Overcurrent Protection
A fundamental problem with any kind of solid-state voltage regulator layout, such as that of Figure 5.3(a), is that if the output is short-circuited, the only limit to the current that can flow is the capacity of the input power supply, which could well be high enough to destroy the pass transistor (Q1). For such a circuit to be usable in the real world, where HT rail short-circuits can, and will, occur, some sort of overcurrent protection must be provided. In the case of Figure 5.3(c), this is done by putting a resistor (R2) in series with the regulator output and then arranging a further transistor (Q2) to monitor the voltage across this. If the output current demand is enough to develop a voltage greater than about 0.65 V across R2, Q2 will conduct and will progressively steal the base current from Q1.
In the inverted stabilizer circuit shown in Figure 5.4, R1 monitors the output current, and if this is large enough to cause Q1 to conduct, then the output voltage will progressively collapse, causing the PSU to behave as a constant current source at whatever output voltage causes the load to draw the current determined by R1. (I know this protection technique works because this is the circuit I designed for my workshop bench power supply 20 years ago,1 which has been in use every working day since then, having endured countless inadvertent output short-circuits during normal use, as well as surviving my son having left it on overnight, at maximum current output, connected to
a nickel-plating bath that he had hooked up, but which had inadvertently become short- circuited.) In the particular layout shown, the characteristics of the pass transistors used (Q3 and its opposite number) are such that no current/voltage combinations that can be applied will cause Q3 to exceed its safe operating area boundaries, but this is an aspect that must be borne in mind. Although I use this supply for the initial testing of nearly all my amplifier designs, it would not have an acceptable performance, for reasons given earlier, as the power supply for the output stage of a modern hi-fi amplifier.
However, there is no such demand for a completely unlimited supply current for voltage amplifier stages or preamplifier supply rails, and in these positions, a high-quality regulator circuit can be of considerable value in avoiding potential problems due to hum and distortion components breaking through from the PSU rails. Indeed, there is a trend in modern amplifier design to divide the power supplies to the amplifier into several separate groupings: one pair for the gain stages, a second pair for the output driver transistors, and a final pair of unregulated supplies to drive the output transistors themselves. Only this last pair of supplies normally needs to be fed directly from a simple high current rectifier/reservoir capacitor type of DC supply system.
A further possibility that arises from the availability of more than one power supply to the power amplifier is that it allows the designer, by the choice of the individual supply voltages provided, to determine whereabouts in the power amplifier the circuit will overload when driven too hard since, in general, it is better if it is not the output stage that clips. This was an option that I took advantage of in my 80-W power MOSFET design of 1984.2