[Elecraft] K2 DSP summary

[Jim Brown]

On Fri, 19 Dec 2003 10:35:27 -0800, Lyle Johnson wrote:

>but it rarely makes
>sense to have the AFIL wider than XFIL.

I differ with you on this part of it -- If you have a lot of broadband
impulse noise at your QTH, the more filter slope you have engaged, the
more ringing you will hear, and the narrower it is, the closer that
ringing will be to the frequency of the CW you are trying to copy. Thus
you may find that you have less ringing if you use just enough crystal
filter to narrow the IF bandwidth and get additional noise reduction
from the DNR of the DSP. Both the IF and the audio filters will ring --
how much depends on their slope -- but if the audio filter is outside
the bandwidth where the crystal filters allow the IF to hit it hard,
you'll only hear the ringing of the crystal filter. Of course, if the
principal interference is another signal very close to the one you're
trying to copy, it makes sense to use both audio and IF filters in a
very narrow setting. 

>Finally, some operators prefer to leave everything as wide as possible until
>QRM/QRN make it hard to do so.  For CW use, selecting wide XFIL, AFIL set to
>LPASS (the default for AFIL 0), and then activating the denoiser, makes
>monitoring pretty easy, with signals seeming to just pop out of a very quiet
>background.

Yes,  in the 10 m contest last weekend, I tried staying a bit wider and
depending more on the de-noiser when the QRM wasn't too heavy, and was
pleased with the result both on CW and SSB.  

I also concur with others who suggest not cutting LF too high unless
you need to. Telco is the real expert on this, and their design
bandwidth has been 300-3,000 Hz for nearly a century.  

IMO, a more critical issue is the relative flatness of the overall
response within the passband. The transmit filters are very reasonably
flat (measured on receive), but the crystal filters (at least in my K2
s/n 3007) are quite bumpy (+/- 7 dB as measured on a good audio
spectrum analyzer) when aligned for 1.8, 2, and 2.2 kHz response, and
they degrade the intelligibility rather significantly as compared to my
Omni V. This is something I need to work on. :)  By contrast, both the
2.4 kHz and 1.8 kHz filters in my Omni 5 have a very nice "textbook"
flat response within the passband measured with the same analyzer. 

Jim Brown
Audio Systems Group, Inc.
Chicago
http://www.audiosystemsgroup.com