Tube amplifier

A tube amplifier is an electronic amplifier in which electron tubes are used to amplify low-frequency electrical signals . With the advent of semiconductor electronics , tube amplifiers for audio amplifiers were largely replaced by amplifier circuits with semiconductor components ( transistors ). However, electron tubes are still used today in guitar amplifiers and some high-fidelity amplifiers, especially in the so-called high-end segment.

Push-pull stereo HiFi tube amplifier without power supply electronics - the device is equipped with 4 tubes per channel, one EM83 tube is used for the level display. The very large output transformers for loudspeaker adjustment can be clearly seen.

Circuit principles of audio tube amplifiers

Two basic examples of LF amplifier circuit technology with tubes provide an insight into how they work.

Single ended amplifier mode A

Single ended audio amplifier with pentode
Characteristic curve of an amplifier tube with different operating points

The circuit diagram shows a single-ended amplifier with an end pentode in a cathode-base circuit, both the positive and the negative half-wave of the input signal are processed by this one tube. The only possibility of an approximately linear amplification of the two signal parts is to select the operating point A on the U g - I a characteristic curve (input characteristic curve) of the amplifier tube, which lies in the middle of the straight line part of the characteristic curve, which results in an unfavorable high quiescent current through the tube and a results in less advantageous efficiency of the amplifier - the classification of the operating mode of the amplifier relates to the position of this operating point .

The coupling capacitor C1 separates DC voltage components of the signal to be amplified and thus prevents a shift in the operating point. The very high resistance R1 serves to keep the control grid at ground potential in terms of DC voltage. The cathode resistor R2 is responsible for generating the grid bias ; its value determines the operating point of the tube. The cathode becomes positive due to the cathode current flowing through R2 and the associated voltage drop compared to the grid, the resulting negative grid bias is automatically regulated depending on the cathode current (static negative feedback for stabilizing the operating point). The resistor R2 should for an A amplifier with a pentode EL84 z. B. have the value 135 Ω, which generates a grid bias of −7.2 volts. An incorrectly dimensioned cathode resistor R2 results in asymmetrical operation in which one half-wave of the output signal is limited significantly earlier than the other. This reduces the usable linear dynamic range and creates distortions.

The capacitor C2 is used to bridge the resistor R2 in terms of AC voltage. The limit angular frequency 1 / (R2 · C2) of the negative feedback defines the limit below which the gain is reduced. If C2 is omitted, the amplifier is also fed back in terms of AC voltage, which reduces both the gain and the distortion.

The output transformer separates the loudspeaker from the high anode voltage and transforms the high output impedance of the output tube ( 5.2 kOhm for an EL84 end pentode in single-ended A mode) to the low impedance value of a dynamic loudspeaker.

In contrast to the circuit diagram above, the screen grid of the end tube to limit the screen grid current is usually connected to the anode voltage via a resistor. It serves to increase the dynamic range and the efficiency by maintaining the anode current even with low anode voltages through its electric field.

Advantages of the Class A tube single-ended principle:

• The simplest circuit concept with few components in the signal path.
• No phase splitting required as with the push-pull output stage.
• No power takeover distortion at low volume.

• Lack of linearity if (as in the above circuit) no negative feedback is used and the output transformer is premagnetized on one side by the continuously flowing anode current. The desire for a reduced distortion factor led to the invention of negative feedback .
• Low efficiency and high power dissipation .
• High demands on the residual ripple of the anode voltage, especially with headphone amplifiers (hum sensitivity) .
• A complex and expensive output transformer is required for HiFi operation .

Push-pull amplifier mode AB

Push-pull LF tube amplifier

The example on the right shows the circuit diagram of a powerful and typical tube audio amplifier, the two end pentode EL34 after the push-pull principle work - as opposed to the extremely efficiency-low single-ended amplifier share in the output stage two tubes strengthening work by a tube for the positive voltages, the other tube is responsible for the negative voltages, which leads to an improved power yield: if one output tube conducts, the other blocks and vice versa, the push-pull amplifier can therefore assume different operating points than class A. The far-reaching functional principle was developed in 1912 by the Canadian electronics engineer Edwin H. Colpitts .
In addition to the possibilities of class A operation (high quiescent current) and class B operation ( crossover distortion at the zero crossing of the signal), the operating mode with the favorable position AB of the operating point on the above Ug / Ia tubes has predominantly been used for the audio push-pull amplifier -Characteristic enforced: a low quiescent current allows the tubes to work in class A mode with small signal amplitudes , with increasing modulation the amplifier gradually changes to class B mode, with full modulation the tubes work completely in mode B , which leads to significantly higher levels Output power and better efficiency results.

The power tubes are controlled with a circuit design by the British tube specialist DTN Williamson, which has become known as the split load phase inverter with driver stage : since there are no complementary tube types, analogous to semiconductor components, this part of the circuit must ensure the phase reversal of the input signal - the Both control grids of the output tubes require two amplified signals with the same amplitude , but mirror images that are symmetrical to ground .

The first triode system of the ECC83 manages the necessary voltage amplification of the input signal, the second triode of the ECC83 is the actual concertina (or cathodyn ) phase splitter . At the cathode and the anode of the concertina triode, the two signals in phase opposition are decoupled and passed on to the two driver trodes of the ECC85, which drive the two EL34 output tubes in push-pull.

An output transformer, which, due to the push-pull method, must have a center tap on the primary side, ensures that the signals are brought together and the power is adapted to low-impedance loudspeakers. An adjustable overall negative feedback from the secondary side of the transformer to the cathode of the first triode linearizes the frequency response and reduces the distortion factor .

The advantages of the tube push-pull concept are class AB

• higher output power with good efficiency,
• comparatively simple and inexpensive output transformer,
• which is not premagnetized on one side and therefore generates less distortion as well
• more favorable tolerance to residual ripples in the anode voltage.

• the much more complex circuit design in which a signal is split into positive and negative half-waves , which are then amplified separately and only added back to the overall signal in the transformer
• the AB working point leads to takeover distortions .

Special form ironless output stage

A modern, permanently dynamic loudspeaker with its low impedance can be easily adapted to the low-impedance output of a transistor output stage. The high-impedance output stage of an audio amplifier tube, however, almost always requires a power adjustment with a low frequency transformer ( transformer ).

An exception were concepts of "coreless" OTL audio tube amplifiers (OTL stands for O utput T ransformer L ess) in which the output transformer was saved for reasons of cost, which could however not free himself for various reasons later in the famous peopleâ worked the low-frequency output tube directly onto a cantilever loudspeaker with its high impedance of 2000 ohms, and in the 1950s and 1960s the Philips company later used a circuit with different tube types and the 600-800 ohm loudspeakers required for this in a number of tube radios and televisions . The output signal is picked up via a coupling capacitor and fed directly to the loudspeaker. Philips abandoned the concept after a few years: The high-impedance voice coils often suffered damage (interruption) and no external second speaker could be connected.

Since the shortcomings of the transformer impaired the signal quality, the American Julius Futterman developed the first hi-fi tube amplifier without an output transformer for operation with regular low-impedance loudspeakers in the mid-1950s , the conception of which was later used in the OTL amplifier series from NYAL ( New York Audio Labs ) was continued.

More modern concepts, in which a large number of relatively low-resistance current control tubes are usually connected in parallel in the output stage, still occupy a niche in the market. These amplifiers are very ineffective in terms of their available power, the tube consumption (service life of the output tubes used) and the power dissipation.

The reason - the too great difference in impedance - can be explained using a 50 W amplifier: A suitable 4 Ω loudspeaker draws the current

${\ displaystyle I _ {\ rm {eff}} = {\ sqrt {\ frac {P} {R}}} = {\ sqrt {\ frac {50 {\ rm {W}}} {4 {\ rm {\ Omega}}}}} = 3 {,} 5 {\ rm {A}}}$

on. This corresponds to a maximum current of 5 A, which a power transistor can process without any problems. The voltage on the loudspeaker is

${\ displaystyle U _ {\ rm {eff}} = {\ sqrt {P \ cdot R}} = {\ sqrt {50 {\ rm {W}} \ cdot 4 {\ rm {\ Omega}}}} = 14 {\ rm {V}}}$

The voltage on the loudspeaker must therefore fluctuate between the maximum values ​​of −20 V and +20 V. This is ideal for the direct connection of a transistor output stage with complementary transistors (combination of NPN and PNP). These numerical values ​​show the fundamental weakness of electron tubes for "ironless" power amplifiers:

• The maximum permissible cathode current of power tubes is 1 A, apart from large tubes for radio transmitters. The necessary total current of 5 A can only be achieved by connecting a sufficient number of units in parallel.
• In order to “draw” this cathode current, the tube must be supplied with at least 150 V anode voltage. More would be better. But only 20 V of this is used for the loudspeaker, the rest contributes to the considerable power loss of the anodes and to the poor efficiency:
${\ displaystyle P _ {\ rm {Loss}} = (150 {\ rm {V}} - 20 {\ rm {V) \ cdot 5 {\ rm {A}} = 650 {\ rm {W}}}} }$

Avoidance of unwanted oscillations

Electron tubes work in a very wide frequency range from direct current to around 2000 MHz and can therefore tend to undesirably oscillate in the high frequency range if they are not constructed appropriately. It is characterized by a tendency to oscillate in the manner of the Huth-Kühn circuit at very high frequencies. Whether unwanted vibrations occur is particularly dependent on the wire lengths on the control grid and on the anode (see line theory ): the shorter, the higher the resonance frequency. Since the tube amplifies less and less with increasing frequency, the minimum amplification will be undershot at some point and vibrations can no longer occur.

The operating point or the modulation have an influence on the slope: Often at −15 V (AB setting without signal) no oscillation is observed because the gain is too low. At −1 V the characteristic curve is steeper and the circuit oscillates. This means that from a certain volume in the rhythm of the NF produce such audible distortions that do not exist at low levels.

Detection is difficult without a spectrometer or oscilloscope because the frequency of the wild oscillations is usually not even roughly known. A series connection of a glow lamp with a 50 pF capacitor can be placed between anode and ground as an indicator . The capacitive resistance of the capacitor is too high with NF and the glow lamp does not light up. However, it flickers at frequencies above a few megahertz.

As a countermeasure, damping resistors connected in series to the supply lines have proven effective:

• a 1 kΩ resistor directly on the control grid, which very effectively dampens the resonant circuit quality of the wire. At high frequencies (from 100 kHz) this resistance, however, forms an undesirable RC low-pass with the input capacitance increased by the Miller effect . This low-pass effect can be reduced by a throttle connected in parallel.
• a 100 Ω resistor directly on the screen grid for pentodes

Comparison of tubes and semiconductors

Even after the invention of the transistor , the electron tube was for years without an alternative as an active control element in all areas of electronics. The very low transit frequency , the noise and the temperature problems of the early germanium transistor types limited the possible uses of the transistor. With the use of the semiconductor material silicon , a consistent further development of the silicon transistor and its numerous advantages, the semiconductor components bipolar transistor , field effect transistor and MOSFET increasingly replaced the tube in almost all electronic fields of application.

Compared with the electron tube, the transistor has very decisive advantages to offer, including: a. especially in the following areas: small dimensions, low weight and low price level, mechanical insensitivity, simple and undemanding power supply paired with high efficiency, extremely long service life, hardly any changes in characteristics over the entire usage cycle, extremely good electrical values ​​of current transistor types through permanent and ongoing research.

One of the most serious disadvantages when using electron tubes is the need for a complex high-voltage supply, as there are hardly any tubes that can deliver a significant anode current (and thus a significant output power) at a low supply voltage of around 50 V and also function with little distortion in this area . The considerable power required for heating the cathode and the comparatively large power loss also affect the circuit environment. In power tubes in particular, various chemical-physical processes, especially in the cathode, trigger an accelerated aging process, which is why the tube has to be replaced after a certain period of operation (see also electron tube ).

With tube amplifiers, it is not the tubes, but the output transformer that determine the upper and lower limit frequency of the amplifier. Even with special winding techniques, the range from 30 Hz to 15 kHz (−3 dB) can hardly be exceeded. Transistor amplifiers can easily work well beyond the LF range (e.g. 5 Hz to 100 kHz) with far lower tolerances (e.g. −0.5 dB) because they do not require a transformer.

The output resistance should be as low as possible in order to dampen unwanted resonances from the loudspeaker. Because of the much stronger negative feedback and the different circuit topology, it is in principle significantly smaller in transistor amplifiers than in amplifiers with an output transformer. Because of its phase shift, at least at the limits of the bandwidth, the negative feedback in tube amplifiers must be dimensioned significantly smaller. The circuit topology also plays a major role: Tube output stages are usually implemented in a cathode-base circuit, which naturally has a high output resistance. In contrast, transistor output stages are usually designed as a collector circuit (emitter follower), since this has a low output resistance even without external negative feedback.

There is also an important difference when the connection between amplifier and loudspeaker is disturbed:

• If this connection is accidentally interrupted with a tube amplifier at high AF output, the resulting high induction voltage will destroy the output transformer and / or damage the output tubes.
• Semiconductor amplifiers that have become insensitive to critical impedances (short circuit, inductive loads, capacitive loads) have been common since the 1980s at the latest. For this purpose, the voltage, the current, a power loss approximation and the junction temperature of the current driver are monitored, which in all cases means that the output stage cannot be damaged. This is called SOA operation (Safe Operation Area).

Effective protection of tube power amplifiers against inductive overvoltages, on the other hand, is very difficult and cannot be implemented at all using tube technology, which is why it is hardly used in practice (a snubber network is by no means sufficient protection). On the other hand, overcurrent protection of transistor output stages can be implemented very easily and inexpensively, which is why it is available in almost every transistor amplifier.

Compared to semiconductor components, whose active areas are located in a very small space within a solid, electron tubes are more resistant to short-term electrical overload, radioactivity and electromagnetic pulse (EMP) due to their mechanical construction . Neither of these has anything to do with a hi-fi amplifier.

HiFi amplifier

The hi-fi era began in the 1950s with highly developed tube amplifiers such as the English Mullard 5-10 , which aimed at the most precise possible electroacoustic reproduction of sound events. A few years later, the unstoppable triumph of semiconductor technology in the electronics industry was becoming increasingly apparent, displacing tube electronics from the market with the exception of a few niche applications - but it took some time for the first transistor-based hi-fi amplifiers to reach the high acoustic quality level of the tube amplifier could tie in.

It was not until the mid-1990s that high-quality tube audio amplifiers became socially acceptable again in high-end circles. In a comparison of the technical data, hopelessly inferior to their semiconductor-based competitors in many areas, tube amplifiers often achieve better results when their sonic properties are assessed subjectively. With today's high level of semiconductor technology, however, some critics refer to this as a euphemism of the listener or as a "pleasant falsification" of the sound; the manufacturers of music recordings are responsible for the pleasant sound, not the manufacturers of the playback devices. The latter should be as neutral as possible.

On the other hand, often only sound engineers with their qualified and appropriately trained hearing ability are able to perceive the subtle differences between very good and excellent amplifiers in a differentiated and reproducible manner - they often attest high-quality tube amplifiers an excellent acoustic quality. In an elaborate individual project at the Technical University of Berlin, an attempt was made to give these results an objective basis. Metrological investigations on the Black Cat tube amplifier built especially for the project showed an extremely low numerical value of the difference tone factor , which means that the amplifier has a low tendency to non-linear distortion , the presence of which causes an unpleasant hearing perception even with the smallest signal components.

Due to their typical characteristics, tubes generate different distortion spectra (spectrum of harmonics ) than semiconductors: Triodes, pentodes and field effect transistors have a characteristic curve that is square to the first approximation, while bipolar transistors have a characteristic curve that is exponential in the first approximation. The often heard claim that triodes and field effect transistors as undesirable distortion products mainly produce even-numbered multiples of the fundamental frequency, while bipolar transistors and pentodes produce mainly odd-numbered multiples of the fundamental frequency, is wrong, as measurements on basic circuits show.

In all basic amplifier circuits, regardless of the active component used, primarily harmonics with double the fundamental frequency and only a few higher-order harmonics arise. Even-numbered multiples of the fundamental frequency sound rather “warm” and “brightening” for many listeners, whereas the odd-numbered partial tones are said to have a rather pointed sound.

Instrumental amplifier

In the classical stage amplifiers for the rock music instruments typical electric guitar and electric bass tube technology has for various reasons claims to this day: the special behavior ( soft clipping ) of the tube amplifier, its gradually increasing slide into the signal distortion at The targeted use of an overdrive is to be regarded as an inseparable part of the instrument, which gives the e-instrument sound its individual sound character and does not serve to amplify the tones generated by the instrument as precisely as possible.

The American Aspen Pittman has published in his book The Tube Amp Book (en.) A fairly complete collection of the circuit diagrams of historical guitar tube amplifiers. Many of the circuit details mentioned here can also be found in different variants.

Negative feedback

The electronic negative feedback ( NFB-negative feedback ) is an important development for the circuit technology of hi-fi amplifiers by the American electronics engineer Harold Stephen Black from 1927: Part of the output signal is added in reverse phase to the input signal, which means that the responsible person is carefully dimensioned Switching elements distortion behavior and frequency linearity of the amplifier are positively influenced - depending on the extent of the negative feedback, however, the voltage gain is reduced by the voltages in phase opposition. When designing tube stage amplifiers for electric guitars, the designers and a. also for this reason often on the corrective, but gain-reducing effect of an over-all negative feedback from the secondary side of the output transformer over the entire amplifier branch - the inimitable sound of the special distortion of the amplifier electronics is almost desirable here, the well-known guitar amplifier Vox AC30 is a famous example.

The cathode resistance in the amplifier stages creates its own direct current - alternating current counter-coupling, which ensures a stable operating point, but also a gain reduction - the first ECC83 triode of the above push-pull amplifier is a good example of this. By connecting the cathode resistor in parallel with a capacitor, the negative feedback and thus the gain reduction can be reduced, since its capacitive resistance diverts the AC voltage to ground - an example of this variant of DC negative feedback is the cathode circuit in the above single-ended amplifier.

In many cases, the circuit concept of high-quality hi-fi amplifiers provides for cross-stage voltage negative feedback from the secondary winding of the output transformer to the input tube - since the transformer causes a frequency-dependent phase shift of the signal, excessive negative feedback easily creates the risk of undesired feedback .

With the advent of power pentodes and the increasing mass production of output transformers, negative feedback was experimented with on the screen grid. The aim of the development was to set a triode-like operating characteristic while maintaining the advantages typical of pentodes such as high amplification and adequate efficiency .

This goal was achieved with the development of the ultra linear or distributed load concept by the two Americans David Hafler and Herbert I. Keroes in 1951, which goes back to a patent from the English engineer Alan Dower Blumlein in 1938. In this configuration, the screen grids of the end pentodes are fed part of the anode alternating voltage via a tap on the primary coil of the output transformer - the optimal ultra-linear taps for the screen grid of a push-pull output stage are around 40% of the number of turns of the primary winding, based on the center tap (V ++ in the circuit diagram above a push-pull amplifier) ​​of the transformer. If the position of the tap is shifted in the direction of the anode connection of the end tube, the triode operating mode predominates - a shift of the screen grid tap in the direction of the center tap of the primary coil causes a transition to the pentode setting.

An alternative between traditional current or voltage negative feedback and the ultralinear circuit described above is the cathode negative feedback brought onto the market by the British company Quad Electroacoustics , in which the cathode current of the output tube flows through a secondary winding of the output transformer and the induced alternating voltage is applied to the cathode, that it counteracts the control voltage.

All types of negative feedback can be used equally for single-ended and push-pull output stages.

The extent of the negative feedback used is indirectly proportional to the internal resistance of the amplifier: the transistor amplifiers, which are usually strongly negative feedback, are characterized by a low internal resistance and thus a high damping factor .

On the other hand, tube amplifiers with a rather slight or non-existent negative feedback behave in exactly the opposite way - this leads to the recommendation to use highly damped, yet efficient loudspeakers as sound transducers .

Intermediate transformer

Intermediate transformers in LF amplifiers are transformers with a very high number of turns and inductances . They were often used in older tube receivers prior to 1933 to reduce the number of tubes required. The transformers were mostly designed for a voltage transformation of 1: 3, whereby the tube stage on the primary side had to provide a small amount of power.

With the appropriate design, push-pull intermediate transformers can replace the active phase reversing stage required for push-pull output stages: the high-resistance secondary winding has a center tap connected to ground, and the 180 ° phase-shifted signal can be picked up at the two winding ends. This type of push-pull control was also found in many devices before around 1933.

Today transformers are sometimes used for galvanic isolation and impedance matching and prevent - z. B. with separately installed pre- and power amplifiers - noise caused by hum . However, they are not referred to as intermediate transformers.

Output transformer

Audio push-pull output transformer

Electron tubes are in principle high-resistance components, i. H. their output impedance is much higher (a few kilo ohms in the low frequency range) than that of loudspeakers (4 to 16 ohms are common). For this reason, the operation of loudspeakers with low impedance, which are usually low-impedance, on tube audio amplifiers requires a special low-frequency transformer , the output transformer. The high bandwidths of hi-fi amplifiers and / or high output power can only be achieved through nested windings, which makes them correspondingly expensive.

The following high requirements are placed on LF output transformers:

• Impedance transformation (ratio of turns 1:20 to 1:50)
• high relative bandwidth: the ratio of the upper to lower limit frequency is around 200: 1 (low leakage inductance due to nested windings, high number of primary turns, high permeability of the core)
• Linearity of the core (high saturation induction , air gap for amplifiers in A mode )
• low copper losses : low ohmic resistance of the windings
• low iron losses : low magnetic reversal and eddy current losses of the core material
• good insulation (protection against anode voltage, if overdriven the anode voltage increases periodically to a multiple of the operating voltage)

The impedance transformation is determined mathematically by the square of the number of turns ratio. Because the impedance of a loudspeaker is not constant, but rather dependent on frequency, there are limits to the accuracy of the adaptation in practical operation. The primary inductance must be so high that it only causes a slight drop in level at the lower limit frequency. Several henries are required.

The relative bandwidth is inversely proportional to the transformer leakage flux . A low leakage flux means good magnetic coupling between the primary and secondary windings. The leakage flux can be kept low by nesting the primary and secondary windings, which are divided into partial windings, the choice of a highly permeable core material has a positive effect on this and also on the size of the primary inductance.

Core materials with saturation that only sets in towards high magnetic flux densities (above 1.5 Tesla) have good linearity. However, highly permeable magnetic materials often have a lower saturation field strength, so an air gap is absolutely necessary for single-ended A output stages.

The copper losses are minimized by choosing a suitable core type and material, as this determines the available cross-section for the winding and the specific winding inductance. Effective use of the window area for insulation, windings and shielding enables the use of large wire cross-sections and thus low ohmic resistances.

The iron losses depend largely on the choice of core type and material. Very small sheet metal thicknesses reduce the relative iron content of the core volume, but reduce the eddy currents, especially at high frequencies.

The arrangement of nested windings increases the capacitance between the primary and secondary windings. In order to avoid unwanted capacitive coupling, insulated metal foils connected to ground are often inserted between the winding parts. These foils form a capacitive shield.

Although they require a center tap, push-pull output transformers are simpler to design and manufacture because the quiescent currents of the two output tubes flow in opposite directions through the two primary windings and the DC magnetic flux they generate is canceled out.

Output transformers for single-ended A output stages are more problematic, since in this case the quiescent current of the output tube flows through the primary winding of the output transformer and magnetizes it. In order to reduce the magnetic flux density of the transformer core and avoid saturation of the core material, the core therefore needs an air gap . The output transformer must be carefully dimensioned with regard to the operating point on the magnetic characteristic. The full scale of complex loads must be taken into account.

Another possibility to bypass the DC field in the output transformer of a single-ended output stage is to use a choke coupling. Instead of the primary winding in the anode circuit, a highly inductive choke is connected to feed the anode quiescent current into the anode circuit. This choke, which is provided with an air gap, drops the anode alternating voltage in addition to the unavoidable direct voltage resulting from the primary copper losses. It is fed to the output transformer via a coupling capacitor. Another relatively large and heavy component is required, but the output transformer is kept free from DC fields.