[goodsound]

I am very new to this but almost all diy headphone amps that I have come across use opamps (and is some cases opamps+buffers), even though chips designed specifically for the purpose of headphone amplification are available ? Is there a reason for this ? I'd imagine the audio chips would have been designed with sound quality in mind - for example there are many that are "output capacitor-less" or "filter-less"), unlike opamps which are designed to be used in a wide range of applications (control, measurement, etc...) and no matter how spectacular their specs may be, they might just not be as suitable for audio.

Here are chips from the big three chip makers that are designed to be used with headphones, so why isnt anyone using them ?

From Nat Semi:-
http://www.national.com/catalog/0,,117,00.html
(look under Headphone Amplifiers and Boomer Audio Power Amplifiers)

From TI/BB:-
http://focus.ti.com/paramsearch/docs...3000709|EQ|Yes
OR
http://focus.ti.com/paramsearch/docs...3000576|EQ|Yes

From AD:-
http://www.analog.com/en/subCat/0,28...0%255F,00.html
(look under Portable Audio)

 

[firefox360]

I'm guessing perhaps, costs, ease of implementation, and perhaps the avaliability of op-amps might make people use them more often. Thats my personal idea of why people use them. Could be some totally different reason as far as I know.

 

[goodsound]

could it be because almost all of them are SMD ?

the only other reason - and this is just my guess - is that many of these "amp" chips are built for applications like cellphones, handsets, portable electronics and other such small devices where fidelity is not really the top priority.

<edit:> the only chip that I found was a DIP was the LM4881 and is being used in the PA2V2.

 

[00940]

SMD is not the reason. Increasingly we are dealing with SMD opamps and parts.

It's just that, as you said, their specs are not equal to good opamps. And some recent opamps are damn good wrt to output current, such as the AD8397. The only real good headphone driver I know is the TPA6120A2 from TI and it's a bit hard to use (power pad package).

 

[rickcr42]

Mostly because the amp chips are not high fidelity but meant to be used in cell phones or low end portable audio and sound like it.

there is only so good something can actually sound and still run cool while not sucking all the power out of the batteries in short time.

Class of bias goes directly to idle current and how hot the part will get playing music.bias up higher to class a and the battery life sucks and you will have to allow space for the part to "breath" or it will burn up.
Bias closer to class B and sonics suffer but battery life is good and it will run cool

When something is designed to mount on a multifunction high-density portable gear pc board efficiency and low heat take precedence over unltimate sound quality no matter what the data sheet says to the contrary

 

[Garbz]

Opamps give us a choice. We can impliment them in a Huge variety of different ways. Thoes headhone amplifiers are an all in one implimentation often. We're stuck with however they've designed it.

On the fidelity side of things Opamps are preferred for that reason too. as mentioned above. That said there's plenty of amps on this forum made from TI's TPA6120 headphone driver. This is a very high-fidelity alternative. I have one for my portable use on the way too.

 

[Tomo]

Hi,

Just so you know. Audio Application is NOT a critical application. If it fails, the worst thing that can happen is a whining consumer. ...

It is not that they build lower-grade stuff just for audio. NO ONE can build chips with 100% success rate. A large quantities fail to meet requirements. But majority of those can still be used. Although they may not be used in Space Shuttles or Apache or Missile Guidance, they are fine for everyday-things. Note also you would be a fool not to sell those since the loss is in the order of millions of dollars. (or hundreds of unemployments)

However, don't neglect all "audio"-this or "audio"-that. Some of them can be really good for you. I think OPA134 is few dollars cheaper. And that can really add up.

Ciao,

T

P.S. I thought for a sec and realized some BAD people might push you with AUDIO quality NJM4667's. ... Those you are better off buying commercial-standard at full price.

 

[tangent]

You have to realize how a lot of engineers and marketing folk view "audio". It's that very lowest segment of the EM spectrum, the part that's supposedly the easiest to amplify. Problem is, human ears can be awfully picky. The problem with the problem is that most humans aren't so picky for one reason or another, which validates the position that audio is not a hard problem.

If you see a chip labelled as "for audio", there's a better than even chance that it's because the chip isn't good for much else but basic audio with tolerable levels of distortion.

 

[Ferbose]

Quote:

Originally Posted by tangent You have to realize how a lot of engineers and marketing folk view "audio". It's that very lowest segment of the EM spectrum, the part that's supposedly the easiest to amplify. Problem is, human ears can be awfully picky. The problem with the problem is that most humans aren't so picky for one reason or another, which validates the position that audio is not a hard problem. If you see a chip labelled as "for audio", there's a better than even chance that it's because the chip isn't good for much else but basic audio with tolerable levels of distortion.

I just looked at BB's OPA604 datasheet.
It is of course not a OPAMP built just for audio (20 MHz bandwidth).
But the guys at BB (now TI) are also fully aware of their audio potentials.
They even go as far out as discussing in their datasheet why the subjective sound quality of FET may be better BJT devices.
Clearly, these engineers are quite serious about audio stuff.

 

[GeekGirl]

Quote:

Originally Posted by tangent It's that very lowest segment of the EM spectrum, the part that's supposedly the easiest to amplify...

Don't forget an important parameter: bandwidth. A flat amplitude / phase response from 20 Hz to 20 KHz is 3 decades (8 octaves). Not an easy feat to accomplish, especially for designing filters and linear power amps. RF is easy compared to audio in terms of broadband design.

You should be aware that cell phone (and other) standards define "voice" grade to only work up to 3 KHz or so. It's done to conserve bandwidth which allows more capacity in the system. The chip manufacturers know this so they can design the most cost-effective power efficient parts. Another reason to not use ICs designed for portable devices.

 

[tangent]

Quote:

Originally Posted by Ferbose But the guys at BB (now TI) are also fully aware of their audio potentials.

That is one reason I said "better than even chance". You can't say, "if it says it's for audio, it's bad". It's just a good indicator that you should look into the specs carefully, and try it before deciding that it's good for audio.