How to use quadrature detector to demodulate cw and ssb. I want to use
AD607 in simple project. There is Quadrature detector on "the end" with
two outputs I&Q - and internal phase shifter. How to detest SSB simplest
way?

73

mac

On 2006-07-04, MAc wrote:
How to use quadrature detector to demodulate cw and ssb. I want to use
AD607 in simple project. There is Quadrature detector on "the end" with
two outputs I&Q - and internal phase shifter. How to detest SSB simplest
way?

If it does all that, you just have to choose USB, LSB or AM by chosing
sum, difference or magnitude of the IQ output.

--
Ben Jackson

http://www.ben.com/

Ben Jackson napisa³(a):
On 2006-07-04, MAc wrote:
How to use quadrature detector to demodulate cw and ssb. I want to use
AD607 in simple project. There is Quadrature detector on "the end" with
two outputs I&Q - and internal phase shifter. How to detest SSB simplest
way?

If it does all that, you just have to choose USB, LSB or AM by chosing
sum, difference or magnitude of the IQ output.

Thank you

what frequency should be for Local oscilator?

MAc

It's not that easy Ben.....

The I-Q demodulator in this chip doesn't include the required 90 deg audio
phase shift network. That's usually one of the toughest parts of any
phasing-type detector. Also, the LO phase balance is specified as 1.2 deg
and the amplitude balance is specified as 0.2 dB. These imbalances would
combine to give an opposite sidenabd rejection of less than 40 dB, maybe 35
dB. Not too good by modern standards.

Furthermore, the detector circuits only work at IF frequencies, not RF
frequencies. This implies that the internal mixer would have to be used in a
downconversion mode, resulting in a requirement for an image-reject bandpass
filter at the mixer input and some measure of bandpass filtering at the
mixer output.

On top of all this, the LO noise floor is only -100 dBc/Hz, which would
probably cause pretty bad reciprocal mixing under crowded band conditions.

All things considered, I think that by the time one made this chip do what
it has to do for SSB/AM, one would be better off starting from scratch with
more conventional components.

Joe
W3JDR


"Ben Jackson" wrote in message
...
On 2006-07-04, MAc wrote:
How to use quadrature detector to demodulate cw and ssb. I want to use
AD607 in simple project. There is Quadrature detector on "the end" with
two outputs I&Q - and internal phase shifter. How to detest SSB simplest
way?

If it does all that, you just have to choose USB, LSB or AM by chosing
sum, difference or magnitude of the IQ output.

--
Ben Jackson

http://www.ben.com/

W3JDR napisa³(a):
It's not that easy Ben.....

The I-Q demodulator in this chip doesn't include the required 90 deg audio
phase shift network. That's usually one of the toughest parts of any
phasing-type detector.

What will be if I treat "I" mixer as a typical product detector. Will it
work?

MAc

To clarify - the idea is to build very simple receiver - Not "play with
quadrature detector". AD607 look simply and small

MAc wrote:
To clarify - the idea is to build very simple receiver - Not "play with
quadrature detector". AD607 look simply and small

Andy writes:

While I agree with everything MAC wrote, I think a point is being
missed.
A Quadrature detector is normally used only for FM demodulation. It
accomplishes this by limiting the signal and applying it to one port
of a 90deg detector. If the limited signal is shifted 90 degrees
(narrow
band), the quad detector will detect FM... If the limited signal is
shifted
ZERO degrees, the quad detector will detect AM, and in a much more
linear way than a diode type detector......

With CW, there would be no signal to zero beat with, and no audio
note would be formed.... Exactly the same with SSB....all the output
would be "DC"...

Mac's explanation is correct for an IQ detector... In this case a
local
signal EQUAL to the suppressed carrier of the SSB signal has to
be supplied..... I don't know how a local carrier can be synchronized
to
a SSB carrier that, if done really well, doesn't exist......

In Homodyne receivers, it is easy, since the transmit signal can be

the LO signal..... Otherwise, it has to be "guessed" at..... With
really
really really stable local oscillators, the guess can be really close,
and
the difference will creat an error that is so small it doesn't
matter... But,
since you don't know the incident phase of the received signal, you
have to do both I and Q and shift the IF by 90 degrees and sum (or
difference, depending whether you want USB or LSB) the outputs of
the two channels...... For SSB detection, the RF port is NOT shifted,
the LO port is both 0 and 90 degrees, and the IF port ( I and Q) are
shifted 90 and combined........

It is a hell of a lot easier to "approximate" the LO, and shift it
manually
until Donald Duck starts sounding human.... In older SSB rigs, this was

called a "clarifier" control..... It isn't necessary in many modern
systems
with really really really good TCXOs and stuff....

Just my two cents worth.... If anyone disagrees, I'd be happy to
learn from them where my explanation is wrong..... but I warn you,
it ain't far wrong.... cause it worked for me for many many years
in designing receivers for both commercial and military products
....:)))))

Andy in Eureka, W4OAH ( retired comm/radar engineer and ham for 45
years )

AndyS wrote:
A Quadrature detector is normally used only for FM demodulation.

Yes.

but (from Analog Devices):
"...Applications of the AD607 include narrowband systems with a high
first IF (21.4 MHz to 300 MHz) and a second IF at 10.7 MHz, 455 kHz, or
450 kHz. These include Dual Conversion IS136, GSM, TETRA, and MSAT
Receivers; and Single or Dual Conversion VHF and HF AM, SSB, CW, or QPSK
Receivers... "

But :-) I did not find any SSB application :-( So I need an example.




Just my two cents worth.... If anyone disagrees, I'd be happy to
learn from them where my explanation is wrong..... but I warn you,
it ain't far wrong.... cause it worked for me for many many years
in designing receivers for both commercial and military products
...:)))))


Andy Thanks for your cents, rather dollars.

PS there is internal quadrature pll in ad607.

MAc

Andy Thanks for your cents, rather dollars.

PS there is internal quadrature pll in ad607.

MAc


Andy responds;

Ok.. I admit that I didn't go find the AD607 and study it before
replying
to Mac... However,,, I'll bet you a dollar that the fellow who wrote
the
app note that Mac refers to has a hell of a lot less experience than
either Mac or myself..... not that I would expect a young app note
writer to overstate the applications of his product ( big damn grin
)...
The AD606 really makes a great tie-tac if you glue an alligator clip
to it, tho I've never actually seen one for sale... :)))))))

if the gilbert cell, or single ended long tailed pair multiplier
used as a
quad detector allows external coupling to the ports, it can be used as
a BFO for CW or SSB,..., HOWEVER, having a limiting stage in the
linear path of such a device is NOT a great idea.... Sure, it can be
made
to work, but..... well, heck... I've built a LOT of things that were
really
bad ideas, and I try not to duplicate them again, or tell others about

them.... especially in app notes that I've made input to......
I've seen the MC3372 used as an SSB receiver in either Ham Radio
or some such.... and I never seriously considered doing it myself since
the limitations were so obvious....

An 811A can be used a a product detector also, but the
manufacturers
don't advertise it as such,...... because it is a REALLY BAD idea....
If a fellow wants a simple CW/SSB receiver, a much better idea is to
use something like an NE602..... I've done that, for a home project ,
and
it works OK,..... but certainly not something I'd try to produce as a
receiver for Bendix, Texas Instruments, or Raytheon...... There's a
limit
to how far one should stretch an application.....

Thanks Mac, for your input. I bet we agree on almost
everything.....

Andy W4OAH

Ok.. I admit that I didn't go find the AD607 and study it before
replying.....

.....HOWEVER, having a limiting stage in the
linear path of such a device is NOT a great idea
---------------------

Andy,

You really should go and read the data sheet on the AD607.

1) It's a linear component with no limiter stages
2) It has an "I-Q" detector, not a "quadrature detector".

Comparing it to the MC3372, or any other FM detector chip, is as far off
base as comparing it to an 811.

However, having said all that, I think we would agree that there are better
and simpler solutions for amateur SSB and AM applications.


Joe
W3JDR


"AndyS" wrote in message
oups.com...

Andy Thanks for your cents, rather dollars.

PS there is internal quadrature pll in ad607.

MAc


Andy responds;

Ok.. I admit that I didn't go find the AD607 and study it before
replying
to Mac... However,,, I'll bet you a dollar that the fellow who wrote
the
app note that Mac refers to has a hell of a lot less experience than
either Mac or myself..... not that I would expect a young app note
writer to overstate the applications of his product ( big damn grin
)...
The AD606 really makes a great tie-tac if you glue an alligator clip
to it, tho I've never actually seen one for sale... :)))))))

if the gilbert cell, or single ended long tailed pair multiplier
used as a
quad detector allows external coupling to the ports, it can be used as
a BFO for CW or SSB,..., HOWEVER, having a limiting stage in the
linear path of such a device is NOT a great idea.... Sure, it can be
made
to work, but..... well, heck... I've built a LOT of things that were
really
bad ideas, and I try not to duplicate them again, or tell others about

them.... especially in app notes that I've made input to......
I've seen the MC3372 used as an SSB receiver in either Ham Radio
or some such.... and I never seriously considered doing it myself since
the limitations were so obvious....

An 811A can be used a a product detector also, but the
manufacturers
don't advertise it as such,...... because it is a REALLY BAD idea....
If a fellow wants a simple CW/SSB receiver, a much better idea is to
use something like an NE602..... I've done that, for a home project ,
and
it works OK,..... but certainly not something I'd try to produce as a
receiver for Bendix, Texas Instruments, or Raytheon...... There's a
limit
to how far one should stretch an application.....

Thanks Mac, for your input. I bet we agree on almost
everything.....

Andy W4OAH