COHERENT C.W. NEWSLETTERX Compiled by G31RM 1W __ ro Issue number 26 February 1994 Publicity The article by Bill de Carle duly appeared in the February 1994 issue of QEX. I have added this to the list of articles I have available. COHERENT operation DL1 BT has written to say that he will be active with COHERENT in the near future. Strangely enough I have not heard from any of the members listed last time who have the board and program saying that they are active - what are they doing with the program? AMSAT (UK) has also ordered ten more boards - I wonder what they are doing with them. Bill de Carle has sent me another list of amateurs who have the COHERENT program. They are: DF1AE DL1BT DF2KI3 DJ2ZV DL2AD DL2FI DL6KL EA2SN VE3GF and W6HDO. DF2KB and DJ2ZV both ordered two complete systems, the second copies intended for use by their friends DL6KL and DL2AD, so it looks as though there will be activity from these stations even if they are only working each other. Palmtop computers There are a lot of palmtop devices around but most of them appear to be nothing more than elaborate diaries and schedulers. One of the exceptions is the Sharp PC3000. There is also another version, the PC31 00, which has a larger memory. Apart from the usual scheduler, calculator, clock and to-do list it has a complete file manager and laplink. The latter enables direct transfers between it and other computers. It also has RS232 and printer ports and a bus expansion connector. It can be run at either 5 or 10 Mhz. It is claimed to be an XT model with the added extras mentioned above. It also has a full 6Ox25 display so it can run many programs without any problems. In addition there is a 1.44Mb disk drive as an optional extra. Memory cards can also be used but they are expensive. I first tried to use the PC3100 with the G4BMK program. This program uses a calibration routine to make adjustments for different clock frequencies. Running the program produces two figures which are entered before the program is used. I could not obtain anything like consistent results and, in the end, struck a compromise and used an intermediate pair of figures. Even so the program ran well and I had no problems using it with the BMK terminal. This problem of setting the frequencies made me wonder if they use some sort of cheap oscillator which is not very stable and this led me to wonder how the computer would work with COHERENT which also relies on the computer for timing. I suppose the answer will be try it and see what happens. I am now using the PC3100 for amtor, pactor, cw and rtty using the MFJ-1278B controller. After much searching I found a program called COMMO which I can recommend as a very good terminal program. It is possible to redefine all the keys for special purposes and it is also possible to rewrite the help screen to match the alterations. I have the F-keys set up to change modes and send changeover signals and a number of messages are stored on the Alt-keys. What is more there is not the faintest sign of any interference from the computer in the receiver. I have the latest version of COMMO. The shareware version is a much improved version of the one which is distributed by the major dealers so, if you want a copy to try, let me know. I will even include my macros (as far as I have got with them) for using the program with the MFJ-1278B. The Petit filter Just when I thought that this was becoming obsolete I had a letter from Ray Petit with some interesting suggestions. Here they are (including American spelling and grammar!). It has been apparent to me for some time that further improvement is possible in the CCW receiving processor concept. Clearly the Petit filter is antiquated: VE210's product gets the CCW community abreast of the times. But the sidelobe selectivity of the "pure integrate-and-dump" can be improveddramaticallywhileobtainingawidercenter lobe of the frequency response curve. This can be done, either in hardware, with a revision of the Petit filter design, or with software changes in VE21Q's machine. I write the following as an invitation for any CCW enthusiast to do some more home building and experimenting, and as food for thought for an upgrade from Bill, since the equivalent processing can be done in software. We are all familiar with the "(sine (pi * x))/ (pi * x)" curve of CCW with its zeroes'of response at plus- and-minus multiples of 10 Hz, sidelobes at 13.5 dB down (l 5 Hz), la dB (25 Hz), 21 dB (35 Hz), and so on. But even at 95 Hz it is only 29.5 dB down and at 500 Hz it is only 44 dB. The two "integrate-and-dump" processors in the Petit filter could be replaced with 8-pole Bessel lowpass filters having a cutoff frequency set to 6.4 Hz. At this cutoff frequency, their risetimes are equal to the CCW pulse duration, 100 milliseconds. The frequency response at the cutoff frequency is -3 dB (by definition); at 12.7 Hz it is about -1 3 dB-, at 25.5 Hz it is -51 dB, and at 38 Hz it is -80 dB. The Bessel filter does not "ring", and its frequency response has no sidelobes Commercial switched-capacitor 8-pole Bessel filters are available and inexpensive. Representative part: MAXIM MAX296CPA, listed for $6.02 in the current Digi-key catalog. Using these filters, the RECEIVER selectivity is dramatically improved, but the TRANSMITTED signals still have the wider spectrum of near- rectangular pulses. If the transmitted signals are shaped properly, CCW signals could be spaced 50 Hz apart without mutual interference except in the most extreme circumstances. But the transmitted signals would also sound very different than CCW signals sound today! The output of the CCW filter would still sound the same, but the "on-air" signal would have slow, smooth transitions from "on" to ..off" that would make it harder to read by ear. To take advantage of the selectivity these filters offer, it is essential that no circuits in the radio preceding the filter are overloaded by strong nearby signals. My suggestion is to use your narrow CW filter in the radio, and then adjust the rf gain of the radio such that no signal in the passband causes pumping of the AGC. Then adjust the audio level such that no signal causes audio distortion. Then the radio's headphone output can go to the filter. If the signal you are listening to is weak at the input of the filter, it will be weak at the output too, but no other signals will be present, for practical purposes, unless they are almost on top of yours. So, you can put d.c. gain at the output of the filter and draw your weak one out. You can even put AGC here, on the output side. But if you put AGC on the input side, the strong neighbor signals control it, and your signal goes down. Another development that invites fresh development work is the appearance on the market of single-chip "RISC" microcontrollers which contain virtually everything needed to make a CCW filter, including an 8-bit analog-to-digital converter: MICROCHIP PIC16C71. So perhaps there is still life in the hardware CCW filter! I only hope that if this idea is taken up - perhaps Ray will have time to do it himself now that CLOVER is up and running - it will be compatible with the old Petit filter and also with COHERENT. Information sheets The old list of information sheets I have collected is getting somewhat out-of-date and contains many items which are no longer of interest so I have updated the list and am including a copy with this Newsletter. The COHERENT program I have received updates to the program from Bill. The latest is still experimental and is version 2.3. There are two alterations. The first is the inclusion of two slow speeds - 6 and 4 words per minute. These are obviously intended for use by the lowfers who could well prove to be Bill's best customers as CCW offers so much to those who are restricted to very low power. The second is what he calls the G3IRM mode - fame at last! This is a result of the suggestion I made that I would like to be able to use the headphones in place of the internal computer speaker. Both amendments will be included in future issues of the disk and are optional. You enter the instructions to start the additions when starting the program. All the amendment notes are included in the list included with this Newsletter. Bill de Carle's latest Having mentioned Ray Petit's latest ideas it is only fair that I should mention the latest from Bill from the CEX article. Here is some of it abbreviated. The ideal CW filter would be 0 Hz wide and have an instantaneous response time. How can we design such a filter? How much of the signal is needed? Could we take a momentary snippet out of a waveform, analyze it extensively on a fast computer and figure out its complete spectral content just from that tiny portion we looked at? The answer is yes. For more read the QEX article. 73 de G31RM 2 Briarwood Avenue, Bury St. Edmunds Suffolk IP33 3QF, England.