As I read over the various DIY power supply offerings, I see very little about what goes into designing a schematic for a power supply. I see catch phrases such as "shunt regulated", "C-R-C", "line-load regulation", and "fixed voltage regulator", but no explanations why the design is good or better than another.
Some power supply designs are better suited for certain circuits than others.
It all actually depends on WHAT the power supply must be feeding.
It also depends on what a designer feels is important or thinks others find important (noise, voltage stability, output impedance, temperature stability, impulse response, current capabilities, efficiency, size, heat dissipation, longevity, leakage currents).
It could also be a designer wants to do something different than others or made a design and subjectively likes it better for who knows what reason.
Are there practical applications where one design might be better than the other?
Yes there are, and it depends on WHAT type of circuit needs the power supply and how much immunity that circuit has to certain aspects of the power supply.
Some circuits draw (near) constant currents, others are very dynamic in current draw.
Some circuits have poor PSSR meaning noise/garbage/signals on the power supply line make it (attenuated) into the signal. A low noise and stable voltage is a must.
Some circuits have excellent PSSR and don't need a super duper power supply and won't 'improve' in a technical sense but will certainly improve 'subjectively' to some people even when nothing changed in an electrical sense in the output signal.
SMPS (those lightweight cheap and powerful switching power supplies) may have high leakage currents (due to the way they are built) and these currents and RF output MAY cause problems in some equipment when care isn't taken. This is not of relevance in linear power supplies.
If you need a fixed voltage you can use both. If you need a variable voltage (for experimental purposes) a variable voltage is more convenient.
It is easier to design a high performance power supply for low currents.
High current designs usually have (slightly) less specs.
I can only conclude that "low-current" refers to input and output stages, but I do see headphone amplifiers referred to as "low-current" too. If headphone amplifiers are considered low-current, why, then, does
AMB's σ22 and
Twisted Pear's Placid HD exist?
If you have a headphone amplifier that can only deliver a few hundred mW's you can do with a smaller power supply.
If it also must feed HE-6's at high SPL you will need more current.
Also have a look at these circuits:
sjostromaudio.com/pages/index.php/hifi-projects this guy knows what he is talking about.
People suggest the latter PSU's for "hold nothing back" builds, but why? I'm sure it isn't as simple as heat dissipation since the circuits themselves vary drastically as well.
Most DIY builders want nothing but the best in the interest of 'it can't hurt except the wallet'.
People that have basic knowledge ensure themselves that it will perform optimally.
More experienced/knowledgeable people tend to know what is minimally needed for optimal performance and simply design their power supply so it meets those demands with minimal effort/costs.
For DIY'ers it is fun to 'improve' that circuit by using 'better' power supplies.
In most cases there is no measurable gain, or no gain in blind tests but those who use the 'upgrade' will still benefit because their mind is eased that the performance is higher and was what they are after.
Reality is:
1: A good (sufficient) enough power supply is essential for all circuits.
2: Over-dimensioning won't hurt but there is no need to overdo it (feeding a 1W amplifier with a 100W power supply where 20W will be MORE than enough)
3: How well the PS quality should be in general depends on the PSSR of the audio circuit(s) used in.
4: There are physical boundaries that cannot be broken so a perfect power supply can not be made (no a battery isn't ideal either).
5: Each power supply has their strong and weak points and the designer should select what points are paramount. This should determine the choice of circuit.
So yes... power supply is important
up to a certain point.
Some, however, want the absolute best (and will always be looking for better and are never satisfied) and they will only be satisfied with super-regulated power supplies, whether it is actually needed or not or even complete overkill but can understand the thought.
From a low budget standpoint (this website) it is worth looking into what the circuit needs and comply to the minimal needs and simply use a power supply that exceeds the needs slightly. Might save a lot of money.
Some clearly overrate the improvement a power supply makes, subjectivity lurks here BIG time.
Unless one is familiar with how to test 'correctly' it is easy to hear improvements when they aren't there.
Most circuits are perfectly happy being fed from a simple fixed regulator, some (mostly simple) circuits need better than that and that circuit may benefit from superregs.
A lot of 'audiophile' circuits in the DIY audio scene have poor PSSR (=PSRR) by design and is why you see these circuits used there often.