Op-Amp Challenge: Compare Op-Amps, By Listening To Them

In the world of audiophilia there are arguments that rage over the relative merits of particular components. Sometimes this can reach silly levels as in the high-end ALPS pot we once saw chosen as a volume control whose only task was to be a DC voltage divider feeding a pin on a DSP, but there are moments where such comparisons might have a bit of merit. To allow the comparison of different op-amps in a headphone amplifier, [Stephan Martin] has created a stereo amplifier board complete with sockets to take single or dual op-amp chips.

The circuit is based upon a design from the 1990s which as far as we can see is a pretty conventional non-inverting amplifier. It has an on-board op-amp to create a virtual ground, and three sockets for either two single or one dual op-amp to create a stereo headphone amplifier.

So the burning question is this: will you notice a difference? We’re guessing that assuming the op-amps under test are to a sufficient specification with a high enough impedance input and enough output current capability, the differences might be somewhat imperceptible without an audio analyser or the hearing of a ten-year-old child.

Need more of an audio fix? Try our Know Audio series.


31 thoughts on “Op-Amp Challenge: Compare Op-Amps, By Listening To Them

  1. But the op-amp sockets aren’t gold plated! And what about that questionable power supply? And the different length traces!

    I’ve heard some ridiculous audiophiles in my time.

    1. “…yet it sounds different TO ME, I can hear it !!! ” is the never old last stand for audiophiles, when they ran out of any logical or rational arguments. Like the infamous oxigen free, $1000 wires discussion.

      1. It is not the point if one person can hear the difference
        – but if a few agree on hearing the differences.
        And taste is everywhere, you are just proving it.

    2. The sockets i ordered are gold plated, but i will put a ZIF socket into it, and then maybe another ZIF on top :-)

      When you go down the rabbit hole for Audio it will never end, just start browsing with the Links i put in the Project description, its endless…

  2. TI’s line of exceptionally low THD op amps show an ingenious circuit which multiples the THD of any op amp faster than the gain. By using this circuit, and magnifying the THD, the imperfections of any op amp can be “heard” without the usual debate. See for example, Fig 47 on Pg. 16 of this datasheet:


    When R3 is made very small (down to one ohm or below), THD is magnified greatly even with modest loop gains. As R3 –> 0 ohms, the DC offset of the amplifier can cause it to rail. Adding a small DC offset to the inventing input (or servo-ing the offset to zero with an external zero drift amplifier) can resolve this side effect.

    Tailored use of this circuit can magnify THD up to 100,000x, with loop gains as low as 2. This lets you hear what the THD of any op amp”sounds like”. They truly are quite different.

    1. Tanks a lot for this extremely useful info!! I am involved in a project targeting a low noise/low distortion preamplifier and this clever circuit saves me a lot of time.

      1. Please don’t waste your time “rolling opamps” in your design unless you care less about the best possible sound and just want to be stylish. Just use some LM4562’s and it’ll be as perfect as technology allows.

  3. I wonder, if the people who comment here actually do understand what they are talking about.
    If they have not done an amplifier comparison, or a speaker comparison,
    they are obviously out of their depth – just a guessing game or worse as I can read here.
    And show how they are designed and soun d like.
    I had the opportunity to be part of a group who had access to different amplifiers and speakers at the time. And the 4 of us spent weekends when there were new pieces available on loan to wire them up and compare them.
    It was fun to listen to Take Five in different setups and discuss the perceptions.
    And compare them as well to amplifiers we built ourselves even, or speakers.
    I wonder, if it is possible now, that measurements can beat the sensitivity of the ear.
    And I do not talk about any extremes or specialists,
    just what sounds best for you and compare and discuss it with others
    who listened to the same piece at the same time .
    And actually agree on the results.

    1. Ask yourself this: if you’re comparing the sounds of these devices, what purpose does the group or discussion serve?

      You weren’t comparing amps and speakers, you were comparing how the different amps and speakers, the selection of music, and the ritual of setting it all up made you feel in the context of the group membership. That’s not the same thing as comparing the sound. Discussions about the different combos among the group serve to alter your perceptions based on your particular foibles and insecurities, and your desire to be respected by members of the group participating in this “experiment”. Everyone in the group is under psychological pressure to come up with something to say about the sound that is different from what the others are saying, yet something the others can agree about, thus affirming the value of each person’s contribution to the group.

      If you actually want to compare the sound, you need a properly set up double blind test, which is difficult and expensive. Yes, I know from lurking in audio forums back in the 90s- double blind tests may work for things that are scientific, but listening to music is an emotional experience and you can’t apply science to it… blah blah blah.

      It’s similar to tasting “fine” wine. You drink it when there are others around that you want to impress, or when there are others whose group you want to be considered a member of. You’d better come up with something unique yet acceptable to the group to say about it (“notes of elderberries and dark chocolate”), lest you be considered a poser by the folks you are trying to impress.

      Have you ever noticed how people who suddenly acquire money suddenly also have better and more sensitive taste (and better, more sensitive hearing) than less fortunate people? Just ask a wine (high end audio) salesman…

      If I want to taste elderberries and dark chocolate I’ll just eat elderberries and dark chocolate…

      1. We did these comparisons for fun, not profesionally but still rather seriously
        and there was no pressure at all as you try to insinuate.
        We used the known pieces of music to compare.
        How does double blind work with ears?
        We used our ears as measurement devices and they were sufficient to hear some differences.
        It was as said a fun experience and we did not want to buy or sell,
        just enjoy and discuss and learn a bit,
        and there were a couple of beers as well …
        But it is useless to waste time discussing it with people
        that have not gone through a similar experience so cannot compare
        and probably just post unfounded and untested opinions.

        1. “Double blind” is easy in that case: Sombody else is switching/changing the setup in a way, that you do not see, what is playing at the moment.
          But be careful: I know an anecdote, where the brother of one of the guys did the changing and reconfiguration of the audio setup. And just for fun he used some coathanger wires (steel) as speaker wires in some setups. The “audiophiles” were really embarassed afterwards, that they could not reliably distinguish the coat hangers from real speaker cable.

          1. My comment regarding double blind and acoustics was not serious,
            just a play on words …
            And you can make fun of any group – including many of the programmers here.
            I have been told they mostly wear headphones with loud music while programming not to be disturbed,
            – I wonder, if this would have any influence in respect of sensitivity regarding speaker system test ( including amps and opamps … )??
            The extra coat hangers would either have an influence or not. Who knows …

    2. You can have a lot of fun by listening to a lot of different speakers / systems, if you really care what sound sounds like. They’re all compromises, and some are truly crazy designs, and it’s fun to hear what the people making them are getting at.

      Part of the point of listening to a bunch of different systems is to _learn_ to hear the differences critically. And I think there’s definitely a subset of audiophiles who enjoys that process. Finding words to describe differences in sound is tricky, and there’s no substitute for first-hand experience.

      mrehorst comparing it to wine is apt. There’s definitely no best wine, and there’s no best speaker. But they taste/sound different, and learning to dissect those differences can increase your enjoyment of both. Drink enough different wines, critically, and you’ll know what folks mean by “blackberry”.

      (I keep mentioning speakers b/c it’s easy enough to build an amp that’s pretty much flat and phase coherent these days. That was solved in the late 80s – mid 90s, IMO. The rest is how much of a pathological load you want it to drive, and with how much overhead. Speaker design is where audiophile stuff gets fun.)

      That said, this project is cool b/c it lets you experiment. Maybe you’ll find there is no difference among decently specced op-amps? But I like experiments. You’ll definitely be able to hear the difference between an OPA-series amp and an LM358 — and that’ll teach you what crossover distortion sounds like. And then you can correlate it with what you see on your scope. That’s science, y’all.

      1. Great answer.
        But I have at least learnt now reading the other posts
        that there are no good speakers, no good amplifiers, no good ears, no good wine – all only perception
        and you should probably select your sound system by colour?
        Upps, then the perception changes from acoustic to optics – so the same issue …

    3. You’re brave. Even here, folks think it’s always open season on audiophiles.

      There are certainly some deserving targets. Like esoteric AC cords ;-) . But when you get down to it with serious testing, there ARE differences. They are often very subtle, sometimes subjective, and honestly, usually inconsequential for most of us, if audio is not our hobby/obsession.

      And of course, the distortion and inaccuracies of just about all loudspeakers are still a few orders of magnitude greater.

      Part of my career was in maintaining studios. We were offered the chance to test a $$$ esoteric power amp vs the good commercial-cinema-grade conventional amps we’d been using. The test rig was minimal: CD player to both amps, and switching to select which amp would drive a pair of Tannoy FSM monitors. Levels were matched using a SPL meter. Testing was double-blind, and no participant was told which was which, til all results were in. The test participants were all recording engineers or editors.

      Results: there was a clear preference for the ‘commercial’ amp; the esoteric amp wasn’t considered ‘cleaner’ or more accurate or whatever.

      While we were still in a testing mood, I also modified the test rig for a test of 3 kinds of cable – some cheap stuff we used mostly for MIDI cables ;-), the basic shielded pair stuff we installed miles of, and some esoteric cable with a big ‘M’ on it. Equal lengths (50ft or 100 ft, I believe). Triple-blind testing.

      Results – very very subtle differences, and opinion differed as to who thought which was superior. The esoteric cable was NOT conclusively ranked better. The stuff we already used was about in the middle of the pack. So, yes, tiny differences, but not even close to making us want to rewire the plant.

      Back to opamps – yes, differences too, and a demanding load like headphones is a harsh test. The interaction between the series output resistor and the actual headphones used might be the single biggest variable. For the record, I have a bin half-full of 5532 dual opamps that I use for almost everything (except low-voltage/low-power portable stuff), and they drive most headphones just fine.

      1. Thank you very much for the facts by somebody who actually knows.

        It would be nice, if you could comment as well on this myth:
        that if you cannot measure it, it is not valid.
        At the time we did these comparisons, there was no way ( we could afford or knew )
        so the ears were the best still.

        1. I just wanted to confirm that differences are possible, and sometimes reproducible, and that different isn’t always ‘better’.

          If a difference is reproducible, to a statistically valid level with different listeners, I personally will go with it, even if I can’t quantify it with measurements. I also accept that people have different tastes, and one might prefer the colouration of speaker A over speaker B, and others would disagree, but neither is necesssarily better.

          Caveat – I have some gear I’m fond of, but I’m not an audiophile, and most equipment over a certain level of quality is satisfactory to me.

          1. Great.
            This confirms what I thought.
            Taste is taste and it really does not matter. The same for other aspects in life.
            For us it was fun to discuss the differences we heard.
            Nothing to do with exaggeration or cost.

        2. A few years ago I wanted to do some Speaker DIY again,
          and in the process get some better speakers for my laptop.
          See the link to the one I used.
          First it was the large enclosure as I wanted to hear again how a horn sounds.
          This one is actually easy to build. Get the wood cut and screw it together.
          But then I built the smaller version for the PC;
          as a first step I just put it together using tape to test.
          They sound great and the version I listen to now is still the same taped together.
          Another piece of audio fun for me,
          and others here might want to have a look or even try it out.
          All data required is in the link.

  4. “To allow the comparison of different op-amps in a headphone amplifier …”

    Let op-amps do what they’re made for; they can indeed drive medium impedance loads such as headphones in the tens of ohms, but that is pushing them hard, and just by using a different headphone one could alter the sound significantly. For those who aren’t yet ready to accept that Class D is better just about in every field, I would rather suggest a decent op-amp followed by a mosfet wired in Class A, which doesn’t waste too much power for headphone use.

  5. I built a QRP CW ham radio RX using a stereotypical LM386 headphone amp and I installed a socket (beware, this is a great way to make an oscillator) and compared my stock of 80s to 90s era ‘386 and there was SOME audible variation between mfgrs. I would theorize a ‘stick’ of genuine TI parts shipped directly from Digikey would be very consistent, likely almost identical, but rando parts drops including pulls and old Radio Shack parts had some inter mfgr variation. The timeframe was early 90s (not last week). I would estimate 6 dB difference in hiss levels. Obviously I left the highest SNR (lowest hiss) amp installed.

    Ironically having lower gain lower performance high freq output probably makes a better headphone amp for ham radio purposes… I did not care about audio response much below 500 hz or above 900 hz for morse code use…

    Remember what your specs mean… just because THD is measurably low at 100 mW out doesn’t ‘scientifically prove’ it’ll be subjectively pleasant to listen under weak signals condx.

    Another observation that is “politically incorrect” because we all make fun of audiophiles, is that ‘everyone knows’ LM386 based amps have a low freq oscillation ‘motorboating’ problem and you can swap in and out individual chips all day long and nobody wants to admit the TI products have less of an oscillation problem.

    A final observation is you’ll run into people comparing the -1, -3, -4 models (all pin compatible IIRC) claiming they’re the same, but they’re not, read the datasheets, IIRC the output impedance of the -4 was (is?) about a quarter the output impedance of the -1 which relates to the rated output power being about three times higher. But that’s going to somewhat affect noise and possible oscillation problems regardless how much people make fun of ‘audiophiles’.

    Overall installing a socket and plugging different amps in is worth the effort.

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.