Want to build a headphone amp.

[Richard Crowley]

24-bits can record a dynamic range of 145dB. I can't think of ANY audio gear out in the Real World that has even that kind of dynamic range between its noise floor and full-scale clipping.  You might be able to approach that dynamic range under carefully controlled laboratory conditions.

32-bits can handle a dynamic range of 192dB. That is essentially the absolute limit of sound that can be produced in normal atmospheric pressure at sea-level. Of course no transducer (microphone or speaker) can handle anything remotely resembling that.  Selling 32-bit audio DACs to consumers is essentially the very definition of AUDIOPHOOL.

Most people who live in the Real World can't even appreciate the full 96dB dynamic range of 16-bit. Most of us live in a 60dB (11-bit) world between the quietest we can experience without travelling to the Deep Woods, vs. watching big jets taking off at the international airport.

32-bit audio is simply a manifestation of the "Bit-Depth Wars" and has no practical value in the Real World (unless you have a vivid imagination.)   

[Synthetase]

Haven't had time to read the whole thread, but to the OP I can say I've built this amp and it works very well (it's also NOT a class B amplifier, it's AB). Distortion is too low to measure with my equipment. It drives practically any load with ease and I use it to drive a pair of HD800s (although I modified its output impedance to 4R for this task).

However, it's not portable. If you're looking for portability, you could try P109 from the same site. It's easy to build on a piece of prototyping board and it's cheap.

[Fungus]

I suspect the 'amplifier' problem has been studied, the design compromises have been made and it's now available in a monolithic packages from a dozen different manufacturers. 

In a way, yes. Monolithic amplifiers can be very good and have excellent performance/price ratio. Discrete amplifier designs are still almost always crap that simply won't last (e.g. class AB output stages with Ube multipliers )

In commercial applications class D is extremely important, too, of course. Not just for low end applications, but also for high-end applications. Today you will often find that even ICs designed for low end markets incorporate not simply PWM modulators but actually quite good SDMs. Class D can surpass 0.005 % THD+N today (actually, ten years ago, but let's not nitpick on time lines).

I was wondering about that...

When I looked for amplifier ICs on TI's website there were no class A/AB chips for anything over about 5W. All their power amplifiers are class D. "Class D" is even in the category title for their power amp section ("Mid/High-Power Class D Amplifiers")

I haven't really been paying attention lately but I guess the world has moved on from when class D was only for "subwoofers and cars", and gone are the days when amplifiers were designed around their heatsinks.

Unless you're an audiophool, in which case heatsink design is alive and well:   



nb. I'm not knocking the sound quality of class A/AB amplifiers, it just seems like newer technology is making them obsolete.

[tooki]

When I was talking volume range, I meant rms volume of sine wave. You need more bits to represent sine waveform, don't you? So the total dynamic range should be volume range AND snr of the waveform itself.

Nope. Because it's a bandwidth-limited signal, no additional data is required to achieve perfect (in the literal sense) representation of the sine wave through sampling, nor does additional data add anything. More bits simply reduce noise, but for this, 16 bits is already more than enough. Did you actually watch the video? He demonstrates this very concept.

[dom0]

no additional data is required to achieve perfect (in the literal sense) representation of the sine wave through sampling

Actually any signal band limited to the sampling band.

I suspect the 'amplifier' problem has been studied, the design compromises have been made and it's now available in a monolithic packages from a dozen different manufacturers. 

In a way, yes. Monolithic amplifiers can be very good and have excellent performance/price ratio. Discrete amplifier designs are still almost always crap that simply won't last (e.g. class AB output stages with Ube multipliers )

In commercial applications class D is extremely important, too, of course. Not just for low end applications, but also for high-end applications. Today you will often find that even ICs designed for low end markets incorporate not simply PWM modulators but actually quite good SDMs. Class D can surpass 0.005 % THD+N today (actually, ten years ago, but let's not nitpick on time lines).

I was wondering about that...

When I looked for amplifier ICs on TI's website there were no class A/AB chips for anything over about 5W. All their power amplifiers are class D. "Class D" is even in the category title for their power amp section ("Mid/High-Power Class D Amplifiers")

I haven't really been paying attention lately but I guess the world has moved on from when class D was only for "subwoofers and cars", and gone are the days when amplifiers were designed around their heatsinks.

Unless you're an audiophool, in which case heatsink design is alive and well:   

[img]http://1.bp.blogspot.com/-4JF1O3qiPuw/T3HSCnB_P5I/AAAAAAAAB_o/mQqip-lN-5A/s1600/Pathos%2Badrenalin%2B3.jpg[/img ]

nb. I'm not knocking the sound quality of class A/AB amplifiers, it just seems like newer technology is making them obsolete.

Ah, their parametric search has a few bugs. TI still has a few AB power stages in their lineup, namely LM3886, LM3875, LM3856, LM4780 (=Two LM3886 dice packaged together). They also have a few HV input stages (LME49810, LME49811, LME49830, LM4702, ...), but I think they're starting to discontinue them.

I don't think AB is obsolete just yet, for two reasons. First, class D has become very, very good, but still can't quite reach the best linear power amplifiers. Second, audio PAs are quite often found in laboratories as cheap drivers for stuff - in a laboratory people might not like the additional EMI introduced by a class D amplifier -- although it's inside (inter)national regulations.

--

As I  mentioned above 24 bit DACs mostly make sense when you do digital volume control by scaling the PCM stream, which is a quite common thing to do.

[Fungus]

As I  mentioned above 24 bit DACs mostly make sense when you do digital volume control by scaling the PCM stream, which is a quite common thing to do.

Also for recording or any source where you might not have optimal levels set.

[Richard Crowley]

As I  mentioned above 24 bit DACs mostly make sense when you do digital volume control by scaling the PCM stream, which is a quite common thing to do.

Also for recording or any source where you might not have optimal levels set.

Yes, definitely.  I always use 24 bit when recording live events. Because it gives you the dynamic range to handle the kinds of unexpected peaks encountered especially from amateur performers. And especially when recording multi-track where you can't keep an eagle-eye on every microphone.

And in post-production editing, most DAW products use 32-bit arithmetic internally.  But creating a final delivery mix more than 16 bit is essentially pointless.  For that matter, MP3 is far more often the delivery format of choice.

[Lightages]

The OP wants to build a headphone amp and he has to wade through a flood arguments over digital bit depth and digital theory?

[retrolefty]

The OP wants to build a headphone amp and he has to wade through a flood arguments over digital bit depth and digital theory?

 About par for this site. It's part of the drama that draws us like bees to honey.

[Richard Crowley]

"Topic drift" is par for virtually ANY un-moderated online forum.  Quite often, the "drift" turns out to be more interesting (and informative) than the original question/topic.

[retrolefty]

"Topic drift" is par for virtually ANY un-moderated online forum.  Quite often, the "drift" turns out to be more interesting (and informative) than the original question/topic.

[Fungus]

The OP wants to build a headphone amp and he has to wade through a flood arguments over digital bit depth and digital theory?

The OP's question was answered on the first/second page. He can stop reading there if he wants to.

[AF6LJ]

"Topic drift" is par for virtually ANY un-moderated online forum.  Quite often, the "drift" turns out to be more interesting (and informative) than the original question/topic.

I totally agree; in the case of this thread the original topic was interesting, the thread managed to get better.
This whole bit depth / bit rate discussion has been going on for some time in the software defined radio community, especially since twelve bit DACs come up short in terms of performance.

[Lightages]

So "Fuck the OP and his post, we want to talk about something completely different!"? Sorry, but I don't get it. Why not start another thread if you want to talk about something completely different instead of hijacking? I thought this forum was better than that.

I guess this is one of my pet peeves, thread crapping and hijacking.

[Lightages]

jdraughn:

I would try building the first headphone amp you linked to and see how it goes. It is a good test to see how your construction and packaging techniques are. After that you can maybe improve or change the design, or try another. It is fun to experiment but don't worry about making it small until you have your final design that you like the most.

[Marco]

So "Fuck the OP and his post, we want to talk about something completely different!"? Sorry, but I don't get it. Why not start another thread if you want to talk about something completely different instead of hijacking?

This is off topic for this entire forum, yet you have the audacity to complain about it being off topic for the post

[Richard Crowley]

So "**** the OP and his post, we want to talk about something completely different!"? Sorry, but I don't get it.

Yes, apparently you don't get it. This is NOT "something completely different" it is an analysis of WHY people think they need special magic circuits for rather prosaic applications like accurately driving dynamic headphones.  And furthermore a comparison of what we have available today in terms of monolithic ICs vs. what they had to work with decades ago when those circuits were designed.

[Lightages]

Hi all.

I am looking to build a little headphone amp and am looking at building this one http://sound.westhost.com/project113.htm

Could I get some advice on whether that is a decent design or not? Also, I would like to make it as small as possible and I like to work with 0402 and 0603 size resistors and caps if possible.  Can anyone tell me what parts should be rated for some wattage? I am guessing that the 4.7 ohm resistors should be of higher wattage. How low can I go? I was thinking of running it with 3 or 4 Li-Ion batteries for the top rail and 3 or 4 Li-Ion batteries for the bottom rail to make it portable.

Any gotchas I should be aware of?  I have a pair of Sennheiser 555 headphones that I want to build this for. They sound great just running out the headphone jack on my phone or laptop, but from what I have read they should sound a bit better with some real power behind them.

I am not set on that schematic and am open to any other circuits.  Thanks

jdraughn:

What have you decided to build? I am curious.

[Zero999]

I've recently used a TDA2003 amplifier module to drive several 50R headsets (mono only). It works fine gives reasonably decent sound quality. I did have some problems with it having poor supply rejection but thas was fixed with an RC filter followed by a TIP121 Darlington transistor on the positive rail.

[dom0]

Very poor PSRR is normal with car audio ICs. Lots of people still don't get, that these ICs are unsuitable for use with transformer + filter supplies (and how bad these ICs actually are compared to even the stuff 20 years ago - it's 70s-90s car audio, maximum bang, minimum cost, quality doesn't matter).

[madires]

I've built a small amp based on the TDA1519B for my PC speakers and it includes a headphone output with a 120 Ohms series resistor. It's powered by the PC's 12V rail (LC filter) and runs quite well. As long as I don't turn the output to max the audio is decent, also for headphones.

The TDA2822M might be a good amp for headphones.

[Lightages]

Why, WHY, would you put a 120 ohm resistor in series? High impedance outputs are a guarantee that you will have frequency anomalies.

[dom0]

The response of headphones is optimized for ~1 ? drive impedance and/or 120 ? drive impedance.

[Lightages]

Where did this 120 ohm source requirement for headphones come from? Headphones range from maybe 16 ohms to 600 ohms. A 120 ohm source guarantees problems.

[Lightages]

OK, found the reason people want to use a 120 ohm source. From Wikipedia:

The 1996 IEC 61938 standard recommended an output impedance of 120 Ohms, but in practice this is rarely used and not recommended with modern headphones.

on this page:
https://en.wikipedia.org/wiki/Headphone_amplifier

Read the whole article section on output impedance.

So as I said, a 120 ohm source is detrimental. You want less than 2 ohms.