|
Due to the low grid-anode capacity, screen-grid and HF pentode valves are now practically universal in the HF stages of a receiver. They are not without their drawbacks, however, and it is shown in this article that it is possible to obtain high amplification with a larger undistorted output and a greater degree of stability by using triodes in pairs.
Fig. 1. - Unstable amplifying circuit (triode valve).
It is well known that the attempt to use triode valves for radio-frequency amplification, with a circuit such as that shown in Fig. 1, is defeated by persistent self-oscillation, which sets in when the two circuits are brought into tune with one another. The cause of this instability is the feed-back of energy from the output circuit to the grid circuit through the grid-anode capacitance.
Up to a few years ago this difficulty was overcome by various expedients, some ingenious and more or less satisfactory, others frankly desperate and correspondingly unsatisfactory, but, as far as ordinary broadcast receivers are concerned, all these devices were swept away at one stroke by the introduction of the screen-grid tetrode valve. This attacked the trouble right at its source, and sought to get rid of the instability by eliminating the grid-anode capacitance that gave rise to it. The attempt met with a degree of success that completely revolutionised the design of receiving sets, and paved the way for the rapid advance of recent years in sensitivity and selectivity, not to mention the variously qualified kinds of automatic volume control and other refinements of the modern commercial receiver.
In the face of this triumph of the tetrode it may seem almost perversely reactionary even to mention the triode in connection with radio-frequency amplification. The fact remains, nevertheless, that in spite of obvious advantages which are likely to ensure its continued dominance in commercially produced receivers, the use of the screen-grid valve is not the only practicable method of securing stable radio-frequency amplification, and is not even the best method in some cases.
As already stated, the advantages of the tetrode are obvious, and need not be enumerated. Its limitations, however, are not so obvious, and it is necessary to call attention to them in order to indicate the cases in which an alternative triode circuit, described in this article, may actually be preferable in some important respects.
In the first place, the screen-grid valve amplifying circuit is not inherently and unconditionally stable. It is not possible to reduce the grid-anode capacity to zero, and the small residual capacity, even though it may be as little as 0.005 pF, is sufficient to produce oscillation with really efficient coils. The goodness of a tuned circuit, from the point of view of amplification, is measured by the tuned rejector circuit impedance of the coil sometimes referred to as its dynamic resistance. It is not difficult to realise as high a value as 200,000 Ω or more for this quantity in a good coil. It can be shown theoretically that something like 100,000 Ω is the highest permissible value, consistent with stability, in a screen-grid valve circuit, and in the writer's experience it seems necessary to use even flatter circuits than this to be quite sure of stability over the whole tuning range.
Overcoming Instability
The limitation is not a very serious one in practice, but it does restrict the sensitivity and selectivity obtainable in a single stage, and, moreover, it prohibits the effective use of retroaction for the enhancement of selectivity. In the recently developed single-span receiver, for example, this difficulty was encountered, and overcome by adopting the distinctive feature of the alternative triode-valve arrangement to be described later.
A more serious, but less obvious, disadvantage of the tetrode is due to the secondary emission from the screen-grid. It does not seem to be generally known that both the voltage factor and the internal resistance of a screen-grid valve vary very greatly with anode voltage as a consequence, apparently, of this secondary emission. The variation does not appear in the characteristics as usually exhibited, because it so happens that the mutual conductance, i.e., the voltage- factor divided by the internal slope resistance, varies very much less than either of its constituent elements. The mutual-conductance characteristic is therefore very deceptive as a guide to the dynamic behaviour of the valve, unless the anode-circuit load is small compared with the internal resistance a condition which is not always as closely fulfilled in practice as it is usually assumed to be.
Fig. 2. - Amplification characteristic of a screen-grid valve (two Volt battery type. Anode circuit impedance 180 kΩ.Frequency 1,000 kHz).
The practical consequence of this variation of the valve constants is an appreciable degree of curvature in the amplification characteristic. Fig. 2 shows a typical example of a number of measurements made by the writer with various types of screen-grid valves and variously efficient anode circuits. It will be seen that there is a very appreciable departure from linearity when the output voltage exceeds about 15-20 Volts. (In the case illustrated the anode circuit impedance is higher than would normally be used, but with values as low as 40,000 Ω a similar curvature was apparent at similar output voltages, but the amplification was correspondingly smaller.) Although this effect is not likely to be serious in small amplitude operation, it is likely to have undesirable consequences in cases where a tetrode stage is required to supply large output voltages for 'linear' detection. Of these consequences, the most troublesome is that known as cross-modulation, the effect of which is an apparent loss of selectivity, for which there is no practicable remedy. In one case investigated by the writer it was found that the reception of a fairly strong unmodulated carrier wave, in a receiver embodying a tetrode stage, caused a very noticeable increase in the intensity of the interference due to a transmission on a neighbouring wave-band. This indicated a very considerable cross-modulation effect, for in the absence of it the interference would have been considerably reduced when the strong unmodulated carrier was received, due to the so-called demodulation effect of the detector.
It will not be suggested here that the use of the triode valve in the way to be described later is the only means of evading these undesirable consequences of secondary emission. It is well known that in the output-pentode type of valve the effects on the anode circuit of secondary emission from the outer grid are minimised by incorporating yet one more electrode into the structure - the so-called suppressor grid. The function of this grid is sometimes rather misleadingly described as being the suppression of secondary emission from the outer or screen grid, but in all probability it is only suppressed in the sense that it is prevented from affecting the anode circuit to any appreciable extent. The success of this device in the output pentode led to the development of radio-frequency pentodes. Such valves should theoretically be capable of handling larger output voltages than the tetrode, but the writer has no exact information on the point based on his own measurements, and is not aware of any published data, There is, in any case, no obvious reason to suppose that the other limitation of the screened valve, i.e., the stability limitation, would be materially affected by the inclusion of the additional electrode.
Briefly, therefore, the limitations to screen-grid valve amplification are (a) in- stability if very efficient coils are used, and (b) unsuitability for operation at large output amplitudes, except, possibly, in the case of the pentode type of screen-grid valve.
For the triode valve arrangement, now to be described, can be claimed (a) complete and inherent stability, however efficient the tuned circuits; (b) linearity of the amplification characteristic, even up to output RMS voltages practically equal to the anode direct voltage.
As against the above, it must be pointed out that the complete amplifying unit involves two triodes, as against a single tetrode In this comparison, however, it must be remembered that the screen-grid valve may cost as much as, or even more than, the two triodes.
Detail and Performance of the Triode Valve Radio-Frequency Amplifying Circuit
Fig. 3. - Basic circuit of stable triode-valve amplifying stages.
The actual design was arrived at after considerable theoretical and experimental work, the details of which are given in a paper [★] F M Colebrook: A Study of the Possibilities of Radio-Frequency Voltage Amplification with Screen-Grid and with Triode Valves. JIEE, Vol. 74, pp. 187-198. by the present writer, published in the Journal of the Institution of Electrical Engineers. For the sake of brevity, the present account will be mainly confined to conclusions and results.
The basic circuit is shown in Fig. 3, from which it will be seen that the scheme is essentially an application of the familiar 'buffer' valve technique to the problem. The greater part of the actual voltage amplification is effected by the second of the two valves, the principal function of the first, the buffer or coupling valve, being to protect the input circuit against the effect of the negative input resistance of the amplifying valve. Alternatively, the function of the buffer valve can be described as changing the phase of the feed-back of energy from the output circuit in such a way as to prevent self-oscillation.
The success of the arrangement in its present form is due partly to a theoretical analysis of the most suitable values for the components, and partly to the high efficiency of modern valves, but the idea itself is not really novel, for something very similar was introduced many years ago by J. Scott-Taggart, under the title Tuned-aperiodic-tuned.
For reasons which are detailed in the article already referred to, it is found, rather surprisingly, that the most suitable type of valve for each position is the comparatively low voltage factor, or L type - or even the small power class, such as the PM202 or P215. This implies a fairly high step-up ratio in the output transformer, the design of which in relation to the valve used is quite standard. The coupling resistance in the anode circuit of the buffer valve is not critical, but should not exceed a few thousand Ohms.
As an example of the performance of such a combination of buffer and amplifying valve, the following case may be quoted:-
Valve- Voltage factor - 13/7.
A C resistance - 6,240 Ω.
Transformer- Secondary dynamic resistance - 150,000 Ω at 1,000 kHz.
Ratio - 4:1.
Coupling resistance- (R) - 5,000 Ω
Anode voltage 120.
The voltage amplification from the first grid to the secondary of the transformer was measured for various values of the frequency. At 1,000 kHz, for example, it was 132. The calculated value at this frequency was 3.67 for the buffer stage and 33.5 for the amplifying stage, giving a product of 123, which is close enough agreement in view of the uncertainty of some of the data, - more particularly the grid-anode capacity.
Fig. 4. - Amplification characteristic of triode- valve system at 1 MHz. Dotted curve represents amplification characteristic of screen-grid valve as in Fig. 2.
The variation of output voltage with input voltage applied to the first grid was found to be as shown in Fig. 4. It will be seen that there is no sign of any departure from linearity up to 80 Volts output. The dotted curve represents the performance of the screen-grid valve (see Fig. 2) over the same range.
The performance of this type of circuit in the long-wave range is particularly good. For example, with the same valves and components, and an output transformer having a dynamic resistance of 340,000 Ω at 200 kHz, with a. step-up ratio of 6:1, the measured amplifications were:-
and the response was substantially linear up to at least 110 Volts output, as shown in Fig. 5.
Fig. 5. - Amplification characteristic of triode system at 200 kHz.
Stability
In the experimental circuits the input and output coils were placed about one foot apart, in positions of minimum mutual inductance, but were not closely screened, the only screening being an earthed metal plate set up between the two circuits. In spite of this very rudimentary screening there was no tendency to instability. In fact, both calculation and measurement showed that the input circuit is somewhat over-stabilised by the rather low input shunt resistance of the buffer valve, particularly at the higher frequencies (1,000 kHz and thereabouts). With 5,000 Ω coupling resistance the buffer valve input shunt resistance may be as low as 10,000 Ω or so, rising to about 100,000 Ω as the coupling resistance is reduced to 500 Ω. In its simple form, as in Fig. 3, the arrangement would not, therefore, be very good, either in selectivity or sensitivity, at frequencies of the order of 1,000 kHz. There are, however, various means of removing this excessive damping of the input circuit, and of these the best and simplest is a semi-fixed retro-active arrangement, as shown in Fig. 6.
Fig. 6. - Triode-valve radio-frequency amplifying system, with retro active control of damping due to buffer valve.
If the retroaction-control capacitor is set to the maximum value consistent with stability at the minimum setting of the tuning capacitor it will be found that the damping of the input circuit is satisfactorily low over the whole tuning range. Alternatively, the retroaction control can be used in the ordinary way as a control of damping. The pre-set adjustment is, however, preferable in respect of simplicity of operation.
Retroaction can similarly be applied to the output circuit in any of the usual ways, but if this is done it must be remembered that the effect of retroaction is a great increase in the 'dynamic resistance' of the circuit, and the advantage of this in an amplifying stage is fully realised only if the transformer step-up ratio is made appropriately large.
The Advantages of a Low-Capacitance Triode
The performance figures given above do not represent the best that can be obtained by this system. The principal limitation is the grid-anode capacity. In triodes of normal design this is at least twice as large as it need be, because the anode and grid leads are brought through a small glass pinch to the valve pins in the base. For a normal triode in an ordinary valve-holder, the value appears to be anything up to 7 pF. It was pointed out in the full publication already referred to that if the disposition of electrodes in the triode was that which has been adopted for the tetrode, i.e., if the grid or anode were brought out to a top terminal, the grid-anode capacity could be reduced to 3 pF. or less, a change of considerable advantage in almost all applications of the triode. It is very unfortunate that the present construction became standardised before the harmful effects of grid-anode capacity had been sufficiently fully realised, and it is noted with satisfaction that in a recently announced series of valves, with a new type of holder, this separation of the electrode connections has actually been adopted.
As an example of the potentialities of the buffer-valve amplifying circuit with reduced grid-anode capacity, the following case may be quoted:-
μ = 10.5.
Ra = 3,000
(i.e. mutual conductance 3.5 mA/V)
Secondary dynamic resistance = 150,000 Ω at 1 MHz.
Grid-anode capacity = 3 pF
The valve figures are not unduly optimistic they are, in fact, realisable in existing vales and the low capacity could be realised as indicated above.
Calculation shows that with such valves it would be possible to get an amplification of 230 from the grid of the buffer valve to the transformer secondary at 1,000 kHz. Moreover, the effective shunt resistance of the buffer valve would be nearly four times as high as for a similar valve with the normal grid-anode capacity, an advantageous feature even it the damping effect of this resistance is eliminated as described above.
Practical Realisation
Finally, it is desired to emphasise the fact that the triode valve amplifying circuit here described is a useful and practical one at least for the home constructor if not for commercial construction. Apart from advantages in respect of linearity of operation, it is particularly suitable for the type of receiver in which high selectivity, with single-peak resonance curves, is combined with tone correction, since there is no upper limit to coil efficiency imposed by considerations of stability. The writer, for example, uses a home-made set of this kind, with two loosely coupled input circuits, a triode valve radio-frequency tuned-transformer stage, embodying two ordinary L-type battery valves, and connected as in Fig. 6, diode rectification, a tone correcting stage, one audio-frequency stage, and a mains output valve. Though only three tuned circuits are used, the selectivity is quite sufficient for the best British and foreign transmissions, and the quality is very good in relation to this selectivity.
|
