Let me end the Premp/Amplifier sound debate ...

To date, I've yet to see a tube amp measurement that does not look similar to the Black Line into the Stereophile simulated speaker load. Many speakers' impedance curves make the Simulated load look like a resistor.

Have you looked? As long as the speaker impedance does not go well below 4 Ohms, 15dB should be sufficient feedback to allow most tube amps to act as a voltage source within limits.

If the 15 dB of negative feedback was meant for discussion of the Melton MKT-88 tube amp

@toddalin I did not have the Melton in mind.

If a tube amplifier is able to act as a voltage source, how it does it is by cutting its power in half as impedance is doubled and not the other way 'round. This means that to get proper voltage response with some speakers and some speaker loads (aka Stereophile) you might have to use the 4 Ohm tap. You can see that this is one of the ways that tube power is more expensive than solid state power.

15dB is enough feedback that the output impedance of most transformer coupled amplifiers will be less than 0.1Ohm. Since most speakers do not need a damping factor of more than 10:1 (some need 20:1) you can see that this will work out just fine.

The problem IMO is that 15dB isn't enough feedback- it will add distortion of its own (mostly higher ordered, causing brightness and harshness) as a result. This is why feedback has a bad rap. But if you can get over the hump with enough feedback, then the circuit will be able to clean up those higher orders that feedback otherwise produces.

Have you looked?

For decades...

Most tube amp manufacturers these days seem to eschew lots of feedback.

FWIW, I've never measured a tube amp output voltage into a loudspeaker that was anywhere close to what an SS amp delivered into the same load... going back to the 70's. Admittedly, it's a small sample, and largely irrelevant as measurements mostly don't mean diddly. 😎

FWIW, I've never measured a tube amp output voltage into a loudspeaker that was anywhere close to what an SS amp delivered into the same load... going back to the 70's. Admittedly, it's a small sample, and largely irrelevant as measurements mostly don't mean diddly. 😎

Maybe I am simple fellow… but how can the sound (SPL) be the same, if the current (and voltage) are not the same?

FWIW, I've never measured a tube amp output voltage into a loudspeaker that was anywhere close to what an SS amp delivered into the same load... going back to the 70's. Admittedly, it's a small sample, and largely irrelevant as measurements mostly don't mean diddly. 😎

Its not that the measurements don't mean anything, its that the measurements have to be performed correctly, and the important measurements have to be made (the latter rarely happens, which has lead to the myth that we can hear things we can't measure...).

Maybe I am simple fellow… but how can the sound (SPL) be the same, if the current (and voltage) are not the same?

 

Its a good question. If a given amp is making 10 watts into a certain speaker at a certain frequency and has an output impedance of 0.01 Ohms, and another amp can make the same power at the same frequency and into the same load, while having an output impedance of 1 Ohm, since both amps are making the same power, the voltage and current will be the same.

The output impedance can affect FR and distortion but it won't affect output power since the example has that aspect being the same. If it really is the same the current and voltage has to be the same too.

Pool size and sample variation play a role here.

If you had happened upon Amir’s site in the early days after he had measured just two amps (or DACs) the different measurements would be very noticeable when presented on a graph graphic. After eight or so component tests that graph would have an obvious curve to it.

But, take a look at that graph now (after dozens and dozens of tests) and what do you see: that graph has flattened out. If he continues and eventually tests thousands of boxes that curve will continue to flatten to the point that it will look like a straight line. 
 

One camp could then say “they still all sound alike”. And the other camp could say “they all measure alike”. 
 

Does this all matter? Perhaps only when doing small sample size comparisons, like direct A:B tests at home, which is really the ony practical way of evaluating purchase options…