But Not for Voice

"These phone guys can't hear anything without a super," said a c.w. operator at a recent club meeting. It has been said many times before, in different ways.

Offhand it does seem as if a c.w. man's ability to do good long-range work with a 3 or 4 tube [regenerative] receiver may indicate that he is very bright or skillful. I'm inclined to think that on the average he is both.

However the voice man is not stupid. On the contrary the top-flight 'phone man knows as much about the use of his equipment as anyone in the game, and there is more of it. He has good ears and between them a good set of brains in daily use.

Two Kinds of Sensitivity

The receiver, however, is a great deal "dumber" for voice than for c.w. This is our story and a short one it is.

Suppose we start with the family of simple receivers which are descended from the old Browning-Drake receiver. You've all used them under various names: Pilot Super-Wasp, National SW-5 and SW-3, the "Diamond of the Air," and a lot of others.

Short-circuit the ticker coil of any one of them, and it is dull as mud.[Note 1] In the 80 meter band one must pour into it a terrible signal before standard output results. Anything from 500 microvolts to 2000 microvolts is required; that's 1/2 volt to 2 volts and will wreck a modern receiver's loudspeaker.[Note 2]

With the short removed from the tickler and the detector regeneration put into use, things look better as to sensitivity, but also there begins to appear the difference between c.w. and voice performance.

If the regeneration is advanced to the point where one can barely hear distortion on a high-quality musical signal, the sensitivity has improved so much that an input signal of only 100 microvolts will produce standard output.[Note 3]

Advancing regeneration still further, to the edge of actual oscillation, we find very decent sensitivity indeed, only 5 to 20 microvolts (depending on the particular receiver tested) now being needed to produce standard output. This compares pretty well (at 80 meters) with some superheterodynes of tolerably recent design. Of course the noise is now bad, as is the distortion.

Running the regeneration still higher, into the region of actual oscillation, we find the noise to go down abruptly while the sensitivity rises still further, so that direct comparison with even present-day superhets, on c.w. signals, is not very damaging to the simple set. Which of the sets is the noisier depends on the individual set chosen, and the kind of noise we are dealing with, and the signal level; so that any general answer is probably hasty or opinionated.

Strong-Signal Selectivity

Thus the simple set gave a much better account of itself for c.w. than it did for voice, when sensitivity is considered, removing suspicion as to the defects of the phone man's ears or brain.

On top of this there is the matter of interference. The simple regenerative receiver has a weak-signal selectivity which is very impressive for so simple a gadget; but alas, its strong-signal selectivity is very poor. There are other ways of saying the same thing. One can say that strong off-tune signals break through the selectivity of the regenerative receiver, though weak ones do not; or one can say that the selectivity curve has a very sharp peak but very wide flaring skirts.[Note 4] But it all means that in an amateur band full of phones the thing is out of luck.

The superheterodyne has more selectivity for two reasons, the simpler of which is that it uses more tuned circuits. If we count both the tunable r.f. circuits and the fixed-tune IF circuits, there are 5 to 9, against the 2 tuned circuits of the simpler receiver.[Note 5] In addition to this there is a selectivity gain to the process of converting a signal to the (lower) i.f. frequency. The contrast is a very severe one.

This is altogether apart from the use of crystals as "tuned circuits" in an i.f. amplifier. This device, available to the super but not to the t.r.f. receiver, still further increases the discrepancy in selectivity, and suppresses noise. Whether these gains are enough to pay for the accompanying defects is a matter of personal opinion. The difficulties produced by a crystal filter are: oftentimes a decrease of sensitivity, a tendency to put "tails" on telegraphic dots and dashes, and a quite horrible distortion of voice signals. The latter can be corrected in a large measure and very easily in the audio system, and this was done in the original crystal filter super, the one of Robinson called the "Stenode."[Note 6] It is only fair to point out that the application of a similar correcting-audio-system is not practical for the regenerative receiver as it will there produce a quite overwhelming amount of high-pitched noise.

Notes on the text by David Newkirk, W9BRD:

Note 1: This finding is misleading because merely short-circuiting the tickler coil couples "shorted-turn" loss into the detector's tuned circuit, greatly reducing the Q of its inductor and lowering the signal voltage that can be built up across it. A better test for available non-regenerative amplification would be to entirely disconnect the tickler coil at both ends, temporarily replace it with a short wire, and retune the circuit for maximum.

Note 2: These fractions-of-volts to volts equivalents would hold only if we changed Kruse's microvolts to millivolts. From the context, I think it likely that for the fractionals he indeed intended to say microvolts because the input-signal levels radio communicators encounter in practice are generally in the microvolts to hundreds of microvolts range. It's possible that these equivalancy errors were introduced through subsequent editing.

Note 3: "Standard output" can mean 0.5 W into a standard audio load, but there are other definitions of "standard output" that vary with how much undistorted audio output the receiver under evaluation is expected to be capable of producing. See F. Langford-Smith, editor, Radiotron Designer's Handbook, fourth edition, page 1298.

Note 4: Another reason the selectivity of a regenerative receiver cannot equal the selectivity of a superheterodyne (that does not achieve its adjacent-signal selectivity through regeneration) is that the maximum selectivity achievable with a regenerative detector is dynamic. That is, the maximum selectivity achievable with a regenerative detector varies with the strength of the incoming signal, just as the maximum regenerative gain varies with the strength of the incoming signal.

Note 5: This assumes that the "simpler receiver" has a tuned RF-amplifier stage.

Note 6: Not the H. A. Robinson, W3LW, of "Regenerative Detectors" (a classic article in January 1933 CE QST) and BC-348 (an LF/MF/HF communications receiver of World War 2) fame.

Robert S. Kruse (9LQ, 3ABI, 1OA, W1BAO) was a Director of the American Radio Relay League in 1921 and 1922 CE, and Technical Editor of its journal QST from 1923 to 1928 CE.

Robert S. Kruse's "But Not For Voice" was originally published on pages 68–68 of the April 1936 CE issue of Radio.
This page was revised on August 16, 2014 CE.
.Notes copyright © 2013, 2014 by David Newkirk (DavidNewkirk@gmail.com). All rights reserved.