RF front-ends for GSM mobile handsets continue down path of integration
Scott Smyer
(04/12/2004 9:00 AM EDT)
Integration in the radio-frequency (RF) front end of Global System for Mobile Communications (GSM) mobile handsets is continuing to advance through the use of modules, from front-end modules (FEMs) to RF-transceiver modules. This front-end integration commenced when RF transceivers adopted the direct-conversion,or zero-IF (ZIF) architecture, which eliminated the intermediate-frequency (IF) stage, and subsequently, the IF surface acoustic wave (SAW) filter. As transceiver architectures have evolved, the external synthesis components--namely the voltage-controlled oscillators and phase-locked loops--have been integrated directly into the silicon of the transceiver.
With this integration has come cost savings and ever-shrinking board sizes. The trend toward integration shows no signs of letting up. However, the paths to extreme integration are many and must be given careful consideration.
FEMs on the march
A front-end module integrates the antenna-switch module (ASM) and the receive-side RF SAW filters. Passive components are integrated using low-temperature co-fired ceramic (LTCC) technology. The ASM typically is a PIN-diode type.
ASM and SAW filter suppliers are the primary promoters of FEM products. A FEM can serve as a simple alternative to designing a discrete solution. In the case of a dual-band GSM handset design, a dual-band FEM can replace the dual-band ASM and the two discrete RF SAW filters. The FEM saves some design time and component count.
FEMs also provide advantages when used in platform solutions. A handset manufacturer that is designing three handsets in a platform, such as quad-band, tri-band, and dual-band models, can use a quad-band FEM in all of the designs. The quad-band FEM obviously supports the quad-band handset, but represents overkill for the dual-band handset. However, the design process for the platform can be streamlined by using the one component versus multiple discrete parts for each handset.
While the component cost of using the quad-band FEM in the dual-band handset design may be greater than if employing a dual-band FEM or a discrete solution, employing one quad-band FEM saves supply-chain costs. If the handset manufacturer used a different FEM in each of its handsets, it would add additional line items that would need to be managed, sourced and inventoried. These added steps would multiply if the handset manufacturer decided to use a discrete solution in each of its handsets.
Today, FEMs may or may not command a price premium compared to discrete solutions, depending on the customer. For example, tier-one handset manufacturers are not paying a premium for FEMs, while smaller players are.
FEMs quickly are becoming the preferred solution in mobile handset designs. This proliferation of FEMs will continue as handset manufacturers become more aware of the design and cost benefits of using them.
Integration alternatives
An alternative integration path employs a module combining an antenna switch module with the power amplifier module (PAM), a solution favored by PAM suppliers. These modules use GaAs pseudomorphic high electron mobility transistors (pHEMT) technology for the ASM and GaAs heterostructure bipolar transistors (HBT) technology for the PAM. Passive component integration occurs directly on the GaAs wafer.
More recently STMicroelectronics N.V. (Geneva, Switzerland) released its tri-band RF modules, which incorporate the RF transceiver, bulk acoustic wave (BAW) filters and ASM.
ST uses what it calls integrated passive and active device (IPAD) technology to integrate passive components. Furthermore, it employs a laminate technology that allows the RF transceiver, antenna switch and the BAW filters to be integrated on a single substrate.
The ST module achieves a key level of integration. As opposed to merely placing discrete components on a pc-board and calling it a module--as some semiconductor suppliers have done--ST starts with a highly integrated RF transceiver and adds passive and active component integration, in addition to flip-chip BAW filters.
In ST's RF module solution, the only major external discrete component required is a megahertz crystal. However, microelectromechanical system (MEMS) resonators are being pursued by companies such as Discera Inc. (Campbell, Calif.), which offer the possibility of replacement and integration of the bulky reference crystal.
An integrated future
What are the next steps for integration? One obvious path would be to integrate the ASM with the SAW/BAW filters and the RF transceiver. Another direction would be to integrate the ASM with the SAW/BAW filters and the PAM. The ultimate outcome of all this integration is the creation of a complete module that combines everything.
These paths toward extreme integration will have to be considered carefully. Some technology issues must be ironed out. For example, ASM technology is either PIN diode or GaAs pHEMT, technologies that may not lead to easy integration with the filters and RF transceiver. There have been challenges with the heat dissipation of the power amplifier affecting the SAW/BAW filter's thermal characteristics in an ASM-PAM-filter module.
While there are different paths to integration in the RF front end of mobile handsets, it is clear that the trend of integration will continue. The requirements of the RF subsystem are unchanging and therefore this segment of mobile phones must be simplified to make room for added components and cost in the baseband.
As this integration trend continues, it will remain important to keep in mind the benefit to the end user. Mobile-handset manufacturers would like to streamline designs without adding cost. This desire has been clearly illustrated by the widespread use of FEMs. Will a more highly integrated module be able to accomplish the same? As technology progresses, more highly integrated modules will most likely will make their way into handset designs, offering a more streamlined and cost-effective solution.
Scott Smyser(ssmyser@isuppli.com) is senior analyst for frequency control at the RF and wireless products division of iSuppli Corp. (El Segundo, Calif.).
Related Semiconductor IP
- RISC-V CPU IP
- AES GCM IP Core
- High Speed Ethernet Quad 10G to 100G PCS
- High Speed Ethernet Gen-2 Quad 100G PCS IP
- High Speed Ethernet 4/2/1-Lane 100G PCS
Related White Papers
- Simulating Trade-offs in W-CDMA/EDGE Receiver Front Ends
- Integration by function puts more features, interfaces into handsets
- Making Converter Integration Happen in Mobile Phone SoCs
- RF integration: Changing the face of test
Latest White Papers
- New Realities Demand a New Approach to System Verification and Validation
- How silicon and circuit optimizations help FPGAs offer lower size, power and cost in video bridging applications
- Sustainable Hardware Specialization
- PCIe IP With Enhanced Security For The Automotive Market
- Top 5 Reasons why CPU is the Best Processor for AI Inference