[Elecraft] K3 for CW

[Guy Olinger K2AV]

This applies to CW and particularly to contesting.

On Thu, Jul 5, 2012 at 11:51 AM, Vic K2VCO <k2vco.vic at gmail.com> wrote:

> I say 'much' because the effect of concatenating the DSP and the sharp
> filter will change
> the shape of the bandpass a little.
>

If the DSP width is set to the actual bandwidth of an 8 pole roofer and the
offset of the roofer is carefully adjusted to bring its *** 30 dB down ***
(not -6 dB) skirts to coincide with the 30 dB down skirts of the DSP, the
effect is quite more than a "little".

For the 400 hz 8 pole, this width is 450 Hz, and for the 250 Hz 8 pole this
width is 350, and to get the benefit you must install the "400" Hz as 450,
and the "250" as 350 width with the K3 utility.  It takes some time to get
the combination exactly "skirt-aligned", with the filter offset setting,
but it is worth it.

Until I discovered this procedure, my K3 had the best front end and some
really neat features over my Yaesu FT1000MP.  But the K3 did ***NOT*** have
the best selectivity.  Nor was it particularly close.  My MP had cascaded 8
pole INRAD filters in the 8 MHz and 455 kHz IF's.  I had that arrangement
for 1.8, 400 Hz and 250 Hz.  When someone was high or low but enough into
my bandpass to be a problem, a very small tune-tweak with RIT, with the
offending signal on really, really sharp skirts would put another 10 dB
 rejection on him without my moving my run frequency.

NOW, with the 8 pole offsets set for skirt alignment with DSP at -30 dB,
when I use 450 or 350 width I get the MP dual INRAD style skirt rejection.
 Note it's not so sharp away from the 450 or 350 settings.   That is
because while the DSP uses digital magic to retain it's shape factor with ,
the 8 pole at 400 or 300 is starting the level off process in the curve,
and it is not adding as much rejection to the DSP's rejection.

For my purposes, the rejection at 400 is acceptable, but having the "250"
roofer kick in at 350 really makes 450 to 300 width quite good, even if it
still remains a teensy bit broader than the cascaded INRADs  in the MP.

Now here's the other benefit:  The very sharp drop of the skirt aligned 8
pole/DSP at 450 and 350 converts an up or down frequency signal with key
clicks to LOWER amplitude, but VERY SHARP spikes, viewed on a scope.  Some
people have mistakenly thought this was making clicks worse, and perhaps
from an irritation point of view that was true. That was because the really
spikey part of the sound which irritates most was being held down by the
extra energy let in by broader selectivity.

Why is sharp spikes a benefit?  The K3 is WONDERFUL at handling spikes.
 More so after the spectacular improvements from firmware 4.51, even the
lowest NB setting of IF OFF and DSP t1-7 will simply eliminate all but
louder key clicks, and since 4.51, t2-7 has handled anything.  The AGC
already ignores short spikes, and so unless the key clicks in adjacent
channel are loud enough to get into the hardware AGC (15 or 20 over key
clicks?) , clicks become a non-issue.  In some cases, the clicks go from S7
or S8 to S nothing.  Suppression of five or six S units is more common.  I
sometimes get the clicks supressed only to hear the station's phase noise
or poorly suppressed carrier.

If you have ever held down a run frequency for 4 plus hours, it almost
inevitable to get a clicky signal up or down.  This click-elimination
feature of a K3 can be the difference between staying where all your spots
are, or having to move in self defense and really messing up your rate
until you can find another frequency.  There may not be any open
frequencies, particularly if the band you are on is the only one really
open (or a 160 contest).  Losing a run frequency can cost you 50 or 100
contacts.  Later in a contest, S&P is not an option.

I now just leave NB on all the time at t1-7, IF OFF.  My ear has gotten
used to the very slight truncation of CW bauds at t1-7.

The pair set at 450 and 350 has another use.  While I always settle into a
run frequency where I hear nothing in the 450 bandwidth, It frequently
these days gets squeezed as the K3 phenomenon is transforming traditional
1/2 kHz spacing to 1/3 kHz spacing. More and more frequently, after
starting at 450, I find myself running at 350 bandwidth in self-defense.
 Still with the great key click suppression, and with a 300 setting that
works nearly as good.

I have operated for hours with the (mandatory?) 30 over 9 Italian up 400 Hz
running a pair of 3CX5000's and bad key clicks, and kept it at high rates.
 Something my MP could never do, notwithstanding cascaded INRADs and
super-sharp selectivity.  The Italian still gets a dun after the contest,
but has never responded.  I think he looks for me because I DON'T have any
key clicks and don't respond with squeezing, etc.  With my K3, I just don't
care.  Some low level of clicks is irrelevant, and I can monitor his rate.
 He does turn in good scores.

So I have 5 pole 2.7 for conversational SSB, and for contests, 1.8 8 pole
for SSB (mandatory, but that's a whole other story), and "400" and "250" 8
poles, aligned as above.

This is not everyone's ticket, but  operating in a contest with the roofer
significantly larger than the operating bandwidth is going to cost you, and
there are benefits to matching DSP and roofer precisely.

73, Guy