*Miller effect: the damping at higher frequencies caused by the internal capacitance between the elements within a valve, most apparent in the triode

Nevertheless, many early designs of TRF radio used the triode as a combined RF amplifier and detector. The two main forms employed were Anode Bend detection and Leaky Grid detection. You may come across a variant of Leaky Grid detection, called ‘Power Grid’.

In normal amplification, the working point of the valve is set at the straightest part of its curve. This allows relatively even amplification of both positive and negative-going parts of the AC (RF) signal. Anode bend detection works by selecting circuit components that set the grid bias negative potential so as to bring the ‘central’ point down to a position on the bottom of the valve response curve close to cut-off. Provided the input signal is strong, positive-going cycles are amplified normally, but negative-going signals are ‘crushed’ out of existence. The main drawback to anode-bend detection is the need for a very large input signal to prevent unacceptable levels of distortion. In practice the circuit is not suitable for long-distance reception, though obviously better than detection by diode alone as amplification is also provided, which the diode cannot provide.

The leaky grid detector offers much greater sensitivity and was used extensively in the TRF days. A typical circuit is shown in the diagram above. The capacitor C and the resistor R form a time-constant circuit. Let’s recap just what a time constant circuit is. The time-constant is the length of time a capacitor needs in order to charge up to approximately two thirds of its maximum value when a steady voltage is applied to a series resistor/capacitor network. Capacitors charge and discharge exponentially and in theory take an infinite time to become either fully charged or fully discharged.

Bearing that in mind, consider the leaky grid circuit shown in the diagram. Assuming C to be discharged and therefore offering a low resistance, the incoming signal tries to charge C to develop a potential difference across R. This potential difference would, of course, be the same as the incoming alternating signal except that on the positive-going half cycle of signal the valve grid switches the valve into harder conduction, thereby removing the positive potential. The additional grid positive created by this signal causes electron flow – opposites attract, remember - to occur through R, which charges the plate of C that is connected to the grid, or more simply put, charges C. Note that R is returned not to chassis but to the positive side of the 1.4V filament source. This is to provide a standing grid bias on the valve and is not directly connected with the detection process as such.

At the end of the positive-going half cycle C remains charged negatively so that the grid remains negatively charged also. When the negative half-cycle arrives, the grid becomes still more negative. At this time the electron flow to the grid stops, allowing C to discharge through R. This rate of discharge is determined by the time constant of R and C and must be, in practice, very small. In fact, the value of C must be small enough to become fully charged on each positive signal cycle, smaller than the period of one half cycle of signal.

Neither must the time constant be of a value that would cause excessive distortion of the shape of the grid signal envelope – though some distortion is inevitable - typical values of C being in the order of 150pF, and R in the order of 1MΩ.

The amplified and by now unidirectional signal current appears at the anode as a modulated current envelope consisting of tuned RF oscillations. This is then passed through a filter, often an RF choke followed by a bypass capacitor, which remove the RF and leave a ‘clean’ envelope of audio-frequency signal. In some very old sets, reaction circuits were employed to increase sensitivity and these would tap off the RF at the anode and feed it back positively to the grid via a coil loosely coupled to the tuning coils. A variable capacitor was used to set the critical level of feedback, which for greatest sensitivity was just before the circuit went unstable and burst into oscillation!

Power Grid Detection was used with indirectly heated AC mains valves. The values for both the grid leak resistor and the capacitor had rather smaller values and the detector valve operated with a very high anode voltage. These modifications provided an almost distortion – free signal rectification (detection) process.

In most TRF designs, RF Amplification is used to boost the received signal before detection. Typically, a pentode valve was employed for the purpose, due to the pentode’s superiority of amplification at high frequencies.

vari-mu valves

A special form of pentode was developed called the ‘Vari-mu’. With this valve, varied spacing of the turns of the grid spiral wire allowed gain control by varying the grid bias level. This method of volume control was used in some TRF designs, but the Vari-mu valve came into its own when used in superhet circuits (see later) when, due to the greater sensitivity of the superhet design, the need for automatic gain control became pressing.

The limitations of TRF reception.

With any TRF circuit there is the difficulty of making the tuned circuits track well over the wide range of received frequencies – and all the tuned circuits must be fully variable over the reception bands. Another is poor selectivity, where two or more stations may be heard at once. There is a tendency for circuits to become unstable and burst into oscillation when high gain is aimed for, yet high gain is essential for good reception. Reaction improves performance, but this can require a certain knack on the part of the operator, especially when striving to receive some distant or weak station that is at the very edge of the set’s sensitivity.

The superhet

Problems inherent in the TRF are largely overcome by the use of the Superheterodyne principle. Superhet receivers employ a method of reception called ‘beat reception’ where the received radio-frequency signals are combined with the signal generated by an oscillator in the receiver. This is called the local oscillator and is often part of a combined valve that does the two tasks, namely generating the local frequency and mixing it with the incoming signal. The result of this is a ‘beat’ frequency, well away from the radio-frequency signals and so not subject to interference by them. We call this the intermediate frequency, or IF and it is this frequency that is amplified before being demodulated for AF amplification. Because the IF frequency is fixed it does not require variable tuning no matter what radio frequency is being received so the IF circuits are less of a problem to design and can offer greater and more stable gain, resulting in a more sensitive and powerful receiver.

All rights reserved. © VRW 2008/2010