[jcx]

it is odd to be fixated on decades buffer old tech when you can use a TPA6120/THS6012 DSL driver CFA op amp today for your output buffer - inside of a multi-loop composite circuit with whatever input op amp tickles your fancy
 
just compare the specs - then consider that if you know what you're doing you can also apply the TPA6120 with local gain - further improving the composite loop performance
 
I've even shown that it can be biased push-pull Class A

 

[Steve Eddy]

Quote:

it is odd to be fixated on decades buffer old tech when you can use a TPA6120/THS6012 DSL driver CFA op amp today for your output buffer - inside of a multi-loop composite circuit with whatever input op amp tickles your fancy
 
just compare the specs - then consider that if you know what you're doing you can also apply the TPA6120 with local gain - further improving the composite loop performance
 

 
What exactly is wrong with a decades old buffer technology?
 
You say "just compare the specs," so are you saying that a well-designed diamond buffer would be so bad as to be audibly distinguishable from what you propose?
 
se

 

[jcx]

there are lots of circuits that are probably adequate - even 5534 op amps still show up in this past decade's flagship audio DAC I/V, output filters manufacturer's data sheet circuits that the distortion specs are taken from
 
but some seem to want to argue minutiae of which buffer is "better" - not just "audibly good enough"
 
then I think we are stuck with comparing specs

 

[Steve Eddy]

Quote:

there are lots of circuits that are probably adequate - even 5534 op amps still show up in this past decade's flagship audio DAC I/V, output filters manufacturer's data sheet circuits that the distortion specs are taken from
 
but some seem to want to argue minutiae of which buffer is "better" - not just "audibly good enough"
 
then I think we are stuck with comparing specs

 
Gotcha.
 
se

 

[tyre]

Quote:

it is odd to be fixated on decades buffer old tech when you can use a TPA6120/THS6012 DSL driver CFA op amp today for your output buffer - inside of a multi-loop composite circuit with whatever input op amp tickles your fancy

 
I can think of a few reasons. The old buffers are relatively basic implementations using few transistors and hence vaguely understandable to someone (like myself) with limited electronics knowledge. They all have a high quiescent current so they are more likely to operate in Class-A. They use metal-can packages that dissipate heat well. The FET-input LH0033 has high input impedance and is tolerant of being used in a stand-alone configuration outside of a feedback loop. It also has pins to adjust for DC offset via an external pot.

With current feedback op-amps, there is the drawback of having to isolate the output from cable capacitance with an output resistor.

 

[Chris J]

Flavour of the month Op Amp and a fancy buffer sound better to the consumer than a plain old TPA6120.

 

[stv014]

.

 

[mikeaj]

Quote:

Well, the TPA6120 is not exactly perfect, and it even costs more than the three op amps in the O2, for example. It seems there was too much focus on making the IC very "fast" (since that is a marketable feature to audiophiles), at the expense of other parameters that are more relevant to a headphone buffer, such as stability, low output impedance, and the ability to drive low impedance loads with low distortion.

 
It was made fast, so it could be a DSL line driver (THS6012).  I don't think they bothered designing something like the TPA6120 from scratch.

 

[jcx]

the TPA6120/THS6012 does have extremely low audio frequency distortion - the market hadn't yet settled on DSL acceptable distortion limits and a concern was distortion, IMD products wiping out adjacent channel's S/N at 100s kHz to low MHz while driving loads down to 25 Ohms with Watts - so it was designed and in fact manages to deliver very low distortion despite the limited open loop V gain
 
the parts per million audio frequency distortion performance was probably a suprise to TI as well - leading to the renumbering, "audio" branding
 
I suspect the output is heavily Class AB biased for a op amp - certainly enough to qualify as meeting Nelson Pass' "first watt" principle as translated to many headphone loads, sensitivities
 
the output impedance is quite low, with feedback it is much lower than most classic buffers - in a multiloop the practical limit of amp output Z can easily be the wiring/connectors
 
the output does have to be isolated from capacitive load at 10s of MHz, this can be done with a series inductor as is near universal in power audio amps - the audio frequency output Z can then be very low

 

[Chris J]

Quote:

the TPA6120/THS6012 does have extremely low audio frequency distortion - the market hadn't yet settled on DSL acceptable distortion limits and a concern was distortion, IMD products wiping out adjacent channel's S/N at 100s kHz to low MHz while driving loads down to 25 Ohms with Watts - so it was designed and in fact manages to deliver very low distortion despite the limited open loop V gain
 
the parts per million audio frequency distortion performance was probably a suprise to TI as well - leading to the renumbering, "audio" branding
 
I suspect the output is heavily Class AB biased for a op amp - certainly enough to qualify as meeting Nelson Pass' "first watt" principle as translated to many headphone loads, sensitivities
 
the output impedance is quite low, with feedback it is much lower than most classic buffers - in a multiloop the practical limit of amp output Z can easily be the wiring/connectors
 
the output does have to be isolated from capacitive load at 10s of MHz, this can be done with a series inductor as is near universal in power audio amps - the audio frequency output Z can then be very low

 
Plus it can output up to 700 milliAmps.
 
i.e. up to 2 Watts into a 32 Ohm load.
 
pretty damn good to me for a mere headphone driver.

 

[Chris J]

Quote:

 
Well, the TPA6120 is not exactly perfect, and it even costs more than the three op amps in the O2, for example. It seems there was too much focus on making the IC very "fast" (since that is a marketable feature to audiophiles), at the expense of other parameters that are more relevant to a headphone buffer, such as stability, low output impedance, and the ability to drive low impedance loads with low distortion.

 
Just wondering, what makes you think all this?
The O2 comparison is not really fair; the TPA6120 outputs far more current into low Z loads.

 

[wakibaki]

The TPA6120 does have something important the NJM4556s in the O2 don't have. Short circuit and thermal protection stated in the datasheet.
 
Lord V expresses confidence that they will survive a short circuit and states that he performed tests on which he based his conclusions. Perhaps he's right. In my experience, however, datasheets are quasi-legal documents now as well as being promotional material, and in the nature of things, if a device really has a particular feature, it will be stated in the datasheet, and vice versa.
 
I'm retired now, but I do know that when I was working, my line manager would not have accepted my assurances about a device surviving short circuits unless it was stated in the datasheet.
 
Consequently, if I'm designing a piece of equipment, even for amateur construction by enthusiasts on a forum like this, if I want to be able to give the assurance that it will survive a short circuit, I don't employ NJM4556s in the output stage.
 
w

 

[Chris J]

Quote:

The TPA6120 does have something important the NJM4556s in the O2 don't have. Short circuit and thermal protection stated in the datasheet.
 
Lord V expresses confidence that they will survive a short circuit and states that he performed tests on which he based his conclusions. Perhaps he's right. In my experience, however, datasheets are quasi-legal documents now as well as being promotional material, and in the nature of things, if a device really has a particular feature, it will be stated in the datasheet, and vice versa.
 
I'm retired now, but I do know that when I was working, my line manager would not have accepted my assurances about a device surviving short circuits unless it was stated in the datasheet.
 
Consequently, if I'm designing a piece of equipment, even for amateur construction by enthusiasts on a forum like this, if I want to be able to give the assurance that it will survive a short circuit, I don't employ NJM4556s in the output stage.
 
w

 
Who's Lord V?
What is the reference?

 

[wakibaki]

The designer of the O2.
 
You must surely be able to identify him from that. Unfortunately since he's banned from this site, I can't post a direct link to his blog without breaking the rules. I'm probably skating on thin ice as it is.
 
w

 

[Chris J]

Oh, NWAVGuy!
Now I get it!
I don't think the moderator's mind naming him, but I wouldn't post a link either! LOL!