> just yesterday and came across that W9CF page. I must admit, it's way
> over my head. I can't even tell what his findings are. Maybe you can
There was a long discussion about this on a newsgroup. If you read
Kevin's web page, you'll see he says his detailed analysis supports
my practical experience and measurements in my lab.
I'm less mathematical and more hands on, so maybe an
explanation from me will help.
> give us a condensed version. Now for this question of ...Is it a click
> or is it the filter...?? I can testify that on some cw signals, even
> ones that are 30db over S9, I can tune within 200hz of them, sit the
> 870's dsp for a 100hz bw and not hear a single click. However on most
> signals a click will indeed be heard. I'm sure some of it has to do
> with cw speed and faster code produces a wider bw etc.
There are many signals that are excessively wide Steve, and it is a
major problem. Basically it works like this....
1.) The slope of the envelope controls the frequency distribution at
any short period of time (the "bandwidth")
2.) The change in level at a given slope controls the amplitude
3.) The click extends the same bandwidth and amplitude
regardless of speed, as long as the rise and fall times do not
become large compared to the time of envelope amplitudes at the
"flat" parts of the waveform (zero and maximum). What you hear
are more clicks per second, not a different overall bandwidth!
Bandwidth is needed any time we change the amplitude of a
signal. How often you repeat that change (CW speed) does NOT
matter because neither the receiver or you ears "store" sidebands
over a period of many off-and-on cycles (dots and dashes and
spaces). The only exception to this is if you key so fast the
envelope never fully rises to maximum or falls to zero someplace in
the system.
To prevent excessive bandwidth, we must use enough time to
change the level and the shape has to be controlled. We can not
allow any "sharp directional changes" in the envelope at any point.
The ideal waveshape would be one that produces a signal width
during rise and fall that matches the width of a signal that can get
through the filters of the receiver. With a 300Hz filter, that is about
5mS with a sine shape. The envelope would look like it balloons
out gradually and gently rolls over at the top, with no sharp
transitions at any point. The rise time and fall time would be around
5 milliseconds.
That would be how one of the signals you hear without clicks looks.
Of course you can have a good shape and still have clicks if the
VCO is moving around, but if you have ANY sharp points in the
envelope or if the rise and fall is much faster than 5 ms you will
have excessive clicks.
73, Tom W8JI
W8JI@contesting.com
|