PART IV Power Amplifiers expected to produce, but the performance is repeatable, predictable, and definitely does not require large global feedback factors. Distortion 1 cannot be totally eradicated, but its onset can be pushed well above 20 kHz by the use of local feedback. Distortion 2 (VAS distortion) can be similarly suppressed by cascoding or beta- enhancement, and Distortions 4­7 can be made negligible by simple topological methods. All these measures will be detailed later. This leaves Distortion 3, which includes the intractable Class-B problems, i.e. crossover distortion (Distortion 3b) and HF switch-off difficulties (Distortion 3c). Minimizing 3b requires a Blameless amplifier to use a BJT output rather than FETs. A Blameless Class-B amplifier essentially shows crossover distortion only, so long as the load is no heavier than 8 ; this distortion increases with frequency as the amount of global NFB falls. At 4 loading an extra distortion mechanism (3a) generates significant third harmonic. The importance of the Blameless concept is that it represents the best distortion performance obtain- able from straightforward Class-B. This performance is stable and repeatable, and varies little with tran- sistor type as it is not sensitive to variable quantities such as beta. Blamelessness is a condition that can be defined with precision, and is therefore a standard other amplifiers can be judged against. A Blameless design represents a stable point of departure for more radical designs. This may be the most important use of the idea. REFERENCES Ball, G. , 1990. Distorting power supplies . Electronics & Wireless World , 1084 . Cherry, E. , 1981 May. A new distortion mechanism in Class-B amplifiers . JAES 6 , 327 . Feucht, D. , 1990 . Handbook of Analog Circuit Design . Academic Press , p. 256 (pole-splitting) . Oliver, B. Distortion in complementary-pair Class-B amplifiers . Hewlett-Packard Journal , 1971 February , 11 . 306