Anyone tried the OPA 2111 ?

[IMEP]

In my never ending quest for knowledge, I just re-read the HeadWize paper titled "Designing an OPamp Headphone Amplifier". On page 2 it recommends a slew rate of 5uV/sec or more. If this is a good rule of thumb, the 2111s slew rate of 2V/uS would be more than enough. Right??

Is the HeadWize article's 5uV/sec sugestion correct or a misprint? 5uV/sec seems really slow to me, but what do I know!

IMEP

Edited: Removed my erroneous decimal conversion

 

[Magic77]

I'm no expert on Slew Rates; but maybe op-amps with very high slew rates are not really necessary for headphone amps?

Another words; a high slew rate is ok, but it may or may not yield better quality sound or performance?

I think we may need a "Slew Rate" Thread.

 

[PeterR]

Quote:

Originally posted by IMEP In my never ending quest for knowledge, I just re-read the HeadWize paper titled "Designing an OPamp Headphone Amplifier". On page 2 it recommends a slew rate of 5uV/sec or more. If this is a good rule of thumb, the 2111s slew rate of 2V/uS (2000V/sec) would be more than enough. Is the HeadWize article's 5uV/sec sugestion correct or a misprint? 5uV/sec seems really slow to me, but what do I know! IMEP

µ=micro=1e-6, 2V/µs=2,000,000V/s
At 5µV/s it would take the amp over two days to swing up to 1V, a tad too serene perhaps...

 

[PinkFloyd]

Quote:

Originally posted by Magic77 I'm no expert on Slew Rates; but maybe op-amps with very high slew rates are not really necessary for headphone amps? Another words; a high slew rate is ok, but it may or may not yield better quality sound or performance? I think we may need a "Slew Rate" Thread.

It depends on the application the Op amp is used for Magic77 but measurements count for screw all when music is involved. Something that measures absolutely dire can sound absolutely breathtaking whereas something that measures perfectly can sound utter crap.

Musical appreciation can be measured by one thing (two things actually) your ears. So much of todays equipment is designed purely by measurement alone...... if it "plots" well it goes into production. In the halcion days of Hi-Fi mathematics made up 90% of the design, the remaining 10% involved months (even years) of tweaking and honing to get the sound just "right"

It's NEVER a good idea to employ a component in a design in the final "listening tests" on the merit of "it measures well" The traditional designer was not afraid to use his ears. In the case of the 2111 "my" ears tell me it is the best op amp for the Chiarra amp as it sounds fantastic. I don't give a flying screw what the spec sheets say, this op amp is my choice in my amp.

Let's get back to the trial and error design approach with Hi-Fi and stop regarding TI spec sheets as the bible to audio Nirvana cause good sounding equipment is not something that can be produced by soldering a few "perfectly" measured components together.

Use your ears

Pinkie.

 

[rickcr42]

Quote:

I'm no expert on Slew Rates; but maybe op-amps with very high slew rates are not really necessary for headphone amps?

correct , though everyone seems spec sheet sensitive it comes down to listening

i have used the PMI/AD OP27 in many audio stages and from memory it is a 3V/uSEC slew rate .

it sounds just fine-especially with a discrete output stage for higher current .

another point worth consideration is high speed layout is not that simple , sometimes slower is actually better and settling time is far to often an overlooked spec

 

[IMEP]

Quote:

Originally posted by PeterR µ=micro=1e-6, 2V/µs=2,000,000V/s At 5µV/s it would take the amp over two days to swing up to 1V, a tad too serene perhaps...

Thanks for correcting my error. I must have had "mili" on the brain. Sorry for the confusion. Actually, we went to my wife's cousin's home for lunch/dinner. At the table (bored of course) I was mulling the conversion over in my head and realized I made a mistake. I was hoping I would get back home and correct it before anyone noticed!

No such luck!

Edited: Regarding the HeadWize article's =/> 5uV/s slew rate recommendation - I guess you can't always believe what you read!

 

[ADS]

Quote:

Originally posted by rickcr42 correct , though everyone seems spec sheet sensitive it comes down to listening i have used the PMI/AD OP27 in many audio stages and from memory it is a 3V/uSEC slew rate . it sounds just fine-especially with a discrete output stage for higher current . another point woth consideration is high speed layout is not that simple , sometimes slower is actually better and settling time is far to often an overlooked spec

Agreed, and I was hoping that the slower slew rate would mean a much better settling time which could do wonders for smoothing out treble response, but the settling rate is 12 times slower. I'm not passing judment on this amp until I hear it of course, and it should be hear the 23rd. I'll tell you all what I think of it when I get it.

 

[IMEP]

Quote:

Originally posted by PinkFloyd It depends on the application the Op amp is used for Magic77 but measurements count for screw all when music is involved. Something that measures absolutely dire can sound absolutely breathtaking whereas something that measures perfectly can sound utter crap. Musical appreciation can be measured by one thing (two things actually) your ears. So much of todays equipment is designed purely by measurement alone...... if it "plots" well it goes into production. In the halcion days of Hi-Fi mathematics made up 90% of the design, the remaining 10% involved months (even years) of tweaking and honing to get the sound just "right" It's NEVER a good idea to employ a component in a design in the final "listening tests" on the merit of "it measures well" The traditional designer was not afraid to use his ears. In the case of the 2111 "my" ears tell me it is the best op amp for the Chiarra amp as it sounds fantastic. I don't give a flying screw what the spec sheets say, this op amp is my choice in my amp. Let's get back to the trial and error design approach with Hi-Fi and stop regarding TI spec sheets as the bible to audio Nirvana cause good sounding equipment is not something that can be produced by soldering a few "perfectly" measured components together. Use your ears Pinkie.

Hi Pinkie,
I didn't mean to offend you with my posts. I respect your opinions and knowledge of headphone amps. Your original post is what made me look closer at the OPA2111.

I completely agree that a spec sheet can't tell you how an opamp will sound. It does however tell you how it performs electronically; which should give you a rough idea how it will perform sonically.

Like I said, I respect your opinions and your audio knowledge, but your (or anybody’s) personal impressions of a devices sonic quality is subjective. What you may consider sounds "dark" or "bright" others may consider differently.

A spec sheet on the other hand is pretty darn objective. I am trying to better understand what is and isn't important when reading an opamps spec sheet. When I compared the 2111 specs with the AD8620 the difference in slew rates was one of the things that jumped out at me. I am pretty sure an opamps slew rate has some sort of sonic effect. What I don't know yet is what is the minimum slew rate for good sound and at what point it's a matter of diminishing returns (or even detrimental). Hence my posts.

I am pretty much a mechanical guy with very limited "sparky" experience, so please forgive my ignorance. This hobby is great way for me to learn more about electronics and have lots of fun to boot!!

No harm, no foul? As always, I will continue to look forward to reading more of your opinions and impressions.

IMEP

 

[ADS]

Well, if we assume that we're all listening to redbook CD, we have 44,100 discrete voltage levels per second, and lets say we want to be able to swing from 0V to 12 V in a single sample. To do that we'll need a slew rate of:

12/(1/44100)= 0.5292 V/usec

OPA2111's slew rate=2 V/usec

This means that the OPA2111 should be just fine. Now let's say that we want the output to have settled before the next sample. Within the OPA2111 specs we find the following:

To settle to 0.1%: 6 usec
To settle to 0.01%: 10 usec

A single sample takes about 22.67 usec, which would mean that if the slew rate was infinite the output would always have settled down by the next sample, but due to the fact that the slew rate is not really perfect, it is possible that at high frequencies/output levels the voltage output will not have settled by the time the next sample comes around. What this means in terms of audible effects, I have no idea.

If someone could tell me that would be wonderful. I'm guessing that this means that you may get some high frequency sibilance/distortion if you drive the opamp too hard. At half the volume (6V, which I think is plenty to drive most headphones) none of these problems exist. I guess the conclusion to this long rambling post is that yes, in extreme cases the OPA2111 may be too slow, for most uses it's just fine. I can't wait to hear it.

 

[rickcr42]

ADS-funny thing about specs is they can tell you if something sucks if the specs are really bad or the noise figure is high and the application is fopr low noise and high gain (mic or mc amps) but can never tell you if something spounds good

as for speed ,it is not all it is cracked up to be-some very "slow" tube amps with limited bandwidth sound very convincing while a some state of the art (specs ) can sound very uninspiring.
I guess that is why we listen

BTW-I can remember the TL072/LF357 being called the "ultimate" opamp and we would not even consider its use now unless for a musical instrument effect .

In fact I beleive the LF357 was the Van Alstine 'secret" opamp upgrade for the original Dynaco PAT-4 preamp and the reviewers all touted the quantum leap forward in sound !

It also depends on system matching what will or will not sound good

for example the often disregarded OP275 will sound great in a system with a light bottom end but sound bloated and dull in a bass heavy system.The '275 also has a nice settling time

system integration is more important than flavor of the week or the ultimate spec

 

[ADS]

In a little addition to my above post, here's the time required for the OPA2111 to settle in a discrete voltage sample at maximum extension (12 V and redbook CD just like before)

(12 V) / (2 V/usec) = 6 usec

To reach 0.1%=12 usec
To reach 0.01%=16 usec

Since the time between samples is 22.67 usec it looks like the OPA2111 has plenty of headroom for what we're going to do with it, even at levels it's unlikely to ever see so the possibility I mentioned in my previous post of high frequency problems won't happen. Conclusion? I guess that we're all using parts that are way too fast.

 

[rickcr42]

if it measures good and sounds bad, it is bad; if it measures bad and sounds good, you have measured the wrong thing."

- Daniel R. von Recklinghausen,
former Chief Research Engineer, H.H. Scott

 

[PeterR]

Quote:

Originally posted by ADS Well, if we assume that we're all listening to redbook CD, we have 44,100 discrete voltage levels per second, and lets say we want to be able to swing from 0V to 12 V in a single sample. To do that we'll need a slew rate of: 12/(1/44100)= 0.5292 V/usec OPA2111's slew rate=2 V/usec

Okay, first of all the output of your CD player doesn't look like discrete steps (neglecting for a moment sake induced esoteric DACs by designers who decided reconstruction filters are for wimps, but even in this case you wouldn't necessarily want to reproduce them). You need to look at the waveform of the signal you want to reproduce. The steepest slope of a sinewave is Vpk*2*Pi*f, so for the CD's 20kHz upper limit and 12V you'd need a slew rate of 1.5V/µs.

Settling time is the time it takes the output of the amp to settle to a value within a certain percentage of the target value when you apply a step signal at the input.
So in this case, if you instantaneously change the input from 0V to 10V, the amp will try to change its output as fast as its slew limit allows to -10V (G=-1). At 2V/µS this takes 5µS. The output may then overshoot (could be also seen in the frequency response) and you'd have a damped oscillation around the final value. The 2111 needs an additional 1µS for its output to stay within 10mV of -10V. For our application, this figure is pretty insignificant.

 

[PeterR]

Quote:

Originally posted by ADS Agreed, and I was hoping that the slower slew rate would mean a much better settling time (...)

Slew time is included in the settling time. Slower slew rate -> longer settling time (all else being equal).

 

[bigcat39]

Heh, I work for a really major research lab .... as in 100+ EE PhD level researchers. About half of them are analog guys. (BTW, I'm a metrologist, so I really BELIEVE in measurements). The engineers I have shared my plans for PPA building with are all intrigued by the circuit .... and totally mystified by the super high performance opamps we all choose. Anything you can't measure is indistinguishable from magic ... and there has never been an instrument devised, IMHO, that can tell what an audio circuit will sound like subjectivly.
It's interesting to get insights from all this high priced talent (You should hear what they say about super expensive power cables for high end audio

), but I their advice all boils down to one thing: I'm going to let my ears be my guide. I have made a lot of electronic measurements over the years, and how an audio circuit behaves is often unrelated to anything you can measure. Measurements, even those made by the manufacturer, follow the laws of uncertainty. What they don't tell you in spec sheets is that sometimes the uncertainty of the measurement is LARGER than the measurement itself. What is important is PERFORMANCE, relative to the ears to be stimulated.