[HomeBrew] Single sideband generation by the Weaver method

[Ray Anderson]

WB9IPA wrote:
>
>Does anybody have (or know where I can get) a copy of "A Third Method of
>Generation and Detection of SSB Signals" by Donald K. Weaver?  This appear=
ed
>in the December 1956 issue of the Proceedings of the Institute of Radio
>Engineers, pages 1703 - 1705. =20
>
>I am in the process of building a Weaver-type transceiver and would like t=
o
>have a copy of this for reference.  I am willing to pay reasonable costs f=
or
>copying and postage.  Any help would be greatly appreciated.
>
>
>Paul Alexander
>WB9IPA
>E-mail:  paul.alexander@respironics.com
>

Paul-

I may have a copy of the original Weaver article somewhere, I'll
need to look around, however the following bit of info is something
I put together a while back when I was looking at tradeoffs between
the Weaver and the Phasing method. It may be of interest (or not :).

A quick google.com search on "weaver sideband" comes back with a
gazillion hits on the subject, but a fast scan of the results
didn't show any electronic copies of the original 1956 article.

If I find the original article I'll see to it that you get a copy.

73's,=09
=09Ray=09WB6TPU
raymond.anderson@sun.com

---------------------------------------------------------------------
           Phasing Vs Weaver methods of SSB Generation


     Phasing method and filter method both have sideband offset=20
from suppressed carrier, while Weaver method has suppressed carrier=20
centered in middle of signal.
    =20
     In Weaver method, imperfect suppression of unwanted sideband=20
energy shows up as inverted speech on top of the desired sideband,=20
while in phasing and filter methods the unwanted sideband shows up=20
on the other side of the carrier.

     Weaver method- input audio is mixed with locally generated=20
subcarrier centered in audio passband. i.e. for 300-3000 audio the=20
band center is 1650 [(3000+300)/2] . So when audio input is mixed=20
with this subcarrier the following results:

     1650+300=3D1950
     1650+3000=3D4650
     1650-300=3D1350
     3000-1650=3D1350

     An upper sideband is created from 1950 to 4650

     A lower sideband is created such that components from DC=20
to exactly 1/2 the audio passband (1650) appear from 0 to 1650 and=20
those audio components from 1650 to 3000 are aliased into the same=20
range but are inverted and are 180 degrees out of phase with the other=20
components.
    =20
     This signal is low pass filtered at a freq. between 1650=20
and 1950 to scrape off the the sum product. This leaves a signal from=20
DC to 1350 that contains the audio information for the full 300- 3000=20
Hz input audio signal. There is a band whose width is equal to twice=20
the lowest frequency input audio component where no signal exists=20
centered on the 1650 subcarrier. Therefore the LPF requires a  ratio=20
of 1650/1350=3D1.22 between the passband and the stopband. An eight=20
order SC filter can handle this.

     The above process transpires for both cos and sin=20
subcarriers. The result is the Ia and Qa signals as in the=20
phasing method.

     When these Ia and Qa signals are mixed with sin and cos=20
versions of the carrier oscillator the nonfolded component forms=20
a sideband above the carrier and the folded (inverted phase)=20
component forms an independent sideband below the carrier. To an=20
external receiver this appears as a SSB signal centered on the=20
suppressed carrier.
  =20
  Weaver/Phasing Differences:
  --------------------------=20
    =20
     *Unwanted components appear in-band with Weaver method.
They appear out of band with phasing and filter method.

     *Weaver signal is centered on suppressed carrier.=20
Filter and phasing signal is offset above or below.

     * Unwanted sideband energy appears as inverted speech=20
on top of desired signal. -10dB is not disruptive, -20 is not bothersome.=
=20
It is reported that even 0 dB suppression allows intelligibility.=20
Since it is in band it does not need to meet FCC spec of -40 dB for=20
out of band energy. Phasing and filter method signals with imperfect=20
opposite sideband suppression do not effect intelligibility but may=20
cause interference to other stations , and is a regulatory issue.

     *Carrier leakthru in Weaver method due to imperfect=20
circuit balance in the mixers etc. cause interfereing carrier to be=20
inserted in center of audio passband. In TX applications this may=20
be attenuated enough to be legal and not effect signal, but in rx=20
application it will limit dynamic range of receiver due to spur=20
in center of passband.

     *Bandwidth of Weaver signal is set primarily by LPF after=20
first set of mixers. In Phasing method it is determined by audio BPF=20
prior to phase splitter. Filter method BW is determined by SB filter=20
(xtal or mechanical usually.

     *Critical elements- Weaver: mixer balance(DC offsets),=20
LPF steepness. Phasing: goodness of hilbert transformer (amplitude=20
balance and quadrature) goodness of audio BPF and mixer balance. Filter:=20
goodness of SB filter and mixer balance.



Sideband suppression for phasing method:
     Sup (dB)=3D 20LOG((200+E)/E)
     E is % amplitude imbalance

1% =3D 45 dB
2% =3D 40 dB
4% =3D 34 dB

Sup (dB)=3D 20LOG(COT(A/2))

A is degrees deviation from quadrature (90)
1 =3D 40 dB
2 =3D 35 dB
3.5 =3D 30 dB

-------------------------------------------------------------------   =20
=1A=1A=1A=1A=1A=1A=1A=1A=1A=1A=1A=1A=1A=1A=1A=1A=1A=1A=1A=1A=1A=1A=1A=1A=1A=
=1A=1A=1A=1A=1A=1A=1A=1A=1A=1A=1A=1A=1A=1A=1A=1A=1A=1A=1A=1A=1A=1A=1A=1A=1A=
=1A=1A=1A=1A=1A=1A=1A=1A=1A=1A=1A=1A=1A=1A=1A=1A=1A=1A=1A=1A=1A=1A=1A=1A=1A=
=1A=1A=1A=1A=1A
=1A=1A=1A=1A=1A=1A=1A=1A=1A=1A=1A=1A=1A=1A=1A=1A=1A=1A=1A=1A=1A=1A=1A=1A=1A