1. ESSENTIAL KNOWLEDGE [aka t00bs for n00bs 
]
1.1 What is a tube?
To know where we are now, we have to have a reference point - where we began. Hence this section will start out with a quick history lesson (with a quick science lesson embedded within).
The vacuum tube as we know it (short form: `tube') (also known as a 'thermionic valve' or just ‘valve’ to our European friends) descended from the first light bulbs. In those, air was evacuated from a glass cylinder tube in order to prevent the rapid oxidation (i.e. burning) of a white-hot filament via reaction with atmospheric gases. The glass was then sealed to maintain the vacuum.
Thomas Edison first noticed that as a bulb neared the end of its useful life, its glass envelope would darken with a thin layer of tungsten (you can observe the phenomenon today in some halogen bulbs). The tungsten filament was evaporating and depositing on the inside of the glass envelope. To rectify this, a positively charged `plate' (positively charged relative to the filament) was introduced within the bulb in order to `attract' the tungsten vapour over to it instead. But something unexpected happened - a current flowed across the vacuum.
In 1904, John Ambrose Fleming invented what is essentially a diode - a device that allows current to flow in only one direction. He named his invention the `electrical valve'. Two carbon electrodes were placed in a vacuum tube - one being heated till it glowed white-hot (the `filament') and the other left cold (the `plate'). When a source of AC was connected between these 2 filaments, current would only flow in one direction.
Science lesson: Electrons are negatively charged particles. They are strongly attracted by positively charged objects. All conductors have free electrons within their structures that are bound by attractive forces to adjacent (and positively charged) atoms. If enough heat energy is applied to the conductor, some electrons gain enough energy to escape these attractive forces. Thus electron emission increases extremely rapidly with a corresponding rise in surface temperature. Hence, a hot glowing carbon filament will emit a large amount of electrons relative to a cold carbon filament. Now link this with the `electrical valve' - during one half of the AC cycle, electrons escape off the hot filament, fly through the vacuum and enter the cold filament, but during the next half-cycle, nothing escapes the cold carbon filament to enter the hot filament. So there is an electron flow in a single direction, which means a current flow in a single direction.
In effect the hot filament has becomes a cathode (negative electrode) and the cold plate, an anode (positive electrode). If we keep the plate at a positive electrical potential relative to the filament, and ensure that the filament is heated sufficiently so that it emits electrons reliably, we can consistently pass a large current through a vacuum.
Lee de Forest patented the Audion in 1908. A platinum wire was placed between filament and plate, creating a triode (contrast this with the diode, which has 2 elements). By applying a weak AC voltage to this platinum wire, he realised that he could modulate the large (and previously uncontrolled) flow of electrons from filament to plate. Lee de Forest had invented an amplification device, albeit a crude one - the end signal at plate was roughly the same as that applied to the platinum wire, the main difference being that it's magnitude (i.e. amplitude) was much larger.
Lee de Forest's Audion
While comparing a modern tube to a light bulb is akin to comparing Modem Man to Homo Erectus, the Audion can be said to be, at most, the grandfather of the modern tube. Obvious physical differences apart, they are identical in both idea and implementation.
1.2 Tube types and general structure
Directly vs indirectly heated
As mentioned above, in earlier tubes the filament was also the cathode. In other words, the cathode was directly heated. Such tubes glowed white-hot at 3000K, even though a lighting effect was not really necessary to their proper function. These tubes were known as bright emitters.
The next evolutionary step was to add minute amounts of the element thorium to the tungsten filament, creating the thoriated tungsten filament. Not only was running temperature reduced significantly to -2000K, electron emission was also improved by a factor of roughly 10. These tubes are known as (drumroll!) dull emitters, and are still in use today in audio - examples include the 300B and 2A3 tubes.
But a real marked improvement came about with the invention of the oxide-coated cathode. The cathode was now a thin coating of a mixture of various metal oxides which was heated by a completely separate heater filament. The 2 were separated by a thin layer of aluminium oxide insulator - hence creating an indirectly heated cathode. Why a metal oxide cathode? Such cathodes run cooler at roughly 1100K, yet have an emission efficiency of more than 100x that of bright emitters! Most of the tubes in audio use are indirectly heated types.
Despite their vastly improved emission efficiency and cooler running temperature, indirectly heated tubes have their weaknesses - amongst other things, they are extremely fragile, and the oxide cathodes are easily poisoned. These will be covered in more detail in the section "The proper care and feeding of tubes".
The triode
Triode structure has already been covered briefly above. There are 3 main active elements in a triode - the cathode, anode and control grid.
The control grid is the direct descendent of the platinum wire in the Audion. It has been refined to what is typically a helical coil of extremely fine wire (<l0um thick; a human hair is about 100um thick) that encircles the cathode.
Most audio tubes in use today are triode types. Many have two triodes in a single glass envelopes, and are known as twin-triodes. Example of triode types commonly seen in audio include the 6SN7, 6DJ8 family and the 6080/6AS7.
The tetrode
The tetrode differs from the triode in just one way - it has an additional grid, between the anode and control grid, known as the screen grid.
The purpose of the screen grid is to accelerate even more electrons toward the anode. To this end it is kept at an even more positive potential than the anode, so that electrons that would otherwise spiral off into space after leaving the cathode would be strongly attracted towards it. The coils of screen grid wire are wound more coarsely to avoid absorbing these electrons, so the most of the electrons pass right through the screen grid and straight into the anode.
As can be imagined, tetrodes usually provide more voltage gain than triodes. Yet, they are relatively uncommon compared to triodes or pentodes. Why? Read on!
The pentode
The pentode in turn differs from the tetrode in just one way - there is yet another grid between the screen grid and the anode. This grid is known as the suppressor grid.
While tetrodes provided higher voltage gain, an unfortunate side effect of increasing electron velocity was uncovered. At higher anode voltages, an electron may be so sped up by the screen grid that when it hits the anode it dislodges two electrons. The anode has now effectively emitted one electron, which leads to a net fall in anode current. This phenomenon is known secondary emission. As can imagined, secondary emission detracts somewhat from the linearity (the ability to amplify without distortion) of the tetrode.
The pentode (invented by Philips/Mullard) solved this problem. The suppressor grid is kept at an electrical potential that is very slightly more negative than the anode. Hence electrons emitted via secondary emission are repelled by the suppressor grid, and back towards the anode. Voila! Examples of pentodes commonly seen in audio include the EL34 and EL84.
The beam tetrode
While the pentode solved the linearity problems associated with tetrodes, it was under patent to Philips/Mullard. In an effort to circumvent this, the beam tetrode was invented by MOV.
The beam tetrode works just like the tetrode, but with a few small differences. The cathode is formed in such as way as to emit electrons in parallel `beams' focused towards the anode. The control and screen grids are wound at the same pitch (coils per unit of vertical height) so as to be optically aligned. The cathode is aligned so that the electron beams pass between the grid wires (still influenced by grid potential), but never intersect with them (hence never absorbed).
This reduces grid current (current generated when an electron hits and is absorbed by a grid) by around 20%. As a result, more electrons reach the anode, contributing to higher anode current and higher overall tube efficiency. Examples of beam tetrodes commonly seen in audio include the 6L6, KT66 and the KT88.
Tube rectifiers
Rectification refers to the process of converting AC to (pulsating) DC. Tube rectifiers are simply tubes that accomplish this.
Please refer to the section `2.3 Regarding tube-based rectifiers...' for more on tube rectifiers.
Other Tubes
There are other tubes used in radio transmitting. These (typically) directly heated tubes are sometimes watercooled or may have a ceramic substrate replace the vacuum. These tubes are irrelevant to consumer audio and will not be covered within this guide.
1.3 A pictorial guide to tube structure
Please click
here for this section.
1.4 Tube naming conventions
There are 2 main types of tube naming conventions – American and European. They will be covered separately (and in detail) below.
European
European tubes follow a letter-letter-letter-digit -digit system. E.g. ECC88, ECC83, ECL82.
Letter 1, heater type:
A – 4V
B – 180mA
C – 200mA
D – 0.5V to 1.5V
E – 6.3V
F – 13V
G – 5V
H – 150mA
K – 2V
P – 300mA
Letter 2 (and subsequent letters), valve type:
A – Small-signal diode
B – Double small-signal diode
C – Small-signal triode
D – Power triode
E – Small-signal tetrode
F – Small-signal pentode
H – Hexode or heptode
K – Heptode or octode
L – Power tetrode or pentode
M – Fluorescent indicator
N – Thyratron
Q – Nonode
X – Gas-filled rectifier
Y – Half-wave rectifier
Z – Full-wave rectifier
Digit 1 (ignore digit 2 unless specified otherwise), base type:
1 – Use 2nd digit
2 – B8B
3 - Octal
4 – B8A
5 – B9D
8 – B9A
9 – B7G
Examples:
1. The ECC81 has a 6.3V heater, and is a twin small-signal triode using a B9A base.
2. The ECL82 has a 6.3V heater, and has a triode and a power tetrode/pentode within the same envelope. It has a B9A base. One would have to refer to the datasheets to verify that the ECL82 is a triode and a pentode.
3. The GZ34 has a 5V heater and is a full-wave rectifier using the octal base.
A few European companies however, ignored this naming convention completely, preferring to use their own proprietary naming systems. Ediswan/Mazda, Marconi Osram and STC are known to have different naming systems.
British tubes had their own separate naming system, distinguished by their ‘CV’ (Common Valve) prefix.
American
American tubes for the most part followed the Radio Electronics Television Manufacturers Association (RETMA) naming system. This system followed the digit-letter(s)-digit naming pattern.
1st digit: Approximate heater voltage (A 7 or 14 here refers to Loktal base types)
Letter(s): Arbitrarily assigned. Related to individual valve design.
2nd digit: Number of electrodes (heater included)
For octal tubes, an additional suffix system was used. These suffixes can be tacked on to the end of any American octal tube.
G: ‘Glass’ indicating likely usage of the ST14 (Shouldered Tube) glass envelope. Shoulder Tubes resemble soft drink bottles
GT: ‘Glass, Tubular’ later glass envelops were simply cylinders with a hemispherical end.
Examples:
1. The 6SN7GT uses a 6.3V heater and being a double-triode, has 7 electrodes (2 triodes + heater). It has a cylindrical glass shape with a hemispherical end.
2. The 12AT7 uses a 12.6V heater and being a double-triode, has 7 electrodes (2 triodes + heater).
3. The 5AR4 uses a 5V heater and being a full-wave rectifier, has 4 electrodes (1 cathode, 1 heater, 2 anodes).
There were also military tubes, either having a ‘JAN’ (Joint-Army-Navy) prefix, following a VT-digit-digit-digit (e.g. VT-231) system or a 4 digit system (e.g. 5692, 6080). In the later 2 both cases, the digits were chosen arbitrarily, hence attempting to define any logical pattern is futile.
Western Electric followed its own numbering system - usually 3 digits followed by an alphabet. E.g. 300B, 300A, 421A, 417A.
A note on American JAN/VT designated tubes:
While these tubes were subjected to a more stringent set of quality control criteria to attain military 'approval', more often than not these tubes were completely identical to non-military tubes of the same vintage. Often both civilian and military versions were made in the same factories in the exact same production line.
A single tube may have many different designations – for examples, the 12AX7 tube is also known as the ECC83 or CV492. Please remember to contact your amp manufacturer before doing any substitutions.
1.5 Why tubes?
There are many people out there who would lead you to believe that tubes are the greatest. Then there are people who think the same of solid-state equipment. Both sides have valid points to bolster their arguments. In the author's experience, the difference between high-end solid-state and tube equipment is typically imperceptibly small.
Tubes are fragile, bulky, produce excess heat, have more noise, objectively give more measured distortion when compared to solid-state, and are more susceptible to stray capacitances. They are also subject to the vagaries of production by hand, are more expensive to make and are also dwindling in supply. Taking all these into account, why even consider tubes? In short - their superior clipping characteristics and different distortion characteristics.
Clipping occurs whenever the amplification device literally runs out of juice to amplify a given signal with an initial amplitude that is too large. This occurs often during musical instances known as transients - examples include the crack of a snare on a drum or a full-out symphonic crescendo.
Imagine a sine wave:
Now we shall amplify it till it clips. For solid-state devices in general, the output would look like the sine wave A: the top section is lopped off completely, forming a plateau that is parallel to the horizontal axis. Not only does this sound terrible, the plateau is in effect DC applied to your headphone drivers (very bad).
sine wave A
Compare this now to sine wave B: the top of the sine wave, while not like the original signal, is now a curve with a slightly different rise/fall gradient.
sine wave B
This sounds a lot less offensive and is certainly easier on your headphone drivers than the `hard' clipping offered by the above solid-state example. It is also far less audible than the audibly obvious solid-state clipping. This has led to the popular myth that a`tube watt' is twice as powerful as a 'solid-state watt'. Clipping still occurs - you're just less likely to hear it.
As mentioned earlier, tubes have more measured distortion. However this distortion is primarily 2"d order, lessening greatly as we go through the higher harmonics. Distortion characteristics for solid-state however, tend towards higher odd order harmonics (5"' 7`n etc), albeit in smaller amounts.
Scientific studies have shown that humans perceive even order distortion as being musically consonant while odd order distortion is perceived as musically dissonant. Anecdotal evidence shows that while up to -5% of 2°d order distortion is audibly tolerable, only -0.5% of 5`h order distortion is audibly tolerable.
Tubes win!
What that means in plain English is that a tiny bit of odd order distortion is going to sound a lot worse than large amounts of even order distortion to the average human ear. Compare this to the distortion characteristics of tubes and solid-state.