In learning about Class D, it appears that after the linear signal is
destroyed and turned into a saw tooth wave form, the output filter?
tries to put the signal back together...doing an incredible job
considering the complexity of this task...but ultimately is unable to
restore micro details to recreate the original signal...which people
such as myself notice very quickly, while others either don’t notice or
enjoy the new coloration.
The opening statement here is false. Here's how class D works:
There are essentially 3 building blocks:
1) the triangle wave generator (sets the 'scan' frequency)
2) A comparitor, the compares the incoming audio signal to the triangle wave
3) the output section. This is the power transistors that switch on and off, and any circuitry needed to drive them
The sawtooth or triangle wave generator is running all the time. The comparitor is too- taking the audio signal and turning either on or off depending on the state of the triangle wave and the audio signal at any given time- the output is a series of pulses of varying width, which are used to switch the output transistors either on or off. This technique is called 'Sigma Delta' and is a means creating the output pulses of varying width so is also known as 'pulse width modulation'
The scan frequency is a function of the triangle wave generator. Some ICs can't make a good triangle wave at higher frequencies so quite often this can be the reason the scan frequency is limited. The other reason might be the output transistor's ability to switch (these days its pretty easy to find inexpensive parts that can switch at well over 1 or 2 MHz...).
There is usually a filter at the output of the amp to filter out the scan frequency. Once that is filtered out, all that is left is a much higher current version of the input signal.
The scan must be kept constant to maintain fidelity. The output devices can sometimes stay on longer than they are supposed to (typically they take longer to turn off than to turn on) so they can both be conducting at the same time. If this happens, you get a phenomenal called 'shoot-through current' which can heat up the outputs really fast, so sometimes additional circuitry is used to make one transistor wait until the other is off before it turns on. This wait time is called 'dead time'. The longer the dead time the higher the distortion. Dead time is usually needed at higher switching frequencies, so you can see that the need to go to higher frequencies to reduce distortion (and increase resolution) is hampered by the fact that more dead time might be required, which increases distortion. So you can see that the designer has to weigh options!
This is it in a nutshell. Some things are glossed over and others omitted entirely, but if you know what is written here then you have a pretty good idea of how they work.
