Class E RF Amplifier Theory of Operation

The idea behind class E is to reduce or eliminate the effects the various capacitances within the MOSFET have on efficiency and operation at high frequencies. The major operational condition is that the MOSFET is only switched (turned on) when there is no voltage across the device. This eliminates switching losses, a major loss component of most RF amplifiers.

There are three capacitances at work within the MOSFET itself; the input capacitance, the output capacitance and the so-called "transfer" (drain to gate) capacitance. The effects of the capacitances within the MOSFET can be reduced by making the capacitances part of resonant circuits rather than "forcing" energy into and out of the capacitances, and by controlling the timing of the switching of the MOSFET such that the device is switched on only when the output capacitor is discharged. Let's look at the various elements.

The element we must consider first, as far as class E operation is concerned is the drain, or output capacitance. This capacitance exists from drain to source. In normal switching arrangements, this capacitance is simply charged and discharged by the MOSFET(s). However, as the frequency is increased, more and more current is required to quickly charge and discharge this MOSFET capacitance. If this current flows through the MOSFET, the MOSFET's internal resistance will dissipate power. The efficiency will drop dramatically as the frequency is increased. In class E, the output network values are chosen such that the output capacitance is part of a total resonant circuit. The capacitor is "charged" by the flyback effect of the tuned circuit.

The diagram below shows a basic class E RF output stage, and the drain and gate voltage waveforms when properly adjusted. The DC voltage applied to the drain in this example is 50Vdc. Notice the peak RF drain voltage rises to almost 200v.

The tuning and circuit values are chosen and adjusted such that the drain capacitance (and shunt capacitor connected from drain to ground) will fully discharge (drain voltage falls to zero) before the MOSFET is turned on. In this way, the MOSFET is only switched on (by the gate voltage) when there is already no voltage across the MOSFET, drain to source. When the MOSFET is switched on, it isn't actually "doing" anything at that moment, voltage-wise.

The gate, or "input" capacitance will prevent the MOSFET from being driven easily at high frequencies. This capacitance is very high in most MOSFETs - in some cases, in the order of thousands of picofarads for a single MOSFET. Values which we would consider to be a "short circuit" to RF in the vacuum tube world are commonplace operating values in the MOSFET world. There are several ways to deal with the input capacitance. One of the easiest ways to drive the gate is to use a very low impedance driver integrated circuit. These devices are available at very reasonable prices, and work very well. The IXYS IXDD614 is one such device.

All of the energy which is put into the gate is lost in the form of heat, caused by the charging and discharging of the gate capacitance, and by heating of the resistive component of the gate impedance. This resistance is internal to the MOSFET, and varies from device type to device type. MOSFETs with metal gate structures are much better in this respect, and are capable of operating at higher frequencies, as opposed to MOSFETs which use a silicon gate. However, the metal gate type MOSFETs are considerably more expensive, for a given device rating.

It is only necessary to drive the gate to about 12v (positive). The MOSFET will be fully saturated at this point. It is possible to "drive" the MOSFET with a square wave, however as the frequency is increased, the amount of power required to force a square wave into the gate capacitance becomes excessive. A trapezoidal waveform is generally the best compromise between good switching times and driver power.

The reverse-transfer (drain to gate) capacitance effects the ability of the MOSFET to be switched at high frequencies when high voltage is present at the drain. Ideally, you want to choose a MOSFET which has a low a reverse-transfer (also called the Miller capacitance). The reverse-transfer capacitance causes the drain voltage to "work against" the gate voltage. Improvements in technology and manufacturing techniques have dramatically reduced reverse transfer capacitances over the past few years. Be aware of this value, along with the related Gate Charge value when choosing MOSFETs for RF applications. The lower the gate charge, the better is the MOSFET for RF.