CCWN 75:57 INTRODUCTION TO COHERENT CW BY CHAS. WOODSON Morse code CW signals have generally been handled as analog signals and decoding has generally been done by skilled human operators who distinguish between the signals at the desired frequency and the accompanying interfering signals and nolse. Filtering has generally been with filters having a bandwidth of around 2100 Hz or 500 Hz centered near the desired frequency. The short term stability and resetability of the transmitter and receiver have generally been such that variations of 100 Hz are common durinq communication. Morse code sent is a very regular digital fashion can be analyzed into a binary (on-off) pulse code with a pulse length equal to the length of a dot. The length of the key units of the Morse code, dots, dashes, one-unit spaces, three-unit spaces, and longer word-spaces, all have in common the unit of time equal to the dot length. Each Morse code message therefore converts into a series of signals of the pulse unit length. This allows the use of digital methods of processing Morse signals. If Morse code is transmitted in a regular way, the resulting coherent CW (CCW) may be detected by a coherent digital matched filter. This technique is considerably more effective than the usual analog methods. Coherent digital methods require that the pulse periods be equal and begin at sufficiently regular times that this timing may be obtained by the receiver (pulse phase acquisition) or is known so the receiver filter can be set to be in shase with these pulse periods. As transmitted the CW signal can be conceived of as a binarx,~ coded signal, that is a series of on/off pulses. If the pulse l~ngth of .1 second is selected, the operating speed is about 1- words per minute. Note that the signal is not of zero b?ndwidth. The modulation produces sidebands of energy, though they are much narrower than those produced by voice modulation. For example, if a series of dots is being sent, the sidebands are at 5 Hz and harmonics of 5 Hz as the signal is being modulated by a 5 Hz off and on signal. If we are sending a series o' t's as individual words, modulation would be a 1 & 4/6 Hz. Clearly if we could design a filter which was only a few cycles wide and did not ring, a great deal of improvement could be made in CW reception. The above paragraph described the CW signal as transmitted. As received, it is changed considerably. Obviously it is much weaker. Secondly, other signals, QB1 and noise have been added. Frequently the QRM and noise are stronger than the desired signal. Thirdly, less well known, the signal is subject to changes produced by the propagation path such as phase modulation. This modulation which may be conceived of as FM modulation I have found to often be about .5 Hz per skip by HF signals using the F layer, although it can be much less or more. I conclude that a CW signal may be considered to have a bandwidth of something like 7 or 8 cycles.