DomainKey-Signature: a=rsa-sha1; c=nofws;
        d=gmail.com; s=gamma;
        h=mime-version:date:message-id:subject:from:to:content-type;
        b=XsqzVBzUgjw9J8G33M+mBxykSKR8GCPnvsqSbmKSUkREmvsG+94qrNMRpKXPO3rt0m
         vssPI+xiiuUXheGpIZvrtbs7fY53IH1uOdt0a0GSIFpUJC7EttH3UFbDj7KYyA9VOrtT
         iyH+pJ+RIgTaBmHTReRNe5x/YSOXQZSTlVSqs=
MIME-Version: 1.0
Date: Mon, 7 Feb 2011 23:21:54 +0000
Message-ID: <AANLkTimkLcqnM1M=dzDNYDiXNmnd0nRmgOV2-C+jn-+V@mail.gmail.com>
From: Andy Talbot <andy.g4jnt@gmail.com>
To: rsgb_lf_group@blacksheep.org, sub9khz@yahoogroups.com
DomainKey-Status: good (testing) 
Subject: LF: Frequency Shift keying types
Content-Type: multipart/alternative; boundary=20cf3054a17152da3a049bb97f5a
Sender: owner-rsgb_lf_group@blacksheep.org
Precedence: bulk
Reply-To: rsgb_lf_group@blacksheep.org


--20cf3054a17152da3a049bb97f5a
Content-Type: text/plain; charset=ISO-8859-1

Wolf  brought up the subject of coherent Frequency Shifting earlier...

See http://www.g4jnt.com/FSK_Types.pdf   for spectral plots comparing both
types of FSK.   These were generated using the PIC DDS source with both
software versions in turn.    The waveform is made by switching
between tones of 1200 and 2200Hz using 25WPM CW from a beacon keyer
module to simulate random data.

The lack of any structure for the narrowest frequency resolution using phase
continuous keying is very obvious.  This is the more usualy type of FSK seen
in practice, usually generated by altering an oscillator frequency with the
input data. The disadvantage when anyone tries to receive one of the tones
using ultra narrow band techniques is very obvious.     The slightly higher
out of band levels due to the hard switching of the coherent waveform are
just about noticeable, but don't appear that much worse.

Phase coherent (or Frequency Exchange) Keying is more complicated to do in
hardware, requiring two oscillators to run continuously and to select one or
the other with the input data.   However, if implemented using DSP probably
the same complexity for either,  with two simultaneous NCOs, and selecting
one or the other for the tone output  but removing the need to change the
frequency word regularly.

Andy   G4JNT



> ... !
>
> One suggestion for VLFers: A PIC doesn't cost a fortune. Build two of these
> generators, and clock them with the same source (GPSDO or OCXO). For the FSK
> simply flip between both generators. This has the *big* (cannot emphasize
> this enough..) advantage that if the signal is to weak to be seen as DFCW
> etc, we can still dig it out through a long FFT because each of the two
> carriers remains coherent, and can be coherently averaged as long as
> necessary (much longer than a single DFCW symbol). I noticed this on
> Stefan's signal today : With the very slow settings, all you see is a
> rounded hump in the spectrum because the signal isn't coherent. No chance to
> detect it for "DX" receivers.
>
> Of course flipping the (FSK-) switch may cause a heavy click then
> (depending on the momentary phase shift between the two carriers) but I
> think a robust PA will survive this.
>
> Have a nice sunday,
>    Wolf .
>

--20cf3054a17152da3a049bb97f5a
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable

<div>Wolf =A0brought up the subject of coherent Frequency Shifting=A0earli=
er...</div>
<div>=A0</div>
<div>See <a href=3D"http://www.g4jnt.com/FSK_Types.pdf">http://www.g4jnt.c=
om/FSK_Types.pdf</a>=A0=A0 for spectral plots comparing both types of FSK.=
=A0=A0 These were generated using the PIC DDS source with both software ve=
rsions in turn.=A0=A0=A0 The waveform is made=A0by switching between=A0ton=
es of 1200 and 2200Hz using 25WPM CW from a beacon keyer module=A0to simul=
ate random data.</div>

<div><br>The lack of any structure for the narrowest frequency resolution=
 using phase continuous keying is very obvious.=A0 This is the=A0more usua=
ly=A0type of FSK seen in practice, usually generated by altering an oscill=
ator frequency with the input=A0data.=A0The disadvantage when=A0anyone tri=
es to receive one of the tones using ultra narrow band techniques is very=
 obvious.=A0=A0=A0=A0 The slightly higher out of band levels due to the ha=
rd switching of the coherent waveform are just about noticeable, but=A0don=
&#39;t appear=A0that much worse.</div>

<div>=A0</div>
<div>Phase coherent (or Frequency Exchange) Keying is more complicated to=
 do in hardware, requiring two oscillators to run continuously and to sele=
ct one or the other with the input data.=A0=A0 However, if implemented usi=
ng DSP probably the same complexity for either,=A0=A0with two simultaneous=
 NCOs, and selecting one or the other for the=A0tone output=A0 but removin=
g the need=A0to=A0change the frequency word regularly.</div>

<div>=A0</div>
<div>Andy=A0=A0 G4JNT</div>
<div><br>=A0</div>
<div class=3D"gmail_quote">
<blockquote style=3D"BORDER-LEFT: #ccc 1px solid; MARGIN: 0px 0px 0px 0.8e=
x; PADDING-LEFT: 1ex" class=3D"gmail_quote">
<div text=3D"#000000" bgcolor=3D"#ffffff">... !<br><br>One suggestion for=
 VLFers: A PIC doesn&#39;t cost a fortune. Build two of these generators,=
 and clock them with the same source (GPSDO or OCXO). For the FSK simply=
 flip between both generators. This has the *big* (cannot emphasize this=
 enough..) advantage that if the signal is to weak to be seen as DFCW etc,=
 we can still dig it out through a long FFT because each of the two carrie=
rs remains coherent, and can be coherently averaged as long as necessary=
 (much longer than a single DFCW symbol). I noticed this on Stefan&#39;s=
 signal today : With the very slow settings, all you see is a rounded hump=
 in the spectrum because the signal isn&#39;t coherent. No chance to detec=
t it for &quot;DX&quot; receivers.<br>
<br>Of course flipping the (FSK-) switch may cause a heavy click then (dep=
ending on the momentary phase shift between the two carriers) but I think=
 a robust PA will survive this.<br><br>Have a nice sunday,<br>=A0=A0 Wolf=
 .<br>
</div></blockquote></div>

--20cf3054a17152da3a049bb97f5a--