EMRFD Message Archive 6532

Message Date From Subject
6532 2011-08-06 19:39:26 qrp.gaijin Single-signal DC rx without AF phase shift
Greetings,

At the following web page, a design for a single-signal DC receiver is proposed that uses no AF phase shifting. 90-degree phase-shifted RF signals are mixed with 90-degree phase-shifted LO signals, and the resulting audio is simply summed without requiring AF phase shift.

http://www.qsl.net/g3cwi/dc.html

To quote the original web page:

>The signal comes in from the aerial and is fed into a phase shift
>network that gives outputs of 0 degrees and 90 degrees. These
>outputs are fed into two mixers. The local oscillator is also fed
>into a phase shift network giving 0 and 90 degrees shifts. These
>outputs feed the oscillators [ed. note: "mixers" is probably
>intended here instead of "oscillators"]. The output from the
>oscillators [ed.: mixers] is fed to a balancing potentiometer and
>then through to the AF stage. By switching the phases correctly
>you can get USB or LSB.

I haven't seen a practical implementation of this design, but it sounds like it would work well. Do you see any inherent problems with this approach?
6533 2011-08-07 07:45:43 Ashhar Farhan Re: Single-signal DC rx without AF phase shift
Won't work.
Let's imagine a signal at 7000khz and our local osc. at 7001khz. Now,
the two 'sidebands' are going to be at 14001khz and 1khz. The scheme
will suppress the other sideband at 14001khz that is anyway filtered
out by the audio low pass.
On the other hand, imagine a signal at 7002khz that you don't want.
The signal that will emerge from this architecture is local osc -
signal freq. In this case, it will be at -1 khz (meaning, it's phase
is inverted).

On the other hand, if you just shell out a dollar more for an NE5532,
you will have a fabulous phasing receiver that rivals the audio of an
AOR. We must all thank rick for showing us the way!

- farhan

On 8/7/11, qrp.gaijin <qrp.gaijin@yahoo.com> wrote:
> Greetings,
>
> At the following web page, a design for a single-signal DC receiver is
> proposed that uses no AF phase shifting. 90-degree phase-shifted RF signals
> are mixed with 90-degree phase-shifted LO signals, and the resulting audio
> is simply summed without requiring AF phase shift.
>
> http://www.qsl.net/g3cwi/dc.html
>
> To quote the original web page:
>
>>The signal comes in from the aerial and is fed into a phase shift
>>network that gives outputs of 0 degrees and 90 degrees. These
>>outputs are fed into two mixers. The local oscillator is also fed
>>into a phase shift network giving 0 and 90 degrees shifts. These
>>outputs feed the oscillators [ed. note: "mixers" is probably
>>intended here instead of "oscillators"]. The output from the
>>oscillators [ed.: mixers] is fed to a balancing potentiometer and
>>then through to the AF stage. By switching the phases correctly
>>you can get USB or LSB.
>
> I haven't seen a practical implementation of this design, but it sounds like
> it would work well. Do you see any inherent problems with this approach?
>
>
>
>

--
Sent from my mobile device
6534 2011-08-07 11:06:16 Alex P Re: Single-signal DC rx without AF phase shift
Farhan,

Had a look at the NE5532 and I get a TI opamp. Can you take a look and check this ?
Thanks.

Alex

6535 2011-08-07 13:43:21 Chris Trask Re: Single-signal DC rx without AF phase shift
>
> Won't work.
> Let's imagine a signal at 7000khz and our local osc. at 7001khz. Now,
> the two 'sidebands' are going to be at 14001khz and 1khz. The scheme
> will suppress the other sideband at 14001khz that is anyway filtered
> out by the audio low pass.
> On the other hand, imagine a signal at 7002khz that you don't want.
> The signal that will emerge from this architecture is local osc -
> signal freq. In this case, it will be at -1 khz (meaning, it's phase
> is inverted).
>
> On the other hand, if you just shell out a dollar more for an NE5532,
> you will have a fabulous phasing receiver that rivals the audio of an
> AOR. We must all thank rick for showing us the way!
>
> - farhan
>
> On 8/7/11, qrp.gaijin <qrp.gaijin@yahoo.com> wrote:
> > Greetings,
> >
> > At the following web page, a design for a single-signal DC receiver is
> > proposed that uses no AF phase shifting. 90-degree phase-shifted RF
signals
> > are mixed with 90-degree phase-shifted LO signals, and the resulting
audio
> > is simply summed without requiring AF phase shift.
> >
> > http://www.qsl.net/g3cwi/dc.html
> >
> > To quote the original web page:
> >
> > >The signal comes in from the aerial and is fed into a phase shift
> > >network that gives outputs of 0 degrees and 90 degrees. These
> > >outputs are fed into two mixers. The local oscillator is also fed
> > >into a phase shift network giving 0 and 90 degrees shifts. These
> > >outputs feed the oscillators [ed. note: "mixers" is probably
> > >intended here instead of "oscillators"]. The output from the
> > >oscillators [ed.: mixers] is fed to a balancing potentiometer and
> > >then through to the AF stage. By switching the phases correctly
> > >you can get USB or LSB.
> >
> > I haven't seen a practical implementation of this design, but it sounds
like
> > it would work well. Do you see any inherent problems with this approach?
> >
>

Farhan is right, this won't work. The first mixer has a cosine input
and a cosine LO, resulting in a cosine output. The second mixer has a sine
input and a sine output, which also results in a cosine output.

I spent a lot of years trying to devise a method for eliminationg
signals that were not symmetrical about a fixed carrier. One of the
circuits I came up with was for demodulating an input into sum (I) and
difference (Q) channels without using baseband phase shifters. It can be
done with four mixers, but to eliminate phase and amplitude erros it
requires six, arranged in pairs.

The first pair of mixers (#1 and #2) receive the input signal in the
same phase. The LO signals are 90º out of phase, the result being the
typical I/Q demodulator. A little deëmphasis filtering gets rid of the
image and LO leakage.

The second pair of mixers (#3 and #4) receive the #1 and #2 mixer
outputs. The two LO signals are in phase. Now comes the fun part. The #3
and #4 mixer outputs go to the third pair of mixers (#5 and #6), whose LO
signals are 90º out of phase. This converts the signals back to baseband
with the additional quadrature phase shift in the Q channel (mixer #6
output). Now you just add or subtract them to get LSB or USB.

A little busy perhaps, but it gets rid of the annoying baseband phase
shifter and is good down to DC if everything is direct coupled.

Chris Trask
N7ZWY / WDX3HLB
Senior Member IEEE
http://www.home.earthlink.net/~christrask/
6536 2011-08-07 13:48:16 Chris Trask Re: Single-signal DC rx without AF phase shift
>
> Farhan is right, this won't work. The first mixer has a cosine input
> and a cosine LO, resulting in a cosine output. The second mixer has a
sine
> input and a sine output, which also results in a cosine output.
>

Ooops! That was supposed to be "The second mixer has a sine input and a
sine LO, which also results in a cosine output."

Chris Trask
N7ZWY / WDX3HLB
Senior Member IEEE
http://www.home.earthlink.net/~christrask/
6537 2011-08-07 16:18:10 Ashhar Farhan Re: Single-signal DC rx without AF phase shift
alex,

yes and no. i was goofing off with a TL084 (which has 4 op-amps) and i
mistakenly assumed that the NE5532 is also a quad amp (it has two op-amps in
the package).

my intention was to say that four op-amps (using two NE5532 chips) and a
couple of resistors and capacitors can easily make audio wideband all pass
filters that can achieve between 30-40dB sideband suppression. which is as
good as what you may expect from a nominally designed ladder crystal filter
based ssb transmitters. look up the article on microR2. sure, the circuit
looks a little elaborate. but you will realize that it is all at audio
frequencies and it uses all ordinary parts available in any junkbox. the
exotic resistor values obtained by adding regular resistors in series. not
complicated at all.

- farhan



6772 2011-11-05 00:46:27 drmail377 Re: Single-signal DC rx without AF phase shift
The scheme described here by Chris Mr. Trask is interesting. I'm wondering if there has been any thought give to reversing this process some way to transmit single sideband as well?

Rgds, David

6773 2011-11-05 01:30:40 drmail377 Re: Single-signal DC rx without AF phase shift
I think I answered my own question. The method of transmission is similar (a bit simpler actually); it is the classic Weaver's Thhird Method of SSB Generation. I looked at this long ago.

It is my opinion that any complexity removed with the elimination of all-pass networks Weaver's method is offset by the critical need for quadrature phase and amplitude balance. I would think the same applies to the method outlined by Mr. Trask.

However, one advantage of Weaver's method is that by adding yet a bit more complexity, the sideband rejecti