Audio PA
3/11 2019
Recently running into Class A audio PA. I spend lots of time exploring how to increase the output power. There are many factors affected it , i thought at first.
It turns out i'm wrong, loss in details and ignore the basic. refer to that journey : From Crystal Set To TRF
The quiescent current and supply voltage defined the total available output power we can achieved.
Typical efficiency is 20% even as low as 10%. from first one and efficiency we know the maximum power it could provide
Don't even bother to trying other thing like change transformer etc. don't work. the maximum efficiency,50%. consideration the voltage swing available might just 10Vpeak as the VCC is 12 the maximum efficiency(not considering the resistance and efficiency of the transformer) might only 30~35%.
The transformer is just for impendance mach: which provide a high impendance to collector, thus achieved high gain.
Given a quiescent current, regardless whatever you do, then only thing you could change is improve the efficiency as near as 50%.
While changing transformer, the quiescent current could amplified by transformer thus driven speaker with more current, but at same time the voltage at primary is raised, thus the room for voltage swing is limited.
Considering thermal stability, emitter need a resistor, high current flow it raise the base bias voltage, further squash the voltage swing. The driven voltage should not reverse bias the b-e and b-c junction, this is another wall limit for voltage swing. The driven signal also could not pull all current from the transistor, this could happen as same time as reverse bias the b-e junction.
And another voltage swing limitation is the highest voltage at collector, while the current in transformer almost reach to zero, the voltage reached to maximum, at it just reach to this point , the output distorted. we should leave some room for that for better amplifier.
All these individual factors limit us choices, limit the voltage swing and driven level, then limit the efficiency we could achieved. And to get a non-distortion waveform, the maximum efficiency as low as 20~25%.
The Push Pull PA
With Push Pull PA, the puzzle seems resolved. The Most freedom is that output power not limited by the quiescent current anymore. The current pull by the transistor as needed only if the transistor could handle.
Quiescent current is critical, too much or too little leading to distortion.
The function of the output transformer is synthesized output waveform, and change impendance to get voltage gain.
If we don't provided the voltage Gain by the push-pull state itself, we could use transformer present proper impendance to transistor, then we could get power as much as we want only if the transistor could handle.
Large signal drive limited by the voltage swing available on collector, almost 2*Vcc. The output transformer could then change this voltage swing to load for getting power we want, don't worry the current while the transistor is powerful.
OTL Push Pull
No transformer.
Output voltage exceeding 2Vpeak is hard, around 250 mW, mainly because the T1 driven current is inefficient.
bootstrap or a constant source replace 2k resistor is better.
Output wing limited by 2 factor: the lowest voltage not possible lower than the T3 base, most likely 1 V to 2 V(considerate T3 different bias method ). The highest voltage not exceeding vcc- 3 V due to the 2 k resistor voltage drop. Thus the maximum output power is limited, around 250 mW for 8 Ohm load.
Another limited factor is the driven current provided to T1. cause it's totally provided by a pull up 2k resistor , T1 barely could not exceeding around 250 mA. this limited hardly the output swing to maximum ~2 Vpeak. use smaller pull up resitor help a little , it's not worth to do so cause it's required too much quiescent current of T3. We must keep T3 collector at around half VCC voltage.