Amateur Radio's NewState of the Art? BY RAYM0ND C. PETIT,* W7GIIM IT HAS BEEN KNOWN for many years that the bandwidth used for cw reception in most all amateur applications is much,much wider than is needed for efficient reception of the signal. Today it is very common for an operator to use a receiving filter that is 100 times broader than needed and this huge amount of extra bandwidth permits a great deal of QRM to pass which otherwise would be rejected. There are several reasons why extremely narrow-bandwidth receiving systems have not been more widely used. The first is frequency stability If a filter only 10 hz wide were to be used in the receiver the desired signal would soon drift out of tile passband unless both tile transmitter and the receiver were manufactured to standards greatly exceeding present designs. Even getting the signal tuned correctly in the first place would be difficult because of inadequate dial resolution and backlash So far no commercial manufacturer of amateur equipment has considered it worthwhile to provide frequency accuracy or stability exceeding about 100 hz. Another problem with narrow filters is ringing. In a typical audio-filter design. Qs must bc kept low (and hence bandwidths broad) to prevent random noise and even tile signal itself from producing so much ringing that it masks the desired dots and dashes. A Bold Leap in Frequency Stability Recent advances in integrated-circuit tech- noloqy now make it possible to synthesize all tile local-oscillator signals needed for both transmission and reception - at a cost that is attractive to the amateur. With synthesis techniques only One is independent and highly stable crystal oscillator used as a reference to generate all the other signals needed. The signals so generated are just as accurate and stable as this one master standard and the standard can be set easily to within about 1 hz of WWV at 10 Mhz. With a synthesized transmitter and receiver you can tune to the frequency you want and know that you will be. within a few hertz of that frequency. If both stations in a QSO are operating with synthesized rigs the problem of drift is eliminated. Nine Hertz bandwidth and No Ringing If enough is known about the incoming CW signal. It is possible to build what is known as a ''matched filter'' for this signal at the receiving end. Such a filter has the narrowest bandwidth possible (and therefore the best possible signal- to-noise ratio) For that particular signal and it will not ring when correctly adjusted . For Morse code speed of 13 wpm such a filter is 9 hz wide and it's skirt selectivity can best characterized as spectacular . What needS to be known about the incoming signal? Naturally, you must know it's frequency which simply means that you must have the signal centered in the passband of the filter to within a few hertz. Here is where frequency-synthesized rigs are necessary. Operation of the matched filter also requires that the transmitted Signal be synchro- nized in time. To do this it is only necessary to generate a suitable timing signal by frequency division from the master-frequency standard. By Using thiS precise clock signal as the speed-control signal in an electronic keyer, all transitions from mark (dots and dashes) to space and vice versa occur at extremely well defined instants. For a 12 Wpm morse code signal the basic time interval is 100 milliseconds and all (dots and dashes begin and end at instants which are exact multiples of this interval. All dot, dashes and spaces can be thought of as being "blocked" into time segments of precise duration. The matched filter recovers the signal by analyzing the receiver output, block by block,and presenting the receiving operator with a cw tone corresponding to block. The intensity of each tone is proportional to the total signal energy that the filter received within its bandwidth during that time block. Adjustment of the phase Of the filter timing signal the receiver compensates for the propagation delay between the transmitter and the receiver. Experiments Show 20 - db Signal Boost Over QRM [f a receiver using a 2-khz bandwidth is provided with a matched filter having a 9hz bandwidth, theory indicates that 23.3-db improvement in signal to noise ratio will result. If your receiver uses a 500Hz filter the improvement is 17.4 db. What does this mean in practice, in crowded conditions on the air? The author built an experimental matched Filter for 12 wpm and a suitable Morse code test generator. A very weak signal from the generator was combined with severe 80 meter QRM including RTTY signals, other cw signals and static. The test signal was buried in the QRM. When the matched filter was switched in the signal stood out distinctly, almost like a code practice oscillator alone. Clearly here is a technique which could be a great help on our congested bands! Similar experiments by WA7ZVC have given equally worthwhile results. Coherent Cw on the Air Several amateurs, mostly in the U S Northwest are building coherent-Cw stations Using simple designs worked out by the author. Nets are being formed on 8O and 20 meters for on the air experi- mentation. and contacts via coherent CW. The first complete CCW station to go on the air has ben built by Andy McClaskey, WA7ZVC. As shown in Figs l & 2. he assembled his station from a Ten-Tec PM-2 transceiver and a number of home built models especially for ccw. Andy's frequency standard generates thee 4-Mhz,500, 4-, and 1 khz and 40 hz clock signals that are required for the various components. For transmitting the Ten-Tecs VFO is phase locked to the Standard and operates at 3550.000 khz. For receiving the VFO is phase locked to 3551.000khz.to provide a 1 khz beat note for the incoming ccw signal. The isolation amplifier is required to prevent a coherent spurious signal at 3550.000khz. from getting back into the receiver and blocking the system. The ccw keyer uses CMOS IC's Instead Of an internal Clock. The clock is a 40-hz Signal derived directly from the 4-Mhz standard, and this pre- serves the timing required for ccw operation. The experimental ccw filter Andy constructed is shown in fig 3. It provides a bandwidth of less than l0 hz at a 1-khz center frequency and has no ringing in the output. A 10-position phased setting switch is visible in the lower right corner of Fig 3. Ready ~ made circuit boards for a similar Filter are now available, as are kits of parts. This filter uses two quad op amps and eight CMOS IC', plus a lot of resistors and capacitors. For more information write to Charles Woodson, W6NEY 2301 Oak 5t. Berkeley CA "Woody" also edits the Coherent CW Newsletter (CCWN), which contains full technical information and diagrams, and complete ccw information. A current subscription will be provided to any amateur who is willing to build his own ccw station.