[QRO] ssb refresher course needed

[Dave Haupt]

rb,

In the spirit of ham brotherhood and fraternity, would
you be kind enough to tell us your callsign?  Then
when we have a QSO on the air, we'll both know that
we've encountered each other online!

There are a few things you touched on.

First, the 3kHz is a "nominal" standard.  There is no
such restriction in the FCC rules.  Various
gentlemen's agreements about how the HF bands are used
assume that SSB signals nominally occupy 3kHz or less.

If everything is working correctly, the nominal
bandwidth of your signal will be determined by the IF
filtering system in your exciter.  Stock SSB filter
bandwidths range from 2.1kHz (rare) to 2.7kHz.  Most
commonly, radios come equipped with a 2.3 to 2.4kHz
wide filter.

But there is a risk in consider that to be the rig's
bandwidth.  Filters are never "brick wall".  That is,
outside the designated bandwidth, they are still
passing energy.  The "banner bandwidth" of a filter in
ham radio usage is usually the distance between the
two points where the response is 6dB down from peak. 
And, there is quite a tolerance on that.  In my
measurement of rigs with stock crystal filters, I have
seen bandwidths vary from 2.4 to 2.8 kHz when the
filter is stamped "2.3kHz".

So, there is always energy outside the expected range,
simply because filters have non-vertical skirt
selectivity.

The linearity of the so-called linear stages in radios
is really not all that good.  It is common to see an
HF exciter whose intermodulation distortion figures
are 30dB down.  This means that if you are
transmitting with 1kW in the intended part of the
signal, there are distortion products being delivered
that have one watt's worth of RF power.  As our QRP
brethren can attest, that power into a good antenna is
capable of great world-wide communications - or QRM. 
Fortunately for us, the majority of the IMD is 3rd and
5th order.  The distance from center frequency of
these spurious responses is approximately one to five
times the fundamental frequency of the human voice
being used to modulate.  In practice, a rig being
operated properly, with a nominally 2.3kHz filter, can
be seen on a spectrum analyzer to be delivering energy
outside of a 3kHz window that is more like 50-60dB
down.  Again, this amount of energy into a good
antenna is definitely going to be heard as QRM in some
areas - but statistically far fewer areas than if the
energy were only 30dB down.

One normally valid assumption is that most of the
intermodulation distortion in a rig is produced by the
final amplifier.  In a well-designed rig, this will be
the case.  And, in the normal case, the distortion
level will be reduced if you ask for less power from
the radio.

However, not all rigs are designed so well.  I have
two Icom IC-735s, and the spectrum analyzer
measurements are a bit shocking.  On the spectrum
analyzer, with the stock SSB filter, the nominally 6dB
down points (visible bandwidth when modulated by my
voice) seems to be about 5kHz - and this is not a
radio intended for hi-fi audio.  Yet, the SSB filter
itself does measure nominally correct (about
2.4-2.5kHz at the 6dB points).  Clearly, something in
this radio is generating intermod products rather
badly.  Still within FCC specs, though.  If I reduce
power, there is no difference seen at all in the
distortion - in other words, it is not coming from the
final amplifier, but an earlier stage.  I have an
early '735 and a late '735 and they behave the same. 
A fellow engineer reports the same results with his.

On the other hand, my Kenwood TS-850SAT, on the
spectrum analyzer, is seen to have a 6dB bandwidth
that correlates reasonably well with the SSB filter's
performance, as long as I do not overdrive the rig.

When you add a Power Amp, you will ALWAYS widen the
signal.  However, if things are adjusted correctly,
the bandwidth widening will be imperceptible. 
Specifically, if the distortion performance of the
amplifier is at least 10dB better than the exciter,
the bandwidth widening will, in fact, be virtually
impossible to measure on a spectrum analyer.

I have worked with a few contesters in the local area,
who were heard as having very wide signals on the
band.  It turns out they were all making the same
mistake.

Amplifiers have ALC signals, which are routed back to
the HF exciter.  If the amplifier is overdriven, the
ALC signal goes back to the exciter and "commands" the
exciter to reduce power.  But at the inital ramp-up of
the RF signal from exciter to amplifier, there can be
a brief condition of overdrive, before the ALC circuit
can take effect.  In other words, the ALC corrects the
problem AFTER it has happened.

In a recent case, a ham was told he had key clicks all
over the place.  After a brief study, it was
determined that he ran his exciter at 100 watts when
barefoot.  When he clicked in his amplifier, he did
not change any settings on the exciter, and allowed
the amplifier's ALC system to control the exciter's
power level.  As it turned out, the amplifier (a solid
state unit) only needed 25 watts of drive to deliver
1kW.  The rig would begin to transmit, and the
envelope ramped very quickly OVER 25 watts,
overdriving the amplifier, until the amp's ALC circuit
could force the exciter to reduce power.  As an
experiment, he reduced his exciter's RF gain control,
until the ALC meter barely moved.  Before doing this,
I could hear his keyclicks 25-50kHz above and below
his intended signal.  After he did this, I could not
hear any keyclicks at all.  The SSB experiment was
similar - intermittent splatter was heard many kHz
above and below the intended signal, until he adjusted
the exciter to deliver only the required amount of
power to the amplifier.

This is a long answer, but the topic of IMD,
bandwidth, and driving an amplifier is not as simple
as one might hope.

73, and good luck,

Dave W8NF

--- rb <rbig@bellsouth.net> wrote:
> I need a brief refresher course in HF SSB.  It's
> been awhile now since I've
> been into this stuff....
> 
> Assuming a modern, commercially produced SSB final
> amp is operating in
> linear mode, is not overdriven by audio or rf gain
> on the input side, and is
> not distorting, can the transmitted signal still be
> wider than the 3khz
> standard?
> 
> Based on the above conditions, if an amp is putting
> out a SSB signal that is
> too wide, what practical means are there to narrow
> the sideband down?
> 
> _______________________________________________
> QRO mailing list
> QRO@mailman.qth.net
> http://mailman.qth.net/mailman/listinfo/qro


__________________________________
Do you Yahoo!?
New Yahoo! Photos - easier uploading and sharing.
http://photos.yahoo.com/