- Buyers Guide
Pat Hindle, MWJ Editor
Clash of the Titans - GaAs vs CMOS PAs
Will CMOS finally conquer GaAs in handsets?
Having traveled to Barcelona for Mobile World Congress (MWC) in late Feb, heading back out to Beijing for EDI CON 2013 in early March and then returning quickly to get the May IMS 2013 issue in ready, I am only now catching up with some of the items of note from these shows. The one item that sticks out in my travels is the Qualcomm announcement. Although there has been some commentary about Qualcomm announcing the RF360 Family just prior to MWC, it is surprising that it has not gotten more attention as it could turn the industry upside down. Up until this point in time, CMOS PAs have only been successful in 2G applications with some limited success in 3G handsets. But they have yet to achieve performance that would rival GaAs in 4G applications where high peak to average powers are required.
According to the announcement, “the Qualcomm RF360 Front End Solution, a comprehensive, system-level solution that addresses cellular radio frequency band fragmentation and enables for the first time a single, global 4G LTE design for mobile devices.” Qualcomm further states that “the RF front end solution includes the industry’s first envelope power tracker for 3G/4G LTE mobile devices, a dynamic antenna matching tuner, an integrated power amplifier-antenna switch, and an innovative 3D-RF packaging solution incorporating key front end components.”
According to Strategy Analytics, Qualcomm's announcement ignited a flurry of activity at MWC, with no fewer than 10 companies releasing new details about how they will support envelope tracking (ET) to enable better LTE devices with multimode, multiband PAs (MM-PAs). They stated that “adoption of ET in LTE handsets has suddenly moved about 18 months closer to reality, with first products likely to appear by late 2013 or early 2014 in production mobile devices.” Strategy Analytics added that new PAs and ET from GaAs-oriented competitors will dilute Qualcomm’s new presence in the RF front end, but widely available ET will also improve prospects for CMOS PAs, challenging the dominance of leading GaAs-based PA suppliers Skyworks, RFMD and others. It is interesting to note that RFMD bought CMOS PA designer Amalfi last year so now has CMOS PA capability to potentially protect them from this threat.
This is the showdown we have been talking about for several years between GaAs and CMOS PAs in the handset market (see our previous article “3G Power Amplifiers: Moving from GaAs to CMOS”). But I have some doubts because after years of research by CMOS PA experts like Javelin, Amalfi, Black Sand and others, CMOS has yet to compete in high performance applications even though they are getting closer. I just wonder what innovation(s) Qualcomm discovered that allowed them to attain 4G performance. I also think it is mis-leading that CMOS PAs with ET are constantly compared to GaAs PAs without ET. The efficiency of GaAs is much improved with ET so shouldn’t we compare apples to apples? But if anyone can accomplish this in CMOS, Qualcomm would probably be the best positioned to do it.
We should find out more information soon as the Qualcomm product will be formally released later in the year, and Microwave Journal will be publishing an article written Qualcomm and one or two GaAs leaders in June addressing this very subject. The handset RF front-end is undergoing rapid innovation with ET and antenna tuning being coupled with better PAs.