With apologies to Clint Eastwood and any fans of the Italian cowboy genre known as Spaghetti Westerns, let's take a look at what's good, bad and downright ugly about IC op amps and high end audio now that we're all experts on their design.
Remembering what we picked up in our last few posts: that op amps are not necessarily IC's - they are functional amplifier blocks that can be built from any type of amplification device such as a tube, a bipolar transistor, a FET, etc. - and that most of us incorrectly assume when we say "there's an op amp inside our equipment" that it must be the IC version. Now, I admit, chances are really good that in fact it probably is true but my point is let's no longer make that assumption. For example in all of our DACS, from the PerfectWave DAC to the DLIII, we use op amps but we don't use IC versions of op amps. There's a big difference and now let's jump in and find out why.
First, the good news. IC's have two things that discrete versions do not have and never can: uniformity and short paths between devices, Uniformity as it concerns temperature stability and parts matching is important. For example, remember our diff pair? The closer to identical these two transistors and their associated components are the better their performance - and you can't do better than putting them on one substrate as you do in an IC. Interconnection of devices is also optimized in an IC because the entire circuit is built on the smallest of scales and the distance between components isminuscule. Compare that with discretes who suffer from greater distances, individual packages, leads, solder connections, PCB traces - none of which is suffered by an IC.
In fact, with few exceptions IC's should be better than their discrete counterparts. Unfortunately, in the majority of cases they are not as they relate to high-end audio. There are multiple reasons why they are not and so now let's take a look at some of the bad.
IC's are particularly well suited to making transistors and diodes. They are not great at resistors and capacitors over a certain size and even then are very limited in the types you can use. They are also not great at higher voltages and perhaps the biggest damming factor of IC op amps, as applies to high-end audio, is that rarely are IC's ever designed with what high-end audio designers are looking for. Certainly there are several attempts at making good sounding audio op amps, such as National's LM4562, but even this great effort doesn't address many of the issues we're concerned with. Let's discuss what those might be.
First and foremost nearly all IC op amps try to get as close as possible to the ideal model of an op amp: that of having "infinite" gain. Remember in our discussion of how op amps work we showed you how feedback works? Feedback is where we take the output and feed it back into the unused input to set the gain of our amplifier. But what happens if you take the feedback off? In an op amp with "infinite" gain as soon as any signal comes into the input (even a tiny bit of noise) the output of the op amp slams to what we call a saturated state - totally clipping - 100% distortion. In fact, most op amps cannot operate at all without feedback. Why is that important in our quest for better sound?
Let's grab that as tomorrow's topic and then we'll move on to high voltage.
