Design Notes No.5.
Power rating, Safe Operating Area etc.
As mentioned elsewhere, one of the major problems of high power amplifier design is the variation in the possible load impedance. If we only had to worry about 8 ohm resistive loads life would be much easier, but in reality the impedance may fall to 3 ohms or less at some frequencies and be highly reactive also. If the impedance falls to low values at frequencies where the full output power is not required then this may not matter too much. Electrostatic speakers have this property, and may fall to 2 ohms or less at 20 kHz. The input signal voltage swing at this frequency is unlikely to exceed a tenth of the maximum, so in practice this need not be a serious problem.
At lower frequencies with conventional speakers a minimum impedance of 3 ohms seems a reasonable assumption. A few speaker designs do exist with lower impedances but my own feeling is that these are not sufficiently common to be worth worrying about.
In what follows I use the term 'single ended' to refer to amplifiers driving one end of the load. The term is now commonly used for non-push-pull output stages.
The problem with single ended amplifiers is not simply the need for double the power rating in the output power transistors, but worse than this the fact that this power dissipation takes place at higher collector voltage, where the second breakdown effect may limit the power rating. The choice of output device is more limited, and it may be necesary to resort to parallel pairs of devices. The use of a bridge amplifier is one good way to avoid these problems.
The following diagram shows the load lines for the output devices in a bridge amplifier of the type shown in Design Note No.3. The different curves are load lines for different values of the reactive component of the impedance. The maximum levels of these curves form a straight line representing the maximum current possible at every value of collector voltage. The peak power dissipation of a little over 110 watts occurs at a collector voltage around 22 volts, and there are plenty of transistors available able to cope with this power level at this voltage.
At 32 volts the peak dissipation is still 80 watts, so it is necessary to ensure that the power rating has not fallen too far at this voltage. For a single ended amplifier the load line diagram would be almost identical but the voltage scale would be doubled, so that a power about 220 watts is dissipated at 44 volts, and 160 watts at 64 volts. It is not simply a matter of using two parallel 110 watt transistors of the type used in the bridge version to achieve the 220 watts dissipation because the dissipation occurs at higher voltage where the 'second breakdown' effect is important. It is necessary to use more expensive types with better high voltage power ratings. The apparent saving in complexity of the single ended design may not after all be such a clear advantage.
Return to Index