From The Editor | November 21, 2012

New Heights In Amplifier Efficiency


By Paul Kruczkowski, Editor

Improving efficiency has long been and will continue to be a fundamental challenge in amplifier design for wireless infrastructure. Higher efficiency base stations require less power and heat dissipation, which reduces operating costs. Today’s base stations utilize Doherty amplifier designs that boast 45% to48% efficiency, but as impressive as that may be, more than half of the power applied is still lost or dissipated. However, I recently learned about a new supply modulation technique that could soon lead to new heights in base station efficiency.

The technique, known as asymmetric multilevel outsourcing (AMO), was developed by MIT electrical engineering professors Joel L. Dawson and David J. Perrault, who also co-founded the startup Eta Devices to commercialize the technology. AMO increases the efficiency of Eta Devices’ class-E gallium nitride (GaN) amplifier design to approximately 70% for LTE bandwidths.

AMO amplifiers are similar in concept to envelop tracking amplifiers. However, AMO doesn’t suffer from the same modulated bandwidth limitations of envelop tracking. “The fundamental advantage of AMO as a supply modulator is that we don’t demand as much from the DC/DC converter block as envelop tracking does, so AMO’s efficiency curve stays flatter as bandwidth increases,” Dawson explained to me during a recent conversation.

This past year, Eta Devices has been focused on the difficult work of transforming the marketable research into a marketable product. The company laid the foundation by selecting reliable class-E operation, and paved the future by opting for cutting-edge GaN technology — to ensure efficiency and appeal to many industries. “Although we have chosen to focus on the base station and handset market, AMO is a platform technology and can be applied to defense, satellite, broadcast TV, and MRI applications,” Eta Devices CEO Mattias Astrom told me.

Eta Devices tackled the hard work of eliminating technical risk when they effectively increased linearity and achieved adjacent channel power ratios (ACPR) of -50 dBc without sacrificing the efficiency gains that they had demonstrated in their research.  In addition, they quadrupled output power levels to 80 W peak /16 W average to reach power levels for small cell applications.

The AMO amplifier is poised to make its debut at Mobile World Congress in Barcelona, Spain, this February. So what can we expect to see? I have been told that the Eta Device solution accepts digital I/Q inputs and will provide RF at the antenna output at the levels I previously indicated. I expect the ASIC control and AMO supply modulator to support LTE bands between 700 MHz and 2.6 GHz, and I have been told that Eta Devices plans on presenting a 1.8 GHz reference design at the show.

A lot can change between now and February, but double-digit gains in efficiency deserves attention. I have only experienced two such double-digit gains of efficiency in my 17 years in the broadcast industry. One was the advent of the multi-collector inductive output tube that catapulted efficiency percentages from the 30s to the 50s. The other was the rise of predistortion techniques that enabled the transition from inefficient class-A solid-state amplifiers at 10% to 15% efficiency, to class-AB LDMOS technology that achieved 30% efficiency. The AMO’s 70% efficiency claim has certainly got my attention, and I’ll be sure to keep you informed as the technology develops and as Eta Devices scales it for use in the handset market, which holds the promise of doubling battery life.

Please share your thoughts on this technology in the comments section below.