[Elecraft] K3 Filter question - SSB

[Bill W4ZV]

Bruce this is from a similar question on the Yahoo Groups K3 list:

-- In Elecraft_K3 at yahoogroups.com, Hector Padron <ad4c2006 at ...> wrote:
>
> The 1.8 roofer today with so much band noise and fool ops close to your
> freq disrespecting the spectrum, its a must.This 8 poles filter together
> with the DSP makes brick rx easy to work dx or contesting.The trick to
> recover the lost audio quality when using it is to move counterclockwise
> the shift control down to 1.2 and inteligibility is back.

This is very misleading and I must strongly disagree. I preface this by
saying
I made ~2500 QSOs on 10m single band in the CQ WW (high-claimed USA SOSB/10
score) and was never able to effectively use the 1.8k filter even though I
tried
in vain several times.

Here's what a 1.8k will and will not do:

1. It *WILL* keep your AGC from pumping if there's another strong station
within the 1.8 kHz passband. However, do you really think you could copy a
weak
signal while a S9+30 interfering signal is inside your 1.8k bandwidth? I
don't
think so. With any typical SSB filter bandwidth, AGC pumping is not a
practical
issue (it IS a big deal for CW however).

2. It *WILL *NOT* keep splatter from adjacent signals out of your passband.
If
an interfering signal is 5 kHz wide and partially falls within your
passband, NO
filter can remove it. Splatter is a real signal which NO filter (XTAL or
DSP)
can remove.

3. It *WILL* require very careful tuning for intelligibility. With callers
that are off frequency by only 100 Hz, you'll miss off-frequency callers the
first time which will slow your run rate. I had one caller even 500 Hz below
my
run frequency and I'm certain I would never have heard him if I was using
the
1.8k.

The most effective use of a 1.8k is probably for copying an extremely weak
DX
signal in white noise (not strong splatter or QRM). By tuning VERY
carefully,
you may slightly reduce the noise floor by reducing the noise bandwidth
(potential reduction of 1.8k BW versus 2.1k BW is 10*[log (1.8/2.1)] = -0.67
dB). I really doubt many of us can detect a 0.67 dB improvement in
signal/noise.

The 1.8k is mainly a DXing tool...not a contesting tool. It cannot magically
overcome the basic laws of physics.

73, Bill W4ZV

--
View this message in context: http://elecraft.365791.n2.nabble.com/K3-Filter-question-SSB-tp6981339p6981505.html
Sent from the Elecraft mailing list archive at Nabble.com.