Posted by
Ron D'Eau Claire-2 on
May 28, 2005; 10:10pm
URL: http://elecraft.85.s1.nabble.com/K2-XTALS-tp378841p378845.html
Geoff, GM4ESD, wrote:
Of course the frequency response of the microphone and Tx audio circuitry
needs to be included in this consideration. But the group delay
characteristics of the filter should also be thought about. A typical Cohn
ladder filter, for example, has wicked gyrations at the corner frequencies
which requires that the carrier frequency be moved up the filter's skirt to
reduce distortion at those audio frequencies that contribute to
intelligibility. This of course increases the level of unwanted sideband
unless at the same time, the Tx audio rolls off very sharply at the lower
audio frequencies. To avoid this, Gaussian, Gaussian to 6db or 12db or the
equiripple families of ladder filter which approximate constant delay can be
used, and are the most useful from the distortion viewpoint. Although not
relevant to this discussion, a narrowband Cohn ladder as might be used in a
CW receiver will ring.
Without seeing the delay and frequency response plots of the
modified filters, any further comment would be guessing on my part.
-----------------------------
Those are good points that I haven't addressed since I'm not a filter
specialist. As an operator I've heard first-hand some of the awful stuff a
filter can produce in response to signals on the edge of its bandpass.
The K2 filters went through a long cycle of testing and optimization at the
hands of John Grebenkemper, KI6WX, with consultations with Wayne at Elecraft
and possibly others. I presume these issues were considered at that time.
Obviously, a complete redesign that would involve more crystals or a
different filter topology was out of the question since such changes
wouldn't fit on the SSB module. So I won't speculate on what compromises he
may have made in his choices. He did publish a detailed "how-to" description
of his recommendations in a paper that can be downloaded:
http://tinyurl.com/8emeu. This is a document written for the typical
Elecraft builder and does not go into the analytical work behind his
recommendations.
If you're curious, I'd suggest visiting John's web site
(
http://tinyurl.com/8cs8w) and perhaps dropping him an e-mail. I've always
found him to be a very knowledgeable and patient in describing his work.
Ron AC7AC
_______________________________________________
Elecraft mailing list
Post to:
[hidden email]
You must be a subscriber to post to the list.
Subscriber Info (Addr. Change, sub, unsub etc.):
http://mailman.qth.net/mailman/listinfo/elecraft
Help:
http://mailman.qth.net/subscribers.htmElecraft web page:
http://www.elecraft.com