9. The Coherent Detector: Detection of Pulse Signals

In all the published work by other people on the homodyne receiver, one particularly important idealization is made, namely, it is always assumed that the local oscillator gives a pure oscillation even though it is synchronised to a component of a complex input spectrum. It has been pointed out by the present author in several papers that the local oscillation will not, in general, be a pure tone of the frequency of the wanted carrier. Many other, unwanted, components of the input spectrum will be present, although, if the circuit is correctly designed and adjusted, these will be largely discriminated against, and the local oscillation can usually be made substantially pure while retaining adequate synchronisation. Some analysis of the mechanism of the discriminating properties of a synchronised oscillator has been published, and all that need be given here is a reminder that the difficulties of synchronising are the ultimate limitations of the performance of the synchrodyne system. However, sometimes it can be arranged that a pure tone of the wanted carrier frequency is available, and then this tone can be used (in the correct phase relationship) to operate the frequency-changer of what is otherwise a synchrodyne receiver. Moreover, since the local oscillation is no longer dependent on the received carrier for synchronisation, the carrier can be intermittent--i.e., the receiver can be used for pulse signals. This arrangement has become known as the "coherent detector", and has been fully discussed in published papers; it has many important properties, those of most practical use being associated with the reception of pulse signals against a high level of noise background. If a comparison is made with the detection of pulse signals by an ordinary so-called linear detector (the usual diode detector with a large signal amplitude approximates to this), then it is found that as the input signal/noise ratio is decreased from about 3 to small fractional values, the improvement in detection given by the coherent detector increases rapidly. If the detection is measured by the ratio

increase in d.c. output when signal is applied
r.m.s. noise voltage when signal is absent

then, calling the input signal/noise ratio R it is found that for the coherent detector the detection is proportional to R, but for the linear detector it is proportional to R^2, provided R is small compared to unity. These properties of the coherent detector appear to have been noticed only comparatively recently.

10. Synchrodyne or Homodyne?

In view of the fact that the homodyne developed through the years into a more refined system which was identical with the synchrodyne, there seems to exist a difficulty as to what name the system should be given. It is clearly unsatisfactory that one system should have two different names. In discussing the coherent detector, Smith appears to identify it with the homodyne. This suggests a very nice way out of the difficulty. Since the prefix "homo" suggests "same", while the prefix "synchro" suggests "synchronised", the logical distinction of meaning is the following: ''homodyne" should signify a system where the local oscillation is, in fact, the same as the wanted carrier, being obtained by some direct process (e.g., by direct transmission from the transmitter), while "synchrodyne" should signify a system in which the local oscillation is obtained by synchronising an otherwise free oscillator to the incoming signal. This nomenclature has the big advantage of making obsolete the somewhat vague name "coherent detector"; the word "coherent" seems to lack an adequate technical definition. Whether it is now too late to make these changes remains to be seen.

11. Conclusions

It is hoped that this review of the somewhat obscure history of the homodyne, and its subsequent development as the synchrodyne, will have cleared up many previously doubtful and mistaken matters. The subject is one which concerns the author rather personally, but his position in relation to it should now be unambiguous. Although many interesting and important properties and principles have emerged from the work described here, nevertheless no commercial application of them in the radio field appears to have been made until the advent of colour television.

Last updated
17th September 2001