"Frank Gilliland" wrote in message
...
On Sat, 26 Mar 2005 11:40:53 -0500, "Leland C. Scott"
wrote in :

snip
Come to think about it, how about just a 567 or 4046? Either chip
would probably do it as long as the radio doesn't drift too much. And
if I'm not mistaken, both have a VFO on the chip. So a CW 'adapter'
could be a single-chip project -- no DSP required.

Those two chips are just simple PLL building blocks. There isn't a way to
do
any tone filtering as such unless you make the VFO frequency range vary
narrow which would have the same effect.


That's the idea.


The 4046 has a range and frequency
offset capability using just two resistors and a capacitor. You would
want
to use the phase-frequency detector, not the simple XOR phase detector
which
could lock on harmonics of the audio tone. That would be the chip to use.
The lock detect circuit could then be used as the visual signal device
for
CW. Something this simple you could breadboard in an evening to see how
it
works.


Actually, it only took about 15 minutes, but it doesn't work. Capture
and release times are too erratic within the bandwidth, and any other
CW signals within about half an octave make the thing go spastic.

Hummmm...., what did you use for the bandwidth? I would have picked
something like 800Hz at the VCO center frequency and a VCO range of 700 to
900 Hz. One thing the data books don't tell you is slapping a capacitor
across each of the two resistors for frequency span and offset helps a great
deal in settling down the loop. Seems like the VCO is some brands of chips
have stability problems. I've had to use that trick on some of my company's
high power inverter controls to get the switching frequency to lock on to
the load's tuned frequency when using the 4046 as a phase locked loop. Try
using something like a 0.001uf or smaller capacitor.

It
-might- work if the loop filter Q could swing with the VCO freq, but
that ain't gonna happen with just a couple chips.

They use those chips for some data transmission applications and they work
fine at rates higher that you would expect somebody to be sending CW. I
wouldn't give up on using the simple circuit just yet.


Setting a narrow capture bandwidth, and using the clarifier (or BFO)
for tuning, works pretty well. I also tried a state-variable with a
schmitt trigger. Discrimination was better; but high Q, cumulative
drift and fading made it difficult to get a clear copy.

What was drifting, the filter, the frequency of the tone being transmitted,
or the receiver's frequency causing the tone's frequency to change? Before
giving up on anything it would be prudent to find out what the problem
happens to be before saying it doesn't work. The problem may be easy to fix,
or it could be the equipment you used for the test isn't representative of
what could be used now days.


I should add that it's been quite interesting trying to read Morse
with an LED..... after a few minutes it's almost hypnotic. I tried
different colors, and while a big green LED was the best it was still
irritating after about 10 minutes.

There are hearing impaired Hams doing something like that for years. I guess
they just get use to it.


--
Leland C. Scott
KC8LDO

Linux - The alternative OS to Micro$oft Windows