Wireless 3G/4G technologies, Wi-Fi, Bluetooth, GPS, and radio-frequency identification (RFID) rule in today’s mobile devices ranging from mobile handsets to inventory-control handheld computers and machine-to-machine (M2M) terminals. The onslaught of new technologies and applications has, in turn, heightened the complexity within the RF front end. These front ends, consisting of tuned circuits, filters, switches, low-noise amplifiers, and power amplifiers, are critical in terms of multi-band wireless-device performance. 

Simultaneously, keeping pace with the rising tide of wireless data transmissions across the globe, coupled with the move to 4G/LTE networks, will hinge on better signal quality and higher efficiencies in the mobile device. Specifically, adding more bands to the network to support data-capacity demands requires wireless-system components to be able to support wider bandwidths with lower loss.

As a result, the latest mobile terminals require more filtering and RF front-end components with enhanced linearity. The good news is that by deploying mobile devices with tunable components, operators can augment the link between the radio and the basestation. Furthermore, they will secure a better data connection, reduce the number of basestations, and improve network management. 

Over the past several years, numerous technological approaches have shown promise in the ability to tune everything from mobile handset antennas to filters and frequency synthesizers. In the current wireless market, designers can immediately reduce circuit complexity via tunable matching networks. They can extend the frequency range and performance of a single antenna, or implement tunable low-pass, bandpass, and notch filters with selectable rejection ratios. A viable tunable technology would be positioned to dramatically improve wireless radio performance.

Beyond improving radio performance, tuning circuitry helps reduce radio complexity, which is particularly important for military radios, software-defined radios, and UHV/VHF radios. Modern radios require multiple RF filters to support various bands. Therefore, these applications can benefit from tunable low-pass and bandpass filters, tunable notch filters, and production centering.

In the past, tunable filters were designed in with a switchable filter bank to select the desired frequency band. Now, a single digital tunable capacitor (DTC) filter can replace that large, switchable filter bank, reducing radio complexity and cost.

Success in the high-volume wireless market depends on the selected technology being low-cost, repeatable, and reliable while consuming low power. One such technology, used by Peregrine Semiconductor, produces DTCs that are based on the company’s high-volume UltraCMOS process. These monolithic, solid-state, digitally controlled variable capacitors operate in the 100-MHz to 3-GHz range.

Tuning Advantages

Though wireless devices would have always benefitted from tuning, it’s no longer optional due to the addition of multiple radios (Bluetooth, RFID, Wi-Fi, and cellular) and greater number of frequency bands. Multiple modes and bands require an antenna with broadband coverage—a near impossibility in a small form factor due to the tradeoff of RF performance for form factor in handset antennas (the smaller the antenna, the lower the performance).

Poor handset antenna performance leads to reduced network coverage area, low data rates, dropped calls, and poor battery life. Unfortunately, small antennas often have high-impedance mismatch loss on band edges, which reduces the power delivered to the antenna and the radiated power. This, in turn, degrades RF performance.

Moreover, the manner in which users hold a mobile handset may affect antenna performance. Potential issues include increased filter losses, reduced power amplifier output, and a jump in current consumption. Adding tuning circuitry, such as a tunable matching network, boosts radio performance (“more bars” on the mobile handset). It also simplifies radio complexity by cutting the number of required components, reducing cost and footprint.

For example, a mismatch at 700 MHz can cause an antenna to lose 84% of the power delivered to it (Fig. 1). If this cellular device must radiate at a certain power level, it has to boost its output power to compensate for the mismatch-induced energy loss. The result is diminished transmission efficiency and battery life. Fortunately, tuning technologies provide a mechanism that reduces energy lost due to mismatch.


1. This plot shows the power delivered to the antenna with and without the impedance-matching tuner. The tuner significantly increases the output power over a wider frequency range.

Adding a tunable matching network boosts the power delivered to the antenna for a fixed output power for the cellular device, demonstrating an improvement of 0.5 to 5 dB (Fig. 1, again). This translates to better radio efficiency and enhances battery life. Furthermore, at 700 MHz, a device with a tunable matching network can achieve nearly three times the data rate as one without a matching network—at the same distance from a basestation.

Tunable Technology Options

Historically, tunable RF components have been large, mechanical, and costly. The earliest variable capacitors were sizable mechanical capacitors rotated by a motor (or manually). Then, trimmer capacitors were used to fine-tune circuit performance in a production environment.

Today’s growing need for tuning technologies did not take the industry by surprise, so several alternatives are being developed to satisfy anticipated needs. The most promising tuning technologies include microelectromechanical systems (MEMS) switched-capacitor banks, barium strontium titanate (BST) ferroelectric capacitors, and DTCs based on FET switches.

Because tuning may be new to many designers, an understanding of the figures of merit is essential to its successful implementation. Applications such as mobile handsets and military radios, for instance, require a tunable capacitor with demanding performance specs:

  • High power handling: +33 to +42 dBm
  • High linearity: Harmonics ≤ 36 dBm, IMD3 ≤ 105 dBm
  • Low-loss: Q = 30 to 100
  • Capacitance: 0.5 to 20 pF
  • Tuning ratio 3:1 to 10:1
  • Low total power consumption: <1 mA
  • Fast switching speed: 10 μs
  • Reliable, rugged, and suitable for high-volume production

In terms of antenna applications, the tunable component must provide a wide impedance tuning ratio and a high quality factor, handle RF power levels up to 2 W, and meet the stringent harmonics and intermodulation distortion (IMD) requirements for 3G/4G operation.1 Given the maximum GSM (Global System for Mobile Communications) transmit power of +33 dBm and the potential voltage multiplication in matching networks under mismatch conditions, the tunable component must also linearly tolerate RF voltages of up to 30 V p-p, corresponding to over +39 dBm in 50 ω.

To support cellular GSM RX/TX tuning, the tunable component must be reliable over 1012 switch cycles and have less than 10-μs switching speed. The cellular handset application environment also requires small size, low cost, high reliability, and high yield for economical high-volume mass production. Until the arrival of UltraCMOS DTCs, the lack of electrically tunable reactive components meeting these tough specifications prevented the adoption of tuning matching networks in mobile wireless devices.

Integration is another important consideration when selecting a tunable component technology. In addition to economies of scale, the silicon-on-sapphire process employed in UltraCMOS DTCs enables multiple tunable components to be monolithically integrated on one die together with analog circuitry and a shared digital control interface. Then die can be flip-chip-mounted on module laminate.

This process represents a significant improvement over GaAs processes, which require a separate CMOS control chip and wire-bond assembly. A module containing the single-die implementation of multiple DTCs with a serial interface, coupled with surface-mount device (SMD) inductors, leads to a low-cost implementation with a reduced module footprint.

The DTC is based on a solid-state, monolithically integrated CMOS switched-capacitor bank (Fig. 2). Each switch is realized as a stack of field-effect transistors (FETs) instead of a single FET. Such a stacking approach, which fosters high RF power handling and high linearity, uses the same RF switch technology (with an insulating sapphire substrate) that’s experiencing high-volume production in 2G/3G/4G handsets.


2. In this generalized diagram of the DTC integrated circuit, the series MOSFET switches of the five integrated capacitors are connected in parallel. A 5-bit digital code turns switches off or on, depending on the desired capacitance. CMOS control circuitry operates the switches.

The DTC is programmed through a two- or three-wire serial interface, designed using existing UltraCMOS switch FETs combined with metal-insulator-metal (MIM) capacitors. The UltraCMOS DTC can be used either in a series or shunt configuration, which makes it a fit for a host of tunable circuits.

Tunable Matching Networks

In one example of a single-path, tunable matching network using UltraCMOS DTCs, three tunable components, biasing, and control circuitry are integrated onto a single die (Fig. 3). This design controls the DTCs through a serial interface. By employing a reconfigurable coupled resonator topology (widely used in bandpass filters and impedance-matching networks), this tunable matching network provides wide impedance coverage in the 698- to 960-MHz and 1710- to 2170-MHz bands.


3. Shown is an antenna impedance-matching network consisting of three discrete inductors and three DTCs. The network is tuned for minimum VSWR using digital commands through the MIPI interface.

Each DTC in Figure 3 is implemented using 4-bit resolution as a tradeoff of resolution versus the number of individual tuning states available for the network (4096 for 3- by 4-bit DTCs). The device measures approximately 3 mm2, so it’s suitable for integration into a 3.5- by 3.5-mm module that includes passive components, such as high-Q wirewound inductors.

The tunable matching network’s performance in the system can be evaluated with a metric called the power delivered improvement (PDI).2 PDI measures how much the cellular device’s performance will improve (in decibels) with a tunable matching network for a given antenna voltage standing-wave ratio (VSWR) and phase angle.

When antenna impedance is at high VSWR (12:1), using the UltraCMOS tuning circuitry improves power delivered to the load by 4 dB or more. The break-even point in power delivery occurs when the load VSWR is approximately 2:1. When it drops below that, the tuner dissipative losses are higher than mismatch loss without the tuner.

In summary, antenna performance can improve significantly by adding a tunable matching network. This will increase the power delivered to, and radiated out of, the antenna. Furthermore, the reduced radio complexity through tuning will undoubtedly benefit military radios, software-defined radios, and UHF/VHF radios.

References

  1. Ranta, T., Ella, J., and Pohjonen, H., “Antenna Switch Linearity Requirements for GSM/WCDMA Mobile Phone Front-ends,” 8th European Conference on Wireless Technology Proceedings, Paris, France, Oct. 2005, p. 23-26.
  2. Whatley, R., Ranta, T., and Kelly, D., “CMOS Based Tunable Matching Networks for Cellular Handset Applications,” IMS2011 Proceedings (in press).