There are two ways to avoid crossover distortion. One is to avoid crossover by running in Class A, either single-ended or push-pull. That means that the peak signal current cannot exceed the quiescent current (twice for p-p).
The other way is to use the correct quiescent current for Class B (or AB). This won't be perfect, but it can be quite good. The correct quiescent current depends on the output topology and could be anywhere between about 10mA and 250mA.
Some people prefer to move the crossover region away from the zero crossing, and instead accept a switching transition at intermediate signal levels, by increasing quiescent current significantly above the optimum Class B value but nowhere near as high as the optimum Class A value. I think this is sometimes known as "First watt" bias. It is theoretically wrong, but could be OK in practice provided there is sufficiently heavy negative feedback to smooth out the corner.
The other way is to use the correct quiescent current for Class B (or AB). This won't be perfect, but it can be quite good. The correct quiescent current depends on the output topology and could be anywhere between about 10mA and 250mA.
Some people prefer to move the crossover region away from the zero crossing, and instead accept a switching transition at intermediate signal levels, by increasing quiescent current significantly above the optimum Class B value but nowhere near as high as the optimum Class A value. I think this is sometimes known as "First watt" bias. It is theoretically wrong, but could be OK in practice provided there is sufficiently heavy negative feedback to smooth out the corner.
Last edited:
hm.... so to say, by using Class A, we can AVOID crossover distortion at the cost of double peak current. (at cost of heat wastage)
Or 'tune' to correct Iq, but according to my understanding, it is easy to get out of 'perfect' alignment cause by temperature or other factor. (If here use special biasing method, such as 'emitter feedback bias' circuit, is it able to provide best bias ?)
Or 'tune' to correct Iq, but according to my understanding, it is easy to get out of 'perfect' alignment cause by temperature or other factor. (If here use special biasing method, such as 'emitter feedback bias' circuit, is it able to provide best bias ?)
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.