Topband: K3 & some interesting noise lessons in the ARRL 160.

[Guy Olinger K2AV]

This may be specific to a K3 and no joy for other rigs. But there might be
parallels elsewhere.  Definitely YMMV.

Usually in a 160 contest I try and get a run frequency down around 1815-20.
For some reason in my locality that range is, and for a long time has been,
a general garbage minimum in the noise floor around here, sometimes 2-3 dB
better, so there is no incentive to CQ elsewhere. I have not had a problem
establishing down around 1815-20, even when running 100 watts.

But this year, off and on since late summer, and of course without any
rhyme or reason or published schedule, I have had a really bad intermittent
power line style buzz that was S7 to ten over S9, depending on only heaven
knows what. I have not had any luck localizing it, largely because of it's
variable and erratic nature. A lot of the time it has irregularly separated
fast bursts, almost like it's trying to send Morse with its buzz.

On the first night of the ARRL 160 contest, buzz was entirely absent. And I
actually had the ARRL 160 weekend clear of family conflict for the first
time in recent memory. Oh Joy! By the end of the first night I already had
a personal best for the contest. The second night the noise struck half way
through with a solid buzz. Ten over S9 in my usual hangout 1815-20-ish. A
little less noise up higher but still covering all but the louder signals.
No real help using my newly repaired and pattern-verified NE RX antenna.
Loud there, too. Earlier work had ruled out a source in my or neighbor's
houses. No quickie fixies.

I was unable to hear anything except the louder signals, which I had pretty
well worked out the previous day.  Forget operating. So I decided to
experiment with the noise mitigation methods/settings on the K3.

Usually in these buzz circumstances, you can't find a persistent weak
signal on 160 far enough into the noise to let you experiment with noise
mitigation settings based on signal to noise. You wind up using the reduced
noise level as the only clue for settings.

Reduced noise level method is fine if you are working strong signals and
you just don't want to hear the noise. But to pull out weak signals what
you really need is to restore signal to noise separation all the way down
to the weak signals. It turns out best signal-to-noise and best
level-of-noise do not always generate the same weak signal readability.
Some methods/settings reduce noise well but also trash the weak signals in
the process.

But being the contest, 160 was *loaded* with weak signals to test with. So
I spent a few hours experimenting with K3 settings on weak signals,
optimizing for S/N. Came up with NB only (no NR) DSP T1-7 + IF NAR4 using
"250" 8 pole filter, which clearly gave the best separation between weak
signal and noise, without the usual weak signal obliteration from
traditional noise blanking in a contest.

Usually the buzz gating the NB will add a 180 Hz raspy modulation (center
carrier and +180, -180), an irritating buzzy noise, and can mush the wanted
signal.  Narrowing the CW width to 250 or 200 Hz (+/- 100 Hz) cleans off
both the modulation (3 x 60 Hz), and signal-covering "hashy fuzz" caused by
the irregular shape of the buzz waveform.

After the determination of those settings, I scanned the band with those
settings in NB. I discovered a narrow "null" in the noise or sweet spot
between 1831 and 1833, sounding almost normal, with clear rendition of weak
signals, which was up to 2 s-units better than the NB improvement in the
least effective spots, which included my normal hangout of 1815-20.  Don't
ask me how that works, because I haven't a clue.

1831.5 was unoccupied, as was 1832.5 in a later stretch, where operating
with the blanking on, it was as if the noise was not there at all. (See an
earlier post about my unfortunate adventure with the DX window rule.)

That narrow sweet spot in the blanked noise was still there Tuesday, at
1830-1832. With the noise back solid, and using the NB settings above,
Tuesday night I was hearing LZ2DF on 1832 clearly, at what I would call 559
or 549. He was not hearing me running 1.5 kW, so this was clearly a normal
state of affairs, controlling noise was on his end, not mine, even though
the buzz was full on.

Wednesday the buzz was there with separated bursts, but not quite so loud,
and the aforementioned NB settings killed it at least semi-decently across
the band. Go figure. I was able to clearly pull out way-down birdies and
such, simply not there without the NB.

Thursday the noise-blanking sweet spot was at 1.838, and a little broader
than during the contest. The RX antenna clearly hears a weak birdie at
1.838 much stronger than on the TX. Yes, Virginia, the RX antenna has been
working correctly the whole time.

Today the buzz is gone at my noon-time opportunity for driving around and
trying again to locate it with my K2.

Tonight the buzz is back, mostly steady. The sweet spot is at 1.828 tonight.

We will be renewing the search for the noise when it decides to come back
steady in the daytime.

But regardless, now I have noise blanker settings for the Stew and CQ 160
which actually do improve signal to noise for weak signals in power line
noise -- just in case the noise is still with us.  And I know now to look
for the wandering-around sweet spots where the K3 does some real magic,
before I go for a run frequency.

73, Guy