Audio power amplifier frequency compensation: Is PLIL an Alternative to the Miller Capacitor?

摘要

The transistor audio amplifier can be generalised as having an input stage, which may be a single transistor or differential pair, a voltage amplifying stage (VAS), a driver pair and an output pair as shown in Fig. 1. As there are more than two time constants encompassed by the feedback loop, negative feedback can cause instability that requires frequency compensation to put right. A common method to stabilise amplifiers is to use a "Miller" capacitor Ccomp, as this causes a monotonic roll-off at 6dB/octave, which is, as Nyquist, Bode and others have shown, stable. Many audio amplifiers, ranging from the decidedly mediocre (typically from the 60's and 70's) to good ones from the 90's, use the Miller capacitor. An inherent problem with the Miller capacitor is that it delays the output voltage response and thus the negative feedback to the input stage. If the input signal is fast enough, this will cause transient distortion due to overloading in the input stage, until such time as the feedback "catches up" with it. The problem can be eliminated by using resistors in series with the input transistor emitters which are large enough to prevent overloading in the input stage, or using moderate resistors and increasing the current in the input stage to achieve the same voltage margin, such that the input transistors do not cut off for any input within normal limits. This approach to "fast slew" was exploited by Stochino. I hasten to add that for ref. 1, transient distortion from slewing does not arise for normal audio-band signals usually taken to be 20kHz maximum, as it has a good margin to at least to 150kHz. Nevertheless, the potential for input stage overload is not desirable, and the Miller capacitor method of compensation is not one which I would use by preference. However, the search for an alternative is not easy but phase lead compensation with input lag (PLIL) seems to be a possibility. This article reports on the investigations I have undertaken in the PLIL approach.