Op amps vs. chips

September 26, 2021
 by Paul McGowan

I wince when people equate op-amps with chips.

An op-amp describes a topology, not a chip.

Yes, the op-amp topology is often built onto a silicon wafer, but we should be careful not to forever bind the two together unless we’re clear about the difference.

(If you’re technically minded there’s a great paper describing in detail the nature of op-amps from some of my engineering heroes you can read here or, for those less technical but still interested, my friend and audio superstar, Nelson Pass published this great and easier to understand work).

Op-amps have been around since the 1930s. They were originally built using vacuum tubes. The term itself, op-amp, is short for Operational Amplifier.

There are multiple types of op-amps, but for our purposes, we’ll focus on the standard voltage amplifier so often used in audio circuits.

A simple description of an op-amp is an analog circuit block that takes a differential voltage input and produces a single-ended voltage output. Here’s a simple diagram.


Where you see V1 and V2, those are inputs, and where you see Vout that’s the output. Rf is the feedback value.

Using this gain block one can easily make all sorts of amplifying circuits from balanced inputs, to phase swappers, to buffers, to phono preamplifiers.

(Here’s a link to an ancient paper showing how to make 41 audio projects using a chip op-amp. Note: the author uses a 741 op-amp in his designs. The 741 op-amp is perhaps the single worst choice in the entire world for building audio circuits. Perhaps there’s something worse, but in my near-50 years, I have yet to find one. Use instead the venerable NE 5532 op-amp, which sounds great).

Many of PS Audio’s products are based on op-amp topology though we rarely rely upon the chip versions because for best sound quality, we require more control over the amplifier’s performance. Instead, we often build discrete versions of the operational amplifier topology using higher voltage rails and hand-selected devices to best suit our needs.

The point of this post is to set the record straight. Op-amps are not necessarily chips and chips are not necessarily op-amps.

I’m just protecting the good name and reputation of a core fundamental gain block of audio.

Subscribe to Paul's Posts

44 comments on “Op amps vs. chips”

  1. So I went to the Nelson Spassky link for Dummies and half way through I dropped out on the basis of Sunlk Cost.

    Today I was sent a link explaining Quantitabe Easing. I gave up half way through.

    Yesterday I listened to three superfi headphones through three highly related headphone amps and couldn’t choose any as best.

    Am I alone? Life: I’m too old

    1. You’re not too old Peter and you don’t need to understand these circuit topologies if you’re not engineering minded. I don’t have any idea what quantitatable easing is but I’m going to look into the subject. It doesn’t sound sophomoric and I would not get down on myself for lack of understanding if I were you.

      The one part of your post that I think you need to look into is why your comparison of headphones and headphone amps yielded no discernible sonic differences. Perhaps you should think about visiting an audiologist and ENT in a medical environment not in a retail store to see exactly what the status of your audible range is at present. I am 75 and my hearing is definitely not what it used to be but I went back into headphones listening about two years ago and I quickly jumped up from a reasonable headphone and headphone amp combo to a much higher quality combination and the difference was staggering. Hopefully it’s only earwax that can be removed and correct your issue.

  2. So I went to the Nelso Pass link for Dummies and half way through I dropped out on the basis of not chasing after any Sunk Cost.

    Today I was sent a link explaining Quantitative Easing. I gave up half way through. I don’t need to understand everything any more. I trust people whose job it is.

    Yesterday I listened to three superfi headphones through three highly related headphone amps and couldn’t choose any as best.

    Am I alone? Hopeless?
    Life: I’m too old

    1. Aha! Stimpy2 It’s all down to Method. Previously I heard clear differences in mid-range phones; Senn 650, AKG 701, Hifiman 400s (best).

      Yesterday I used Keith Jarrett trio. Wrong! Harman have discovered that jazz trio does not distinguish fidelity well. Which might be why we love Waltz for Debbie at hifi shows?

      No, it takes big symphonic Eg Berlioz Telarc or female voice to discriminate.

    1. Hello Jazznut,

      I’ve been looking over a few external electronic crossovers that use transformers to go from unbalanced to balanced on some models. So I like it when a little ‘wiggle room’ is given when words like ‘usually’ are used. ✌️

      1. Good morning Mike!
        I don’t know if you’ve ever heard of these or not.
        But however, a company that’s called Rolls, makes a little black box, that they refer to, as a hum eliminator.
        It has a pare of transformers inside of it.
        But not only does it get the humming noise out of the audio signal, you can also get a balanced signal from an unbalanced signal.
        I guess this is why Rolls calls it the HE18 Hum Eliminator Buzz Off.
        I have four of them working for me.

          1. Hi again Mike!
            Thanks for your concern!
            And yes, things are just grate over here!
            What made me brake down and order four of those Rolls HE18 Hum Eliminator Buzz Off boxes, is this.
            Last year, I ordered a pare of JBL LSR310-S power subwoofers to add some bottom end to my quod of Avantone Pro CLA-10 studio monitors.
            They don’t have speaker level inputs on them.
            Just a pare of XLR inputs and outputs on both of them.
            They also have a pare of 1.4th inch inputs on them as well.
            I thought that I could get away with feeding them unbalanced audio signals.
            But I quickly discovered that, they don’t like unbalanced audio signals.
            And so, that Rolls box, was the solution to my problem with that.
            But also, at the same time, I found other uses for those boxes.
            To tell you the truth, they work really well!

          2. The original product I believe was called the HUM-X. There are several others besides the one John recommended as well. Not the best solution but it does work.

        1. Good afternoon, John.

          I recall an old cartoon (in Stereo Review, I think) of a guy connecting a black box to his stereo labeled “Glenn Gould Hum Filter”. It made it into a compartment of my brain that I refer to as “The Weird Stuff* Files”, apparently among the most robust long term memory in my aging slushware.

          *I actually use a scatological term for fecal material, but I will keep it family friendly in this public forum of Paul’s. Mr Rat can provide confirmation.

    2. No, not an op amp chip necessarily. But a balanced input which is often found on an op amp chip.

      One of the points I probably failed to get across is that opp amps aren’t necessarily chips. In many of our products we use op amps but not in chip form. Big difference actually.

      Take the M1200 monoblock amplifier. It has a balanced input but that’s accomplished through a differential input built around a vacuum tube. Or, take the M700 which also has a balanced input and it is based upon discrete MOSFETs.

      No chips.

      Balanced inputs.

      1. Thanks! So I get that what you did in the Stellars is much better than a chip based balancing, but the difference between what you did in the Stellar products and a true fully balanced design is, that in the latter, most of the circuit is doubled?

        I don’t confuse it with double mono, but I’m not sure what’s exactly needed for a completely balanced design of a piece of equipment.

          1. Thanks again!

            Often those differences don’t get clear for the customers, as they are not always mentioned in the technical description. A double mono design is obvious optically, but if no double mono, how to know?

            E.g. for the BHK pre you write „fully balanced input to output“, so I assume it’s having all the circuitry of a „true“ fully balanced design…but otherwise there’s no hint to it for a non technician I guess.

              1. Saw it and want to have a listen, but didn’t do yet.

                This band of him (other than e.g. the one with Miroslav Vitous and Roy Haynes) never had much to tell to me personally to be honest. I rather had the impression of them as of an all star band covering some famous tracks. Maybe this one’s better.

                1. It sounded really good through my iPhone and with Chick’s passing I want to have as much of his work as possible because to me he is one of the greatest pianists and Innovators I have ever heard. This is another extension of chicks genius which you may or may not like. I don’t like some of his work because I can’t understand it but this is definitely unique to my ears. The link will give you not only 15 seconds of each track but the entire recording so you can get a really good idea if you like it before you make a purchase.

                  I have seen Return to Forever live more than three times and each of the four musicians in the group on many other occasions especially Al Di Meola and Solo
                  recitals by Chick over the span of 50 years.

                  1. I heard it today on the stereo…didn’t like it, just as the other acoustic ones with Patituchi and Weckl…the music leaves me cold unfortunately. Just my taste 😉

                    I have about 35 Cores albums and I really like about 5-8 of them. The Trio Music ones, a few other acoustical albums and a few newer fusion ones (with Patituchi and Weckl).

                  2. Yes, no problem.
                    For me music needs the emotional touch or the groove or something special for the head. Much of his music falls in none of those categories for me. It’s still perfectly played jazz, but unfortunately doesn’t touch me. Many see this very differently and that’s fine.

  3. Paul, What is the triangle symbol that is light blue and has a + and – inside the triangle in the diagram that you posted. What is it called? A Gain Block? What circuitry is inside it? Is this what used to be built with vacuum tubes?

    1. Tony,

      The triangle is the op amp. Quite often drawn like that when it is chipped based, but also on some higher level block diagram schematics of discreet component op amps.

      I’ll let Paul go into the details of the circuitry inside.

      You could think of it as any other chip you may be familiar with. All that really matters are the goes inta’s and the goes outta’s. What’s in the middle you can’t change (if it’s a chip design op amp) unless you’re growing / creating your own.

      1. Yes, Mike’s correct. The triangle is the symbol on a schematic denoting an op amp.

        Inside could be any number of ways of building the op amp. The triangle does not denote a chip, but rather the op amp functional block. Inside every chip is a different circuit, or inside every operational functional gain block can be a different topology as long as it conforms to what I wrote about: a differential input and a single ended output.

      2. Mike, when I was in engineering school the triangular symbol was what was drawn on the blackboard the first day we discussed the topic of Op Amps and we were being Introduced to the theory of it’s basic circuit design and not IC chips. That’s the way my instructor led us in to the topic. More than likely that’s why you see the symbol regardless. The Nelson Pass paper seems less straightforward than the Walter Jung paper to me as a former engineer and student and I would say to you to take a look at that paper and go through it very slowly at the beginning where it may make more sense to you. I have seen Nelson on videos speaking to other engineers and he is so ingrained in his encyclopedic knowledge of audio engineering that he doesn’t present these papers In an academic way for students. It’s his writing style. The other paper is more textbook.

        I was sort of amazed that after many decades of not being an engineer that understanding this paper was so natural for me to recall and comprehend even with some higher order {???} math. Normally I don’t have much to contribute to the daily posts because it seems that you guys have so much more knowledge than me but when it comes to academic papers like this I feel at home.

        1. Hey stimpy2,

          I’m no design engineer, but have a strong background in electronics, and now that’s supplemented with optics. The op amp diagram just seemed normal to me… I’ll be checking out the papers later, just because it sounds like a good read. It may be a good thing to do on the plane tomorrow.

          As far as the so ‘much more knowledge’…. I put myself at the bottom of the heap compared to most who post here. I’ve gleaned room acoustic tips, isolation tips, and some AC power tips from reading here. Not necessarily exact instructions, but hints that I was able build upon. Some of the esoteric thinking topics get to me, but obviously there are many who like them.

          So I’m under no illusion that I have any sort of reference system, or that 2 channel in its current form will sound ‘live.’ Rather I take in the different perspectives – sometimes find the time to ‘tinker’…..
          – but most of all enjoy the music and 2 channel presentation I have – I still occasionally fall into the critical listening pit, and start to think what if… it’s taken some time and perseverance but now I’m much better at climbing out without falling back in…

          1. Great response Mike. I have a much better understanding of where your head is at now. I have intense school knowledge from 50 some odd years ago and worked as a digital engineer for only four years but my engineering school knowledge is so vast across the entire engineering spectrum that I couldn’t possibly be that knowledgeable about each type of engineering. I hardly ever did analog design although I have two friends from my former job that are masters of analog engineering. They seemed to gravitate in that direction where digital was like building blocks and timing diagrams for the most part and I found it easier to do this type of design. I was also the designated assembler language go to guy to compare microprocessor instruction sets and write certain routines for our projects using some of the first 8 bit microprocessors. Everyone has their own strengths and weaknesses and that’s why I commented about my staying out of certain conversations because I felt that I was not confident enough to throw my two cents in. A lot of the guys/gals who post here are so deeply entrenched in what’s going on with audio cables, crossover design, speaker construction, galvanic isolation, etc that I tend to shy away.

            I made a comment several weeks ago that I just was interested in enjoying the music now just like you and I don’t want to fall back into the pit either. So far so good. Any comments I make don’t go into real detail and are just remembrances from the past.

  4. As an aside, National never developed the NE5532 for audio usage. Having said that, I find it interesting that you seem to be advocating the NE5532 for usage in audio, when the choices for much better performing choices are on the market. Who wants the 10% overshoot, the slow recovery, the long ring time… in other words, why advocate a very poor performing 40 year old design, and then claim you prefer the much better sounding discrete choices you have made. Choices on the market can prove as solid a performer as even discrete. ESS in their reference designs and performance results material states the performance degradation when you utilize the NE5532, and is the bottom of the bottom of the barrel of choices to utilize. I would suggest rather, the OPA-1612 for audio usage. (however bean counters would prefer to spend $0.10 for the NE5532 than $3~$4 for the Burr Brown OPA-1612)

    1. The 5532 is an excellent sounding chip and one of the backbones of the audio industry. Our Studer 32 channel mixing console out of Switzerland is based completely on the 5532 as are more audio circuits than one can count.

      National makes some better performing chips and I am not suggesting we don’t want to look at those but the facts are the facts. The 5532 is the prerred workhorse of audio and is a great sounding chip. Some of National’s “audio specific” chips they tout sound like crap. I wouldn’t trust a National engineer to maker a sonic judgment.

      1. I stand behind what I have said. The NE5532 is a great choice for table-top radio performance. Have you forgotten you said this?

        Sounding harsh and compressed – July 19, 2012 – by Paul McGowan

        “So what does one hear when you use an IC op amp with all this gain? It really depends on the op amp. Bipolar based op amps like the classic NE5532, used by so many designers of audio equipment, can sound a bit brittle on the top end – transistory and slightly compressed overall. The best way I can describe it is impressive and powerful sounding always – but the harmonics and overtones are slightly etched sounding and overall dynamics sound a bit like trying to squeeze the music through cheesecloth. I know, this description is going to freak out the engineers reading this. Sorry. I just have spent too many hours listening to all these variations to really care.”

        Your assertion from 2012 is more accurate than 2021.

        1. No, I think we’re mixing apples and oranges. There’s no doubt the 5532 is not perfect. No op amp is regardless of how it’s made.

          I don’t remember when I wrote that but I suspect that I would have said that in reference to replacing the 5532s with discrete versions which in fAct sound considerably better.

          Our Studer analog board is stuffed full of 5532. I can hear them and don’t like what they do. Would I replace the 5532 to solve that? Actually, no, I would (and will) replace the entire board and chain to fix this problem.

          My point was that when it comes to chip op amps used in audio circuits, there’s probably no other chip used more than the 5532. I would also stand by my statement that in many cases, that venerable old circuit sounds better than many of the more modern ones.

          I would point out that isn’t an over riding proclomation. I know Darren’s found some newer ones that are far better and I am sure they exist. I just don’t spend much time playing with them.

          1. I don’t think I’m mixing fruits, I stated in my original post that some of the choices we have available on the market today can perform as well as discreet. Certainly the NE5532 can’t, but that’s my point. Why advocate poor performing, 40 year old designs, or use it as a benchmark when there are so many other good options to be had. I heard Vince Gill say, “thank God everyone doesn’t always like what Grandpa liked, or else there would be a line of 20 year olds at Grandma’s door looking to park and spark…”

            The market continues to utilize the NE5532 because of it’s low cost and the willingness of the consumer market to accept it as some sort of defacto-standard, and then fall back on the rational that since it has been used in mixers for decades then it’s great for reproduction, and that’s where the train runs off the tracks. Even you did it in this thread. I oft say that the creation of music is an art, but the reproduction of music is a science. That’s the apple vs. orange.

            The consumer market and the average Joe doesn’t have the ability to change what is given to them over the counter. Some, offer services to improve what has already been built with upgrades to very effectively create a much better listening experience. And until we as the consumer market speak out and say, “we don’t want the NE5532 in our equipment, we want what today’s technology can provide…” manufacturers will continue to build what they feel is acceptable. (or that the market will bear) There are some applications where discreet isn’t a good option. It’s time to let go of the NE5532. I offered a drop-in replacement, but it isn’t $0.10 either. Value can be found in other ways and is not only in low cost parts, but as you seek to do with the mission statement of your company, good sounding audio.

  5. In the 1960’s Electronics World had a series of articles on building a tube based opamp, I think it used 12au7’s but it was a long time ago. I priced it out and it would have cost $100 and that was a fortune for a high school student in 1964(?).

    I remember Burroughs used to have a line of tube based modules in the 50’s that could be configured as op amps but they were to expensive for most hobbyests.

  6. Great post today Paul. The papers you referenced were also classic for someone like me and I can’t wait to go through the140 page PDF file. When I was a student in 55 years ago, I never conceived that Op Amps would be such a crucial circuit topology in so many areas of electronics engineering as they are today. I really enjoy straight out posts like this one.

    1. Good morning Neil!
      Right after I read Paul’s posting, I looked all of it up on the internet this morning.
      When Paul said something about the very first Op amp being made in the 1930’s using tubes, that peaked my curiosity.
      And so, I looked up the history of OP amps.
      I read a very lengthy write up about the whole entire thing.
      It actually dates all the way back to the mid to late 1800’s.
      They said something about doing the DA AD conversion, started with water mains instead of any electronic means to get it done.
      And how they used that to make an electronic mottle of what they were doing with the water thing.

  7. Right on Paul, it’s good to clarify this basic tool. It’s also great that you identify the core element should NOT be demonized based on cost minimized IC devices but many discrete approaches such as your team’s work or Burson, Sonic Imagery, and local Sparkos Labs sound excellent. Thumbs up Paul!

  8. I appreciate this perspective, and the references! My introduction to op amps came from testing discrete 990s at Boulder Amplifiers (on Sterling Dr.) around 1990. I had a part-time job there while in college, and over the course of a couple years managed to pick the brain of the owner, Jeff Nelson, on analog design. He helped me refine my microphone preamp design (1:10 input tranformers and twin gain stages using NE5534s, like a twin servo design without the servos or the 990s) based on ideas he himself had learned from Deane Jensen. It holds its own to this day. I found my way to instrumentation amplifiers (three op amps) as a building block from there.

Leave a Reply

Stop by for a tour:
Mon-Fri, 8:30am-5pm MST

4865 Sterling Dr.
Boulder, CO 80301

Join the hi-fi family

Stop by for a tour:
4865 Sterling Dr.
Boulder, CO 80301

Join the hi-fi family

linkedin facebook pinterest youtube rss twitter instagram facebook-blank rss-blank linkedin-blank pinterest youtube twitter instagram