The Triode as Voltage Amplifier
In the triode a third electrode, the grid, is inserted between the cathode and the anode. In passing, it is worth remarking that the step of adding the grid to the diode valve was made by Lee De Forest in 1907, and that it represents what is probably the most significant advance in the whole history of what is now known as electronics. This is because the introduction of the grid resulted in a device capable of amplification, whereas previously no such device was in existence.
In the triode, a grid is interposed between cathode and anode. In this diagram, which illustrates an indirectly-heated triode, the grid consists of a spiral of fine wire.
The grid may, in practice, consist of a spiral of fine wire, as shown in the diagram above. In this diagram the turns of wire which make up the grid are welded to two stout support wires on either side, the external connection to the grid being made by way of either of these support wires. Other types of grid construction are also used, these employing wire mesh or similar materials, and they all have the basic feature that apertures, or gaps, are provided to allow the passage of electrons from the cathode to the anode.
The three-electrode assembly of cathode, grid and anode in an evacuated envelope constitutes a triode, or triode valve, or triode tube, and it is represented in circuit diagrams by one of the symbols shown below.
(a) The circuit symbol for a directly-heated triode
(b) The symbol for an indirectly-heated triode.
(a) gives the symbol for a directly-heated triode (ie one in which the filament acts also as the cathode), whilst (b) shows the symbol for an indirectly-heated triode (in which the cathode is heated by a separate filament or heater). The dashed line between the cathode and the anode represents the grid.
A circuit which may be employed to obtain IaVg curves for a triode valve.
In the next image, presented above, we connect a triode such that its cathode connects to the negative terminal of a high tension battery whilst its anode connects, via a current reading meter, to the positive terminal. At the same time, we apply to the grid a voltage obtained from a potentiometer connected across two batteries in series. These batteries are so connected that, when the slider of the potentiometer is at the top end of its track the potential applied to the grid is positive of the cathode, and when the slider is at the bottom end of the track the potential applied to the grid is negative of the cathode. A voltmeter connected between grid and cathode indicates the grid potential. As was pointed out in the first article, valve electrode potentials are always, unless otherwise stated, assumed as being with respect to cathode. Since the voltmeter has to measure both positive and negative voltage it may conveniently be an instrument having a centre-zero scale. A second voltmeter indicates the voltage from the HT battery. It is assumed that the cathode is heated by having the heater connected to a suitable source of supply.
Let us now see what happens when we vary the potential on the grid of the triode. If we adjust the potentiometer so that the slider is at the top end of the track, the grid becomes positive of the cathode and attracts electrons from the space charge. In doing so it partly neutralises the space charge, whereupon a large number of electrons are able to travel from the cathode to the anode, these passing through the apertures in the grid. If the potentiometer slider is now moved down towards the position corresponding to zero potential, the positive potential on the grid reduces, as also does the neutralising effect exerted on the space charge.
The latter then imposes increased limitation on electron flow and the flow of electrons from cathode to anode decreases. When the potentiometer slider is moved further downwards it brings the grid negative of the cathode, with the result that the limiting effect due to the space charge is augmented by the negative grid. Electron flow is, therefore, decreased further. As the grid potential is made continually more negative, the flow of electrons continues to decrease until a negative grid potential is reached at which electron flow ceases altogether. This negative potential is known as the cut-off potential for the valve. It varies for different values of anode voltage because, as we saw last month, the latter also has a neutralising effect on the space charge and therefore affects the electron flow limitation given by negative grid-plus-space charge.
We may therefore see that, with a positive potential on the grid, a large number of electrons may pass through the grid apertures to the anode.
Thus, a positive grid potential corresponds to a large anode current. As the positive potential on the grid reduces to zero the electron flow and, hence, anode current, reduces. The process continues after the grid passes through zero potential and goes negative, the anode current continually reducing as the grid becomes more negative until we reach the cut-off potential at which anode current ceases completely.
A set of typical IaVg (anode current - grid voltage) characteristic curves, as may be obtained from the circuit, are shown below.
Representative IaVg curves for a triode at anode voltages of 100, 150 and 200. For the purpose of explanation these curves show anode current for positive grid voltage, but it should be pointed out that it is normally desirable to operate triodes with negative grid voltage only, and that it is possible for a valve to suffer damage if excessive positive voltage is applied to its grid.
These demonstrate the fact that anode current decreases as grid voltage goes negative. Three curves are given, these corresponding to anode voltages of 100, 150 and 200.
An important point which we have not yet discussed is that, when the grid is positive of the cathode, some of the electrons from the cathode flow to it. This occurs because the positive grid acts as though it were the anode of a diode. On the other hand when the grid is negative of the cathode no electrons flow towards it, since it then offers repulsion to such electrons. Thus, there is a flow of grid current when the grid is positive, and none when it is negative. The consequence is that, if we keep the grid of the valve always negative of the cathode, we can control anode current by grid voltage only, since no current need flow in the grid circuit at all. Apart from any other factors, this is an extremely useful condition because it means that high impedance circuits (That is, circuits which can be represented by a generator having a high internal impedance.) may be connected to the grid without risk of loss of voltage due to grid current. In practice, it is customary to operate a triode such that its grid is always negative of the cathode, and it is only in certain specialised circuits that the grid is allowed to go positive and, thereby, allow the flow of grid current. (It is possible for a small current, due to effects inside the valve which have not yet been mentioned, to flow in the grid circuit even when the grid is negative of cathode. However, such currents are, proportionately, very low. For many applications, and in the present discussion, they can be considered as being negligible.) Even in these circuits, however, the flow of grid current has to be kept at a low level, because excessive grid current may cause overheating of the grid and consequent damage to it. Valve manufacturers published IaVg curves normally cover grid voltages from the negative cut-off potential to zero voltage only, these being similar to the curves to the left of the zero grid voltage point seen above.
The Triode as a Voltage Amplifier
We have seen that we may vary the anode current of a triode by varying its grid voltage. Let us take this a stage further by applying an alternating voltage to the triode grid. The resultant effect on anode current is illustrated graphically below.
Applying an alternating grid voltage to the V}a}=150 curve of the previous diagram.
It reproduces the '150 Volt' IaVg curve from the graph. In this latest diagram the alternating voltage is applied at the -3 Volt point on the Vg axis, so that it causes the grid to swing by equal amounts positive and negative of -3 Volts. The resultant anode current (produced on the right by drawing vertical lines from the alternating voltage up to the curve then, horizontally, to the right) also has an alternating waveform. Thus, by applying an alternating voltage to the grid we have caused an alternating current to flow at the anode. We can, in consequence, look upon the valve as a 'generator' of alternating current, in which case it needs a load. Such a load can be provided by a resistor or by an impedance, and we shall now examine the case where the load is a resistor.
Connecting an alternating voltage generator to a triode voltage amplifier. The 3 Volt battery ensures that the alternating voltage is applied to the Vg axis at the same position as occurred in the last diagram.
A resistive load is shown above and it consists quite simply of a resistor connected between the anode of the triode and the positive terminal of the HT supply. It is referred to as the anode load.
Applied to the grid of the triode is an alternating voltage generator whose output is the same as the alternating voltage shown before. Initially we positioned the alternating voltage at the -3 Volt point along the Vg axis. We do the same in practice by inserting a 3 Volt battery between the negative HT supply line and the lower terminal of the alternating voltage generator, this battery making the lower generator terminal 3 Volts negative of the cathode.
The alternating voltage generator causes, as is to be expected, corresponding changes in anode current. But the anode current now flows through the anode load resistor, with the result that the voltage dropped across this resistor varies according to the current which flows through it. In consequence an alternating voltage appears at the anode. Assuming that a satisfactory value of resistor has been chosen for the anode load, the alternating voltage at the anode may be many times greater than the alternating voltage applied to the grid; which means that the valve has caused the alternating voltage applied to the grid to be amplified. A circuit of the type shown, whose function is to produce an amplified version of an alternating voltage, is described as a voltage amplifier circuit.
Although the grid circuit corresponds to the grid voltage conditions of the diagram, there are no further points of similarity between the two diagrams. This is due to the fact that the curve in which we first applied an AC voltage defines anode current when anode voltage is constant. Anode voltage is quite patently not constant in our circuit, because we have purposely inserted a resistor in series with the anode to obtain an alternating voltage output. It would seem from this that there is little point in drawing characteristic curves for a triode with a fixed anode voltage because, when we want to use the triode as a voltage amplifier, the first thing we require is a varying anode voltage! As we shall see later, however, it is possible to find the alternating voltage appearing at the triode anode from curves which give much the same information as is given by the representative family of curves.
In our circuit we positioned the alternating voltage at the -3 Volt point along the Vg axis by the simple expedient of inserting a 3 Volt battery in series with the alternating voltage generator. This battery provides a grid bias of -3 Volts and may be referred to as a grid bias battery. Alternative circuit devices can be employed to position the alternating voltage at a desired point along the Vg axis, and these are also described as applying bias. (In American terminology, grid bias voltage is referred to as the 'C voltage'. The 3 Volt battery of then becomes a 'C battery'. The letter 'C' differentiates the grid bias supply from the 'A supply' (the supply for the filament or heater) and the 'B supply' (that for the anode, and which is described in this country as the high tension or anode supply).)
A further point in the circuit is that the current which causes the amplified alternating voltage to appear at the anode is provided by the HT battery. The grid may be considered as a controlling device which 'turns on' more anode current as it goes positive, or vice versa, and which consumes no power (to be exact, negligible power in the present context.) in the process provided that it does not go positive of the cathode. It should be remembered that, in the circuit under consideration, anode current is always present, and the alternating voltage only appears because this anode current is made to vary by the alternating voltage at the grid.
We have, up to now, looked upon the alternating voltage applied to the grid of the triode as being provided by a generator. In radio work we handle a number of alternating voltages, these including the electrical signals which correspond to sound. We have seen, for instance, that if we connect up a carbon microphone in the manner shown in (a) below we obtain, from the secondary of the transformer, electrical signals corresponding to the sound which reaches the microphone diaphragm.
(a) Obtaining electrical signals from a carbon microphone
(b) Applying the microphone output to the voltage amplifier.
The secondary of the microphone transformer can replace the alternating voltage generator, giving us the circuit shown in (b). In this circuit, amplified electrical signals corresponding to the original sound appear at the anode of the triode. It should be noted, in passing, that the microphone transformer of (b) could, in the present instance, have quite a high step-up ratio, this being of the order of 1:100. A practical carbon microphone is a low impedance instrument, whereas the grid input circuit of the valve presents a high impedance. The step-up transformer helps to match these two impedances to each other.
Voltage Amplifiers in Cascade
It frequently happens that a particular item of equip-ment requires more voltage amplification than a single triode can provide. This requirement may be met by adding a second triode to amplify the alternating voltage at the anode of the first. How may these two triodes be coupled together?
Two triode voltage amplifiers in cascade. Resistance-capacitance coupling is employed between the two valves.
The simplest method of providing the coupling is shown above. In this diagram the alternating voltage is applied to the first triode, now designated V1, in the same manner as occurred previously. It is only the alternating voltage at the anode of V1 which needs to be fed to the second triode, V2, and this alternating voltage is transferred by way of a capacitor whose value is such that it presents a low impedance at the frequencies being handled. If (as is usually the case in circuits employing triodes) these are audio frequencies, the value of the coupling capacitor would be of the order of 0.01μF to 0.1μF. The second triode has a grid bias battery (which, we may assume for the purpose of explanation, is also 3 Volts), and this is coupled to the grid via a high value resistor referred to as the grid resistor. The grid resistor normally has a value of some 250kΩ to 1MΩ. The output alternating voltage is then taken from the anode of the second triode.
To fully appreciate the manner in which the capacitive coupling functions it is helpful to commence with the condition where the AC generator offers zero output (as would occur if the alternating voltage were obtained from a microphone on which no sound impinged). The voltage on the grid of V1 will then be a steady -3 Volts, and V1 anode will take up a steady voltage which corresponds to this grid voltage. The capacitor has its upper plate connected to V1 anode and its lower plate connected, via the grid resistor of V2, to the negative terminal of the second grid bias battery. The capacitor thus holds a charge in which its plates are maintained at these two potentials and no current flows in the grid resistor (although, of course, the charging current needed to bring the capacitor to its present condition would have flowed when the HT supply was initially switched on). Since no current flows through the grid resistor there is no voltage drop across it, and the -3 Volts bias from the second grid bias battery is applied to the grid of V2.
If, now, the generator produces an alternating voltage, the voltage on the grid of V1 will swing positive and negative of the -3 Volts provided by its grid bias battery. The anode of V1 will, also, swing positive and negative of its previous steady voltage by an amount depending upon the amplification provided. The amplified positive and negative excursions at the anode of V1 are passed via the capacitor, which tends to retain its charge, to the grid of V2 which similarly swings positive and negative of the -3 Volts provided by the second grid bias battery. A twice-amplified alternating voltage then appears at the anode of V2. As may be seen, the situation at the grid of V2 is, so far as bias is concerned, just the same as if an alternating voltage generator had been inserted between the negative terminal of the bias battery and the grid.
When a valve amplifies the output from a preceding valve, the two valves are described as being in cascade. If a third valve amplified the output of the second, the three valves would also be referred to as being in cascade, and so on. The term 'cascade' similarly applies when amplifying valves other than triodes are used or when other methods of inter-valve coupling are employed. (Alternative inter-valve coupling circuits will be discussed in later articles.) The method of coupling shown is described as resistance-capacitance coupling, or RC coupling, because it is achieved with the aid of an anode resistor and a coupling capacitor.
Other Bias Circuits
In the circuit we have been considering grid bias for the two triodes was provided by two grid bias batteries. In practice it would be a little expensive and inconvenient to have a grid bias battery for each valve in an amplifier, and one method of simplification could consist of employing a single battery instead, as shown in (a) below.
(a) Employing a common source of grid bias voltage instead of the two batteries shown before
(b) If the internal resistance of the battery is high, part of the alternating voltage applied to the grid of V2 becomes fed back to the grid of V1
(c) The situation in (b) may be obviated by inserting a decoupling circuit, given by Rd and Cd, in the bias supply to V1.
This arrangement is quite practicable but, for correct operation, it is important that the internal resistance of the common grid bias battery below. If its internal resistance were high, the situation shown in (b) could arise. In this diagram, part of the alternating voltage applied to the grid of V2 appears across the internal resistance of the common battery and is, therefore, applied to the grid of V1. This is not a desirable state of affairs, and will not offer results corresponding to those given by the circuit considered previously. In more complicated types of equipment the use of a source of grid bias offering a high common resistance (or impedance) to several grid circuits may give rise to quite serious faults, such as instability. ('Instability' defines the tendency of an amplifier to break into oscillation, and may be caused by an unwanted coupling between an output point and an input point (ie two circuit points between which amplification occurs).
The difficulty may be overcome by inserting a resistor and capacitor in one of the grid bias supply circuits, as is shown in (c). In this diagram Rd has a high resistance and Cd a low reactance at the frequencies being handled so that, even if part of the alternating voltage applied to the grid of V2 appears across the grid bias battery, only an extremely small proportion of this unwanted voltage is actually applied back to the grid of V1. Rd and Cd are known as decoupling components because they de-couple (ie break the coupling) between two circuit points.
Some circuits have a common bias supply which is not derived from a battery but from an alternative source. Decoupling circuits similar to those of (c) are, then, frequently employed to prevent unwanted couplings between stages. A decoupling circuit may, indeed, be inserted in each grid circuit.
An alternative method of providing bias for a valve dispenses with the grid bias battery concept, but it may only be employed with indirectly-heated valves in which the cathode is insulated from the heater. As we have already seen, the alternating voltage at the anode of a voltage amplifier is caused by changes in its anode current, this current being provided by the HT battery. If we look upon the current in terms of a flow of electrons, we can say that electrons leave the negative terminal of the HT battery, pass to the cathode, are emitted to the anode, then flow back to the positive terminal of the battery. Clearly, therefore, the anode current flows, also, through the cathode circuit.
If a resistor of the requisite value is inserted in series with the cathode of the triode, the cathode becomes positive of the lower terminal of the generator by the required bias voltage.
We insert a resistor in series with the cathode of a voltage amplifier triode. We also apply our alternating voltage generator, but in this case its lower terminal is connected to the negative HT line. No bias voltage is, in consequence, applied to the grid at all. The anode current flows through the resistor in the cathode circuit, causing a voltage to be dropped across it, with the result that the cathode goes positive of the HT negative line and the lower terminal of the alternating voltage generator, by that voltage. The cathode going positive of the lower generator terminal is exactly the same as the lower generator terminal going negative of the cathode, as occurred previously when we employed the grid bias battery. If, therefore, we give the cathode resistor a value which causes the required grid bias voltage to be dropped across it, then we will achieve the same result, so far as direct voltage is concerned, as was given previously by the grid bias battery.
Unfortunately, the cathode resistor, on its own, does not completely meet the requirement for a bias circuit. This point may be readily understood if we examine circuit operation over a cycle of the alternating voltage offered by the generator.
The disadvantage with the cathode resistor as used above is that the cathode tends to 'follow' the alternating voltage on the grid. The grid voltage for a single cycle is shown in (a) and the corresponding cathode voltage in (b) Both voltages are with respect to the HT negative line.
Considering the graphs above, (a) illustrates a single cycle as provided by the generator, whilst (b) shows the voltage appearing at the cathode of the valve when connected as in the previous circuit. Both voltages are shown with respect to the HT negative line. At point A, the generator voltage is zero, whereupon the grid of the valve is at the same potential as the HT negative line and the lower terminal of the generator. As is shown in (b), the required bias voltage is then dropped across the cathode resistor. At point B, the generator output is at its peak positive potential. This results in an increase in anode current and, in consequence, in cathode current. A greater current flows through the cathode resistor and a greater voltage is dropped across it. The cathode, therefore, goes more positive. At point C of (a) the alternating voltage at the grid is zero and the voltage dropped across the cathode resistor is the same as for point A. At D, the generator voltage reaches its maximum negative excursion. This causes a reduction in anode current and in cathode current, with less voltage dropped across the cathode resistor. The cathode goes, therefore, negative of the potential it held at points A and C.
We have examined cathode voltage at the peaks B and D but, as is shown in (b), this voltage will similarly change, by corresponding amounts, for grid voltages between zero and the peak values. A low-amplitude version of the grid alternating voltage appears, in consequence, at the cathode of the valve, this going positive when the grid goes positive and negative when the grid goes negative. Now, the voltage which controls the flow of electrons in the valve is that between grid and cathode, and it may be seen from inspection of the graph that the grid-cathode alternating voltage is less than that provided by the generator. The cathode voltage changes in sympathy with the grid voltage and the actual grid-cathode alternating voltage is that provided by the generator less the alternating voltage appearing on the cathode.
Adding a capacitor having a very low reactance (at the frequencies handled) in parallel with the cathode resistor maintains the cathode at a steady potential and prevents the effect shown in the graph from taking place.
By inserting the cathode resistor we have obtained a useful bias voltage but we have caused the overall amplification provided by the circuit to be reduced. The solution consists of connecting across the cathode resistor a high-value capacitor having a very low reactance at the frequencies being handled. (See above.) The added capacitor prevents the cathode voltage from 'following' the alternating voltage on the grid, whereupon the circuit provides the requisite bias for the valve and behaves, when alternating voltage is applied to the grid, in the same way as did the previous circuits using grid bias batteries.
The circuit we now have is, in practice, a very convenient one, because it enables the bias requirements for the valve to be met by means of two simple components - a resistor and a capacitor.
The bias applied in our final circuit is described as cathode bias, and it may be employed with amplifying valves other than triodes. For audio frequency amplification, it is usual to give the cathode bias capacitor a value of the order of 25μF. An electrolytic capacitor can be employed here and, since the voltage dropped across the bias resistor is normally low, it can be a small component having a low working voltage rating.
As a point of terminology, the loss of overall amplification which results when the cathode resistor is employed on its own is referred to as cathode degeneration.