[Elecraft] [Elecraft_K3] curious behavior of the K3/P3, etc.

[Scott]

Thanks Rich, that is a very clear and thorough explanation!  I cannot relate all to my level of understanding, but your explanation is complete so I can study it.  In really basic language, what I understand is that adjusting the width will only affect the upper limit until I get it narrow enough so that both are in an acceptable range.

However, I am still left with some doubts: particularly, why does the upper limit change when I adjust the lo-cut, and the lower limit change when I adjust the hi-cut?  Thus, if I want to narrow my bandwidth to avoid a station above the desired one, I must adjust the lo-cut (I can see this happen on the P3).  This occurs on all bands on CW.

You are good at giving explanations—I hope I am not trying your patience!

Thanks, 
Scott  AA0AA


Date: Tue, 11 Feb 2014 14:06:56 -0500
From: Richard Ferch <ve3iay at storm.ca>
To: "elecraft at mailman.qth.net" <elecraft at mailman.qth.net>

OK, if you're talking about CW, the K3's normal CW is on lower sideband.
In CW the filter behaviour gets tangled up with the sidetone pitch
setting, which is also the default FC (centre frequency) for the filters
in CW mode. Suppose the sidetone pitch/FC is 500 Hz and the BW
(bandwidth) is 2000 Hz. Given those numbers, the calculated hi-cut and
lo-cut frequencies would be 1500 Hz (FC + BW/2) and -500 Hz (FC - BW/2).
That's a minus sign on the lo-cut frequency, meaning it's on the other
sideband. The firmware should not allow that; it needs to place the
lo-cut frequency on the proper sideband far enough above zero to give
adequate carrier and opposite sideband suppression. IIRC, the low-pass
cutoff is adjustable in the latest firmware - I believe it can be set to
100 Hz, 200 Hz or 300 Hz.
So, with the low-pass cutoff set at 200 Hz, FC at 500 Hz and BW at 2000
Hz, the actual filter bandpass will be from 200 Hz to 1500 Hz. Adjusting
the Width control would only affect the upper limit, as the lower limit
is fixed by the low-pass cutoff. If the Width setting is sufficiently
narrow for the calculated lo-cut frequency to be at or above the
low-pass cutoff (100, 200 or 300 Hz), the effect of the Width control
becomes symmetric about the centre frequency, as you might have
originally expected.
Because CW is on the lower sideband, adjusting the Width control only
affects the lower RF (=higher audio) edge of the bandpass - the upper
edge is fixed by the low-pass cutoff.
Does this help make sense of your observations?
73,
Rich VE3KI

>Scott Monks <cq_dx_de_aa0aa at yahoo.com> wrote:
>
>Now before you guys think I too am complaining--no, I just have a curious mind!
>
>Two questions that I am sure you more experienced can help me with:
>
>1. When I begin narrowing the DSP down, on filter 1 (2.7, std, offset -0.84) only the lower freq side moves.? The DSP switches to the 2.1 filter, reacting the same, but when it narrows to about 1.10 (within filter 2; 2.1, offset 0.0) it changes so that both low and high move equally.? Even narrowing continues through the 700 and 400 Hz filters.? The reverse occurs in the same way.?? I have checked the offsets with the K3 program, and they are set as required.? On the K-3 display the "bar-thing" always seems to move equally on both sides.? Is this normal?
>
>2. When seen on the P-3, on hi-cut-lo, if I move the lo side, the higher freq side moves, and if I move the hi side, the lower freq side moves.? On the K-3 display the expected side moves.? Again, is this normal?
>
>I have the last production software installed.? This was first noticed when I got the P3 and a "big screen", when in a contest I wanted to shift the receiver up to avoid a disturbing signal and found I had to move the hi side to move the lo side!? Now I am accustomed and just watch my screen to make sure I am moving outside of the interfering peak.
>
>These only "bother" me because I know there must be a reason but I don't why it happens!? If curiosity is going to kill the "cat", don't tell me and I will just continue to be happy with my partial K-line!
>
>Thanks for your patience,
>Scott? AA0AA


Scott  AA0AA; XE1/AA0AA
cq_dx_de_aa0aa at yahoo.com