No doubt that "crossover distortion" in A/B amps is a problem.
How big a problem is open to debate. In a well designed power amp it tends to be a small problem that may or may not be easy to identify except by really serious measurements. One poster mentioned Doug Self, his book really goes into deep detail on this subject, although his conclusions may or may not be correct.
So-called "Pure" class A is the solution to the switching problem in ultimate terms. But the price is cost, excess power dissipation, heat, and size (you get 25% efficiency)with 75% of the input power to the amp going away as heat and nothing else.
I use what is called a "high bias" class A/B in my large Mosfet amp where it is running around 35 watts of class A power sitting around, so in practice you are listening to a 35 watt class A amp 90+% of the time, with "free" headroom.
There is *still* a gain non-linearity where this switching takes place. In a typical AB design this non-linearity is in the 1 watt region, where much listening takes place.You can argue that it would be virtually impossible to detect at the 35 watt level.
Tube amps have the same issues, and more once a transformer is added into the equation...
Switching power supplies are another matter entirely.
How big a problem is open to debate. In a well designed power amp it tends to be a small problem that may or may not be easy to identify except by really serious measurements. One poster mentioned Doug Self, his book really goes into deep detail on this subject, although his conclusions may or may not be correct.
So-called "Pure" class A is the solution to the switching problem in ultimate terms. But the price is cost, excess power dissipation, heat, and size (you get 25% efficiency)with 75% of the input power to the amp going away as heat and nothing else.
I use what is called a "high bias" class A/B in my large Mosfet amp where it is running around 35 watts of class A power sitting around, so in practice you are listening to a 35 watt class A amp 90+% of the time, with "free" headroom.
There is *still* a gain non-linearity where this switching takes place. In a typical AB design this non-linearity is in the 1 watt region, where much listening takes place.You can argue that it would be virtually impossible to detect at the 35 watt level.
Tube amps have the same issues, and more once a transformer is added into the equation...
Switching power supplies are another matter entirely.