[Elecraft] Change to Elecraft Keying Mod

[John, KI6WX]

Guy;
The plots shown in the note are the spectrum of the keying envelope.  They
do not include the effects of the local oscillator phase noise which is a
significant contributor to receiver noise when you are talking about offsets
from a carrier of a few hundred Hertz.  Measurements I did make of the
spectrum of a keyed waveform that do include the K2's phase noise show less
than a 1 dB difference at a 250 Hz offset between the Elecraft version of
the mod and my changing R21 to a 180K resistor.

Your method of computing the comparison numbers is mathematically incorrect.
You can not take a peak value at one point in a communications channel and
do the comparison at that point.  You need to integrate both measurements
across the entire bandwidth of the communications channel to compare the
relative noise generated by each keying waveform.  This means for your
comparison that you need to integrate both spectrum from 250 to 750 Hz.
After you do this, I doubt whether you will find much difference between the
intensity of the two waveforms in the adjacent channel.

Finally, you state in your last paragraph that there have been other
measurements made of the keying bandwidth for the Keying Modification, and
you believe that my blue curve is somewhat high.  Other than subjective
listening of key click intensity, I haven't seen any measurements of the
spectral bandwidth of a keyed K2 before and after the mod.  It is a
difficult measurement to do correctly for a number of reasons including the
wide dynamic range involved in the transmit signal and its sidebands.  I'd
be very interested to see any other graphs of the spectral bandwidth of the
keyed waveform in a K2 before and after the mod.

So I disagree with your conclusion that changing R21 to 180K is WORSE for
the adjacent channel.

Secondly, the Elecraft Keying Modification was intended to fix a problem
that the K2 with a high power amplifier would generate key clicks over a
several kHz of bandwidth; it was not intended to fix a problem with ADJACENT
channel interference.  This version should work even better for the problem
that Wayne was intending to fix.

So far I have two reports from people who have implemented the R21 change,
and their impression is that it does work better.  I'd like to get a few
more reports, particularly on bands other than 20 meters.

-John
 KI6WX

----- Original Message ----- 
From: "Guy Olinger, K2AV" <olinger@bellsouth.net>



> I have looked at the waveforms you show on the website below.
>
> I believe the best way to interpret the results would be to presume
> 500 Hz channels, and look at the highest power presented in each
> channel to get an idea of the click power in each. Seeing how close in
> clicks disappear is a slyly deceptive indicator for best improvement.
>
> The adjacent channel would then be 250 - 750 Hz, with the co-adjacent
> channel at 750-1250. For a sloping curve across the bandpass, the
> total power in a channel will be controlled by the high end of the
> curve.
>
>
> That gives the following results:
>
> mod                250-750       750-1250
>
> orig (red)          -32            -54
>
> elec (blue)         -46            -62
>
> WX   (green)        -40            -73
>
>
> While you have improved the results at some points by ten db, these
> are in the LESS AFFECTED co-adjacent channel.
>
> In the most affected channel, the adjacent channel, results are WORSE
> by 6 db.
>
> Since the worst click problem for any rig occurs in the ADJACENT
> channel, that is where the FIRST effort must go, as Wayne has
> targeted.
>
> What would be more useful is if you could improve anywhere WITHOUT
> trading away anything in the 250-750 range. The next guy up or down is
> the one getting the maximum discomfort from our bandwidth.
>
> I would also point out that I think some of the modified K2's rigs I
> have heard are better than your blue curve. Wonder if there is an
> alignment issue, like for the little chirp.
>
> 73, Guy.