CCW NEWS Issue number 10 Compiled by G31RM February 1991 New Petit CCW filter I have almost completed a new CCW filter based on the W6NEY circuits in OST. This was prompted by the availability of the EPIL devices from Colin Horrabin. There are now three of these available and they are all priced at F-8.50 each. The devices are a synchronous keyer (of which there is more than one version), an audio synthesiser and a replacement for part of figure 15 in the QST articles. I haven't quite got the whole thing completed but preliminary tests using a simulated CCW signal show the correct responses. I feel sure it will be satisfactory. Copies of the work done so far are available and the final circuits will be added to the list in due course. If you would like to try these devices and don't want to wait until I have everything completed, let me know and I'll send copies. A 4 MHz clock is required. The circuits have been built using just a single five volt supply in place of the dual supplies in the original filter. Taking figure 13 first, the four op-amps used have been replaced by a single TLO84 and the 74C74 deleted. The remainder of the circuit is the same. In place of the 74C74 there is an EPLID which generates quadrature signals at 1 KHz which can be varied up and down by 64 Hz for fine tuning. The EPLID only produces two signals so one third of a 74HC 1 4 is used to give the other two phases. in figure 14 the four op-amps have again been replaced by a single TI-084 and the remainder of the circuit is the same as far as the outputs from the 10K resistors. This output is fed to a summing amplifier and a low pass filter both in a single TLO92. A CD4047 is used as an audio square wave oscillator variable between about 2KHz and 4KHz to give- an output tone from the filter of 50OHz to 1KHz variable as required. Figure 15 can be split in half. A divider chain using two 74HC390s and a 74HC74 provide output from the 4MHz clock at 1 KHz which can be used to drive the G3SBl keye I r. The same 1 KHz is used to clock a second EPLID which replaces the remainder of the circuit to generate the sample and dump pulses. No provision has been made for CCW at speeds other than 12 wpm. The phase shifting is done by two push buttons - one to advance and the other to retard. Fine adjustment is also provided in Colin's EPLD. Similar push buttons are used to increase and decrease the audio synthesiser frequency and to reset this frequency to 1 Khz if required. Although the total number of devices used has not been reduced to any large extent, the additional features should be useful. It is certainty much more compact than the earlier versions. New members DL2RW Hagen Henniger, Herrsching G4XOP Terry Cooper, St. Austell Y24XO Hans Bartz, Berlin Full list of members For -the benefit of those who do not have the set of Newsletters, here is a list of all amateurs on my mailing list: AKOB GOITIM G3XVR KBOR DF3CT GOIYY G48ZC NQ6Z DG7YFW GOLPL G4FKH OE1 KYB DJ3AM GOMIF G4GMV OE6DUG DJ4SB G3CCH G4HCC OE6WTD DJ7HS G3CW1 G4LRA PA3EZ1 DK1 10 G3EGQ G40KO SM6FPC DL2AAX G3FMW G4SYC VE1 LKM Di 2NI G3GRT G4VSO VE2KN DL2RW G31RM G4XOP VE3DPB DL3YDZ G3JPP G4YND VE30BE EA3FYZ G3LWM G7CBX W7GHM EA8YU G3MBN G14KIS WU01 F1LC1 G3RHl GM4PUQ Y23VH F61VT G3SBI GWOKZN Y24XO GOFVE G3TFV GW3DEX ZF1 HJ GORIN G3TMQ HB9DDO ZL1 BTT GOIKS G3XHQ K81XZ 3A2LF Keyer and monitor As mentioned earlier there are now three versions of the G3SBI keyer chip. The one I used is the mark I version and it is hoped that the circuits around this will not only form a synchronous keyer but also a monitor and signal generator. The chip accepts a 1 KHz square wave and produces correctly formed dots, dashes and spaces. The output is a square wave with elements of 100 mSec. as required. The 1 KHz- square wave is also tumed into a sine wave in an op-amp and this is swhched by a mosfet to produce the output tones. By ck)sing the key and holding it closed, perfect dots are provided for use as a CCW signal generator for checking the filter. Low level sine wave output is fed into the input of the filter. I intend to try using the square wave from the CD4047 in the filte in place of the 1 KHz square wave to give a monito frequency the same as the signal received. Another suggestion Danny Higgins G3XVR has been preparing some training courses for the Piccolo modem and came across what he thinks may be a solution to the CCW synchronisation problem which should also speed up the throughput at the same time. The modem uses 12 tones spaced at 20 Hz to send data in a manner very similar to CCW i.e. matched filters with tones lasting exactly 50 mSec. Each character is represented by a pair of tones, giving 144 combinations. These are used to send the ITA 5 character set (128 characters) plus one pair for standby tones and 15 invalid sequences which are not transmit- ted. With a two tone system (which Danny prefers to an on/off system) there are 4 combinations of tone pairs. If we call the tones A and B, then the combination AB could represent SPACE, AA could be DOT, BB could be DASH and BA is not used. Therefore, when not sending, the combination ABABABAB... would be transmitted conlinu- ousiy. A phase locked loop searching for this combination could automatically synchronise the clocks. The system is used in Piccolo to obtain sync. in less than 3 seconds. Since each dot and dash is always followed by a space, there is no need to send this space. Therefore the letter A would be represented by the sequence AABB, i.e. 200 mSecs. as opposed to 500 mSecs. for conventional CCW. This improvement is even greater on letters like D,v.,J and Z i.e. 400 mSecs. instead of 1300 mSecs. One drawback is that a processor is required to convert the CW to tone pairs and vice versa. Since the RF signal is not readily readable by casual observers, it may require permission from the National Authorities for this mode, although it could be argued on the grounds of experimen- taton. One way to experiment with all these different systems without having to build new hardware each time, is to use a general purpose DSP module and plug in different PROMs for each application (see the RSGB Radio Communication January 1991). There are plug-in modules for PCs and the PC could be used as the inpuVoutput device. High precision frequency standard Jim DeLoach WUOI has sent me a copy of an article by Gardner Johnson published in 73 Amateur Radio Today describing a standard which is locked to the US NTSC TV standard. The block diagram shows a pick-up loop on 15.7 KHz followed by a schmirt trigger and a divide by 63 stage to produce output on 250 Hz. The reference signal is on 1 MHz (not high enough for some CCW circuits) and this is also divided down to 250 Hz to feed a CD4046. This produces the error signal. Only four devices are used and the system could be adapted for other TV signals. It is very similar to the p.I.I. systems used where the long wave Droitwich and German broadcast stations have been used as bases for fre- quency standards. I have one or two designs from various sources for this type of standard though, so far, I have not been abiq to use one as a standard for CCW. Conversion of commercial equipment Kurt Arlt OE1 KYB has sent in some preliminary details of his conversion of the Trio TS-940. This looks like a rr enterprise and not to be undertaken too lightly witi _. expensive piece of equipment. His letter includes circui diagrams though, as yet, these are not complete. If anyone would like to try conversion please ask for copies of the diagrams. Suggestions for conversion of othe commercial equipment would be appreciated. Notes from members Ray Petit is collaborating with Bill Henry of HA Communications for the development and eventual manufacture of Clover equipment. He has also become a columnist for RTTY Journal. Ray also mentions a proposal from Jack ZF1 HJ that, if a computer had a 1 to 4 million word dictionary with each word indexed by a 20 to 22 bit number, with voice recognition and synthesis at each end, 'voice' in real time could be communicated over a Clover link, with a bandwidth only 3% of what voice now requires. How's that for a suggestion!! Geoff G4SYC, together with G4XVF, has been working on a semi-coherent system. AF is injected into the microphone socket of an SSB rig but with phase exchange keying of the tone. They have developed an ingenious way of detecting phase inversion at receiver without the need for an external phase refer( after an inital 'setting' signal. Geoff also suggests trra when using a conventional system it is unrealistic to try to use a p.I.I. to lock on to an incoming signal due to the gaps between words. Suppose CW was inverted in the marlqspace sense then the maximum length of no signal would be 300 mSec. In that case a p.I.I. could be made to remain near the correct frequency and maintain lock. John G3FMW has an oven controlled oscillator and one of the original Petit design filters ready and is modifying the WiCER receiver (ARRL Handbook). My own next move, when I have tidied ur) the latest filte and keyer, is to make a real effort to get a transceive working. Information sheets 40 (5 pages) New version CCW filter (preliminary) 41 (4 pages) High precision standard (73 Mag.) 42 (4 pages) TS-940 modifications (preliminary) Peter Lumb 732 Briarwood Avenue Bury St. Edmunds Suffolk IP33 3QF England