Ecology personified! It's off the drawing board and ready for use. Using less san be more as- you'll find out in Part I of Abe Weiss's two part article on coherent s.w. Coherent C.W. - The C.W. Of The Future Part I BY ADRIAN WEISS*, K8EEG10 During the past two years or so, a new technique for the transmission and reception of c.w. has been developed which promises to revolutionize c.w. communications in a manner analogous to the revolution in voice communications brought about by the development of s.s.b. However, the overall strategic impact of coherent c.w. (c.c.w.) could be far greater because it is effective in circumstances which render conventional c.w. and s.s.b. useless. A comparison of the effectiveness of c.c.w., conventional c.w., and s.s.b. can best illustrate the point. Given current receiving techniques, in which conventional c.w. utilizes a 500 Hz bandwidth, s.s.b. a 2.1 kHz bandwidth, and c.c.w. a 9 Hz bandwidth, a- 5 watt c.c.w.. signal, which can be generated under every imaginable set of circumstances, would be on par with a 320 watt conventional c.w. signal about 18 dB improvement) and a 1250 watt s.s.b.' signal Ed. Note: Strictly speaking, s.s.b. cannot be directly compared to c.w:; the emphasis here is upon relative bandwidths. in perspective, 30 dB separates a 5 watt signal from a 5 kilowatt signal! Indeed, coherent c.w. is a new frontier in communications technology! Until recently, communications systems utilizing time-averaging techniques to produce such signal- to-noise improvements were exotic entities re- served for specialized applications such as moon- bounce work, and functioned at totally impractical pulse-lengths of several seconds and more. Their main purpose was, through an analysis of a chart- printout, to indicate whether or not a signal was received via the moon-bounce path. C.c.w. deals with pulse lengths that are practical as far as nor- mal communications are concerned. Advances in integrated circuit technology have made it possible, furthermore, for the average amateur, gifted with a finite amount of technical knowledge, time and money, to assemble a c.c.w. station, or more practically, to add several complementary units to an existing c.w. station in order to convert it to c.c.w. operation without any basic equipment modification A c.c.w. signal is relatively easy to generate, since it is nothing more than a conventional "on-off" continuous wave, and hence, any gear capable of conventional c.w. can be used as the basis of a c.w. station. The complementary units necessary to convert conventional c.w. gear to c.c.w. operation serve the purpose of achieving the frequency accuracy/stability and time-synchronization which are the essential requirements of c.c.w. These complementary units will be discussed in detail below. First, a discussion of the basics of the coherent c.w. concept are in order. Coherence The term "coherent" in c.c.w. refers to the fact that, in the communications system consisting of the c.c.w. transmitter and receiver, the transmitted signal exhibits distinguishing characteristics to which the receiver is designed to respond. A c.c.w. signal has three distinguishing characteristics: 1.J precise frequency; 2.) accurately established pulse length; 3.) predetermined "turn-on/turn-off" instant, or pulse-phase. Fig. 1 shows the basic components of a c.c.w. communications system. Coherent operation is as follows. A 4 MHz master frequency standard at both ends is calibrated with extreme accuracy and stability against a WWV signal. From this master standard are derived, through digital techniques, signals which determine the precise frequency of the transmitter and receiver, the timing signal which determines the basic pulse length or bit, and the timing signal which establishes the pulse-phase or the exact instant at which each pulse or bit begins and ends. At the transmitter end, the c.c.w. keyer generates a signal whose pulse length and phase are in exact step with the timing signal from the master standard. At the receiving end, the c.c.w. filter samples the received signal in units of time which are in exact step with the timing signal from the master standard. In other words, the keyer and filter are precisely "matched" in a c.c.w. system, and their operation is "coherent." This coherent matching is rendered possible by the fact that the master frequency standards at both ends are calibrated to within a very close accuracy, specifically 10-' or 1 Hz, to the common WWV standard. Because of geographic variances 'Ed. Note: In this paper, the ~&classical" c.c.w. model is used throughout for sake of convenience. The c.c.w. system can be designed to use any combination of de- rived timing signals, integrating periods, frequencies etc. between the two c.c.w. stations with respect to the location of the common WWV standard, propagation delay effects may introduce some small degree of inaccuracy into the absolute frequency of either standard. However, this inaccuracy, in terms of current c.c.w. practice, is insignificant with respect to the c.c.w. signal frequency and pulse length. It is significant with respect to pulse phase. To permit precise time-synchronization between keyer and distant c.c.w. filter, the filter pulse phase is adjustable in 0.01 Hz increments over a 0.1 Hz range, the standard pulse length. This is sufficient to offset propagation delay effects in pulse phase synchronization. In short, current c.c.w. technique employs a "fixed frequency, fixed pulse length, adjustable pulse-phase" approach, although other combinations are possible. Another way of describing what happens in the c.c.w. system of fig. 1 is this: the c.c.w. filter knows that the transmitted c.c.w. signal will be within +1 Hz of a given frequency, will turn on-off at precise instants, and the on-off periods will last a precisely predetermined length of time. The effectiveness of c.c.w. results from the fact that the odds against a conventional c.w. signal exhibiting these three precise characteristics are astronomical. Likewise, random noise is rejected by the filter. C. W. Speed/Bandwidth/ Stability The remarkable improvement in signal-to-noise ratio achieved by c.c.w. is attributable to the fact that c.c.w. applies the classical principle which states that, for the most effective reception of a signal, the receiver bandwidth should be no narrower or wider than the bandwidth of that signal. If the bandwidth is wider, then the receiver passes unwanted energy or interference in addition to the desired signal; if narrower, some of the desired signal is lost. Next, the bandwidth of a signal is proportional to the amount of information contained per unit time. The 3 dB bandwidth of an a.m. phone signal is about 5 kHz; with s.s.b., a 2.1 kHz bandwidth is current practice. In regard to c.w., a 60 w.p.m. signal exhibits a 45 Hz bandwidth, while a 12 w.p.m. signal is 9 Hz wide. At this point, the frequency/timing accuracies required for use of narrower bandwidths become impractical. Cur- rent c.c.w. practice is to operate at this threshold of readily attainable accuracy w.p.m. and 9 Hz bandwidth,though anything faster and wider is feasible. Obviously, as the bandwidth narrows, a trade- off occurs between the information rate and susceptibility to interference: a 12 w.p.m. c.w. signal has a far lower information rate than an s.s.b. 3550KHz Fig. The WA7ZVC Frequency stabilizer circuit. signal, but permits narrowing the bandwidth to 9 Hz, making it possible to eliminate much of the interference energy that could pass through the 2.1 kHz s.s.b. bandwidth. The total noise energy passing through a 9 Hz bandwidth is 9/2100 of that which passes through a 2.1 kHz filter, and hence, the energy of a 12 w.p.m. signal needed to successfully compete with that noise is correspondingly diminished. The result is the improvement in signal-to-noise ratio. In short, the effectiveness of the c.c.w. technique is due to the fact that it operates at the proper bandwidth for the most efficient reception of a c.w. signal of a given speed. The addition of "processing information" in the form of accurate pulse length and pulse-phase increases the selectivity of the c.c.w. system beyond that which the narrow bandwidth would allow on its own. See fig. 2 for a graphic comparison of c.c.w. vs c.w. vs s.s.b. bandwidths. The frequency stability required to utilize the 9 Hz bandwidth of a 12 w.p.m. signal clearly is far beyond levels achieved in current practice by either homebrewers or commercial producers of amateur c.w. gear. Generally speaking, stabilities of 100 Hz (10~5) are considered adequate for conventional c.w. in view of the fact that receiver c.w. band- widths rarely are narrower than 200 Hz or so. Even with IC technology and active audio filter tech- niques, "ringing" of narrowband filters can hardly 28 • CQ • June, 1977 be avoided unless the c.w. signal bandwidth is a fraction of the filter bandwidth. In contrast, the Petit c.c.w. filter requires drift be limited to 2 Hz at most, or an accuracy/stability of 10'. While this magnitude may seem like science fiction to those accustomed to stabilities 100 times less, it can be obtained in the hamshack with care. W7GHM feels that accuracies of 10 l° are not beyond amateur practice, but would require a great deal more effort. Thus far, c.c.w. experimenters have relied upon the frequency standard design by K4EEU (ham radio, February, 1974), for the frequency and timing sig- nals used in the c.c.w. station. Pulse Length/Pulse-Phase A codespeed of 12 w.p.m. is based upon a "pulse" or "bit" length of 0.1 second, or 10 bits per second. In c.c.w. practice, the "dit" or "dot" of Morse Code is the basic pulse of 0.1 second, the "dah" or "dash" is exactly three times that length, or 0.3 seconds, and "spaces" are off-periods 0.1 seconds or a multiple thereof. The c.c.w. filter is set to process incoming signals in blocks of 0.1 second, and the keyer generates bits and spaces of the exact same duration. The timing signals de- rived from the master frequency standards insure an extremely high degree of precision in the estab- lishment of the pulse length. Hence, a c.c.w. signal exhibits precisely formed Morse characters. Con ventional keyers will produce well-formed charac- ters, but these will appear of random bit length to the c.c.w. filter. It is highly unlikely that a conven- tionai keyer will either be set to precisely 12 w.p.m. by the operator, or that a conventional keyer will be forming bits precisely at the 0.1 second rate, or that each bit will start at a precise predetermined instant. As a result, the c.c.w. filter will not respond to conventional c.w. signals. Finally, the c.c.w. signal is distinguished from a conventional c.w. signal in that a pulse will begin and end at highly predictable, precise, time points. The c.c.w. keyer employs as its "clock" a timing signal from the master frequency standard for the generation of both "dots" and "dashes" in contrast to a standard keyer, which employs either single or separate clocks for the formation of the two basic character elements. The conventional keyer-clocks are activated at random instants selected by the paddle operator. In the c.c.w. keyer, the clocking signal runs continuously, and the operator must adjust his manipulation of the paddle to the clock- ing signal, which determines the precise instants at which a pulse will begin and end. The c.c.w. filter processing periods, similarly, are controlled by a timing signal from the master frequency stan- dard that precisely duplicates that controlling the generation of pulses in the keyer. As noted earlier, the Petit Filter permits a pulse phase adjustment to offset the slight inaccuracies between the two timing signals caused by propagation delay effects. The adjustment of phase at the c.c.w. filter brings the initiation of filter sampling intervals into pre- cise step with the initiation timings of the distant keyer so that time-synchronization is achieved. The Petit c.c.w. Filter permits a small error in phase of a few hundredths of a second before significant intelligibility is lost. Summary A c.c.w. signal differs from a conventional c.w. signal in several respects. First, the frequency accuracy/stability of the c.c.w. signal is about 100 times greater than that of the conventional c.w. signal. The accuracy/stability figure of 107 is necessary to locate and keep the c.c.w. signal in the 9 Hz passband of the c.c.w. filter. Secondly, the pulse length and pulse-phase of the c.c.w. signal are precisely controlled. Each bit will begin and end at a precise instant, and each bit will last a predetermined length of time. With respect to re- ceiving c.c.w., the c.c.w. filter analyzes the energy present in the 9 Hz passband in 0.1 second blocks, and will indicate the presence of a signal that is at zerobeat, beginning and ending at predetermined instants, and lasting the predetermined length of time; conversely, spaces indicating the absence of a c.c.w. signal will conform to the same time dis- cipline. In comparison to the c.c.w. signal, all other r.f. energy passing through the filter will exhibit a randomness and be ignored by it. The function and operation of the various units of a c.c.w. station will be discussed next. The CCW Station The c.c.w. station consists of several units as follows: 1.) A master frequency standard which is capable of the accuracy/stability figure of 10-7 or 1 Hz, and which provides proper outputs for the various tim- ing signals required by the c.c.w. system. 2.) A transmitter/receiver capable of maintaining the required frequency accuracy/stability to hold the received/transmitted signal within the 9 Hz passband of the Petit c.c.w. Filter. 3.) A c.c.w. keyer that can generate the processed Morse characters in synchronization with the mas- ter frequency standard. 4.) A c.c.w. filter which is matched to the c.c.w. keyer and which can perform the time-averaging function required in order to reconstruct the re- ceived c.c.w. signal. These units will be discussed in detail in the fol- lowing paragraphs. CCW Keyer A conventional keyer is easily modified for c.c.w. operation. The modification consists of disconnect ing the keyer's internal clock, and substituting a timing signal from the master frequency standard as the new clocking signal. The clocks of standard keyers begin upon command from the operator as he depresses the paddle. The c.c.w. keyer is differ- ent in this respect. The clock signal is running con- tinuously and provides the commands for the be- ginning of a bit in precise step with the timing signal from the frequency standard. The operator must therefore fit his manipulations to the keyer. Since the clock in the c.c.w. keyer initiates a pulse period every 0.1 second in the classical c.c.w. system, and since the paddle may be depressed by the operator at any point during that 0.1 second period, the actual keyer output may lag behind the paddle up to .0999 etc seconds, depending upon how early into the previous period the paddle was depressed. Experimenters report that this aspect of using the c.c.w. keyer is a problem at first, but that after a few QSO's, one becomes readily accus- tomed to it. W7GHM and W6NEY are exploring various methods of artificially removing the clock- paddle lag, so that the operator hears an audio signal in step with the paddle while the transmitted signal is actually lagging an average .05 seconds behind. rWA7ZVC's modification of the CMOS keyer by WB2DFA (ham radio, June, 1974) for c.c.w. opera- tion is shown in fig. 3. SW1 permits selection of either conventional or c.c.w. operation. Fig. 3B shows mods for the popular "Accu-Keyer." Master Frequency Standard The frequency standard is the heart of the c.c.w system. All signals for operation of the receiver/ transmitter, c.c.w. keyer, and c.c.w. filter, are de- rived from it. The classical c.c.w. system described in this paper uses a 4 MHz standard, but, as noted earlier, this is entirely arbitrary and relative to the designer's desires. The standard currently used by c.c.w. experimenters is described by K4EEU (ham radio, February, 1974). It is a high accuracy, tem- perature compensated 4 MHz crystal oscillator with iC dividers to obtain 18 precision frequency out- puts in the 1 Hz-2 MHz range. In order to provide timing signals not derived in the K4EEU standard, an additional divider circuit employing five IC's is added to the unit, as shown in fig. 4. Furthermore, the use of one standard is optional, W6NEY uses three separate ones in his setup. Accuracy and stability is the essential factor. Receiver/ Transmitter Stabilization No available amateur gear is capable of the frequency accuracy/stability of 10-7, or 1 Hz, re- quired for c.c.w. operation. However, it is possible, through synthesis techniques, to achieve that figure with existing equipment using signals derived digitally from the master frequency standard. Gen- erally, a frequency reference signal from the standard is phase-locked to transmitter/receiver local oscillators through additional circuitry and corrects any drift that occurs during operation. Two approaches have been successfully used and are similar: first, WA7ZVC's modification of a TenTec PM-1 with its single local oscillator; and second, W6NEY's modification of a double-conversion Heath SB303. Their operation follows. 1.) WA7ZVC System. The first practical c.c.w. sta- tion was assembled by WA7ZVC and used for the first on-the-air c.c.w. contact in January of 1975. A Ten-Tec PM-1 was used because of its low cost and the fact that only one free-running oscillator required stabilization. Also, WA7ZVC wanted to illustrate that, with the c.c.w. technique, even the simplest of gear could be converted to c.c.w. Fig. 5 shows the WA7ZVC c.c.w. station. What is of interest to us here is the frequency stabilization unit, shown in fig. 6. Its operation is described in the following paragraph. A sample of the v.f.o. output is squared up by Q1 and U1 and passed on to U2, a digital IC mixer. The 3550.000 kHz harmonic signal from the master frequency standard is mixed in U2a with the sampled signal to produce a 50 kHz + (undesired- drift) signal, which is then passed to the transmit offset and receive offset chains. That signal is mixed in U3a with a 50.000 kHz signal from the standard, and if there is a difference in frequency, U3a generates a control voltage proportionate to the amount of difference. That control voltage is applied to the varactor tuning diode which has been added to the TenTec PM-1 v.f.o. board, caus- in it to pull the v.f.o. back to a 3550.000 kHz fre- quency. The same process occurs in the receive offset chain, except that in U3b, a 1 kHz difference with respect to the 50 kHz reference signal and the 50 kHz > (undesired-drift) v.f.o. sample signal is desired. This 1 kHz difference is important. The re- ceiver will be 1 kHz away from the desired 3550.000 kHz c.c.w. signal and produce the necessary 1 kHz audio output to the c.c.w. filter. If the mixed signal from U2b does not exhibit the 1 kHz difference be- tween the v.f.o. sample and the frequency standard reference signal, then U3b produces a control volt- age that causes the varactor tuning diode to pull the receiver frequency to produce the desired 1 kHz offset at 3449.000 kHz or 3551.000 kHz. The phase-lock approach used by WA7ZVC in the simple PM-1 is applicable to more complex, multiple conversion receivers such as W6NEY's SB303. We will continue our discussion of stabil- ization techniques, as well as the Petit c.c.w. Filter itself, in the next installment of this paper. (To Be Continued)