for [S.sub.21].

Assuming that the same reference impedance applies to the source and load ports, IL is the power loss caused by inserting the switch into the circuit and is proportional to the square of the voltage gain or attenuation [S.sub.21]. From the graph, the voltage gain of the switch at low frequencies is about 0.75, reducing to just above 0.50 at 16 GHz.

A high insertion loss is one of the major drawbacks of FET-based switches in addition to their limited power capability. The U9397C can handle +27 dBm maximum RF input power or 0.5 W. In comparison, the Radiall R594 range of SPnT coaxial switches is rated for hundreds of watts at lower frequencies, reducing to from 20 to 100 W at 18 GHz depending on the model with further derating for high VSWR values. The typical insertion loss is minimal as shown in Figure 2.

Nevertheless, FET-based switches have virtually unlimited switching life vs. a guaranteed 1E7 operations for the R594 and a much faster operating speed of 350-ps for the U9397C vs. the coaxial relay's 15 ms.

Return Loss

Return loss is related to the magnitude of the voltage reflection coefficient [GAMMA], whether at the input or output, expressed as a power. A return loss value is a negative dB quantity, so the larger the value, the higher the ratio between the input or output and the reflected power. For the U9397C Switch, Figure 3 shows that the ON input reflection coefficient [GAMMA] remains below about 0.2, which means that the fraction of incident power entering the switch 1 remains above 96%.

VSWR = ([GAMMA] + 1)/([GAMMA] - 1) and is plotted along with the ON return loss and [GAMMA] for the U9397C Switch. VSWR is the ratio of the maximum to the minimum of the voltage standing wave caused by interaction of the incident and reflected waves. It is another way to express the quality of the impedance match, in this case at the input of the switch.

The R594 Coaxial Relay has return loss similar to the U9397A above about 4 GHz, but for lower frequencies shows a much lower reflection coefficient and correspondingly greater return loss. Clearly, none of the three quantities behaves in a simple manner with increasing frequency. In particular, VSWR tends to increase with increasing frequency as the quality of the impedance match degrades.

Switch Technologies

Many relay-based switches operate by interrupting the circuit. They change the series impedance by a very large ratio. For example, the ON resistance of a relay typically is <1 [OMEGA], but the OFF resistance could be 1E10 [OMEGA] or more. Series switches are easy to apply at low frequencies, but at high frequencies, the capacitance across the relay contacts reduces the OFF impedance and limits a switch's usefulness.

Some RF/microwave switch modules can be programmed to terminate OFF inputs to a multiplexer, for example. This means that the signal sources driving those switch ports do not see reflections that would occur if the ports simply were left open. Of course, termination involves more relays and complexity.

In narrowband applications, multiplexers may use 1/4-wavelength sections to intentionally reflect power from open ports, diverting it to the one port with a matched termination. The 1/4-wavelength lines transform the low impedance of closed shunt switches at the line outputs to high impedances at the line inputs.

All switch elements provide a change of impedance. To achieve the highest ratio of ON to OFF performance, some modules use both series and shunt elements. For example, an OFF series element might be followed by an ON shunt element. This approach improves the OFF state isolation.

Switches are subject to a number of secondary effects. One of these, video leakage, refers to the disturbance in the signal channel caused by a change of state in the switching control line. As shown in Figure 1, coaxial relays and FET-based switches have very low video leakage; PIN diode switches have more, sometimes a great deal more.

The main reason PIN diode switches perform so poorly in this respect is the shared path for the control and RF signal. The diode resistance is controlled by changing the DC current. But, when the current level changes abruptly, DC blocking capacitors in the RF path and RF blocking inductors in the DC path have to charge. Especially when fast switching is needed, transients are created by the steep leading and trailing edges of the control signal. These transients are within the signal-path bandwidth so they appear as a signal disturbance.

An Agilent application note describes the process in detail and explains why the separate control and signal paths in a FET-based switch largely eliminate video leakage. (1) Some applications such as power amplifier testing must limit input power to 10-dBm max and cannot use PIN diode switches because transients may exceed 2 V or +19 dBm. In contrast, a FET-based switch may produce only a 20-mV transient, a factor of 100 smaller, equivalent to - 21 dBm. In contrast to a coaxial relay, FET-based switches provide much greater throughput and unlimited life.

MEMS switches are not yet widely available, a number of companies finding the technology more difficult than first anticipated. RadantMEMS is one of the more successful MEMS switch suppliers, having demonstrated products with cold-switching lifetimes >1E11 operations and >1E9 with a 500-mA current. The company's SPST RMSW100HP [TM] RFMEMS Switch has <0.20-dB insertion loss, >20-dB isolation, and >24-dB return loss at 4 GHz.

Within the same sealed MEMS assembly, series and shunt switches can be combined to provide higher isolation: 50 dB at 10 GHz has been achieved. Typical of many MEMS switches, RadantMEMS switches use electrostatic actuation, which requires voltages between 40 and 120 V depending on the specific design.

Industry Input

Electromechanical Switches

Mike Dewey, senior product marketing manager at Geotest-Marvin Test Systems, said his company favors relays rather than electronic switching for a couple of reasons. "We use subminiature relays with specified RF performance whenever possible. If we need the bandwidth, we will then look to coaxial relays that are good into the tens of gigahertz. The subminiature relays offer higher density and require less volume, which is attractive for the PXI architecture," he explained.

Regarding electronic switches, Dewey had reservations. "The challenges associated with electronic-based switches include the limited operating voltage ranges that can be accommodated by these devices. In the world of test, you can encounter a broad range of operating and fault voltage conditions, which could very well go beyond the bounds of an electronic switch. We are seeing more promise with many of the new miniaturized relays that offer bandwidths into the low gigahertz region. We still have concerns about contact power rating, which tends to be 10 W or less, but the capability to accommodate high stand-off voltages, coupled with low resistance and flat bandwidth, makes them more attractive than an electronic device," he said.

Pickering Interfaces' Business Development Manager David Owen agreed that "coaxial switches are the switch of choice for applications requiring the widest bandwidth and the highest signal levels. They exhibit the best linearity and the highest levels of performance." He added that subminiature relays were limited to about 3 GHz and above that frequency generally were not price-competitive with coaxial relays.

Jeff Lum, CTO at Giga-tronics, also noted the high-frequency limitations of subminiature relays, but said, "These are still the most commonly used switches up to 100 MHz. There have been some subminiature switches for up to 6 GHz applications. The problem with them, like all electromechanical switches, is short life. Also, isolation drops to 30 dB above 3 GHz."

Although National Instruments makes a number of FET-based switch modules, including the recently released Model NI PXI/PXIe-2543 Multiplexer, Jake Harnack, switches product manager, acknowledged that there are performance characteristics that limit use in some applications. "Coaxial relays are capable of handling higher bandwidth and power applications while maintaining high isolation and low insertion loss. These benefits make coaxial relays well suited for many RF and microwave test system in the aerospace and defense industries, where switches are designed to have a minimal impact on RF signals," he said.

"Electromechanical relays are still the standard in these industries, where projects can require three to five times the bandwidth provided by current FET relays. To help manage the finite lifetime of electromechanical relays, NI PXI switches have built-in relay count tracking," Harnack added.

As data rates and carrier frequencies continue to increase, test systems have to keep pace. Agilent recently extended the company's N1810/1/2 Series of 67-GHz switches by adding the N1810TL/UL terminated and unterminated SPDT latching coaxial switches and the N1811TL/12TL four - and five-port bypass switches. In spite of the higher frequency range, these switches retain 0.05-dB insertion loss repeatability, isolation >70 dB at 67 GHz, and a guaranteed 5E6-operation operating life.

VTI Instruments' has developed the Broadband Integrated Design Wizard as an aid to customers needing to design RF and microwave interface units. The user is prompted to provide data in a logical and complete manner. The design wizard then creates the bill of materials, interconnect diagrams for path level programming, wire lists, and custom graphical web-based front-panel displays. Designs are based on VTI's LXI EX7000 Series platform and include a range of multipole coaxial relays for operation to 26.5 or 40 GHz.

Electronic Switches

Ni's Harnack highlighted the need to look at the overall switching requirement in a test application. "FET-based switches are an excellent way to improve speed in automated test systems, but these benefits are minimized without first reducing software overhead by implementing hardware scanning and triggering. To fully optimize test times, deterministic hardware scanning and triggering are supported on a wide range of PXI FET-based as well as electromechanical relay-based switch modules," he said.

Owen from Pickering commented on the range of 40-88x solid-state switches the company recently introduced. "FET-based switches offer excellent solutions to 6 GHz and provide broad frequency cover: They are good solutions for broadband switching with long operational life, provided signal levels are within their rating," he said. Owen explained that at higher frequencies, FET-based switches must use smaller geometries, which further limits their power-handling capabilities.

Carlos Fuentes, principal engineer at Giga-tronics, agreed, noting that GaAs FET switches had good linearity at microwave frequencies even though their power handling was limited. He added that new GaN switches have the potential to have outstanding linearity at power levels that until now had to be handled by electromechanical switches.

For More Information

Agilent Technologies

Geotest-Marvin Test Systems


National Instruments

Pickering Interfaces



VTI Instruments


(1.) Video Leakage Effects on Devices in Component Test, Agilent Technologies, Application Note, 5989-6086EN, 2007.

by Tom Lecklider, Senior Technical Editor
COPYRIGHT 2012 NP Communications, LLC
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2012 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Lecklider, Tom
Publication:EE-Evaluation Engineering
Geographic Code:1USA
Date:May 1, 2012
Previous Article:Digital MEMS microphone market expands, partly driven by new iPad and iPad 2 designs wins.
Next Article:Memory, Processor, transceiver testing gets boost.

Terms of use | Privacy policy | Copyright © 2021 Farlex, Inc. | Feedback | For webmasters |