Is zero input impedance good or possible when considering impedance matching? For RF applications, we are usually considering of impedance matching to choose load impedance as the conjugate of the source input impedance for the maximum transferred power.
But how about design the source having zero input impedance? Is it even possible or good? A naive thought is that all the power will be on the load if so. I might think it is true for DC case and for amplifier design to have a small output impedance and big input impedance. I lack the knowledge to find the missed chain to put all the knowledge together.
How do you think? Thanks!
 A: The aim of impedance matching is (usually) to achieve the maximum transfer of power to the load. This is often found to be not an intuitive solution.
A simplistic view is that in a closed system with a fixed voltage and a fixed total impedance, the power is dissipated/radiated/applied in both the supply and the load. 
There is a limiting factor in how much current the supply can deliver, based on its internal impedance. If the load impedance is zero then all of the power is dissipated in the supply because P=I2R and the only R is in the supply.
As the load impedance is increased, the total current will drop but the power dissipated at the load increases, because V stays constant and so I decreases (inversely) proportionally as R increases. If you were to graph this it would show the increasing power at the load, and what happens as the impedance of the load becomes higher than the impedance of the supply, and that the "sweet spot" is when the load impedance matches the supply impedance.


In this, it looks like having zero impedance in the source would mean all the power is delivered to the load. In a DC circuit nobody cares much (joke) :-)  but otherwise what happens is that only the load dissipates power, but not all the power will be transferred to the load. 
The reflection coefficient also needs to be considered, which is calculated by Z2 - Z1 / (Z2 + Z1). When Z2 = Z1 then the reflection coefficient is zero and all the power is transferred. You could also take the transmission coefficient as 1 - reflection coefficient. But when either the source or the load is zero then all the power is reflected and doesn't get dissipated in the load. 
And not to forget, but only the resistive component of the impedance dissipates power (not the reactive component) but the reactive component contributes to the impedance and is frequency-dependent.
** edited to expand answer
