I came into the audio business as tubes were going out. I was in the last class in my college where tube theory was mandatory for EE students, and the studios I started out in were in the process of discarding their old tube gear for newer transistor gear
Much of the old gear they discarded I took home with me, because I liked the sound of it. Some of the tube gear gave a transparency and a “rightness” that I just didn’t hear in the harsh-sounding early transistor gear.
So with the renaissance in tube equipment going on today, I would expect to be delighted. I’m not though, because it’s a mixed bag. Too many people are using tube gear and even designing tube gear without understanding what the gear can or should sound like, and without any understanding of the vast body of knowledge that has been built up over the years.
This makes me sad, because I like tube gear and believe that there is still a place for it in the studio—even if it is hard to find at times.
‘Distortion’ is probably the most misused word in the industry. People are always using the words “clipping” and “distortion” interchangably. Distortion is anything that changes a signal in any way other than changing its level. “Clipping” is when a piece of equipment is being asked to produce too high a level, and is therefore “clipping” off the top of the waveform—the nice sine waves turn into squareish-looking signals on a scope, and the result is that the signal sounds harsh and buzzy. So, clipping is an instance of distortion, but only one of many possible instances.
It may not seem like it, but when you are using an equalizer, you are deliberately distorting a signal by changing the frequency spectrum. Changes in frequency response are one sort of distortion and nonlinearities are another one.
We can generally think of nonlinearities as having two symptoms. First, when you put a pure tone into a system, it generates harmonics. This is, needless to say, harmonic distortion. We can measure a lot of things about this distortion; we can measure the spectrum of these harmonics and their level.
Now, a THD (total harmonic distortion) measurement tells you very little, because it doesn’t tell you which harmonics are dominant. Something with a high THD that produces a lot of low-order harmonics may sound rich and pleasant, while something with the same THD that produces high-order even harmonics may sound harsh and gritty.
And of course when you start dealing with transient signals, the harmonics produced are different (for which discovery we can thank Marshall Leach). By looking at the distortion spectrum we can tell an awful lot about how a piece of equipment is going to sound, believe it or not.
Another symptom of nonlinearity is intermodulation distortion. That is, when there are two tones sent into a piece of equipment, they generate beat notes at their sum and difference frequencies. On complex music, high IMD tends to make all the instruments blend together into a general mush.
The truth of the matter, though, is that good models of distortion characteristics are hard to make, so you’re really going to have to use your ears. But I just want to point out firmly that all equipment produces some distortion on any signal, and that distortion colors the way the signal sounds. Nothing is completely transparent sounding, and the job of the engineer is to decide what sort of coloration to avoid and what sort of coloration can be lived with.
Why tubes are different
Tubes themselves have different distortion characteristics than transistors, but you can’t just lump all tube circuits together. There are a lot of different tubes, and a lot of different ways they can be used.
The first significant difference we see is that tube stages operate at much higher voltages than transistors; it’s routine to see 250V or even 450V on the plate of a small signal tube. This means you can swing comparatively high voltages with a tube stage, and even given the additional noise (caused primarily by the higher temperatures at which we’re working) you can get fantastic amounts of headroom from a properly designed circuit.
The second significant difference we see is that the tube requires hardly any current on the input; it has an extremely high-impedance input. This is an incredible advantage for DI boxes that have to be driven by guitar pick-ups and the like. But it also means that a conventional 600 ohm source requires a transformer input stage to step the voltage up and the current down, for good noise performance.
Likewise most tube circuits can’t produce a lot of current, so they need an output transformer to drive a low-impedance load. These transformers are the primary cause of coloration or distortion in most tube circuits, not the tubes themselves. The characteristic sound of a lot of that vintage tube gear comes almost entirely from the sort of transformers used in them.
Incidentally, the one exception to this general rule is the piece of tube equipment that most studio folks are most familiar with: the guitar amp. Being a part of the whole instrumental system, guitar amps are designed and voiced to produce one or more types of distortion or coloration at different stages from input to output. While the people making mic preamps worked to make them as clean as possible, which left the transformers as the primary sources of coloration, the guys making guitar amps worked to make them dirty so the electronics’ coloration swamped that of the transformers in many cases.
The simplest tube circuit we see is a gain stage, just one tube acting as an amplifier to increase the voltage of a signal. Now, depending on the tube chosen, the designer can choose the sound. A simple 12AX7 triode stage tends to have a mellow, lush sort of sound, but we can use a pentode to get higher gain. The distortion characteristic of the pentode is different, producing more of a crisp edge on the music.
For years I have seen fights over which sort of circuit is better, and the fact that the fights have gone on for so long indicates that they both have their place. But you, as the user, need to know that there is a subtle but important difference in the coloration.
There are a lot of other different possible circuits, though. Voltage follower (i.e. cathode follower) circuits give you the ability to drive a low-impedance load and often appear at the output of a tube device, and they have a radically different sound to them. Totem pole or SRPP stages sound even more different, though you aren’t apt to encounter them on typical everyday gear.
You need to realize, though, that this same sort of coloration is also there in transistor systems. The process of circuit designers manipulating the coloration by choosing different circuit designs is not specific to tubes. Transistor circuits also have their own characteristic sound, and it’s in many ways less subtle than the characteristic sound of the better tube gear.
In the past, engineers were forced to use tube systems because that’s all there was. Today we use tube systems because we like the lack of transistor-sounding distortion.
If you listen, you will find that many of those old tube preamps (if properly maintained) sound extremely clean and transparent. This is not because they are adding “tube coloration,” it’s because they aren’t adding “transistor coloration.” While there are cases where “tube coloration” might turn out to be a good thing, you need to realize that the actual coloration these things induce is very subtle.
A typical mic preamp
Figure 1 shows a typical if somewhat stripped down tube microphone preamplifier. It’s not the best performing design but a typical and reasonable one that’s worth inspecting for demonstration purposes. You could build it if you wanted to (I did in fact throw it together over my lunch break just to make sure I wasn’t overlooking anything) although I neglect to show the power supply, which is half the secret to a good desgn.
1. The Transformer Stage
You’ll see that the input transformer takes in the low impedance microphone output, and steps the voltage up to match the high impedance input of the tube more closely. The input transformer is the major source of coloration in the mic preamp, believe it or not, and it adds a lot more coloration than either of the tube stages. Most of this is a very characteristic sound that varies a lot between transformers.
In addition, much of the noise in the preamp comes from the transformer. The small windings on the transformer secondary have a relatively high resistance and therefore an inherently high thermal noise. On top of this, the transformer core is also apt to saturate before the tube stages bottom out, so the transformer also affects the amount of headroom available.
The secret to a good preamp is the input transformer as much as anything. It needs windings with very low capacitance for wide bandwidth, a large core that doesn’t saturate easily and restrict headroom and bass, and heavy enough secondary windings to reduce the noise level.
In this case, we use a 1:15 step-up transformer. Higher step-up ratios will give you more gain and reduce the noise contribution of the tube stage, but they also tend to be noisier themselves because of the added number of turns on the secondary. Balancing all of these factors for optimal performance is a difficult design decision.
2. The Voltage Amplifier Stage
The input stage is a single triode design that amplifies signals by a factor of about 70. The triode produces primarily low-order odd harmonic distortion, and if properly biased this coloration can be minimized so that it adds only a slight bit of lushness to the sound.
For a different characteristic sound we could use a pentode, which would give a brighter sound with more even harmonics. Or we could change the bias voltage by adjusting the value of the cathode resistor, which would change the coloration. Or we could use trickery like driving the plate from a constant current stage to further reduce coloration. This, however, is the most common configuration you’ll see.
3. The Cathode Follower Stage
The output stage is a cathode follower. This is an amplifier configuration that has no voltage gain, but only increases the current drive. It permits the relatively low current output from the tube plate to drive a low-impedance input.
In the old days this would invariably be followed with a 1:1 transformer for a 600 ohm balanced input, though with modern equipment that is more trouble than it is worth, and that additional transformer is just one more source of coloration.
The cathode follower also adds some coloration, although it’s more of a gritty sound caused by even harmonics. Some people like it, others don’t, and it gets more pronounced the lower-value a cathode resistor you use to drive a lower impedance output. In this case we choose a fairly high value and get a very neutral sound at the expense of current drive. Other designs may make other compromises.
Some folks don’t like the cathode follower at all, and will use an additional gain stage with a step-down transformer, claiming that the coloration this provides is preferable.
Of course, if everybody liked the same sound there wouldn’t be so many options.
Some fake (and real) tube gear
There is currently a lot of gear out there that uses a tube stage as a gimmick. Most of these systems give you tube coloration overlaid with transistor coloration, giving you the worst of both worlds. On the other hand, there is also some really fantastic gear out there that produces a very airy and accurate sound.
The most popular of these hybrid devices, like the ART Tube MP and the PAiA TubeHead, generally have a transistor input stage followed by a tube stage that operates at a very low B+ (voltage). Often you’ll see a 12AT7 or even a 12AX7 with only 50V on the plate. This gives you a lot of nonlinearities because there is basically no portion of the tube curve that is linear down at these voltages.
The effect is interesting—the high end becomes slightly muffled and most of the detail in the midrange disappears. These devices can actually be useful because of the high IMD they produce. They tend to make sounds blend together and eliminate detail, which can occasionally be useful to add a sense of ensemble to the sound.
It’s not the sound that a normal tube preamp makes, and in fact the first time I heard one of these units I thought it was broken because it makes that sound that is clearly characteristic of a misbiased tube stage. It’s an interesting sound and a useful one, but the coloration is nothing like that of a conventional tube stage.
The Aphex Tubessence circuit is another interesting one. This one is unusual, with a 12AX7 run at a very low plate voltage but with a transistor current mirror in the plate circuit. This allows the tube to run much more linearly at the low plate voltage. The effect here is subtle, but again the coloration isn’t very much like that of a conventional tube amp. Interestingly, there is a feedback loop inside that can be adjusted for varying amounts of coloration, and I find it sounds better when set as clean as possible. Others disagree.
The Rocktron Saturator is another interesting gadget. It runs the tube with a much more linear 250V on the plate so it stays fairly neutral, but it drives the tube very hard to get into the clipping range. This idea is for this to simulate the saturation characteristic of analogue tape. It doesn’t, really, because the curves are quite different. But it still adds an interesting limiting effect that again you might like for special effects.
The Bellari mic preamps are completely different. They do have a transformer input and a 12AX7 gain stage running at 250V. However, the input transformer is a budget model that induces a different sort of coloration than a good input transformer would, especially on higher level signals that tend to saturate things. Still, it is the closest of the ones we have discussed to the classic tube preamp circuit.
A real surprise is the Peavey VMP-2, which for about $700 is actually a well-designed conventional tube dual mike preamp, with Jensen input transformers, a reasonable 12AX7-based circuit, no weird starved plate stuff or semiconductors in the signal path. I was amazed when I saw it, considering that you’d be hard pressed to build something like that for the price.
There are a lot of different tube circuits and there are a lot of different possible tube sounds. Anyone who tells you that something sounds “tubey” is confused, because there are so many different possible sounds that can be obtained from a tube design that anything in the world can be said to sound “tubey.” Many of these sounds have interesting coloration, and if done right they are free from transistor coloration.
All of these different circuits have merits of their own, and all are definitely worth exploring. Experiment and learn.