Class-G Topology

Class-G Topology
A very common arrangement used in many commercial amps using Class-G is to operate the power transistors in series, as shown in Figure 2 (note the drawing shows only one polarity, and highly simplified. Figure 3 shows a similarly simplified schematic of a Class-G amp, but using both polarities. The two transistor sets will be referred to as the inner pair - those connected directly to the output and powered via diodes from a low voltage supply, and the outer pair - those connected to the high voltage rails.

This arrangement is popular because it's relatively simple to achieve, and if done properly gives very good results. The remainder of this article will concentrate on this topology, although there are others that can also be used. These will be discussed later, but not in great depth.

Figure 3 - Basic Principle of Class-G Amplifier

The output stage above will be used for all further analysis of the Class-G output stage. The front-end and VAS (voltage amplifier stage) are virtually identical to any normal Class-B amplifier. The VAS would normally be used in place of one of the current sources shown. The arrangement above is convenient for analysis because it is quite straightforward.

Because the inner transistors (Q2 and Q5) are supplied with ±35V, both inner output transistors and drivers must be rated for at least 105V breakdown voltage. The voltage across each will vary by the full inner supply voltage plus the difference between the inner and outer supplies of each polarity. As the output swings between positive and negative, the inner transistors will therefore get a maximum voltage of 70 + 35 = 105V, however prudence suggests that a higher rating is preferable, and ideally one would use transistors rated for the full supply voltage. Once the signal calls for a voltage exceeding 35V of either polarity, the outer transistors (Q6 and Q8) boost the supply voltage in the direction required, allowing the output voltage to swing by almost the full ±70V. Even a small miscalculation (in design or implementation) may cause large amounts of magic smoke to escape from expensive devices, and a great number of rude words to be uttered.

The voltage across the outer transistors can (in theory) never exceed 35V, so low voltage, high power transistors may be used. If higher voltage devices are used, their SOA (safe operating area) should be very good - depending on the devices chosen of course. A good SOA is necessary, because by the time Q6 (for example) turns on, it will be expected to supply 8.75A with the full 35V across it. This is an instantaneous power of 306W, far more than any one transistor can withstand without failure. At the point where Q6 turns on, Q2 (the inner transistor) is turned on fully, and remains so until Q6 turns off again. Consequently, the dissipation in Q2 and Q4 will remain quite low whenever the output voltage is greater than 35V in either direction.

With the supply voltages shown, the diodes providing the ±35V supplies must be rated for a continuous average current of at least 5A, preferably more. The voltage rating is not a problem, since the maximum reverse voltage is 35V with the supplies shown.

 

http://sound.westhost.com/articles/class-g.htm