Why not Class-A?

There are Class-A headphone amps. Check out CEntrance.

Integrated amplifiers are usually class AB, which, if implemented properly, has low enough distortion that replacing it with class A does not make a practically useful difference. Also, at least in portable devices, power consumption does matter. Even if the amplifier is not portable, if the wasted power increases demands on the power supply and cooling, that adds more cost that could have been spent on things that make more difference to the audio quality.

Nothing! There's absolutely nothing wrong with class A amps and if i had it my way all amps should be class A due to the lack of distortion mostly. However in some smaller packages, class A is simply not feasible due to the heat issues linked to its 40ish percent efficiency. The price also comes into play as well as the power to size ratio that is less than that of a class D amp for example. 

Quote:

Originally Posted by Mauricio 

Yes, I know there are Class-A amps.  The question is, given the low power requirements of headphone amp, why would you want to go with anything other than Class-A?

Why would you want to go with Class-A? They had a real advantage once in that they were the simplest type of decent amp to construct, but modern electronic components have entirely changed what complex and simple mean. And because of the low power needs you refer to, it's easy to engineer an AB headphone amp that stays in its A region all the time.

Crossover distortion can be minimised in a decent class AB design: even cheap opamps that are suited to audio applications have negligable crossover distortion. When you talk about simplicity of signal path are you talking about single-ended amplifiers?

Think again about the power requirements for most headphones (well under a watt except in rare cases).  It's common to use just an op amp or dedicated headphone driver IC (which are cheap) for the headphone output, which is possible considering the low power required.  Even if the internals of those devices are more complicated than a relatively simple discrete class A output, the cost is lower for this kind of implementation, and you needn't bother with doing much more than putting that IC in the proper circuit.  It even takes less parts on the board than putting together a few discrete transistors, and with the right part and configuration (as pointed out before), there's no crossover distortion.  As mentioned earlier, with the lower power consumption, the demands on cooling and the power supply are lower, so the costs can be further reduced.

There's just not really a real disadvantage in cost, complexity, or performance.  A lot of popular desktop headphone amps run class A though.

Both my Little-Dot II+ and Little-Dot II++ (full tube headphone amps) are Class-A.  My Little-Dot I+ Hybrid amp states that "The vacuum tube stage is run in Class-A" presumably meaning that the transistor output stage is not.

Quote:

Originally Posted by Mauricio 

I am not sure I follow.  Sure, an op amp takes less room on the circuit board, but it takes more transistors in the signal path for the innards of an op-amp consists of an array of several tens of transistors, resistors and diodes.  So, once again, if you believe in the simplest design possible--and by simple, I do not mean the fewest components on the circuit board, but rather the fewest transistors along the signal path--you will tend to shy away from op-amps.  The simplest, least distortion-inducing design is a pure Class-A configuration consisting of discrete transistors, diodes and resistors, not a design based on a biased or unbiased AB configuration based on op-amps.

Okay, well you didn't follow because I didn't follow your question in the first place.  Indeed I was addressing simplicity in terms of design work and low cost, rather than the simplest actual equivalent circuit configuration, when you consider all of the things inside these ICs.  IMHO simple may or may not be better, depending.  I think it's better to be a believer in whatever works.

Have you seen the performance of various class A headphone amps and compared them to various class biased or unbiased class AB designs based on op amps or buffers or headphone amp ICs?  Regardless of the internal simplicity or lack thereof, with op amps costing under $1 in classical barebones configurations, you can get performance like 0.005% THD+N with typical headphone operating voltages at 16-600 ohms (with a better 0.001% or lower THD+N typical), with no crossover distortion artifacts.  To me this begins to beg the question of the necessity of class A, unless you are pursuing a philosophical ideal or some kind of circuit purity.

Quote:

Originally Posted by Mauricio 

Pardon my ignorance, but concretely what are these modern electronic components that have done away with the fundamental advantages of Class-A (i.e. no crossover distortion, simpler signal path)? 

These aren't advantages, they're features. An AB,  perhaps the commonest sort of modern amp, can eliminate crossover distortion by combining the results of stages. Electronic components are now cheap, so why go with a Class A design that burns more power and needs more heatsink? The rationale for class A's, eliminating cd at a low cost in components, is now obsolete.

a potential advantage of Class A circuits is that lower skilled designers can implement them with a few less concerns in layout, specifically ground return and power trace common impedance coupling need less attention since the "contamination" is a linear signal it only degrades crosstalk # at meaningless low levels - but doesn't give distortion

but the "simple circuits are better" is a philosophical/religious proposition - not supported by the "real world" -  Audiophile commentary doesn't count since they never use perceptual controls in listening "tests"

and some memes become outdated by technology advance - "simple discrete" as always better than the best monolithic is past its "use by" date in Audio

Quote:

Originally Posted by jcx

early 70's-80's monolithic op amps were compromised by not being able to use equally good npn and pnp transistors or jfets together

 

by the 90's complementary symmetry processes became available

 

in the late 90's thru to today there has been an explosion in demand for higher performance op amps and huge investment in research, design, and semiconductor manufacturing processes for high performance analog signal processing

 

early monolithic op amps were easily beaten in performance by better discrete op amp modules of the day - the JE990 is the classic discrete op amp for audio

 

now days individual designers using catalog discrete parts can't do that much better than the JE990 - but today's monolithic op amp designers have unprecedented control over transistor parameters, matching and use thin film high accuracy resistor on the chip - they have better modeling, design tools, experience and institutional knowledge than any lone wolf designer

 

and they have applied those advantages competing for market share in telecomm A/DSL drivers that are perfect for dynamic headphone amps

 

medical ultrasound demands low noise, distortion and high bandwidth too - modern monolithics far outperform any discrete op amp at audio frequencies for anything except moving coil phonograph cartridge preamps or high Voltage output

 

 

that doesn't mean you can easily buy the best of today's monolithic op amps in a great headphone amp - exactly because of the deep rooted market prejudices that cause most to dismiss "op amp" audio

 

another bad meme here is that amplifiers should be chosen to "tune" your sound - logically all "good enough" (for all errors to be inaudible) amps of any technology Should "sound" the same - series impedance can be added to SS amps to tune their response to more closely resemble tube amplifiers - but it doesn't increase the amp's technical accuracy

Quote:

Originally Posted by jcx

early A/DSL driver chips were designed with a goal of extreme low distortion into 12-100 Ohm loads at 100KHz-1MHz - kind of a good match for dynamic headphone drive

 

the 1st few generations of chips used high internal bias, including output Q bias so these amps can be "deep(A)" Class AB for dynamic headphone loads - and can be externally biased for full Class A output stage operation with many headphone loads

 

more recently the industry has concluded more modest distortion goals are acceptable and low power is a bigger selling point so recent A/DSL driver chips specs don't look as great for audio

 

the TPA6120 is a "re-badged" THS6012 - so reading both data sheets can be informative

 

as 100 MHz CFA op amps they do have some different requirements than most hobby project diy amp op amps - but at ppm distortion, >1000V/us slew rate, 100 MHz BW into 25 Ohm load its hard to argue that they do anything "wrong" to the sound