I've never seen a chart, but that would be handy.
This Ohm's Law Calculator will teach you a lot. (the formula is there, too)
Remember though that will be peak power, not RMS. Divide your rail voltage by the square root of 2 (1.41) to get the RMS voltage that those rails can supply. For example:
50V rail/8 ohm load = 312 watts peak. But RMS value is lower. 50V/1.41 =35.46V or 157 watts RMS into the 8 ohm load.
You'll have to add a few extra volts to the rails to make up for losses in the output devices. And that should get you close enough. It's all Ohm's law.
Welcome to the forum!
That is largely correct if you have positive and negative voltage rails. If it's single ended, divide the voltage by two.
In addition, realize that at maximum power the supply voltage will sag after a few milliseconds, so you must decide whether you want to know the maximum you can get in a pulse, or steady state. The filtering matters too, as the dips in the mains cycles will be visible, more or less, in the power supply output when loaded. You need to use the lowest momentary voltage in your calculations. To make it more complicated, the resistance of the transformer winding changes with temperature and so a hot transformer delivers lower voltage than a cold one. Further, the mains voltage must be held constant during these tests.
As an example of the overhead required for rail voltages, I'll cite the amplifier I'm designing at present.
I was looking for high-ish voltage to use as a class A unit possibly for a current dumping project that could be amended fairly easily to class AB for higher power.
my spec. was that the class A should be usable to the maximum rail voltage I would need for the power amp and obviously be as high quality as possible.
However, all was going welll 'till I found the simulation showed distortion creeping in when the output voltage was about 80% of my rail voltage, i.e. with rails of +-52V, I was only getting output of +-41V at 0.003% THD. Below this, I was getting consistant 0.001% THD.
The THD climbed rapidly above this but the virtual scope showed no clipping or any other obvious sign that would account for it.
So it would appear that around 15% - 20% overhead is needed for good performance.
ok here is my contribution this is base on a proaudio amps with a regulated front end/driver stage which is 20% higher than the main rail voltage.
200W @ 8Ohms = 70VDC+/-
400W @ 8Ohms = 90VDC+/-
600W @ 8Ohms = 110VDC+/-
800W @ 8Ohms = low rail 65VDC+/- step 130VDC+/- class H
1000W @ 8Ohms = low rail 75VDC+/- step 150VDC+/- class H
Distortion near power rails could be caused by the VAS. Even with a CCS load it is still feeding more current into the output stage at peaks. Most amplifiers begin to distort when asked to get too near their supplies. There may be advantages in running the VAS from a slightly higher supply voltage than the output stage, although many designs do the opposite. Need to be careful about output latchup, though.