I am always amused when I read in a manufacturer’s marketing description that they don’t use op amps in their audio designs – when what they actually mean is they don’t use the IC versions of them. Customers many times make this same mistake when they ask me “do you use op amps in your designs?” Most haven’t any clue exactly what an op amp is but they’ve read somewhere they shouldn’t like them. This is a dilemma for me when I answer it because I want to say “no” but the truth is “yes” but we don’t use IC versions of op amps – and that’s what you’re trying to stay away from – and rightfully so.

I thought it might be helpful to start a new mini series on these issues and get some answers for folks they can actually arm themselves with to be better informed. Are IC’s bad? What are discretes? What’s an op amp?

In the series we’ll make the distinction between discrete’s and IC’s, we’ll cover why designers would use one vs. the other, we’ll cover common mode rejection – what it is and why it’s important – and along the way I’ll explain what an op amp is, how they are designed, what their value is and why people use them. If there’s time and interest I’ll walk you through designing one yourself in our imagination lab.

Tomorrow we’re going to start the series out with a bit of history – a story about a man’s man. A hard driving, chain smoking, kamikaze skiing farm boy made good who changed the world as we know it forever.

Tomorrow, meet the Mayor of Silicon valley.

When we first started PS Audio back in the dark ages of the early 1970′s there was no such things as email, the World Wide Web or, for that matter, personal computers either. The IBM PC was launched a decade after PS Audio was founded and the computers of the day were big cabinet sized beasts that you could communicate with through a teletype machine – a remarkable device in itself because it was a network connected electric typewriter – but that’s another story.

In high-end audio tubes ruled the day and pioneers in solid state designs like Bob Carver with his Phase Linear products, Bob Hafler with Dynaco, James Bongiorno with SAE and PS with its phono preamp and line stage were amongst a small handful of solid state devotees who would continue to forge ahead in the face of a lot of opposition and naysayers of the day.

Now let’s travel back in time a bit more recalling our posts about the invention of the transistor and its rise to fame in the 1950′s. One of the more flamboyant of its inventors, William Shockley, left Bell Labs and founded his own company bearing his name. Shockley had a great nose for sniffing out brilliant talent and inspiring that talent to follow him and join his company. Unfortunately Shockley was also incredibly hard to work for and apparently a lousy manager – not a winning combination for a successful business and true to form, Shockley’s semiconductor business was never a success. One of the men he hired, Robert Noyce, was later to become one of the inventors of the integrated circuit and the microprocessor – two related inventions many would argue has had a far more profound impact on the world than even Shockley’s co-invention of the transistor itself – and decades later change the face of high-end audio.

Noyce, a physicist, was not happy working for Shockley but also had a temperament of trying to fit in and not make waves: keep it steady and make it work. Others around him were not of the same mind and in the mid 1950′s convinced Noyce, Gordon Moore and six other engineers working for Shockley to mutiny and leave the company to form their own. The outcome became known as Fairchild Semiconductor named after the owner, Sherman Fairchild – a company that built cameras and airplanes.

While at Fairchild the business flourished building individual transistors mainly for military and space programs. To keep history in perspective, this was the time of the Cold War, Sputnik, the great space race and the growth of the computer industry – although there was no such thing as a personal computer at the time (Bill Gates was 2 years old at this point).

During these heady days of discovery and emerging industries the pace was fever pitched and discovery was big – but actually figuring out how to fabricate these new devices was equally critical and Fairchild hired an extraordinary technologist that could figure out how to actually make the products: Andy Grove. Noyce was the idea man – prolific in his inventions and ideas. He absolutely loved spewing out note after note of ideas, but was never much to follow through with them. Gordon Moore and Andy Grove worked tirelessly at implementing Noyce’s ideas – and to make Fairchild a success.

One of the ideas Noyce and Grove worked on was a means of connecting together more than one transistor on a piece of silicon to form a circuit that was “integrated” meaning the multi-component circuit was together in one package. Companies like Fairchild had been making integrated circuits called hybrids for years – but these were actual tiny handmade circuits put together with painstaking labor that were then packaged in a single enclosure. No one had fabricated an actual circuit one one piece of silicon or germanium, the material transistors are made of.

Their project to integrate a circuit onto a single piece of transistor material was a side project – a novelty they hoped would work and if it did – they hoped it could be applied in such a way as to interest potential customers. Then came a big surprise. Another smaller company, Texas Instruments, announced and patented the world’s first integrated circuit. It was the brainchild of Jack Kilby an engineer at TI. It was crude, it was based entirely on germanium but it worked and for this effort, Kilby was eventually awarded the Nobel prize. Panic struck at Fairchild as the engineers struggled to get second best and, a year later in 1960, Fairchild introduced the world’s first silicon integrated circuit (a flip flop) and we were off to the races. The Fairchild integrated circuit solved many of the limiting factors of the TI version and much of what they invented is still used today.

The story meanders through the building of solid state memory IC’s at Fairchild, the departure of Noyce, Grove and Moore to form Intel Corporation and the subsequent invention of the microprocessor by Ted Hoff (at the insistence of Noyce) working to show a Japanese calculator company a better way to build a hand held calculator. Noyce himself was a great manager and believed in empowering employees at every step of the process in building a company. Stock options, employee participation and startup mentalities in Silicon Valley were mostly inventions of Robert Noyce who became know as the Mayor of Silicon Valley over time.

Our story stays the path with the integrated circuit and how it applies to audio amplification. It wouldn’t be for another 30 years where the paths of Intel’s microprocessor, the integrated circuit and high-end audio cross paths once again with digital audio.

• Member Quote

In yesterday’s post I mistakenly called the great David Hafler “Bob”. Oops, I think I had “Bob” on the brain. Sorry. But speaking of history, I think the stories of the people in any industry are perhaps more interesting to me than the actual products and technologies they design. I hope you’ll indulge me yet one more.

Yesterday I told you the story of Robert Noyce the man behind the microchip or integrated circuit and today I want to tell you a little of the story of the man who designed a part of just abut any audio product you have ever listened to. Bob Widlar.

In 1973 PS Audio launched our very first product: a phono preamplifier. That preamplifier used a passive RIAA network sandwiched between two linear amplifiers – both of which were IC op amps. The op amps we used at the time were called the 709 and were designed by Bob Widlar at Fairchild Semiconductor which, as you’ll remember from yesterday, was the company run by Bob Noyce, Gordon Moore and Andy Grove, soon to start Intel. During Widlar’s tenure at Fairchild he battled the growing forces of digital in favor of analog once telling Gordon Moore “anyone can count to one!” Widler felt it was beneath him to consider anything not analog and in those days, analog was king.

Widlar was a rebel by anyone’s standards. The tales of his antics, mostly drunken rip snorting antics coupled with a bohemian lifestyle and a disdain for anything orderly, are those of legend. Unhappy with his first linear IC op amp, the 701, he locked himself in a room for 170 hours and reappeared with the legendary 709 op amp as a result. Among his accomplishments Widlar invented the push pull output stage, emitter degeneration to match components, the active current source, use of FET’s inside of linear IC’s, coupling PNP’s with NPN’s in linear designs, the integrated circuit regulator, the idea of an external capacitor to trim the compensation. In fact, it can perhaps be safely said that during the 60′s and 70′s when analog was king, this man single handedly designed more than half of either the actual circuits or the technologies within them for the entire industry.

When we entered into the fray in the early 1970′s the choices for operational amplifiers were extremely limited: there was the 709, the 301 and the 741 as the main contenders. The 709 and the 301 were both designed by Widler and the 741 was a copy of the 301 design just done poorly. A quick audition between those three candidates was eye opening. The 301 and the 741 were slow, cumbersome dogs while the 709 was brilliant.

It was an unfortunate fact then (and still is today) that audio designers didn’t listen to their designs they just designed with meters and textbooks. Had they actually listened to their creations as designers in the high-end have done for years, much of the great and forever precious music created during those same heady days would sound good even today. But alas, the vast majority of mixing consoles in the recording industry were designed with the ubiquitous 301 or the even worse 741 and music through those early op amps was harsh, bright and congested. So too were the recordings they made.

So during this amazing time of converging lines we had the birth of the IC, the birth of analog linear circuits, the explosive growth of the consumer electronics market as well as high-end audio and an explosion of musical energy the likes of which haven’t been repeated many times throughout the course of history. The energy was high, the rush to succeed was great and in the middle of it all was Bob Widler.

If you have listened to a piece of solid state consumer electronics from today to 30 years ago and every moment in between, you have listened to and enjoyed music through the designs of perhaps the greatest pioneer and inventor the history of analog electronics has or will ever know. Widler died at an early age no doubt due in part to his binge drinking and crazed lifestyle but while here he touched the lives of every person who loves music and that’s saying a lot.

Tomorrow, let’s find out what this mysterious op amp is.

I promised yesterday we’d get started understanding op amps. I know some of you already get them so I hope you can tolerate some simple explanations.

This subject is actually very interesting because this class of amplification device is at the core of almost everything in audio today and while simple on the surface, there’s lots to know. Therefore I will break this up into a bunch of little posts to help you assimilate the info one day at a time. First let’s start with a bit of history.

Op Amp stands for Operational Amplifier and despite the fact it permeates nearly every audio design in solid state electronics it was never designed to be an audio amplification device. In fact, operational amplifiers were designed to work as the core building blocks of analog computers.

“What’s an analog computer? I thought that computers were all digital?” Just like everything we think of as new digital wonders they all started with analog versions and, of course, there are many types but a slide rule is a good example of one. The electronic equivalent or a slide rule is what op amps were originally charged with doing and they did it well.

Here’s an example of how we might make a calculation using an op amp. Because op amps are essentially perfect voltage amplifiers you can easily set their gain or multiplication factor with a simple set of resistors. Let’s say I want to multiply a number by 10. I simply set the gain of my op amp to a gain of 10. Now when I input a voltage, let’s say 1 volt, the output of the op amp is 10 volts. Volia! We just did a math function and all I need to do is connect a meter to the output to know the answer. This sounds too simple to be valuable because you can do the math in your head but now let’s try a harder math problem. Multiply Pi 3.414 times 8.215 and the op amp merrily (and instantly) spits out 28.04601 volts. You can divide, add, subtract and perform complex math of any kind. Limiting you say? Consider that when we sent men to the moon the calculations needed were performed on slide rules (mechanical analog computers) – and I’ll bet setting foot on the moon took some heavy math!

So op amps using vacuum tubes and then transistors were running computers years ago but then something happened: Bob Widlar (from yesterday’s post) integrated this circuit onto a single piece of silicon, then packaged this perfect voltage amplifier into a small and, soon to be, affordable package with 8 pins for a single op amp and 14 pins for a dual op amp. This changed everything for every engineer in the world including, you guessed it, audio designers.

The popularity of the packaged IC op amp was immediate and overwhelming. The original Fairchild 701 and later their 709 had a two year waiting period for advance orders so swamped were they with orders – and they were $100 each! By the late 1960′s and early 1970′s when most audio designers starting using these little gems, the price had dropped to around $1 each. So imagine purchasing two perfect voltage amplifiers in one package for $0.50 each. This was the Garden of Eden and we all took a bite out of that apple.

Why? Well obviously the low cost but perhaps most important is the ability to use a building block rather than having to build the block itself. There are only 5 pins to connect on an op amp and two of those connect up to the battery or power supply. This meant that you could design a preamplifier for an audio product by connecting a few resistors up to three pins – two inputs and one output and you had a product. Design time for the schematic? Less than 1 minute and you were guaranteed it world work.

Enticing? You bet. Tomorrow let’s look a little deeper.

• Member Quote

Yesterday I explained why use of the (then) new category of IC op amp was so darned attractive back in the 1970′s and still is today: easy to design with, reliable consistent performance and low cost. Good combinations!

Something else was also attractive to high-end audio designers like Stan and I – they were more like tubes in that they used a standard pinout for their enclosure. If you’ve ever played with tubes you know that not every tube sounds the same even though they are the same style. A 12AX7 tube is perhaps the most popular audio tube ever made. It fits nicely into a socket and it’s easy enough for designers and consumers alike to pop in different brands of 12AX7 tubes and listen to the differences – that’s because the entire amplifying device is contained in one enclosure, just like an op amp.

Now, for the first time in solid state history, audio designers could (if they had a mind to – and most didn’t) pop in any number of op amps into an IC socket and listen to the differences. This took the tube swapping idea to all new heights. With tubes you’re pretty much stuck with swapping out like tubes – 12AX7 was pretty much all you could throw in – but with these new op amps, the sky was the limit. The convention for packages was standardized: 8 pins DIPS (Dual Inline Package meaning two parallel rows of 4 pins) were used for single op amps and 14 pin DIPS were used for dual op amps and all adhered to a standard pin configuration. Wow.

But that’s when the trouble started for those of us who actually listened to our designs in addition to watching our meters. The audible difference in a simple preamplifier or phono stage between a 709, a 301 or a 741 were jaw dropping. Holy baloney! And it got worse: Motorola 709′s sounded decided different than National 709′s and they advertised the same specs! What was happening?

Let’s take a look at a schematic of the internal construction of the famous Widlar designed 709, the first truly great op amp.

Now, let’s compare that schematic to the ubiquitous 741 (not designed by Widlar)

Notice any difference? Of course you do they are very different designs – and yes they had very different specs as well – but the point is these seemingly simple building blocks were internally very, very different and they sounded different.

Here is a schematic of a modern day favorite for audio designers and found in many designs, the NE5532. Notice also that as time goes on in the progression of designs the number of components increases? Just look at the amount of parts in the 709 compared to the 5532.

OK, so tomorrow, let’s show you how all these complicated looking schematics can be easily understood and broken down into simple terms. I think you’ll find tomorrow’s post interesting and easy.

Let’s start today’s post by showing you how easy it is to design an amplifier circuit with an op amp.

When an op amp is used in a circuit as an amplifier, all you need is two resistors, the op amp and a couple of batteries. You can have the op amp amplify as either an inverting or non-inverting amplifier – meaning that whatever signal you put in can have its polarity or phase intact or flipped over – the same way you can reverse the +/- of your speaker terminals.

So let’s roll our sleeves up and design a preamplifier together with an op amp. We will want to have 20dB of gain in our op amp, meaning whatever we put into it will come out ten times louder.

So here is what an op amp looks like in a schematic: it’s a simple triangle with two audio inputs, one output and the two power supply inputs.

The two audio inputs are marked + (non-inverting if we want to preserve the phase) and – (inverting – if we want to flip the phase over) and the output is, well, the output. Simple eh? You have no idea just how simple this thing is. Let’s keep moving.

Here is a drawing of an IC op amp package – a diagram of where each of these terminals are located on the actual device itself. You’ll see some other stuff marked on this one – like offset null – don’t worry, we won’t need to use that or have anything to do with it.

The -V and the +V are the battery inputs and between the two batteries we might use (like a couple of 9 volts) we have what’s known as ground – or the center between the two power supplies. Here’s what the two batteries tied together might look like and ground is in the middle between the two where it is written “0V”.

The picture I grabbed shows 6 volts – but imagine 9 instead – which BTW was exactly the circuit I first used to power the prototype of the PS Audio phono preamplifier – yep, two 9 volt batteries and a couple of op amps in a Roi Tan cigar box for a chassis.

OK, now let’s make our preamplifier. I think let’s start out making an inverting version, then we’ll go tackle a non-inverting one. As I mentioned all we need is two resistors and because the gain of an op amp is set by the ratio of the two resistors, then all I need to do is have one resistor 10 times bigger than the other. So I could use a 1K (1000 Ohms) and a 10K (10,000 Ohms). These are just little bits I could buy at Radio Shack if I wanted – maybe a whole $0.10 each. Here’s what our circuit would look like.

We would put the 1K on the in and the 10K between the out and the in. The little triangle symbol you see on the + input is ground (the meeting point between the two batteries). That’s it. There’s not much more. Just connect an RCA input to the input side and another RCA connector to the output and you’re done. Whatever you plug into the input will come out 10 times bigger on the output RCA without any distortion or problem.

If I used quality parts, a PC board to hold everything, a power supply instead of a battery and a fancy chassis I have a commercial preamp. Want a volume control? Just stick a potentiometer on the input and you have it. Want it “high-end”? Use a good op amp and a better power supply and fancy connectors.

But maybe we want to have a non inverting amplifier instead. OK, all we have to do is change where we connect ground up and where the input is. Here, take a look.

See? R1 is our 1K resistor and R2 is our 10K. Simple, no?

But this is a single ended preamplifier – where we can only put in the output of an RCA connector. Let’s get super fancy and make it a balanced input amplifier – where you would put in the output of a balanced cable with an XLR on its output – balanced out. We’re all familiar with that.

Here’s what that would look like – please ignore the values because I just grabbed this off of the internet.

The little symbol that looks like a rake is ground. Cold is the one side of the balanced signal and hot the other.

OK, this is getting lengthy and the purpose isn’t to teach you how to design an op amp but just how danged simple it is – and I am not simplifying anything here.

Tomorrow let’s see what’s inside of these little buggers that makes everything so easy and so widely used.