Why Do Amps Sound Different?

[Negotiableterms]

I usually don't care for Audioholics, much, as I find them a little too smug about their conclusions. Nevertheless, the current edition has a fascinating article on amps, which might start to bridge the gap between objectivists and subjectivists in the 'All amps sound alike' controversy:

http://www.audioholics.com/audio-amplifier/audio-amplifiers-sound

Comments, thoughts, agreement or disagreement welcome, as long as it stays civil, and conforms to the special rules of this forum.

 

[Pio1980]

Interesting article. Brings to mind Daniel von Recklinghausen at Scott and his statement about sound and measurement.

Sent from my SM-N900V using Tapatalk

 

[Ths364]

Provides some insight "behind the scenes" in manufacturing practice also.

 

[skippy_ps]

Some interesting info there.

Murray

 

[Pio1980]

I've been over at DIY reading some stuff on ADCOM amps and modifications. Interesting comments there on amp performance and design. David Hafler and the null test also illuminates the topic. I'm more interested in the why of how they sound different then the argument that they do or do not sound different, which is imo argument for arguments' sake.

Sent from my SM-N900V using Tapatalk

 

[restorer-john]

The square wave has always been the way to sort out amplifiers faster than any other test. All hifi magazines used to publish oscillograms of square waves, back when people took an interest in actual measurable performance.

It's really simple for me, I won't even listen to an amplifier that cannot reproduce an accurate square wave as it simply is not a high fidelity amplification device if it cannot pass that simple test.

HK always championed the use of the square wave back in the day.

 

[Pio1980]

How do you load an amp to test square wave?
I've used 20 Watt wire wounds for dummy loads.

Sent from my SM-N900V using Tapatalk

 

[Brian]

I must have missed the memo telling engineers to not do square wave testing. Back in the cave man days it was one of the first things we did. It told a lot about what to expect if enough frequencies were checked. Many amps never made it past the test and for those that did we had a database of other tested amps with similar results to start testing against. I can not imagine testing any amp without that test. These tests were not definitive as they did not show results of using different caps, wires, etc.

A result of this long ago was what made realize I preferred amps with capacitor coupled outputs over direct coupled outputs. It helped in identifying potential amps but not the why I preferred it. Transformer/autoformer coupled outputs had output results similar to the capacitor coupled and these I liked best.

 

[BurnBabyBurn]

Back in the 1980's I remember hearing about the Bob Carver T amp story. Here's a sample of the story:

http://www.hifianswers.com/2014/11/amp-modelling-bob-carver-transfer-html/

 

[restorer-john]

Non inductive wirewounds still cause problems at high frequencies with square waves. I use a toaster element I unwound and got a 4ohm and 8ohm straight piece on a piece of plate glass.

It seems to work well enough, but I don't do square wave testing at much more than 1/3 power- you can hear the output transistors 'singing' on the heatsinks on some amps.

Even the square wave response at 1 watt is enough to tell you what the amp is doing.

How do you load an amp to test square wave?
I've used 20 Watt wire wounds for dummy loads.

Sent from my SM-N900V using Tapatalk

 

[BMWCCA]

Interesting read.

I own well-over a dozen amps and I know they don't all sound the same. Heck, I own eight different models of Crown amps and they don't even sound the same. Can I tell the difference between my Crown PS400 and my Soundcraftsmen Pro-Power-Four? Yes. That's why I use the Crown in my main system. Some amps sound better on some speakers and others sound better with others. That's why we play this game, isn't it?

This is a lot like saying all tires are the same. Round and black.

 

[gslikker]

Being no professional, the story does not tell me a thing.
Always thought that if leading edge of the signal different than the trailing edge, slewing and/or overshoot/signal actuation behavior is apparently different up or down indicating non-symmetrical behavior.

 

[Negotiableterms]

Years ago, when I was in college, I took an EE course. Now, I've forgotten most of it, so I'm hoping someone will correct me if the following recollection is wrong.

IIRC, the instructor pointed out that if all you're measuring is the usual THD and IMD, an amp that has a perfect square wave output, except that at some relevant frequency, the top of each square is a straight line 1% lower at the front will measure identically to another amp that is a straight line 1% lower at the back of each square. They both have the same amount of distortion, and might sound identical with some frequencies of sine wave, but will sound quite different when subjected to a real music signal.

Is this correct? And if so, why all the controversy about all amps sounding identical?

 

[davidb1]

The square wave has always been the way to sort out amplifiers faster than any other test. All hifi magazines used to publish oscillograms of square waves, back when people took an interest in actual measurable performance.

It's really simple for me, I won't even listen to an amplifier that cannot reproduce an accurate square wave as it simply is not a high fidelity amplification device if it cannot pass that simple test.

HK always championed the use of the square wave back in the day.

Along with the impulse test, the most important graph you could look at.
Remember the Mac clinics? My Citation 12 was the first the tech had tested. He was quite floored.

 

[mhardy6647]

Stereophile (well, John A) is pretty good about testing an amplifier objectively...

zum Beispeile:

http://www.stereophile.com/content/pass-labs-xa605-monoblock-power-amplifier-measurements

EDIT: on-topic... for an arbitrary batch of (power) amplifiers of the same basic topology (e.g., Class A driver to push-pull Class AB outputs, with or without local or global NFB), I suspect that, in the real world, the single biggest impact on the way the amplifier sounds - with any given loudspeaker - is going to be the amp's output impedance (often described, inversely, at a single frequency, as damping factor).

 

[ConradH]

I test most everything with square waves, including RIAA networks with an inverse network, but you have to be aware of what you're doing. The edge of a decent square wave will have harmonics way up into the RF. Some circuits don't behave well when you slam RF into them, so if the behavior is unexpected you might have to filter the input signal slightly. Some old Interstate generators had adjustable rise and fall time which was very handy. The output networks on power amps can also have a dramatic effect if you load them with capacitors. Sure, I use resistive loads for power testing like most other people, but if you're looking for SQ differences you should be using simulated speaker loads or even speakers.

 

[SoCal Sam]

Science has a place at the table.

 

[SoCal Sam]

A company's signature sound CAN BE tuned in using SW's.

 

[roger2]

A company's signature sound CAN BE tuned in using SW's.

Is that how Bob Carver did it for those Stereophile tests?

 

[Negotiableterms]

EDIT: on-topic... for an arbitrary batch of (power) amplifiers of the same basic topology (e.g., Class A driver to push-pull Class AB outputs, with or without local or global NFB), I suspect that, in the real world, the single biggest impact on the way the amplifier sounds - with any given loudspeaker - is going to be the amp's output impedance (often described, inversely, at a single frequency, as damping factor).

Amps don't really have a single impedance, do they? They have a curve with varying impedance over frequency. If the amp's impedance varies enough, and so does the speaker's, it seems shocking that any two non-identical amps would sound the same. Or I'm confused again...