A Theory on Headphones Amplifiers Power Supplies

[Deleted member 5035]

Here's one for Amir... In the following video, a reviewer posits that the power supply section of smaller headphones amplifiers can't quite keep up, thus affecting the final output. He suggests that this is why expensive, high end amplifiers have much beefier power supplies. I suggested he monitor the power rails of a smaller amp, with an oscilloscope, while playing loud, percussive music through low impedance, hard to drive headphones. He goes on to suggest that this is one reason that measurements aren't everything. This is an interesting theory I'd like to see put to the test. Here's the video...

 

[solderdude]

The power supply in all (analog fed) power amplifiers has always been modulated by current draw of speakers.
In general this power supply wasn't even regulated.

Besides... it's the output signal that counts and not what happens on the power rails as long as there is enough 'headroom'.
Of course, a continuous sine wave under full load will have a slightly different powerrails (this amp has regulated supplies) voltages with peak loads so there might be a very small difference of output power. (considering well designed regulated power supplies with enough decoupling)

In headphone amps it's usually the output devices/circuit that determines max output current.
One of the few exceptions are Geshelli amps, they are limited by the DCDC converters. On maximum load (low impedance) one can be sure the power supply rails is heavily modulated. Haven't heard people complain about this.

 

[wasnotwasnotwas]

Currawrong. His benchmark is Audio-Gd, enough said.

 

[technoian]

In the video you can actually see the markings on the power supply. It states; SMSL Ultra-Low-Noise Power Supply +/- 18V 667mA. I make that out to be a maximum of 24 Watts. The specs for this amp state "6 Watts into 16 Ohm,3 Watts into 32 Ohm Headphones " so the powers supply appears to be more than capable for the intended job.
It's so easy for subjective reviewers to come up with some theory as to why a product sounds like it does to them. Here at ASR we believe in testing equipment and not spreading untested theories.

 

[Mnyb]

This is misleading at best or a red herring or what you call it ? what counts is ultimately whats come out of the amp , if it does it loud enough and drive the load we are fine . Measurement's like this can be used during design or in search for a real problem .

Of course the power supply is going to be modulated in some way and it can never be completely gone even with the best of efforts.
But it fits the audiophile narrative that everything's matters and no solution is ever good enough , hence powers supply can never big or stiff enough in these peoples mindset .

 

[tvrgeek]

More significant is if there is dynamic current available at what is changing. In other words, proper layout and proper use of bypass caps of sufficient size and impedance. Transformer charges the main caps. main caps the board entry cap, board cap the bypass caps. Bypass caps the devices. This is how the parasitic impedances are mitigated. Scoping the rails may not be very informative. First of all, they don't have the resolution and second, where on the rail and where your ground lead is attached maters more. I second post #4, and #5 but do need to add a caveat. Sometimes we don't know what to measure. It is the current through the output that matters. How you get it is the engineers problem.

The board layout may not be the best to support star topology for both power and ground. The kinds of incredible performance, well below human hearing but measurable, it can matter. @ .01%, easy. @ .0001%, well you have to look at that left rear castor and phase of the moon. Arguably irrelevant, but measurable. Probably more significant in AB amplifiers than class A as there is hard switching going on. Don't know enough about class D as there are so many other factors involving switching cleanliness.

And of course, marketing matters. How else can they sell $35,000 DACS when we know a $100 DAC is a couple orders of magnitude better than human hearing and a $5 op-amp even better? Mac could not sell a $35,000 preamp with external supply that can almost meet the specs of my $200 (over priced) Schiit without the mystique of audiophile smoke and mirrors. Look at it as a way that those who have extracted more than their fair share from society can put some of it back as somewhere there are assemblers who have a job building the stuff.

 

[Deleted member 5035]

In the video you can actually see the markings on the power supply. It states; SMSL Ultra-Low-Noise Power Supply +/- 18V 667mA. I make that out to be a maximum of 24 Watts. The specs for this amp state "6 Watts into 16 Ohm,3 Watts into 32 Ohm Headphones " so the powers supply appears to be more than capable for the intended job.
It's so easy for subjective reviewers to come up with some theory as to why a product sounds like it does to them. Here at ASR we believe in testing equipment and not spreading untested theories.

I've been here for a couple of years, and have been involved in dozens of discussions, so when you way "we" in your closing statement, that includes me. I am a person of science, but as such, the one thing I know is that "untested theories" are the absolute basis of the scientific process. If it weren't, we'd all still believe that the earth is flat.

And, my education, and decades of experience are in electronics, including circuit design and testing (what's your background?). A power supply's rated output does not tell the whole story - not even close. How well that output is regulated, its slew rate when hit with a sudden load, etc. are also important parameters.

Again, I am an avid fan of Amir, and this website, and I've made many purchases based on information from this site, but how any audio component responds to sine wave based tests does not tell the whole story, no more than how the horsepower and gas mileage ratings tells the whole story of how a car drives. I wish Amir would add more tests, such as slew rate and output damping tests. Those determine how well an amplifier responds to transients. And that's just one example.

Just as you do, I suspect there is nothing to that reviewer's theory, but I think testing it would be interesting and informative. I enjoy conversations like this, until I start getting insulted, so please don't go there.

 

[Deleted member 5035]

Currawrong. His benchmark is Audio-Gd, enough said.

Good point, but he actually has several very high end components.

 

[Deleted member 5035]

This is misleading at best or a red herring or what you call it ? what counts is ultimately whats come out of the amp , if it does it loud enough and drive the load we are fine . Measurement's like this can be used during design or in search for a real problem .

Of course the power supply is going to be modulated in some way and it can never be completely gone even with the best of efforts.
But it fits the audiophile narrative that everything's matters and no solution is ever good enough , hence powers supply can never big or stiff enough in these peoples mindset .

Never enough - another good point. And yes, what comes out is what matters, but ASR does not test for sudden dynamic loads, and the power supply "stiffness" plays a factor in that. And that plays a factor in how well an amplifier handles transients, which are critical when reproducing music, as opposed to sine waves. Don't get me wrong. I'm a big fan of this website, but I do think there might be more to how an amplifier sounds than what the sine wave tests predict.

 

[wasnotwasnotwas]

Good point, but he actually has several very high end components.

Of course he can hear the difference between them all.

 

[Deleted member 5035]

Of course he can hear the difference between them all.

Well, of course he can! I get your point...

 

[RayDunzl]

And that plays a factor in how well an amplifier handles transients, which are critical when reproducing music, as opposed to sine waves.

I thought maybe so too.

Then I compared the "predicted" transient response calculated from a swept sine to the actual response to the ultimate transients - an impulse, and a step.

The prediction from the sine matched the actual transient signal response from the system to a degree that quenched my curiosity about that subject.

 

[Deleted member 5035]

I thought maybe so too.

Then I compared the "predicted" transient response calculated from a swept sine to the actual response to the ultimate transients - an impulse, and a step.

The prediction from the sine matched the actual transient signal response from the system to a degree that quenched my curiosity about that subject.

That's good to know. Thanks!

 

[RayDunzl]

That's good to know. Thanks!

Links to the experiment

https://www.audiosciencereview.com/forum/index.php?threads/impulse-response.1765/#post-44352

https://www.audiosciencereview.com/forum/index.php?threads/impulse-response.1765/#post-44440

 

[o7_brother]

I don't remember many cases of Currawong preferring a cheaper DAC or amp to a more expensive one.

Before I watch any of his videos, I already know that the expensive stuff he owns is guaranteed to have more "resolution", "smoothness", or soundstage than whatever cheaper device he's featuring.

As is typical of this style of "review" with no blind testing, there can never be a simple solution to a simple problem. Something can always be gained by spending more money on overly-complicated electronics.

 

[Deleted member 5035]

I don't remember many cases of Currawong preferring a cheaper DAC or amp to a more expensive one.

Before I watch any of his videos, I already know that the expensive stuff he owns is guaranteed to have more "resolution", "smoothness", or soundstage than whatever cheaper device he's featuring.

As is typical of this style of "review" with no blind testing, there can never be a simple solution to a simple problem. Something can always be gained by spending more money on overly-complicated electronics.

I've noticed the same thing with most reviewers who are fortunate enough to own the most expensive gear available.

 

[wasnotwasnotwas]

I don't remember many cases of Currawong preferring a cheaper DAC or amp to a more expensive one.

Before I watch any of his videos, I already know that the expensive stuff he owns is guaranteed to have more "resolution", "smoothness", or soundstage than whatever cheaper device he's featuring.

As is typical of this style of "review" with no blind testing, there can never be a simple solution to a simple problem. Something can always be gained by spending more money on overly-complicated electronics.

Bigger, heavier, shinier and more expensive is always better. Entirely agree with your point about no blind testing.

Now, I have zero technical knowledge about all things electro/ acoustic but I can recognise a straw man argument. Its another go at the "measurements cant tell the whole story" narrative. Not all things are measured, so he says that he hears (positive) differences with amps with bigger/better/shinier power components on transients in real music (as opposed to "simple" measured sine waves and sweeps).Therefore it must be the power components in the smaller/cheaper products that are deficient. Without ever demonstrating he actually hears said differences in controlled testing.

 

[egellings]

Feedback in the amplifier will pretty much determine the sound of the amplifier, so long as the supply has enough oomph to supply peak demands. It's just not hard to do. Most symmetrically designed amps have a good PSRR, so if the supply bounces a bit in tune with the music but meets demand at all times, it's nothing to worry about.

 

[restorer-john]

And yes, what comes out is what matters, but ASR does not test for sudden dynamic loads, and the power supply "stiffness" plays a factor in that. And that plays a factor in how well an amplifier handles transients, which are critical when reproducing music, as opposed to sine waves. Don't get me wrong. I'm a big fan of this website, but I do think there might be more to how an amplifier sounds than what the sine wave tests predict.

The CEA2006/490a test amir performs on amplifiers does that exact thing for a 1kHz burst over continuous. ie headroom.

What is way more illuminating however, is a transient overload and recovery test. An amplifier running at half rated power (-3dB) is hit with a +10dB toneburst, putting the amplifier 7dB into overload and then studying what happens to the waveform. That really sorts out the good amplifiers from the poor examples.

 

[technoian]

If Currawong were serious about his theory that the power supply section of smaller headphones amplifiers can't quite keep up, thus affecting the final output then he should have been able to test it then and there.
All he needed to do was compare the sound through the amp using his most sensitive headphones and then his most power hungry headphones. If the amps power supply were a limiting factor then the affect would be far more pronounced on the power hungry headphones. Simple.
Amongst my many headphone amps I own the tiny Topping L30. It can power my headphones to ear bleeding levels with no sign of compression in the sound. To me that's a good enough test of it's power supply.