[Heathkit] Heath audio, rough around the edges.

[Robert & Linda McGraw K4TAX]

I'll jump in and comment here.

The carrier in the SSB generation mode is basically a modulated double 
sideband product and it should reside on either the upper or lower slope of 
the filter.  Not in the center of the filter passband.  If the carrier is on 
the upper slope then the result is the upper sideband is removed and the 
lower sideband is passed.  Just opposite with the carrier on the lower slope 
of the filter where then the upper sideband is passed.  {Should the carrier 
be positioned in the center of the passband the result is a very narrow DSB 
signal, not a SSB signal.}

The position of the carrier on the slope will determine to a large extent 
the low frequency  audio response more so than the upper end.  A filter that 
had a passband defined as 2.8 KHz wide will have the points that are 
attenuated 6 dB located 2.8 KHz apart.  If the carrier is placed at the 6 dB 
point then the lowest audio frequency will be attenuated by 6 db and the 
highest frequency passed will be ~2.8 KHz, attenuated by 6 dB.  Move the 
carrier to the 20 dB point, usually only about 200 Hz away depending on the 
slope of the filter, and the lowest frequency passed will be Af - 200 Hz. 
Thus the net audio passband is 200 Hz to 3000 Hz.  Moving the carrier up the 
slope to say the 12 dB point will result in a audio passband of 100 Hz to 
2900 Hz.  Here one can see that the LF end is affected more so than the 
upper end of the audio spectrum.  Look at it this way, the 2.8 KHz filter 
will simply take a slice that is 2800 Hz wide out of the overall audio 
spectrum.

One must note that additional attenuation provided by the filter, i.e. 20 
dB, is added to the overall carrier suppression.  Hence one can attain 40 dB 
carrier suppression at the balanced modulator and then another 20 dB by the 
loss at carrier frequency by the filter.  Placing the carrier too high on 
the slope will reduce the available carrier attenuation.  But at the same 
time it will improve the low frequency response while slight attenuation of 
the upper end of the spectrum.

One other factor which affects SSB audio quality is the ripple across the 
top of the filter.  No filter is perfectly flat.  A filter that has a 3 dB 
ripple across its passband will have the same effect on the audio response 
at the frequency.  A filter that has a 6 dB ripple will not sound as good as 
one having a 3dB ripple.  Measuring this typically requires a tracking 
generator and a spectrum analyzer.  More importantly, incorrect termination 
of the filter input and output will cause the ripple factor and shape factor 
to be distorted.  Measuring the true response of a filter is somewhat of a 
challenging effort.

An additional factor in audio quality lies in the accuracy of the diodes in 
the balanced modulator circuit.  Here the diode conduction voltage, or turn 
on point, will have great effect on how clean or more likely "how distorted" 
the transmitted audio is presented.  Diodes of extremely close match in 
forward voltage will net a modulated distortion of less than 0.01%THD while 
typical balanced modulator diodes will net only about 1 or 2 percent THD. 
This is why some transmitter "sound better" than others.  Of course one 
would say then why not hand pick diodes for the balanced modulator.  I have 
and in doing so I carefully went through some 1000 pieces to find just 4 
ideal diodes.  Once done, these were put in the circuit and the modulated 
signal measured.  Even at 0.1% THD most ham receivers can't recover the 
difference between 0.1% THD and 1.0% THD.  Was it worth it, nope, not in my 
book.  But it measured very nice.

73
Bob, K4TAX


----- Original Message ----- 
From: "Robert P Ward" <wardbnj at comcast.net>
To: <heathkit at mailman.qth.net>
Sent: Monday, June 19, 2006 7:12 AM
Subject: [Heathkit] Heath audio, rough around the edges.


> Hi Jim
>
> I suppose your method could work here, that should center the carrier
> within the passband of the filter. A bit of experimentation might be 
> needed
> to find that "sweet spot"! Might be interesting to see how you do by 
> moving
> the BFO carrier oscillator the same amount as I have...
>
> As suggested by Grant, looking at the speaker terminals with an AC
> voltmeter might color the response a bit. The comment about -20 dB points
> originated from a 1968 SSB handbook article.
>
> Bob   NQ3N
>
>
> _______________________________________________
> List Administrator: Duane Fischer, W8DBF
> ** For Assistance: dfischer at usol.com **
> $$ See the vintage area on the HCI web site - http://www.w9wze.org $$
> Heathkit mailing list
> Heathkit at mailman.qth.net
> http://mailman.qth.net/mailman/listinfo/heathkit
>