While valves themselves are described under vacuum tube, these characteristics of valve as gain devices have direct implications for their use as audio amplifiers, notably that power amps need output transformers (OPTs) to translate a high output impedance high voltage low current signal into a lower voltage high current signal needed to drive modern low impedance loudspeakers (cf. transistors and FETs (which are relatively low voltage devices but able to carry large currents directly).
Another consequence is that since the output of one stage is often at ~100 V offset from the input of the next stage, direct coupling is normally not possible and each stage needs to couple via a capacitor (or exceptionally another transformer).
The capacitors and transformers have a secondary influence on the performance of the amplifier, but in particular the transformers add dramatically to the cost as well as size and weight.
The basic gain stage for a valve amplifier is the auto-biased common cathode stage, in which an anode resistor, the valve, and a cathode Resistor form a potential divider across the supply rails. The resistance of the valve varies as a function of the voltage on the grid, relative to the voltage on the cathode.
In the auto-bias configuration, the "operating point" is obtained by setting DC potential of the input grid at zero volts relative to ground via a high value "grid leak" resistor. The anode current is set by the value of the grid voltage relative to the cathode and this voltage is now dependent upon the value of the resistance selected for the cathode branch of the circuit.
The anode resistor acts as the load for the circuit and is typically order of 3-4 times the anode resistance of the tube type in use. The output from the circuit is the voltage at the junction between the anode and anode resistor. This output varies relative to changes in the input voltage and is a function of the voltage amplification of the tube "mu" and the values chosen for the various circuit elements.
Almost all audio preamplifier circuits are build using cascaded common cathode stages.
The signal is usually coupled from stage to stage via a coupling capacitor or a transformer,. although direct coupling is done in unusual cases.
The cathode resistor may or may not be bypassed with a capacitor. Feedback may also be applied to the cathode resistor.
A simple SET power amplifier can be constructed by cascading two stages, using an output transformer as the load.
Two power tubes (may be triodes or tetrodes) being differentially driven to form a push-pull output stage, driving a PP transformer load. This output stage makes much better use of the transformer core than the single ended output stage.
A "long tail" is a constant current (CC) load as the shared cathode feed to a differential pair. In theory the more constant the current, the more idealised the performance of the stage, i.e. a long tail linearises the differential stage.
The CC may be approximated by a resistor dropping a large voltage, or may be generated by an active circuit (either tube, transistor or FET based)
A pentode has an additional suppressor grid (g3) to eliminate the tetrode kink. This is used for improved performance rather than extra gain and is usually not accessible externally. Some of these tubes use aligned grids to minimise grid current, these are known as "beam tetrodes".
It was realised (and many pentodes were specifically designed to permit) that by strapping the screens to the grid/anode a tetrode/pentode just became a triode again, as such making these late design tubes very flexible. "Triode strapped" tetrodes are often used in modern amplifier designs that are optimised for quality rather than power output.
Due to the non-linearities, tetrode (and pentode) power tubes are invariably restricted to push-pull amplifiers that have significant NFB.
This type of design results in an extremely simple distortion spectrum comprising a monotonically decaying series of harmonics. Some consider this distortion characteristic is a factor in the attractiveness of the sound such designs produce. Compared with of modern alternative designs, SETs adopt a minimalist approach, and often have just two stages: each a simple triode gain stage. However, variations using some form of active current source or load have become commonplace. Such an active load is not normally considered a gain stage.
The typical valve using this topology in (rare) current commercial production is the 300B, which yields about 5 watts in SE mode. Rare amplifiers of this type use tubes such as the 211 or 845, capable of about 18 watts. These tubes are bright emitter transmitting tubes, and have thoriated tungsten filaments which glow like light bulbs when powered.
The pictures below are of a commercial SET amplifier, and also a prototype of a hobbyist amplifier.
One reason for SETs being (usually) limited to low power is the extreme difficulty (and consequent expense) of fabricating an output transformer that can handle the standing bias current in addition to the music signal, without saturating, while avoiding excessively large capacitive parasitics.
The cancellation of the differential output tubes also largely cancels the (dominant) even order harmonic distortion products of the output stage, resulting is a reduced THD, albeit dominated now by odd order harmonics and no longer monotonic.
NB however that due to tube variations the flux and even order distortion cancellations is not perfect, and PP OPTs will often still have a (reduced relative to an SE OPT) gap to prevent saturation.
Since the 1950s the vast majority of high quality valve amplifiers, and almost all higher power valve amplifiers have been of the 'push-pull' type.
Push-pull output stages can use triodes for lowest Z out and best linearity, but unlike the SET output stage often use tetrodes/pentodes which give greater gain and power. Many of the classic output tubes from the golden age (KT88, EL34, EL84) were specifically designed to be operated in either triode or tetrode modes at the designer's choice, and some of these PP amplifiers can be switched between these modes (when powered off). Post Williamson, the majority of commercial amplifiers have used tetrodes in the so-called "ultra-linear" configuration. This term is a misnomer because ultra-linear is a compromise between triode and tetrode configuration and gives performance between these two extremes, the pure triode mode remaining significantly more linear than so-called "ultra-linear" connected tetrodes.
Class A PP designs have no crossover distortion and distortion levels (excluding noise) reduce asymptotically to zero as signal amplitude is reduced. The effect of this is that class A amplifiers perform extremely well with music that has a large dynamic range, notably classical and acoustic music (among others) ... where the music is many decibels below the peak for most of the time. This is one reason audiophiles favour class A.
Class B and AB amplifiers are more efficient than class A, and can deliver higher power output levels from a given power supply and set of tubes.
However, the price for this is that they suffer from crossover distortion, the amplitude of which remains more or less constant regardless of the amplitude of the signal. This means that class AB1 and B amplifiers produce their lowest distortion figures at near maximum amplitude, and return extremely poor distortion performance at low levels, (regardless of how good the quoted figures (measured at high power) may be.) For this reason class B amplifiers are arguably well suited to eg pop and rock music where high sound levels predominate
Class AB1 and B amplifiers must rely on NFB to attempt to reduce the poor open loop distortion this configuration produces. Measured distortion spectra from such amplifiers show that peak distortion amplitude is dramatically reduced by NFB, but at the expense of the residual distortion having a much more complex, aharmonic and typically non monotonic distortion spectrum.
Most commercial amplifiers operate in Class AB1 (typically pentodes in the so-called ultra-linear configuration), seeking some compromise between these extremes, although a small number of amplifiers run in pure class A to avoid crossover distortion entirely.
The gallery shows a modern, hand wired, fully differential, no negative feedback, pure class A Audiophile amplifier of ~ 15 W, based on 6SN7 and KT88 tubes. This example is a home constructed hobbyist amplifier.
Output transformers are not perfect devices and will always introduce some odd harmonic distortion and amplitude variation with frequency to the output signal. In addition, the excessive phase response of transformers can be problematic when applying overall negative feedback to valve amplifiers, which often requires shelving within the feedback circuit to keep within the Nyquist stability criteria at high frequencies and thus avoid oscillation
However a more controversial side effect is that while the measured peak distortion is dramatically reduced, the distortion spectrum becomes more complex and often contains significant aharmonic components, which some allege audibly "sound worse", even though the design "measures better". It should be noted that in the extreme niches of valve amplifier design in particular there are often heated and religious opinions about the merits and demerits of NFB in the context of high-end audio designs operating in class A that have been designed to deliver very good open loop linearity.
A side effect of NFB is that the output impedance (Z out) of the amplifier is effectively reduced as a function of the level of NFB applied (which may vary as a function of frequency in some circuits). This has two important consequences :
The noise properties of tubes at audio frequencies can be modelled well by a perfect noiseless tube having a source of voltage noise in series with the grid. For the EF86 tube, for example, this voltage noise is specified (see e.g., the Valvo, Telefunken or Philips data sheets) as 2 microvolts integrated over a frequency range of approximately 25 Hz to 10 kHz. (This refers to the integrated noise, see below for the frequency dependence of the noise spectral density.) This equals the voltage noise of a 25 kΩ resistor. Thus, if the signal source has an impedance of 25 kΩ or more, the noise of the tube is actually smaller than the noise of the source. For a source of 25 kΩ, the noise generated by tube and source are the same, so the total noise power at the output of the amplifier is twice the noise power at the output of the perfect amplifier. The noise figure is then two, or 3 dB. For higher impedances, such as 250 kΩ, the EF86's voltage noise is 1/101/2 lower than the sources's own noise, and the noise figure is ~1 dB. For a low-impedance source of 250 Ω, on the other hand, the noise contribution of the tube is 10 times larger than the signal source, and the noise figure is approximately ten, or 10 dB.
To obtain low noise figure, the impedance of the source can be increased by a transformer. This is eventually limited by the input capacity of the tube, which sets a limit on how high the signal impedance can be made if a certain bandwidth is desired.
The noise voltage density of a given tube is a function of frequency. At frequencies above 10 kHz or so, it is basically constant ("white noise"). White noise is often expressed by an equivalent noise resistance, which is defined as the resistance which produces the same voltage noise as present at the tube input. For triodes, it is approximately (2-3)/gm, where gm is the transconductivity. For pentodes, it is higher, about (5-7)/gm. Tubes with high gm thus tend to have lower noise at high frequencies.
In the audio frequency range (below 1-100 kHz), "1/f" noise becomes dominant, which rises like 1/f. Thus, tubes with low noise at high frequency do not necessarily have low noise in the audio frequency range. For special low noise audio tubes, the frequency at which 1/f noise takes over is reduced as far as possible, maybe to something like a kilohertz. It can be reduced by choosing very pure materials for the cathode nickel, and running the tube at an optimized (generally low) anode current.
For high-end audio, where cost is not the primary consideration, valve amplifiers have remained popular and indeed during the 1990s made a commercial resurgence.
Today's circuits in most cases remain similar to circuits from the golden age, however they benefit from dramatic advances in ancillary component quality (notably capacitors) as well as general progress across the electronics industry which gives today's designers a far more powerful insight into how the circuit behaves internally. Many push pull products also tend to be more powerful than products from the golden age, reflecting the needs of inefficient but high quality modern speakers, and the market segment they are targeting.
Taking advantage of modern components (sometimes including transistorised regulated power supplies etc) today's high-end audio amplifiers deliver extremely low levels of noise, distortion and coloration needed to complement today's high quality source material, Frequency response is expected to be essentially flat across the audio band (20–20,000 Hz) to within a fraction of a dB, requiring -3 dB points typically 10–70,000 Hz for power amps (>100,000 Hz for preamps due to there being no OPT) with extremely low distortion levels.
However a small but intense demand remains for tube preamps and filter circuits for studio microphone amplifiers, audiophile phono stages in particular, and exceptionally for active crossovers.
However the far east never abandoned valves, and especially the SET circuit, indeed the extreme interest in all things audiophile in Japan and other far eastern countries sustained great interest in this approach. At the time this was perhaps as much a reflection of eastern philosophy and attitudes as of technical considerations, although the different types of music in the very distinct cultures of the era may also have been significant.
Since the 1990s a niche market has developed again in the west for low power commercial SET amplifies (~ 7 watt class or even lower) single-ended triode, notably using the 300B in recent years which while undoubtedly an excellent tube has taken on a life (and price) of its own driven largely by what can only be called fashion. Even lower power amplifiers based on other vintage valve types such as 2A3 and 45 are also found.
Even more rarely, higher powered SETs are produced commercially, usually using the 211 or 845 transmitting tubes, which are able to deliver ~ 20 watts in SE, but operate at ~ 1000 V making them unacceptably dangerous for many domestic situations. Notable amplifiers in this class are those from Audio Note corporation (designed in Japan), including the "Ongaku", voted amplifier of the year during the late 1990s. This market is very small in commercial terms, yet a very small number of hand built products do sell at very high prices ($10,000 and up, and up). There is always an extreme, and very extreme SET is the Wavac 833 - arguably the world's most expensive hi-fi amplifier, delivering around 150 watts using an 833A tube.
Aside from this Wavac and a very few other high power SETs, SET amplifiers usually need to be carefully paired with unusually sensitive speakers, notably horns and full range drivers such as those made by Klipsch and Lowther, which invariably have their own quirks, offsetting their advantages of very high efficiency and minimalism.
A side effect has been to drive the needed power levels beyond the capability of simple SE circuits, consequently the mainstream for commercial valve hi-fi power amplifiers has since the 1970s moved mainly to Class AB1 push pull (PP) circuits (usually using Tetrode/pentodes, and thus feedback (NFB)).
A minority of commercial PP production continues to use pure class A, or can sometimes be switched between class A and AB; Another (partly overlapping) subset of commercial PP production uses (or can be switched into "triode mode".
Major manufacturers in the PP audiophile market today include:
Indeed not only can hobbyist construct products that equal commercial production, they have a number of advantages, such as:
Selecting a preferred vintage tube version is a luxury denied to commercial producers who have to restrict themselves to tube types and brands that are still in volume production, such that they have a secure supply (and low costs).
One picture below shows circuit constructed using "standard" modern industrial parts (630 V MKP capacitors/metal film resistors). One advantage a hobbyist has over a commercial producer is the ability to use higher quality parts that are not reliably available in production volumes (or at a commercially viable cost price). For example the "silver top getter" sylvania brown base 6SN7's in use in the external picture date from the 1960s.
Another picture shows the exact same circuit constructor using Russian Military production Teflon capacitors and non inductive planar film resistors, of the same nominal values.
The wiring of a commercial amplifier is also shown for comparison
The main problem with very high power SETs is constructing output transformers able to sustain the bias current and resultant flux density without core saturation, at the same time as maintaining the desired wide bandwidth. This problem becomes more difficult and expensive as power levels are increased.
Another problem is that the voltages for such such amplifiers often pass well beyond 1 kV , and such potentially lethal voltages form an effective disincentive to commercial products of this type.
Many modern commercial amplifiers (and some hobbyist constructions) place multiple pairs of output tubes in parallel to increase power while using the same readily obtainable tube types, operating from the same (reasonably) low voltage supplies. Another beneficial side effect of this is that the output impedance of the tubes is reduced as normal when paralleling resistances and thus the turns ratio needed from the transformer is reduced, making it easier to construct a wide bandwidth transformer.
High power commercial products tend to use arrays of standard tubes (e.g. EL34, KT88) in the parallel push-pull (PPP) configuration instead (e.g. Jadis, Audio Research).
Some home constructed amplifiers use high power transmitting tubes (e.g. 813) to yield up to 100 watts (or beyond) per tube pair in class AB1.
One approach to avoid the problems of OPTs is to avoid the OPT entirely, and direct couple the gain device to the output (as is done with most transistor amplifiers for example). Some designs without output transformers (OTLs) were produced by Julius Futterman in the 1960s and '70s, and more recently in different embodiments by others, in an attempt to overcome the problems of transformers.
To do this one must deal with the characteristics of the valve, low current capability for a given anode voltage; consequently very particular valves must be employed. If done with care and the correct configuration, reasonable efficiency and moderate Zout (damping factor) can be achieved.
These effects mean that OTLs have selective speaker load requirements, just like any other amplifier. Generally a speaker of at least 8 ohms is required, although larger OTLs are often quite comfortable with 4 ohm loads. Electrostatic speakers (often considered difficult to drive) often work especially well with OTLs.
The more recent and more successful OTL circuits employ an output circuit generally known as a Circlotron. The Circlotron has about one-half the output impedance of the Futterman-style (totem-pole) circuits; far more important is the fact that the Circlotron is fully symmetrical and does not require large amounts of feedback to reduce output impedance and distortion. The result overcomes the biggest hurdle suffered by the prior art: stability. Thus modern OTLs can be as stable as any amplifier made. Successful embodiments use the 6AS7G and the Russian 6C33-CB power triodes.
A common myth is that a failure (short) in one of the output valves may result in the loudspeaker being connected directly across the power supply, and thus being destroyed (possibly quite spectacularly). In practice, the older Futterman-style amplifiers have been known to damage speakers, not out of shorts but out of oscillation. The Circlotron amplifiers often feature direct-coupled outputs, but proper engineering (with a few well-placed fuses) ensures that damage to a speaker is entirely within the same realm as conventional amplifiers. Certainly it is nearly impossible for the speaker to wind up being across the power supply!
OTLs have been a growing niche, as the Circlotron is so successful that modern OTLs are often more reliable, better performing, better sounding and less expensive than many transformer-coupled approaches. In recent years, the popularity of single-ended triode amplifiers has dramatically increased the range of loudspeakers suitable for OTLs.
Once more there are some safety issues associated with direct drive for electrostatic loudspeakers which in extremis may use transmitting tubes operating at over 1 kV. Such systems are potentially lethal.