I agree that the recommended output resistance of 120 ohm is a great issue. Not only it affects frequency response, but, as already stated here, it dramatically increases harmonic distortion at low frequencies. Really dramatically. I have made numerous measurements on this issue. Only several ohms of output impedance of the headphone amplifier is tolerable, and "zero" output impedance is the best solution. I assume that the recommendation of the 120 ohm output resistance is a relict of deep audio past, when almost no one took care about distortion an resistive impedance of headphones was expected - which is really not the case today.
European norms specify a maximum output impedance of 2R
for the generator when testing headphones sound pressure vs output levels.
http://www2.rohde-schwarz.com/file_12241/1GA48_2e.pdf
for the generator when testing headphones sound pressure vs output levels.
http://www2.rohde-schwarz.com/file_12241/1GA48_2e.pdf
". . . any guestimate on your discrete Class A stage distortion? – I’d guess that -80 dB THD is very good for a complementary follower working at a reasonable fraction of its bias – is your common emitter stage going to do any better (hint: Dr Edward Cherry’s last JAES article lets you answer this easily)"
John, why are you proposing a common emitter output and not common collector for this design?
I'd think with rails at +- 15 you'd have no output swing issues. 3V peak into almost any headphone you care to name would blow your ears out anyway.
John, why are you proposing a common emitter output and not common collector for this design?
I'd think with rails at +- 15 you'd have no output swing issues. 3V peak into almost any headphone you care to name would blow your ears out anyway.
Actually, the alternative suggested by JCX, has just as much peak output. It just does not have as much CLASS A output. In truth, the way you guys are carrying on, JCX, and even the designers at TI should be held equally accountable in destroying peoples ears.
What I am trying to do is to make a very high quality, virtually effortless, headphone amp design, that is both buildable and cost effective. I am pointing out a relatively 'unique' design that was first developed as an App. Note at either Fairchild or National in the late 60's. I built my first version of this to drive a K-horn (temporarily) back in 1970, and then converted it to single ended voltage supply for the road, and used it for about 30 days in a converted school bus driving a JBL monitor, and driven by a Nagra, playing protection masters from the GD and others providing music for a dozen or more hippies. It worked well for the task, and amazingly well for the K-horn. Not perfect, of course, but OK.
One does NOT expect people to drive their ears to destruction. What we want is that the design performs 'effortlessly' like a Bentley, rather than like an economy car with a small engine, or a very high rev'ing one.
What I am trying to do is to make a very high quality, virtually effortless, headphone amp design, that is both buildable and cost effective. I am pointing out a relatively 'unique' design that was first developed as an App. Note at either Fairchild or National in the late 60's. I built my first version of this to drive a K-horn (temporarily) back in 1970, and then converted it to single ended voltage supply for the road, and used it for about 30 days in a converted school bus driving a JBL monitor, and driven by a Nagra, playing protection masters from the GD and others providing music for a dozen or more hippies. It worked well for the task, and amazingly well for the K-horn. Not perfect, of course, but OK.
One does NOT expect people to drive their ears to destruction. What we want is that the design performs 'effortlessly' like a Bentley, rather than like an economy car with a small engine, or a very high rev'ing one.
Last edited:
headphone sensitivity and impedance have quite a wide range
a "universal" headphone amp is impractical on dynamic range/noise floor limitations - sensitive iem can have >135 dB SL re 1 Vrms down to <90 for some "studio monitors"
senstive iem really deserve good step-down xfmr output amps
but most "over ear" headphones can be driven to 110-120 dB SPL with <10 Vrms, 100 mArms
Grado's 32 Ohm, 113 dB re 1 Vrms is a typical "low Z" headphone
Senn's "Audiophile" HD600/650/700/800 are 300 Ohm, 103 dB +/- 2 dB
some new "Audiophile" Orthodynamics are a challenge to standard op amp supply V - however they are wired "Balanced" and can be driven adequately with bridged output from +/-15 Vsupply amps
I zipped a Open Office spreadsheet from a older Head-Fi thread so its not up to date but still has lots of headphone's sensitivities
a "universal" headphone amp is impractical on dynamic range/noise floor limitations - sensitive iem can have >135 dB SL re 1 Vrms down to <90 for some "studio monitors"
senstive iem really deserve good step-down xfmr output amps
but most "over ear" headphones can be driven to 110-120 dB SPL with <10 Vrms, 100 mArms
Grado's 32 Ohm, 113 dB re 1 Vrms is a typical "low Z" headphone
Senn's "Audiophile" HD600/650/700/800 are 300 Ohm, 103 dB +/- 2 dB
some new "Audiophile" Orthodynamics are a challenge to standard op amp supply V - however they are wired "Balanced" and can be driven adequately with bridged output from +/-15 Vsupply amps
I zipped a Open Office spreadsheet from a older Head-Fi thread so its not up to date but still has lots of headphone's sensitivities
Concern about loudness levels aside. The intriguing thing about headphones is that it is much easier to attain concert levels using headphones than loudspeakers. As SY's indicated in the past, I doubt my family and neighbors want to me to replicate the sound of a full orchestra (at concert levels) in the living room. (And I cannot imagine what Nirvana live From the Muddy Banks of the Wishkah might do to the neighborhood.)
.
.
Last edited:
Note: for those who may be wondering, and apologies if I missed the specifics earlier, this is the "Solomon" I believe mentioned by JC: http://www.ti.com/lit/an/snoa737/snoa737.pdf
Brad
Brad
Of course, Johnferrier. We don't want to hurt people's ears, we just want to make sure that the headphone amp doesn't clip. One, 'equalizer' is using a 30 ohm build-out resistor, as I have already suggested. This way, a really sensitive 30 ohm headphone will have its output reduced by 6dB. Not a bad idea. It also keeps the design more class A until voltage clipping is possible, and hopefully not often reached.
Thanks bcarso, yes, that is the article that I keep handy, and have done so for almost 40 years.
John,
My preferred way of making a "low impedance load driver for Op-Amp's" has a similarly low parts cound and is also targeted for 100...200mA Iq.
I like low beta, high Ft Bipolar's as direct follower with a CCS in the tail, an LM317 is surprisingly good (I have seen people use them a anode loads for ECC88, I would not go THAT far).
A total of 3 parts (2SC5171 or 2SA1930 depending if we want to pull the Op-Amp output up or down to put it into SECA, LM317 and one resistor) and we are SE Class A, including the Op-Amp output.
I know it's brutal, basic and simple...
Hang the PNP version of this of pin 5 of a real NE5534 and use a heavily degenerated 2SK/LSK389 with CCS tail and cascode as input stage bypassing the internal one and you will have a surprisingly nice Linestage or headphone Amp using the dreaded NE5534... Or take the degeneration out for a Phono Stage... 😛
Ciao T
My preferred way of making a "low impedance load driver for Op-Amp's" has a similarly low parts cound and is also targeted for 100...200mA Iq.
I like low beta, high Ft Bipolar's as direct follower with a CCS in the tail, an LM317 is surprisingly good (I have seen people use them a anode loads for ECC88, I would not go THAT far).
A total of 3 parts (2SC5171 or 2SA1930 depending if we want to pull the Op-Amp output up or down to put it into SECA, LM317 and one resistor) and we are SE Class A, including the Op-Amp output.
I know it's brutal, basic and simple...
Hang the PNP version of this of pin 5 of a real NE5534 and use a heavily degenerated 2SK/LSK389 with CCS tail and cascode as input stage bypassing the internal one and you will have a surprisingly nice Linestage or headphone Amp using the dreaded NE5534... Or take the degeneration out for a Phono Stage... 😛
Ciao T
Last edited:
OT here, but for anyone else subscribing to or reading the Sound Quality vs Measurements thread, are you experiencing the ads rapidly refreshing at the top of the page and each time returning one to them, and preventing reading new posts? I've PM'd a moderator and as well posted a complaint in the standard place. Maddening! And having donated I understand I'm even supposed to be able to kill ads althogether, although I have no idea how to do this.
This thread at least is unaffected for now 🙂
This thread at least is unaffected for now 🙂
If I can go on, what IC should we use? What do we want to select for? Many IC's have different idle currents, and this effects the Iq. What to do?
For the record, the humble uA741 works pretty good in this amplifier. Amazing isn't it? Why?
For the record, the humble uA741 works pretty good in this amplifier. Amazing isn't it? Why?
JC,
Are you actively building this design in your lab now, are you brainstorming, or is this just a thought experiment for the forum ? FWIW, I thought the OPA627/637 were nice chip op-amps. (Although, as I've posted before, in the end I went with Borbely's design.)
.
Are you actively building this design in your lab now, are you brainstorming, or is this just a thought experiment for the forum ? FWIW, I thought the OPA627/637 were nice chip op-amps. (Although, as I've posted before, in the end I went with Borbely's design.)
.
Last edited:
- Status
- Not open for further replies.