Beresford Bushmaster DAC

[StanleyB]

What I did say is that the headphone amplifiers built into DACs are generally a compromise A DIL IC is not going to be happy to drive low impedance headphones, where as it may drive higher impedances fairly easily. However if the voltage supply is low then it may well not even do that very well.

Hense unless some purpose built headphone amp that doesn't involve the use of a DIL IC (or maybe just as a driver) is built then the headphone output will be a compromise & an external headphone amp would be a wholly better proposition

You haven't been keeping up to date with headphone amp design technology . The TI TPA6120 IC has been used in headphone circuits for many years now by companies like Benchmark in their DAC, and by many others in stand alone headphone amps. Other ICs are also available that can be used for headphone output circuits.

As for your point on voltage output: can you expand on this for the benefit of all the readers of AoS? Maybe you can make a simple list of what current and voltage is required from a headmp in order to produce the maximum permitted power output to drive headphones at various impedances?

[StanleyB]

Definitely guilty along with Mark and Werner
It’s not so much a question of the Dac being inferior because of it having a headphone output, it’s that its likely that the headphone output is likely to be inferior to a stand alone dedicated headphone amp.

Can you expand a bit on that in technical terms? How does the separate case and power supply contribute towards making a stand alone headamp sound better compared to a headamp that shares its casing and power supply with a DAC circuit?

[Reid Malenfant]

You haven't been keeping up to date with headphone amp design technology . The TI TPA6120 IC has been used in headphone circuits for many years now by companies like Benchmark in their DAC, and by many others in stand alone headphone amps. Other ICs are also available that can be used for headphone output circuits.

As for your point on voltage output: can you expand on this for the benefit of all the readers of AoS? Maybe you can make a simple list of what current and voltage is required from a headmp in order to produce the maximum permitted power output to drive headphones at various impedances?

I'm not really interested in ICs A headphone amplifier is going to supply low enough power to make it worthwhile using a discrete circuit in Class A. I don't see many ICs in tube headphone amps so why incorporate one into a solid state version if you want the best quality?

As for expanding on things, sure!

Beyer Dynamic manufacture a 600 ohm version of the DT990. This has a sensitivity of 96Db/mW & a power handling capacity of 100mW or 0.1W. This means it can reach peak levels of 116Db.

Now if you look as this little screen grab you'll notice the extraordinary sound levels possible with symphonic music...

SPL levels at 3m.jpg

Taken from http://www.soundadvice.info/thewholestory/san12.htm

So lets assume you are sitting 12m away from the orchestra The sound will be approximately 12Db down on those figures as it drops by 6Db when the distance is doubled. That still gives peaks of over 120Db.

Now we aren't going to ask the headphones to do what they can't do so we'll take the peak value of 116Db which they can do...

0.1W x 600 ohms = 60. square root of 60 = 7.7459V RMS neaded to produce this. Peak voltage = 7.7459 x 1.414 = 10.95V PK. So the supply will need to be approximately +/- 15V DC allowing for component losses.

Now type 4 ohm headphone into google, you'll find there are quite a few

As I mentioned previously I'd like a no compromise headphone amp that could drive any headphone to full volume... So lets just assume that these hypothetical 4 ohm headphones have the same 96Db sensitivity & the same maximum power handling. Ok I know it's not likely to be exactly the same but there is no point in sucking eggs...

0.1W x 4 = 0.4. square root of 0.4 = 0.6324V RMS needed to produce peak output. Now comes the interesting bit as I want a class A amp

0.6324V RMS x 1.414 = 0.8943V Pk. current needed in a push pull class A amp is 0.8943 divided by 4 ohms = 0.223A. Divided by two again as it's push pull = 0.1118A.

Now as we want to drive any headphone from 4 to 600 ohm to full volume we need that +/- 15V supply & the 0.1118A current which works out at 3.353W dissipation per channel or a total of just under 7W both channels.

I don't see any IC being happy with that kind of power dissipation myself


Hope that's good enough for you Stan


Now I know you may well come back & say the the EU has restricted the maximum sound level that is allowed to be produced by a set of headphones & thus you have your hands tied, but I'm not interested in what the EU says in all honesty

[chrism]

Thanks Mark, very thorough and informative. Just need to remember to turn the vol down on one of yours before switch on!

Regards

Chris

[Reid Malenfant]

[StanleyB]

I'm not really interested in ICs A headphone amplifier is going to supply low enough power to make it worthwhile using a discrete circuit in Class A. I don't see many ICs in tube headphone amps so why incorporate one into a solid state version if you want the best quality?

So it is your prejudice against the use of an IC, rather than its current handling ability that you hold. In a different topic you mentioned that the IC would run out of steam if asked to produce more than 40 or 50mA of current.
However, the 600 Ohms DT990 at 120dB only consumes about 18.5mA .

I don't really care about using a 4 Ohms headphones on a headamp. That sort of low impedance is for speakers. It would also produce an SPL of 140dB from a 11Vpk line, which would of course wipe out one's eardrums in seconds.

If you take the far more popular and common 300 Ohms cans, then a 9Vpk signal will still be able to produce a 120dB SPL output, and a 32 Ohms can would be able to produce a 130dB SPL output. High enough to make you permanently deaf.
And at the end of the day the majority of headamps cater for the most common headphone impedances ,

[Reid Malenfant]

So it is your prejudice against the use of an IC, rather than its current handling ability that you hold.

If I want to live in a prefab then I'll use prefabricated parts yes.

While I'm sure that they measure very well due to having loads of feedback to lower distortion, I'd rather build a discrete circuit that only requires a small amount of feedback to control things.

A triple emitter follower circuit biased into class A on it's own is likely to have about 0.1% distortion (-60Db) driving it's respective load. That's with no front end or voltage amplifier.

In the case of a headphone amp & the worst case 4 ohm load only a two stage output would be needed. That would produce less distortion than a triple. A few transistors making a circuit that gives say 25Db gain open loop & 15db gain closed loop would reduce that output stage distortion to miniscule levels. Yes more than the IC would give, but then it would be a heck of a lot more musical


Now while I can fully understand why people use ICs for DACs is pretty obvious & I'd do the same thing, but for something as simple as an amplifier I think there are better ways of going about things.

[StanleyB]

Sorry, I had to attend to the postman for a minute and today's dispatches.

Coming back to your point about the 120dB dynamic range of the orchestra: where are you going to get that extra 24dB out of the CD format ?

[Reid Malenfant]

There are res formats Stan as I'm sure you are aware.

I don't think any orchestral music has that kind of dynamic range anyway, not whilst playing, so I fail to see what dynamic range has to do with anything?

If you'd have said peak to average ratio of volume of sound I think it would have made more sense. If you look back at that screen shot it shows about 25db peaks (not being silly & giving the 137Db peak figure) over the average content.

[Werner Berghofer]

Stanley,

a 9Vpk signal will still be able to produce a 120dB SPL output, and a 32 Ohms can would be able to produce a 130dB SPL output.

I don’t think that SPL (= sound pressure level, so that other non-techies may be able to fully understand what you’re talking about) is the most important criteria to evaluate the playback quality of a headphone amplifier. While of course a bicycle with an auxiliary engine may reach similar speeds (when driving downhill), there is no doubt that riding a fully equipped motorbike with enough horsepower reserve is much more fun – and more safe, in case the extra power needs to be available unexpectedly and immediately.

When talking about low impedance headphones like the AKG K701 (62 Ω) or the AKG K271 (55 Ω) both you and I know that I’m not the only one who considers their sound rather thin, anemic and weak when they are used on the headphone socket of your Caiman or TC-7520. The moment these headphones are connected to a real headphone amplifier, they really start to shine and bloom, even if one of your DACs is used to feed the head amp.

Werner.