[Elecraft] RE: 2. Re: K2 SSB TX Alignment Tools Proposal

[Ron D'Eau Claire]

Web, KR4WM wrote:

I may be oversimplifying the problem, but here's my experience in aligning
my K2. I happened to remember reading some very old QST articles about a
device used to get stations precisely on the net control's frequency during
net operations. It was a very simple 1KHz tone generator. The NCS station
was supposed to send out a 1KHz tone, and net stations would listen on
frequency while at the same time, playing a 1KHz tone at their location.
This would allow stations to zero-beat the NCS station, thereby being
precisely on the NCS station's frequency. 

-----------------------------------------------------------------

That's great for matching frequencies so that no retuning is needed for
perfect audio when listening to a round table (or net), Web, but it still
leaves the issue of setting the SSB filter bandpass in the best place. 

Assuming the stations tune each other in correctly on the air, adjusting the
BFO frequency has the effect of shifting where the filter bandpass falls
across the audio spectrum.

As the BFO frequency is changed, it has the effect of moving the RF bandpass
up or down so that more low frequencies are cut off in favor of passing more
high frequencies or, cutting off more high frequencies in favor of passing
more low frequencies.

It's a lot like playing with an audio equalizer or set of tone controls in
which they are locked together so that increasing the bass automatically
cuts the treble and vice versa. 

The only way to change both the bass and treble independently is to change
the filter passband. That happens when we choose the bandwidth of the filter
itself. So, when we choose the filter bandwidth we lock the bass and treble
controls together then we move them back and forth for best sounding audio
by adjusting the BFO frequency. 

We tend to choose SSB filter bandwidths that restrict the audio range
transmitted as much as possible while still maintaining good
intelligibility. That reduces the bandwidth needed for the SSB signal while
increasing the "punch" of a given RF power by concentrating that power in
the filter's bandwidth. But, as the bandwidth of the filter is reduced, how
it is positioned to pass the audio spectrum, balancing the "bass" and
"treble" becomes more critical, and more dependent upon individual voices.
The communications "standard" of about 2.8 kHz is a compromise between those
two issues: it's a reasonably narrow bandwidth but wide enough for good
intelligibility if the bandpass allows just the right compromise between
high and low audio frequencies through. 

That's where Spectrogram comes in. It lets us "see" exactly how the filter
passband sits in relation to the audio spectrum: what it's passing and what
it's cutting off. That's great for setting that passband to a specific
place, but it does not indicate what place is best for a particular voice
and microphone. Don gave some good "rules of thumb" that will put the filter
passband in a position that produces good sounding SSB audio under most
conditions, but it may not be optimum for every voice and microphone. That's
why the SSB module book says to do final checks listening to the signal as
transmitted if one wants to do any final "tweaking". 

Just how critical the actual position of the bandpass becomes depends upon
the filter bandwidth. The original K2 used a narrow SSB bandpass of about
2.1 kHz. That's well below the "standard" used by most rigs but still wide
enough to produce good audio. I believe the idea was to give the signal the
most "punch" by putting the signal into a narrower slice of the RF spectrum
while still having enough bandwidth to produce clean, intelligible audio.
Wider bandwidths, such as the 2.7 KHz used by most SSB rigs, are far less
critical to adjust, passing more of both lows and highs for the vast
majority of voices. You can choose either the original narrow or wider 2.7
KHz for your K2 by making minor component changes on the SSB module. Even
wider bandwidths (ESSB) are even less critical, but at the cost of winder
bandwidths and a lower S/N ratio at the receiving end for a given power,
since the RF power is being spread out over more RF spectrum. 

Ron AC7AC