[Elecraft] K3 - LO CUT/HI CUT bugs

[Eric Scace K3NA]

A potentially related question / mis-understanding is the interaction 
between width, shift, lo and hi.  I'm not sure my concept is correct...

Firstly, people talk about 5 frequencies or ranges:
   Fc = audio center of the passband
   Fhi = -3 dB point above Fc
   Flo = -3 dB point below Fc
   Fshift
   Fwidth = passband width between -3 dB points
but I think there are also two other frequencies:
   Ffloor = lowest permitted frequency for a -3 dB point.  Let's assume 
this is 100 Hz.
   Fanchor = a reference frequency that is a conceptual point-of-return 
for the center.  In CW this is Fspot and in sideband this is a passband 
center when the passband is narrow enough that the lower edge is above 
Ffloor.

CW example:
   Start with a spot of 400 Hz and a filter setting of 500 Hz.
   Fc = 400 Hz
   Fanchor = Fspot = 400 Hz
   Flo = 150 Hz
   Fhi = 650 Hz
   Fshift = 0
   Fwidth = 500 Hz
Under my mental picture, Shift moves the passband away from the Fspot 
(Fanchor).  A shift of +100 Hz to move away from adjacent QRM changes 
the picture to:
   Fshift = +100.  Fwidth = 500 Hz
   Fanchor = Fspot = 400 Hz (unchanged)
   Fc = 500 Hz
   Flo = 250 Hz
   Fhi = 750 Hz
If one attempts to shift -100 Hz, Flo runs into the Ffloor value.  Since 
the operator presumably is trying to evade QRM on the high side, what 
should be done?  One approach is to allow the passband to narrow:
   Fshift = -100
   Fanchor = Fspot = 400 Hz
   Flo = 100 Hz (stopped at the floor)
   Fhi = 550 Hz
   Fc = 325 Hz
   Fwidth = 450 Hz.

SSB example:
   Start with 2.8 kHz width, with -3 dB points of 200 and 3000 Hz.
   Flo = 200 Hz
   Fhi = 3000 Hz
   Fc = 1600 Hz
   Fwidth = 2.8 kHz
   Fshift = 0
   Fanchor = 1600 kHz (reference point for the shift)
Now suppose one wants to operate in a contest environment.  Set the 
radio to always use a 1.8 kHz filter.  But, rather than listen to a 
too-low range of 200 - 2000 Hz, center the passband around 1600 Hz:
   Flo =700 Hz
   Fhi = 2500 Hz
   Fc = 1600 Hz
   Fwidth = 1800 Hz
   Fshift = 0
   Fanchor = 1600 Hz
Now increase the width control to 2200 Hz.  The signal still passes 
through the 1.8 kHz filter.  So the parameters are:
   Flo = 500 Hz
   Fhi = 2700 Hz
   Fc = 1600 Hz = Fanchor
   Fwidth = 2200 Hz
   Fshift = 0
This sounds quite acceptable in a contest: very intelligible but not 
troubled as much by adjacent stations as the wider crystal filter 
choices.  Manipulating the shift, width, lo and hi frequencies allow 
deviations from the above... but zeroing the shift always returns Fc to 
the anchor frequency.

   I don't think the radio quite works this way.  And maybe this isn't 
the correct model.  But hopefully it's a constructive addition to the 
conversation.

-- Eric K3NA



on 08 Nov 26 Wed 15:06 Lyle Johnson said the following:
> Hello Joe!
>
>> Up until recently, I had only used the SHIFT/WIDTH functions for 
>> setting the DSP filter parameters.  This morning I decided to try the 
>> LO CUT/HI CUT functions.  I really don't understand the behavior I'm 
>> seeing and I suspect there's something wrong...
>>
>> These oddities aren't apparent in normal day-to-day operation, but I 
>> think they're pretty clearly unintended.
>
> Thank you for your investigation and detailed observations of this. 
> We've taken note of it and put it "on the list" to be corrected.
>
> Hope you have a great Thanksgiving Holiday!
>
> 73,
>
> Lyle KK7P
>