Bombaywalla, You mentioned important point - quality of power supply. Design seems straightforward with just transformer rectifier and capacitors but in reality it is pretty complex. Linear supply is unregulated and because of that requires a lot of capacitance to hold voltage steady. This capacitance is in series with the amplifier's output (circuit closes thru it) so anything that makes capacitor less than perfect plays role. ESR is important, being just series resistance and as such it gets lower when capacitors are in parallel. Same with inductance but then inductance of connections, as you pointed out, might add. Parallel film capacitor might help withe high frequency response or it can get things worse since it is placed across large inductance creating parallel resonance circuit that will ring. Rectifier can be hard or soft. Charging of capacitor happens in narrow spikes that starts at the bottom of the ripple and ends on the top of full wave rectified sinewave. At this peak point voltage across diode changes direction but diode is too slow to react (stores charge) for a moment conducting in opposite direction to snap back and stop conducting. Soft diodes react fast (small overshoot other direction) but snap back slow reducing that way amount of produced HF noise. Larger capacitance usually comes with larger inductance but also can cause startup problems, as you mentioned, and produce noise. Since capacitor is charged only from the bottom of the ripple to peak value, when capacitance is very large ripple is very small and charging time is very short. Such short spikes repeated 120Hz have very low average to rms ratio causing losses in transformer wire while high frequency content heats up the core. It requires transformer to be much bigger. In addition it is creating similar losses in power supply cables. Your amp might take average 1A but it can be in form of 40A spikes that cause voltage drop in power cable (and radiated noise). I mentioned all this to suggest that design of even simple thing like linear power supply, (not to mention amplifier) can be very complicated. Judging amplifier by power doubling might be oversimplifying. We look if amplifier doubles power to see if it can deliver current when speaker impedance drops, but tube amps don't do it (being power source vs voltage source) and we like their sound. In addition, for amp to double means to be perfect voltage source and that might require deeper feedback that can bring other problems. To me "doubling" describes only behavior in extreme situation that might never happen in real life. Your speaker has impedance dip above 100Hz while highest amplitudes are usually well below it, where impedance is higher.
