Solid state mythology
In my previous article regarding the nature of "sterile" sound, I speculated on the mechanisms by which a digital modeler might fail to live up to its potential as a substitute for the archetypal tube amp that it's supposed to model. While writing that article I realized that I have some new insights regarding solid-state amplification.
I'm sure that you're all aware of the conventional wisdom regarding why tube amps sound good and solid-state amps sound bad: that tube amps exhibit "soft clipping" which sounds (based upon the inevitable visualization of the rounded-corner waveform offered as an illustration of soft clipping) "smooth", while solid-state amps exhibit "hard clipping" which sounds (again, using a square-cornered waveform as illustration) "harsh".
The problem with this kind of visual analogy regarding sounds is that our ears are not visual. We don't hear wave shapes; only frequencies. And the relationship between shape and harmonic content - except in a few very simple and limited cases - is not useful as an analytic tool.
I started thinking about solid state mythology recently when I found a Heathkit guitar amp in one of the local music shops. This particular amp is an exemplary specimen of an amp that I built from a kit over forty years ago, played all through high school and college, and eventually sold toward a down-payment on my first house. (Not that the amp was worth much; the down payment, though, was very small by today's standards and I didn't have much in the way of liquid assets.) I always had fond memories of that amp, and often referred to it during discussions of "the gear I wished I hadn't sold".
To be honest, my expectations may be lower than yours. When I was a kid I never played - much less owned - any of the amps that have since become benchmark standards for classic vintage amps. My amps were branded with names like Lafayette, Univox, Silvertone and Heathkit.
But this particular amp - a Heathkit TA-16 - is one that I owned and played for about a decade. I remember the amp as having a warm clean sound at lower volumes and a sweet singing quality when pushed. This amp is - in case you haven't already guess - a solid-state amp.
If you're a modern guitarist fed on internet "wisdom", you've no doubt noticed the clash between my description of this amp and the conventional wisdom regarding solid-state amps. Solid-state amps are supposed to sound "sterile" at lower volumes and are supposed to have "harsh and ugly" breakup when pushed, right?
So, what are the possible explanations for this anomaly? I see three possibilities:
- I have no idea what I'm talking about, and my ears are shot.
- This particular amp somehow magically escaped the rules by which all solid-state amps are supposed to function.
- Conventional internet wisdom is a self-perpetuating myth.
For obvious reasons, I'm going to dismiss option (1). I know what I hear and try diligently to understand the reasons. I'm not always right, but I do the best I can within the (hopefully expanding) limits of my own understanding. I've never been one to blindly repeat what I'm told; I have to know why.
I'm also going to dismiss option (2). While I'm enchanted by the notion of magic in storytelling, I've never in my nearly sixty years in this world encountered any physical phenomenon that can't be adequately explained in non-magical terms.
That leaves option (3): that the `net perpetuates a lot of ill-informed half-truths, one of which is that "solid-state guitar amps always sound bad."
So, if the general "rule" doesn't apply to this amp, what makes it special? Or, from a different perspective: what gives solid-state amps in general a poor reputation for their sound quality as guitar amplifiers?
The things that make a guitar amplifier musically interesting are, in no particular order:
- distortion that varies in amount according to the intensity of the note being played,
- some variation in the quality of the sound as the played note decays, and
- a shaping of the frequency spectrum of the instrument's sound.
The last point, also known as EQ, is entirely independent of distortion and all dynamic behaviors (although the presence of EQ before distortion will have an effect upon the feel of the amp).
Distortion, in and of itself, can be produced by both tube and solid-state devices. Contrary to the most popular `net meme regarding solid-state guitar amps, transistors in and of themselves do not have a linear transfer function bounded only by the limits of the power-supply rails. This is, by the way, the essence of the case against solid-state amps: that the amplifier stages respond linearly until the output bumps up against the limits of the power-supply voltages. That's not necessarily true.
Unfortunately, the above-described scenario can be true, which is sufficient to give legs to this particular myth. A lot of solid-state guitar amps are designed in such a way that some of their amplification stages will exhibit exactly the behavior described by the myth. That's not intentional, but rather a side-effect of either constraints upon the designer or a disconnect between the intended and actual use of the amp.
It took many years for the notion of "good" distortion to take root in the minds of guitar-amp designers. Back in the late `60s and early `70s, tubes were being phased out in favor of "improved" designs based upon transistors. This came at a time when big rock concerts were just starting to take off, long before large-scale sound reinforcement systems had become commonplace in music venues.
In those early days of big rock concerts, very loud amps became very useful. No doubt you've seen the videos that illustrate just how ineffective a few Vox AC-30s were in a venue full of screaming fans when The Beatles played Shea Stadium. It was only months later that Vox came up with the obvious answer in the form of the Super Beatle high-powered solid-state amp. Other manufacturers followed suit. There are, of course, counterexamples in the form of high-powered tube amps like the Marshall Major, the (original) Sunn Model T, the Ampeg SVT and the Highwatt 100-watters. But the industry trend at the time was to embrace solid-state design for its "cool, maintenance-free operation."
Slightly lagging the influx of solid-state designs was the introduction of integrated circuits and - in particular - op-amp chips. These were important, I think, for a slightly different reason than transistors.
If you've ever designed your own audio amplifiers - guitar or otherwise - using a variety of different technologies, I think you'll recognize that it's pretty darned difficult to throw together a tube circuit that won't pass a signal. In technical terms, tubes have a very wide operating range. Tubes are also very robust in that you can violate their operational limits for short periods of time and they'll normally bounce right back.
Transistors are more difficult to use in amplifier designs. If you miscalculate, chances are very good that - in the best case - your amp won't pass a signal at all. In the worst case, because transistors are not as forgiving of electrical abuse as tubes, you're literally going to have to pay for your mistake by buying new parts. There were other problems, too, but the gist is that it's comparatively difficult to design a solid-state amplifier using transistors.
That problem was partially solved with the advent of op-amp chips. Now the design of your voltage-amplifier (e.g. preamp) stages came down to a bit of simple math that was almost entirely independent of the chosen chip. If you needed a certain amount of gain you needed only to do a bit of trivial algebra and pick out a couple of resistors whose values fit a calculated ratio. You didn't even have to design around a particular op-amp chip; that choice could literally be left to the bean-counters. Other functional blocks - like tone controls, tremolo, distortion, and reverb - could literally be baked into the design of the amplifier using a "cookbook" approach. Believe me, this was a lot easier than designing an amp around individual transistors. The designer still had to come up with a suitable power stage, but even that could be tackled using a cookbook approach to obtain acceptable results. (Power amps on a chip didn't make an appearance until well after this era in guitar-amp design had passed.)
The problem with op-amp based preamps is that their behavior exactly matched the internet meme of the amplifier that's completely linear until the output bumps up against a supply rail. This is, in fact, the inherent nature of an op-amp based voltage amplifier. The relationship between the amplifier's input and output voltage is completely determined by the ratio of two resistors. If that ratio is ten, then the amplifier will have a gain of ten. If the ratio of the resistors is one hundred, then the amplifier will have a gain of one hundred. The only time the ratio of the amplifier's input and output voltage doesn't precisely match the ratio of those two resistors is when the output voltage can't swing any further because it's already at the limit of the amplifier's (internal) power supply voltage.
So if the meme is true in this case, what does that mean for solid-state amps...? It means that guitar amps based around op-amp chips are susceptible to the problems described by the meme. The problems are not a certainty. There are design techniques available to avoid or even eliminate these problems, but such techniques always add cost and complexity. Given that a lot of solid-state amps fit the lower tiers of price-differentiated products, it's no surprise that corners are cut and the unsuspecting (and presumed uninformed) consumer is left to pay for a parsimonious purchase with his disappointment in the product's performance.
The benefits of a guitar amplifier design based around op-amp chips (without the necessary design improvements to impart desirable sonic qualities) accrue entirely to the manufacturer. Design cost is reduced using a coookbook approach and component selection is non-critical and can be made solely on a low-cost basis.
But that's not the only approach...
Contrary to popular belief, transistors are not inherently linear devices. What powers the `net meme of "square-cornered transistor clipping" is not the presence of transistors rather than tubes, but rather the use of transistors to form (as in the case of the op-amp designs noted above) circuits stabilized and regulated by large amounts of negative feedback.
You're probably familiar with negative feedback as it applies to the power stage of a tube guitar amplifier. A modest amount of feedback has the effect of "tightening up" the sound. Reduce that feedback and the amp becomes "looser" and "less tame". Increase that feedback and the same amp turns "brittle" or "sterile".
So if too much negative feedback is the real problem, is it possible to build a transistor guitar amp without the problems caused by negative feedback? Or, more precisely, is is possible to avoid the problems caused by too much negative feedback? (No amplifier can function without some negative feedback to stabilize the operating point of the tubes or transistors.) The answer is an unqualified "yes"; my Heathkit amp is an existence proof.
The schematic for this Heathkit amp shows that the design is literally a transistor analogue of budget tube-driven amps of the era during which this amp was created. Each single-ended preamp stage is self-biased by a single resistor, just like a triode voltage amplifier in a tube amp. The volume and tone controls have the same position and topology as in other amps of the era (e.g. the Silvertone 148x series). The reverb uses a single-stage driver, a single-stage recovery amp and a passive mixer, just like the Fender blackface and silverface amps. The tremolo circuit differs from its common tube-amp counterpart only in the presence of a bulb-driver stage to accommodate the current of the incandescent lamp used instead of a neon lamp (because of the lower operating voltage). The biggest difference is in the output stage, which of course doesn't use an output transformer; the basic topology, though, mirrors the tube amp in its use of a modest amount of negative feedback.
The similarities are not entirely topological, either. Just like a tube amp, each amplifier stage of the solid-state Heathkit amp distorts smoothly and asymmetrically. Just like a tube amp, the operating point of each stage may be affected by the presence of asymmetric signals. Just like a tube amp, the power supply may sag under load and affect not only the power delivered by the output stage, but also the operating points of all the preamp stages.
The only thing found in a tube amp that's not present in this amp is an output transformer. It's arguable, of course, that the output transformer contributes something to the tone and behavior of a tube amp. Whether that contribution is essential to a tube-amp sound is - in my mind at least - debatable. My favorite tube amps were always the ones that had "big iron" output transformers. The bigger the transformer, the less its contribution to the tone of the tube amp. So a completely transformerless output stage doesn't detract from my notion of a great guitar-amp sound.
Just to be clear: the preceding analysis of the schematic is given not to theorize about how this amp might sound, but rather to explain why it sounds the way it does. This amp behaves, for all intents and purposes, as we'd expect a decent tube-driven guitar amp to behave. There's a warmth at low volumes; you can hit the front end harder to get more preamp distortion; you can push the volume to bring out a singing quality. This amp - or rather this amp's design, because there's absolutely nothing special about this particular instance of the amp - defies the internet meme regarding "cold and harsh" solid-state amp sounds because it wasn't designed in the manner of the amps (and they're definitely out there as described) that fueled the meme.
This Heathkit, at least according to the criteria by which I evaluate amps, is a pretty darned good amp. But that's not the message I hope you'll take away from this article. What I hope you've learned is that behind every rule of thumb you find regarding MI gear, there's probably a specific case from which the generalization was originally formed. It's important to know not only the rule of thumb, but also its limitations.