COHERENT C.W.
                 NEWSLETTER



Compiled

by
G31RM



Issue number 23



September 1993



Kurt Arlt 0E1KYB
I am sorry to have to tell you that Kurt died on the
31 st of May. He was one of the earliest members to
join the group and was one of the first to have a
CCW station on the air making his first contacts with
DF3CT. Even though he was fighting throat cancer
for over two years he was still talking of the
improvements he was making to his station when he
died. He was a true amateur.
Cassette tapes
I have been somewhat surprised at the lack of
interest in these tapes. So far I have had a request
from one member for copies and two requests from
non-members.
Computer CCW
This must be possible, or so I decided some time
ago. As a result I have spent much of my spare time
recently trying to make a start and have managed to
get part of the way. At least I think I have although
the present design is no where near a CCW system
- it is probably somewhere between CW and CCW.
In the hope that someone else may come up with
some suggestions I am going to describe what I
have done.
Signal generation:
The first thing I decided I needed was a CCW
signal. This was built using a crystal oscillator with
dividers to produce 1 Khz and 800 Hz outputs. The
first of these drives a G3SBI keyer. Dots and dashes
from this key the 800 Hz output to produce the
same output as my receiver except for noise and
interference. I had decided that I would see what
could be done with a perfect signal first before
trying anything else. Ultimately the keyer will be
used to key a very low power signal in one of the
amateur bands which can then be received by the
receiver together with any interference there may
be.
The computer:

The next question to be settled was the choice of
computer. The obvious choice would be to use an
IBM PC but this seemed rather like using a steam



roller to crack a nut and everyone does not have an
IBM type PC. The alternative would be to use a
small home-built device using something like an
8032 and use this as a stand alone computer.
Possibly a few keys would be needed but there
would be no need for a full size keyboard or a
display. The output from the decoder can always be
connected into a PC or other computer running a
CW decode program if a print out is required
Hardware or software:
When building almost any piece of equipment using
a computer it is necessary to decide how much is
to be hardware and how much is to be software.
The best solution seems to be to use as much
software as possible and only use hardware when
the software solution is difficult or impossible.
System development:
I have one of those 8032 based computer systems
described in Elektor Magazine in May 1991 together
with the extension board described in April 1992. A
very useful series of articles on programming the
8032 was in the same magazine from February 1992
to November 1992. This has been used as the
development system, all the surplus devices can be
left out when a final version is built. The 8032 uses
its own crystal but this will be replaced later with the
station standard. The standard will also supply other
waveforms need-
Program development:
As I have said above this is not a CCW system - yet.
For a start I decided to concentrate on the timing
aspect making this comply with the CCW standard.
First a 5 Hz square wave was generated and used
as one of the inputs to the 8032. A second input
was the keyed 800 Hz from the signal generator.
This didn't work. As each 1 00 msec finished another
started so there was no time to process any data
received before the next 100 msec input was read.
It was then that I decided that as one mark frame
contained exactly 80 cycles of audio a 50 msec
frame would contain exactly 40 cycles. The timing
signal was altered and ft worked as expected

counting the input signal for 50 msec leaving the
other 50 msec for processing. This could obviously
be amended by using the 80120 ratio output from a
74C90 counting for 80 msec (64 cycles) and
processing during the remaining 20 msec. Again as
expected it worked

Phasing:

As we all know the transmitting and receiving
stations are using 1 00 msec timing but they are not
necessarily in phase. The Petit filter uses a C04017
for this purpose so I decided I could do the same.
The 74C90 was removed and replaced by a
CD4017. Using the decoded outputs ten phase
positions are available. At the moment these are
switched but I would think that they could be
connected to ports on the 8032 and software used
to decide which output is used. I had to rewrite my
program again as the timing is now 90/10 - the
input reading time is now 90 msec (72 cycles) and
the processing time is the remaining 10 msec. To
test the phasing I drove the signal generator from a
separate oscillator so it now acts the same as would
a signal received via the receiver.
Output:
An output from the 8032 board is used to key an
audio oscillator (a 555) into a speaker. The gating is
from the dividers so that each time a 90 msec count
is complete the oscillator is keyed for 100msec
Keying three frames in a row makes a dash. As the
program runs it counts the number of audio cycles
received. A perfect CCW frame will produce a nil
output for a space and a count of 72 for a mark
frame (three of these in a row for a dash). Obviously
with interference present there will not be an exact
count of 72 every time. Where do I draw the line
between a space and a mark? I wonder if I could
AND the incoming signal with a locally generated
signal on Boo Hz. Could I then assume that ff I get,
say, ten hits in a row the frame was a mark?
Conclusions:
Everything so far is perfect in the sense that I am
only using perfect CCW with no noise or
interference. Perfect CCW can be relied upon during
contacts but there will always be interference. The
filtering which results from the input mixing in the
Petit filter does not exist. The accurate timing of the
Petit filter is there but that is all. Just to see what
happens I now intend to feed an r-f. signal as
mentioned above into the receiver and see how I
can decode weak CCW under interference.
What next:
I have used a lot of this Newsletter describing this
and wonder if anyone can produce any clever ideas



as to where I can go next. I don't quite know where
to go. At the moment there is very little hardware
needed and the program is only about 120 bytes
long so there will be plenty of room to make a
longer program and still get it into the ROM version
of the 8032 making the whole computer out of one
chip though it may be cheaper to use the three chip
version as the ROMIess 8032 is much lower in price.
Commercial transceivers of adequate stability are
now available, a standard is not hard to make and
it would be nice to have a simpler version of the
filter which, when all is said and done, is now at
least fifteen years old.
Expedition to the Arctic Circle
Paul GOBHA has been accepted for this expedition
and will be getting ready. The project he has
decided to do is to operate a CCW station and Dave
Rycroft G40KO of RAFARS is coordinating. He has
appealed in QRV for volunteers to help with
construction. One RAFARS member G3DOP has
written to me but others are needed. Write to me if
you feel you can help (or direct to Dave). If nothing
else please try to be able to receive CCW when
Paul goes on his trip. Better still try and get a CCW
station on the air and and work him
New members
G3DOP G3NCL G3WBI WD4PLI
Correspondence
There is not a great deal to report this time though
I have had one of Cliff Buttschardt's long and
interesting letters. He and a friend WD4PLI are
hoping to get some CCW tests going - they both
have the original Petit boards. David WD4PLI has
offered to lay out and produce printed circuit boards
for a revised filter. As Cliff says we all would like to
see a more up-to-date design available. Two further
letters from Cliff give details of the ideas they are
working on - more next issue. He has also offered
to copy the tapes for U.S.A. members and any non-
U.S.A. articles on my list. I intend to get the more
important ones copied and send them together with
the tapes to him in the near future.
I can always rely on Bernhard DF3CT for a letter
now and again - he is hoping to make a major
rebuild splitting his station into units for easier
construction and testing.



     73 for now
2 Briarwood Avenue, Bury St. Edmunds
Suffolk IP33 3QF, England