[Premium-Rx] Comparing AM Synch Detection to AM Product Detection

[Michael O'Beirne]

Dear All,

Here are some thoughts to expand the debate.

1.    The SSB product detector works independently of the carrier and so
theoretically with a local stable CIO there should be an improvement.  For
listening to voice transmissions you need to tune to within 50Hz of the
carrier, but for music you need to be within about 2Hz which may not be
possible for receivers tuning in 10Hz increments and with the frequency
reference oscillator imperfectly set up.  Not many of us have decent Cs or
Rb frequency standards on the bench to set oscillators to within minute
fractions of a Hz.

2.    The product dectector / SSB arrangement using the SSB filter will
improve the clarity of speech compared with an AM filter but will sound
awful for music unless the SSB filter is relatively wide - say at least
2.7kHz or more at the 6db points.  Most "ham" rigs give around 2.4kHz.

3.    The SSB filter needs to be of very good quality otherwise you get
quite bad phase distortion, which a trained musical ear will detect.  The
2.95kHz-wide ITT crystal filters in my RA1792 are ideal.

4.    Synchronous systems vary enormously, but they generally phase-lock a
local CIO to the carrier and use a product detector.    Most are OK on good
signals but drop out with heavy QRM and QSB - which is when you need them
the most!!!.  Some can select USB or LSB.  This is most helpful where quite
severe QRM affects one sideband more than the other.  I find this sometimes
with the BBC signal on 12,095kHz.  The Surrey Electronics demodulator I
described some months back has unusually elaborate circuitry to deal with
heavy fading and can select sidebands.

5.    In my view the best solution is the classic ISB demodulator comprising
3 separate amplifying and filter chains, one each for USB, LSB and carrier.
The carrier chain will have very narrow bandwidth - eg 50Hz - and this is
fed to the product detector in lieu of the CIO.  In such a narrow bandwidth
the detected carrier will be almost completely free of QRM.  Assuming that
there are no substantial phase delays in the carrier filter, the phase
relationship between carrier, USB and LSB will be preserved (but subject of
course to the usual propagation time delays in the ionosphere which may
affect the sidebands and carrier to different degrees).  What's more, if
there is any drift in the Tx or Rx, the carrier and sideband will drift
together and there will be no change of audio pitch at the output of the
demodulator.

6.    The tightly filtered carrier can also be used via a frequency
discriminator to provide AFC to hold the overall tuning error to a few Hz,
and also via a diode detector to give AGC which can be superior to the usual
sideband-derived AGC systems.

7.    For some years the BBC World Service (and I think CBC) used the Racal
RA98C ISB adaptor with the RA17L for long range HF feeder circuits at the
distant end for rebroadcast of BBC programmes there on local MW or FM.  The
normal -3dB filter bandwith of the RA98 is 300Hz to 6,000Hz.  Using a 6.5
kHz bandwidth in the receiver reduces the overall bandwidth to about 300 -
3250Hz.  In the RA98C version the sideband filter bandwidth is widened to
about 90Hz - 6,000Hz to give a much enhanced bass response, necessary for
rebroadcasting quality.

8.    A transistorised version is the RA298 (for a 100kHz input) which uses
excellent professional Kokusai mechanical SSB filters with a bandwidth of
around 3.5kHz and a very narrow Plessey crystal filter for the carrier.  The
perceived audio quality really is excellent with a distortion spec of around
2%.

9.    Of course the systems outlined above were professional and expensive.
I imagine that some of you clever guys could implement all of this in DSP
nowdays at far less cost.  On the other hand if you can get hold of one of
these demodulators you could use them with the IF output of say a Harris RF
390 and a simple crystal controlled mixer to convert the 455kHz of the IF
output to 100kHz.  I am working on a simple quad mixer using an SBL-1 and a
crystal oscillator on 355kHz.  The hassle is finding a 355kHz crystal at a
reasonable price.  I know that Racal made special converters specifically
for this job but they are like hen's teeth.

73s
Michael O'Beirne
G8MOB


----- Original Message -----
From: "Tom M." <courir26 at yahoo.com>
To: "Ahmet Gundes" <ahmet-m at usa.com>; "Premium-RX"
<premium-rx at ml.skirrow.org>
Sent: Friday, May 20, 2005 1:44 PM
Subject: Re: [Premium-Rx] Comparing AM Synch Detection to AM Product
Detection


> Depending on the rig, the sync detector may offer a bit more advantage as
the
> lock range may as high as +/- 2 KHz or so.  In product detection you're
pinned
> to the frequency of interest.  Adjacent channels that are a problem can be
> partially avoided by the sync detect lock range (i.e. if you're desired
freq is
> 870 Khz and you have adjacent channel on 860 Khz, sync detector may allow
you
> to tune 872 Khz and still remain locked on.
>
> Also, if the frequency generation device drifts at all, product detection
will
> noticeably change freq while the sync detect will stay locked on and you
won't
> notice it in the pitch of the sound.
>
> All this said, product detection still works perfectly fine.  Gives one
the
> opportunity to fiddle with the dials.
>
> The sync detect adds a bit of gee whiz factor.  I have rigs with both and
like
> both.
>