Larissa Marple from Virginia Tech has submitted this month’s question:
Concerning power amplifiers, what is the highest PAE recorded and what design type achieved that efficiency?
From: Tharaka Perera, MillenniumIT
Dear Larissa,
When it comes to power amplifiers "The Sensors Directorate" came up with a monolithic, broadband, high-efficiency power amplifier that was based on indium phosphide (InP) double heterojunction bipolar transistor (DHBT) technology. This power amplifier employs the Class E design methodology to give a peak power-added efficiency (PAE) of 65 percent at 10 GHz with a 20 percent bandwidth. This is the highest reported PAE performance and the widest bandwidth for a Class E design at X-band.
From: Uditha Jayarathna, MillenniumESP
Dear Larissa,
To my knowledge, microwave power amplifiers that have wide bandgap semiconductors such as SiC and GAN achieve 65 to 70 percent power-added efficiency.
From: Suranga Perera, MillenniumESP
Dear Larissa,
The aerospace industry has paved the way for power amplifiers offering a PAE as high as 65 percent.
From: Nick Chopra, UCLA/Northrop Grumman Space Technology
Dear Larissa,
The class of amplifiers known as the 'switching-mode amplifiers' are able to attain 100 percent efficiency - in theory. One of the most common is termed the Class-F amplifier. How is it able to attain 100 percent efficiency? Let's consider an example. At the output of a transistor, say a FET, the ideal Class-F current waveform is a half-sinusoid and the resulting voltage waveform is a half-square wave which is 180 degrees out of phase with the current waveform. Imagine these two current and voltage waveforms overlaid on top of each other and then multiply them. You will see that the product of these two waveforms is zero - zero power dissipation and 100 percent efficiency! However, in practice this is not the case. Why? Here's one of several reasons for this non-ideality: In order to attain a true switched-mode amplifier an ideal switch is required that switches at the desired RF rate. As we move higher and higher in frequencies, our real-world transistors become less and less 'ideal' - due to parasitic effects. Therefore, there is always some physical limit to the maximum attainable efficiency for a given transistor. The trend in Class-F amplifiers demonstrates this point clearly; at the lower frequencies, less than 1 GHz, 90 percent max possible efficiencies have been reported. At 10 GHz, PAEs no greater than 80 percent efficiencies have been reported. This is a complex topic and is an area of intense research. Here are some useful keywords to aid in your research: Class-F, Class-E, switched-mode amplifier, harmonic terminations and harmonic load-pull. Here are two pioneering authors on the switched mode amplifier: Nathan O. Sokal and Frederick H. Raab. And, of course, also recommended are the two editions of RF Power Amplifiers for Wireless Communications by Steve C. Cripps (note that the second edition has a dedicated chapter on class-F amplifiers).
From: Bhavin Shah, Hittite Microwave Corp.
Dear Larissa,
In power amplifiers, the lower biased amplifiers (class B, class C) will provide high PAE at the cost of linearity. The Doherty amplifier is a classic high efficiency power amplifier topology. There are papers that show 45 percent PAE can be achieved at 6 dB backoff from the max output power using the Doherty amplifier topology.
From: Robert Kim, Newgen Telecom
Dear Larissa,
To compare any design method or spec items like PAE, you need to consider other various factors such as operating voltage, operating frequency, modulation type, cost, size, etc., or the comparison may be meaningless because you are talking about two totally different things. If we focused on PAE by ignoring other factors, however, it seems to be 92 percent with Po=+23 dBm at 3.25 GHz using a pHEMT device with class-E topology, which I read in the April 2004 issue of RF Design magazine entitled "Broadband Monolithic S-band Class-E Power Amplifier Design" by Reza Tayrani from Raytheon. Except for Reza's work, there have only been a couple of other works involving X-band frequencies that I know of, and they demonstrated relatively poor efficiency, which was around 60 percent or something, as compared to the expected theoretical 100 percent efficiency with class-E. As for class-E topology, it was known that Sokal first presented this technique in 1975, and demonstrated 96 percent PAE at VHF band, but I would put my bet on Reza's work considering its operating frequency and impressed high PAE of 92 percent. When it comes to amplifier efficiency, the key is to find a method to minimize the power consumption in the amplifier. There are a couple of different methods to achieve these goals such as class-D, class-E/F, class-G and class-H, but I think class-E is the most promising technique to achieve high efficiencies especially in microwave frequency applications. Of course, it also has drawbacks, the most critical one of which is its very poor linearity performance because the active device is working as a switch, and you need other means to compensate in order to apply digital modulation schemes, which requires high linearity mostly, and such an addition will degrade the efficiency as a penalty. Other linearization techniques might be available but you would trade off size, cost and complexity again as it used to be. One interesting thing I found with the class-E technique is that its operational principle resembles that of 'Flyback AC-DC power supply' in power engineering, although I am not sure which one came first in the industry. If you are interested in this, type 'flyback' on the Internet for more information.
