[Elecraft] RE: K2DSP puzzle

[Brendan Minish]

Hi Lyle,

         I think you misunderstand the problem slightly, the pop is caused 
by an abrupt truncation of the RX audio by the K2 mute circuit when going 
into transmit.

slowing down the QSK speed only helps a little since we no longer open the 
RX as often (& by extension have less RX -> TX transitions) but it does not 
solve the problem.

one solution would be to slow down the action of the mute circuit so that 
it does not make quite such an abrupt transition when muting the RX audio, 
one other option would be to delay the mute by a few msec more but this 
would probably require a K2 firmware change and in any case the timing on 
the K2 MCU may be too ' granular '  to allow for this.

in the normal non DSP arrangement the k2 ramps the RX audio down over a few 
cycles about 4 or 5 msec before the mute circuit comes on, the mute release 
is shaped C23 and R17 which is why reducing R17 (to speed up the QSK ) too 
much can cause a similar problem.
The additional delay due to the DSP processing time means that the mute 
circuit comes on before the ramped down RX Audio makes it out of the filter.

Unfortunately the delay in bypass mode is still long enough to cause problems

As for how the K2 sounds on QSK without the DSP installed. it's very 
smooth, the side tone sort of feels like another signal on the band. On my 
k2 you cannot actually hear between the dots much above 25 WPM but you can 
hear between the letters fine.
Smooth QSK is rare and it is one of the (many!) really nice things about the K2



At 15:34 17/07/2003, you wrote:
>Hello Brendan!
>
> > Any ideas for a mod to fix this?
> > Slowing down the onset of the mute circuit slightly should do it but I am
> > not sure how best to go about this.
>
>All the muting is handled externally to the KDSP2.  It has its own muting,
>but that is currently only invoked during during power up, to elimninate a
>loud POP that would otherwise leave your ears ringing for a while!
>
>Signals do take a while longer to propagate through the KDSP2. The sidetone
>is injected after the KDSP2.  Adding delay to the unmute function would
>probably slow down the K2's top speed for effective QSK, but that may be a
>trade-off you're willing to accept to get around the problem you cite.
>
>Since I don't personally operate QSK CW, I am not the right person to work
>out a mod to get around the problem.  WHat sounds right to me might not
>sound right to you, and I have no real frame of reference as to what it
>*should* sound like.
>
>In any event, to slow down the unmute process, increasing the value of C23
>on the control board would probably be the most direct route.
>
>73,
>
>Lyle KK7P