Hi,
I'm having trouble understanding the similarities/differences/advantages of
various AM digital demodulation methods, which makes me incapable of
deciding what to use. I would like to go throug a few methods here and here
your thoughts. 

First, I assume the common goal of all those techniques is to shift the
carrier back down to DC and there are a few ways to do this that have
different behavior. Right?

Method 1)
-Bandpass filter around the carrier.
-Take the absolute value. THIS is what shifts the carrier down to DC. 
-Low-pass filter. 
Now I guess taking the absolute value is not a very good way to shift the
carrier down to DC and is sensitive to anything other signal in the stream.
This is why we need to bandpass? 
Is the performance of the system pretty much only dominated by the bandpass
filter? In other words, is taking the abs() value messing up my data EVEN
if there was only the carrier/data in the stream?

Method 2)
-Multiply incoming stream with sin(wc*t) and cos(wc*t) sample by sample. 
-Lowpass filter each output. This gives I and Q.
-Calculate sqrt(I^2+Q^2).

Here, the carrier is perfectly shifted down to DC no matter what the stream
contains, which is why we don't need the bandpass filter. Can we say that
the performance of such a system is only dependent on the lowpass filter? 
What about the precision of the sin(wc*t) and cos(wc*t)? Keeping only one
bit here would mean using a square wave, which can't be really good. Is
there any way to know how much this affects performance?

And probably most importantly, how can I measure performance of the system?
Is there a standard way. I'm not yet looking for complicated theoretical
explanations...something intuitive is better to learn at first - at least
for me!

Thanks!

gretzteam wrote:
> Hi,
> I'm having trouble understanding the similarities/differences/advantages of
> various AM digital demodulation methods, which makes me incapable of
> deciding what to use. I would like to go throug a few methods here and here
> your thoughts. 
> 
> First, I assume the common goal of all those techniques is to shift the
> carrier back down to DC and there are a few ways to do this that have
> different behavior. Right?

Not quite. There are two methods to demodulate AM: use carrier from 
incoming signal as it is or regenerate carrier locally and demodulate 
synchronously. There is a gazillion of ways to implement either method.

> Method 1)
> -Bandpass filter around the carrier.
> -Take the absolute value.

> Method 2)
> -Multiply incoming stream with sin(wc*t) and cos(wc*t) sample by sample. 
> -Lowpass filter each output. This gives I and Q.
> -Calculate sqrt(I^2+Q^2).

Method 1 === Method 2. Both do non-synchronous demodulation. (1) does it 
in real domain, (2) does it in complex domain.

> And probably most importantly, how can I measure performance of the system?

For analog system, the figure of merit is S/(THD + N). For digital 
system, the figure of merit is BER.


Vladimir Vassilevsky
DSP and Mixed Signal Design Consultant
http://www.abvolt.com

On 4/23/2010 9:53 AM, gretzteam wrote:
> Hi,
> I'm having trouble understanding the similarities/differences/advantages of
> various AM digital demodulation methods, which makes me incapable of
> deciding what to use. I would like to go throug a few methods here and here
> your thoughts.
>
> First, I assume the common goal of all those techniques is to shift the
> carrier back down to DC and there are a few ways to do this that have
> different behavior. Right?

Wrong. The object is to extract the modulating signal.

> Method 1)
> -Bandpass filter around the carrier.
> -Take the absolute value. THIS is what shifts the carrier down to DC.
> -Low-pass filter.

There is a potentially serious flaw here. You must not assume that your 
samples are anywhere near the carrier peaks.

> Now I guess taking the absolute value is not a very good way to shift the
> carrier down to DC and is sensitive to anything other signal in the stream.
> This is why we need to bandpass?

If you bandpassed adequately, the only other signal would overlay the 
signal you hoped to recover. That's called "interference" (hams call it 
QRM) and is outside the scope of your question.

> Is the performance of the system pretty much only dominated by the bandpass
> filter? In other words, is taking the abs() value messing up my data EVEN
> if there was only the carrier/data in the stream?

No. You must not assume that your samples are anywhere near the carrier 
peaks.

> Method 2)
> -Multiply incoming stream with sin(wc*t) and cos(wc*t) sample by sample.
> -Lowpass filter each output. This gives I and Q.
> -Calculate sqrt(I^2+Q^2).

You can think of this another way. Sqrt(I^2+Q^2) gives the magnitude of 
the envelope, exactly what you want.

> Here, the carrier is perfectly shifted down to DC no matter what the stream
> contains, which is why we don't need the bandpass filter.

Of course you need the bandpass filter. You want to operate on the 
carrier and its sidebands only, not on every frequency from power line 
to X-rays. If there is more than one signal in the passband, that's 
still QRM, and the signal is contaminated.
>                                                           Can we say that
> the performance of such a system is only dependent on the lowpass filter?
> What about the precision of the sin(wc*t) and cos(wc*t)? Keeping only one
> bit here would mean using a square wave, which can't be really good. Is
> there any way to know how much this affects performance?

Try it and see.

> And probably most importantly, how can I measure performance of the system?
> Is there a standard way. I'm not yet looking for complicated theoretical
> explanations...something intuitive is better to learn at first - at least
> for me!

Before you can measure or calculate performance, you need a definition 
of it.

Jerry
-- 
"It does me no injury for my neighbor to say there are 20 gods, or no
God. It neither picks my pocket nor breaks my leg."
          Thomas Jefferson to the Virginia House of Delegates in 1776.
���������������������������������������������������������������������

On Apr 23, 9:53=A0am, "gretzteam" <gretzteam@n_o_s_p_a_m.yahoo.com>
wrote:
> Hi,
> I'm having trouble understanding the similarities/differences/advantages =
of
> various AM digital demodulation methods, which makes me incapable of
> deciding what to use. I would like to go throug a few methods here and he=
re
> your thoughts.
>
> First, I assume the common goal of all those techniques is to shift the
> carrier back down to DC and there are a few ways to do this that have
> different behavior. Right?
>
> Method 1)
> -Bandpass filter around the carrier.
> -Take the absolute value. THIS is what shifts the carrier down to DC.
> -Low-pass filter.
> Now I guess taking the absolute value is not a very good way to shift the
> carrier down to DC and is sensitive to anything other signal in the strea=
m.
> This is why we need to bandpass?
> Is the performance of the system pretty much only dominated by the bandpa=
ss
> filter? In other words, is taking the abs() value messing up my data EVEN
> if there was only the carrier/data in the stream?
>
> Method 2)
> -Multiply incoming stream with sin(wc*t) and cos(wc*t) sample by sample.
> -Lowpass filter each output. This gives I and Q.
> -Calculate sqrt(I^2+Q^2).
>
> Here, the carrier is perfectly shifted down to DC no matter what the stre=
am
> contains, which is why we don't need the bandpass filter. Can we say that
> the performance of such a system is only dependent on the lowpass filter?
> What about the precision of the sin(wc*t) and cos(wc*t)? Keeping only one
> bit here would mean using a square wave, which can't be really good. Is
> there any way to know how much this affects performance?
>
> And probably most importantly, how can I measure performance of the syste=
m?
> Is there a standard way. I'm not yet looking for complicated theoretical
> explanations...something intuitive is better to learn at first - at least
> for me!
>
> Thanks!

Method 2 is solid.  After implementing method 2 you can try to do
tricks that may give you some improvements.  One trick I know of is to
sychronize the carrier with a digital PLL.  This is tough, it has some
other aspects to it that can go wrong and you get a maximum 3 dB
improvement over method 2.  If you are in a hurry and have enough
signal to noise ratio use method 2.

On 4/23/2010 11:54 AM, brent wrote:
> On Apr 23, 9:53 am, "gretzteam"<gretzteam@n_o_s_p_a_m.yahoo.com>
> wrote:
>> Hi,
>> I'm having trouble understanding the similarities/differences/advantages of
>> various AM digital demodulation methods, which makes me incapable of
>> deciding what to use. I would like to go throug a few methods here and here
>> your thoughts.
>>
>> First, I assume the common goal of all those techniques is to shift the
>> carrier back down to DC and there are a few ways to do this that have
>> different behavior. Right?
>>
>> Method 1)
>> -Bandpass filter around the carrier.
>> -Take the absolute value. THIS is what shifts the carrier down to DC.
>> -Low-pass filter.
>> Now I guess taking the absolute value is not a very good way to shift the
>> carrier down to DC and is sensitive to anything other signal in the stream.
>> This is why we need to bandpass?
>> Is the performance of the system pretty much only dominated by the bandpass
>> filter? In other words, is taking the abs() value messing up my data EVEN
>> if there was only the carrier/data in the stream?
>>
>> Method 2)
>> -Multiply incoming stream with sin(wc*t) and cos(wc*t) sample by sample.
>> -Lowpass filter each output. This gives I and Q.
>> -Calculate sqrt(I^2+Q^2).
>>
>> Here, the carrier is perfectly shifted down to DC no matter what the stream
>> contains, which is why we don't need the bandpass filter. Can we say that
>> the performance of such a system is only dependent on the lowpass filter?
>> What about the precision of the sin(wc*t) and cos(wc*t)? Keeping only one
>> bit here would mean using a square wave, which can't be really good. Is
>> there any way to know how much this affects performance?
>>
>> And probably most importantly, how can I measure performance of the system?
>> Is there a standard way. I'm not yet looking for complicated theoretical
>> explanations...something intuitive is better to learn at first - at least
>> for me!
>>
>> Thanks!
>
> Method 2 is solid.  After implementing method 2 you can try to do
> tricks that may give you some improvements.  One trick I know of is to
> sychronize the carrier with a digital PLL.  This is tough, it has some
> other aspects to it that can go wrong and you get a maximum 3 dB
> improvement over method 2.  If you are in a hurry and have enough
> signal to noise ratio use method 2.

Use it, but understand it. Understand the implication of in-band 
interference. Understand the need to exclude out-of-band signals from 
the demodulation process. (The baseband low-pass filter can't remove 
aliases.) Replacing sinusoids with square waves adds harmonics that may 
or may not matter. There are some quick-and-dirty ways to calculate 
sqrt(I^2+Q^2) that might be good enough if some distortion is tolerable. 
See http://www.dspguru.com/dsp/tricks/magnitude-estimator.

Jerry
-- 
"It does me no injury for my neighbor to say there are 20 gods, or no
God. It neither picks my pocket nor breaks my leg."
          Thomas Jefferson to the Virginia House of Delegates in 1776.
���������������������������������������������������������������������

On Apr 23, 12:15=A0pm, Jerry Avins <j...@ieee.org> wrote:
> On 4/23/2010 11:54 AM, brent wrote:
>
>
>
>
>
> > On Apr 23, 9:53 am, "gretzteam"<gretzteam@n_o_s_p_a_m.yahoo.com>
> > wrote:
> >> Hi,
> >> I'm having trouble understanding the similarities/differences/advantag=
es of
> >> various AM digital demodulation methods, which makes me incapable of
> >> deciding what to use. I would like to go throug a few methods here and=
 here
> >> your thoughts.
>
> >> First, I assume the common goal of all those techniques is to shift th=
e
> >> carrier back down to DC and there are a few ways to do this that have
> >> different behavior. Right?
>
> >> Method 1)
> >> -Bandpass filter around the carrier.
> >> -Take the absolute value. THIS is what shifts the carrier down to DC.
> >> -Low-pass filter.
> >> Now I guess taking the absolute value is not a very good way to shift =
the
> >> carrier down to DC and is sensitive to anything other signal in the st=
ream.
> >> This is why we need to bandpass?
> >> Is the performance of the system pretty much only dominated by the ban=
dpass
> >> filter? In other words, is taking the abs() value messing up my data E=
VEN
> >> if there was only the carrier/data in the stream?
>
> >> Method 2)
> >> -Multiply incoming stream with sin(wc*t) and cos(wc*t) sample by sampl=
e.
> >> -Lowpass filter each output. This gives I and Q.
> >> -Calculate sqrt(I^2+Q^2).
>
> >> Here, the carrier is perfectly shifted down to DC no matter what the s=
tream
> >> contains, which is why we don't need the bandpass filter. Can we say t=
hat
> >> the performance of such a system is only dependent on the lowpass filt=
er?
> >> What about the precision of the sin(wc*t) and cos(wc*t)? Keeping only =
one
> >> bit here would mean using a square wave, which can't be really good. I=
s
> >> there any way to know how much this affects performance?
>
> >> And probably most importantly, how can I measure performance of the sy=
stem?
> >> Is there a standard way. I'm not yet looking for complicated theoretic=
al
> >> explanations...something intuitive is better to learn at first - at le=
ast
> >> for me!
>
> >> Thanks!
>
> > Method 2 is solid. =A0After implementing method 2 you can try to do
> > tricks that may give you some improvements. =A0One trick I know of is t=
o
> > sychronize the carrier with a digital PLL. =A0This is tough, it has som=
e
> > other aspects to it that can go wrong and you get a maximum 3 dB
> > improvement over method 2. =A0If you are in a hurry and have enough
> > signal to noise ratio use method 2.
>
> Use it, but understand it. Understand the implication of in-band
> interference. Understand the need to exclude out-of-band signals from
> the demodulation process. (The baseband low-pass filter can't remove
> aliases.) Replacing sinusoids with square waves adds harmonics that may
> or may not matter. There are some quick-and-dirty ways to calculate
> sqrt(I^2+Q^2) that might be good enough if some distortion is tolerable.
> Seehttp://www.dspguru.com/dsp/tricks/magnitude-estimator.
>
> Jerry
> --
> "It does me no injury for my neighbor to say there are 20 gods, or no
> God. It neither picks my pocket nor breaks my leg."
> =A0 =A0 =A0 =A0 =A0 Thomas Jefferson to the Virginia House of Delegates i=
n 1776.
> =EF=BF=BD=EF=BF=BD=EF=BF=BD=EF=BF=BD=EF=BF=BD=EF=BF=BD=EF=BF=BD=EF=BF=BD=
=EF=BF=BD=EF=BF=BD=EF=BF=BD=EF=BF=BD=EF=BF=BD=EF=BF=BD=EF=BF=BD=EF=BF=BD=EF=
=BF=BD=EF=BF=BD=EF=BF=BD=EF=BF=BD=EF=BF=BD=EF=BF=BD=EF=BF=BD=EF=BF=BD=EF=BF=
=BD=AD=EF=BF=BD=EF=BF=BD=EF=BF=BD=EF=BF=BD=EF=BF=BD=EF=BF=BD=EF=BF=BD=EF=BF=
=BD=EF=BF=BD=EF=BF=BD=EF=BF=BD=EF=BF=BD=EF=BF=BD=EF=BF=BD=EF=BF=BD=EF=BF=BD=
=EF=BF=BD=EF=BF=BD=EF=BF=BD=EF=BF=BD=EF=BF=BD=EF=BF=BD=EF=BF=BD=EF=BF=BD=EF=
=BF=BD=AD=EF=BF=BD=EF=BF=BD=EF=BF=BD=EF=BF=BD=EF=BF=BD=EF=BF=BD=EF=BF=BD=EF=
=BF=BD=EF=BF=BD=EF=BF=BD=EF=BF=BD=EF=BF=BD=EF=BF=BD=EF=BF=BD=EF=BF=BD=EF=BF=
=BD=EF=BF=BD=EF=BF=BD=EF=BF=BD- Hide quoted text -
>
> - Show quoted text -

Agreed.

I am  assuming that he is properly prepping the signal prior to the
multiplication by sin/cos and will pick appropriate filters at
baseband.

>> Use it, but understand it. Understand the implication of in-band
>> interference. Understand the need to exclude out-of-band signals from
>> the demodulation process. (The baseband low-pass filter can't remove
>> aliases.) 
>
>I am  assuming that he is properly prepping the signal prior to the
>multiplication by sin/cos and will pick appropriate filters at
>baseband.


Ok I must admit that I'm more confused than before! Why do you still need a
bandpass filter for method 2? Isn't multiplying by sin/cos shifting the
carrier frequency to DC? 

About method 1 having the problem of peak values not being close to full
scale, can we say that this is not a problem when fs >> carrier?

Thanks.

On Apr 23, 1:52=A0pm, "gretzteam" <gretzteam@n_o_s_p_a_m.yahoo.com>
wrote:
> >> Use it, but understand it. Understand the implication of in-band
> >> interference. Understand the need to exclude out-of-band signals from
> >> the demodulation process. (The baseband low-pass filter can't remove
> >> aliases.)
>
> >I am =A0assuming that he is properly prepping the signal prior to the
> >multiplication by sin/cos and will pick appropriate filters at
> >baseband.
>
> Ok I must admit that I'm more confused than before! Why do you still need=
 a
> bandpass filter for method 2? Isn't multiplying by sin/cos shifting the
> carrier frequency to DC?
>
I was talking about a bandpass filter prior to the mixing process to
assure that you do not get "out of band" "out of interest" signals
that alias into your baseband signal.  If your signals are clean ,
with no outside interference outside the band of interest, then you
can ignore the BPF comment.


> About method 1 having the problem of peak values not being close to full
> scale, can we say that this is not a problem when fs >> carrier?
>

Perhaps I do not fully understand your method 1, but if you are
fishing for peak values to construct an envelope , this is a real pain
in the butt.  If you build an I/Q signal at baseband, then you know
the envelope of the signal AT EVERY SINGLE SAMPLE by doing the I^2 +
Q^2 thing.  No fishing or interpreting or interpolating required.
> Thanks.

On 4/23/2010 1:52 PM, gretzteam wrote:
>>> Use it, but understand it. Understand the implication of in-band
>>> interference. Understand the need to exclude out-of-band signals from
>>> the demodulation process. (The baseband low-pass filter can't remove
>>> aliases.)
>>
>> I am  assuming that he is properly prepping the signal prior to the
>> multiplication by sin/cos and will pick appropriate filters at
>> baseband.
>
>
> Ok I must admit that I'm more confused than before! Why do you still need a
> bandpass filter for method 2? Isn't multiplying by sin/cos shifting the
> carrier frequency to DC?

What Brent said. Keep in mind that you not only shift the carrier to 
baseband, you also shift everything else down by a similar amount. Where 
do the aliases of the out-of-band signals go?

> About method 1 having the problem of peak values not being close to full
> scale, can we say that this is not a problem when fs>>  carrier?

When the carrier is adequately oversampled, method 1 works. I leave it 
to you to determine what "adequate" means. How many samples per carrier 
cycle are needed to ensure that one is at least 95% of either peak? Is 
that a reasonable expenditure of resources?

Jerry
-- 
Engineering is the art of making what you want from things you can get.
�����������������������������������������������������������������������

On Apr 23, 2:19=A0pm, Jerry Avins <j...@ieee.org> wrote:
> On 4/23/2010 1:52 PM, gretzteam wrote:
>
> >>> Use it, but understand it. Understand the implication of in-band
> >>> interference. Understand the need to exclude out-of-band signals from
> >>> the demodulation process. (The baseband low-pass filter can't remove
> >>> aliases.)
>
> >> I am =A0assuming that he is properly prepping the signal prior to the
> >> multiplication by sin/cos and will pick appropriate filters at
> >> baseband.
>
> > Ok I must admit that I'm more confused than before! Why do you still ne=
ed a
> > bandpass filter for method 2? Isn't multiplying by sin/cos shifting the
> > carrier frequency to DC?
>
> What Brent said. Keep in mind that you not only shift the carrier to
> baseband, you also shift everything else down by a similar amount. Where
> do the aliases of the out-of-band signals go?
>
> > About method 1 having the problem of peak values not being close to ful=
l
> > scale, can we say that this is not a problem when fs>> =A0carrier?
>
> When the carrier is adequately oversampled, method 1 works. I leave it
> to you to determine what "adequate" means. How many samples per carrier
> cycle are needed to ensure that one is at least 95% of either peak? Is
> that a reasonable expenditure of resources?
>
> Jerry
> --
> Engineering is the art of making what you want from things you can get.

This is a wonderful example of how a trivial method in analog (diode
and RC filter) is not really how you want to do it in digital. Now if
he has an analytic signal ....

Clay

On Apr 24, 1:53=A0am, "gretzteam" <gretzteam@n_o_s_p_a_m.yahoo.com>
wrote:
> Hi,
> I'm having trouble understanding the similarities/differences/advantages =
of
> various AM digital demodulation methods, which makes me incapable of
> deciding what to use. I would like to go throug a few methods here and he=
re
> your thoughts.
>
> First, I assume the common goal of all those techniques is to shift the
> carrier back down to DC and there are a few ways to do this that have
> different behavior. Right?
>
> Method 1)
> -Bandpass filter around the carrier.
> -Take the absolute value. THIS is what shifts the carrier down to DC.
> -Low-pass filter.
> Now I guess taking the absolute value is not a very good way to shift the
> carrier down to DC and is sensitive to anything other signal in the strea=
m.
> This is why we need to bandpass?
> Is the performance of the system pretty much only dominated by the bandpa=
ss
> filter? In other words, is taking the abs() value messing up my data EVEN
> if there was only the carrier/data in the stream?
>
> Method 2)
> -Multiply incoming stream with sin(wc*t) and cos(wc*t) sample by sample.
> -Lowpass filter each output. This gives I and Q.
> -Calculate sqrt(I^2+Q^2).
>
> Here, the carrier is perfectly shifted down to DC no matter what the stre=
am
> contains, which is why we don't need the bandpass filter. Can we say that
> the performance of such a system is only dependent on the lowpass filter?
> What about the precision of the sin(wc*t) and cos(wc*t)? Keeping only one
> bit here would mean using a square wave, which can't be really good. Is
> there any way to know how much this affects performance?
>
> And probably most importantly, how can I measure performance of the syste=
m?
> Is there a standard way. I'm not yet looking for complicated theoretical
> explanations...something intuitive is better to learn at first - at least
> for me!
>
> Thanks!

Synchronous demodulation using a PLL will give you 3dB improvement
over ordinary envelope detection.
The problem arrises when you want to do synchronous demod and the
carrier isn't there! What I mean by that is that when you have
double sideband supressed carrier. There is no power at the carrier
freq then and nothing to lock onto.

Solution...among otehr things you need to square the received waveform
and lock into twice the carrier then divide down (missing some other
crucial steps).


Hardy

Clay <clay@claysturner.com> wrote:
(snip)
 
> This is a wonderful example of how a trivial method in analog (diode
> and RC filter) is not really how you want to do it in digital. Now if
> he has an analytic signal ....

I was almost going to mention that square wave demodulation is
commonly used for the usual laboratory "lock-in amplifier", but
again that is analog.  The higher harmonics are normally just
filtered out, but maybe not in the digital world.

-- glen

Jerry Avins wrote:
> On 4/23/2010 1:52 PM, gretzteam wrote:
>>>> Use it, but understand it. Understand the implication of in-band
>>>> interference. Understand the need to exclude out-of-band signals from
>>>> the demodulation process. (The baseband low-pass filter can't remove
>>>> aliases.)
>>>
>>> I am  assuming that he is properly prepping the signal prior to the
>>> multiplication by sin/cos and will pick appropriate filters at
>>> baseband.
>>
>>
>> Ok I must admit that I'm more confused than before! Why do you still 
>> need a
>> bandpass filter for method 2? Isn't multiplying by sin/cos shifting the
>> carrier frequency to DC?
> 
> What Brent said. Keep in mind that you not only shift the carrier to 
> baseband, you also shift everything else down by a similar amount. Where 
> do the aliases of the out-of-band signals go?
> 
>> About method 1 having the problem of peak values not being close to full
>> scale, can we say that this is not a problem when fs>>  carrier?
> 
> When the carrier is adequately oversampled, method 1 works. I leave it 
> to you to determine what "adequate" means. How many samples per carrier 
> cycle are needed to ensure that one is at least 95% of either peak? Is 
> that a reasonable expenditure of resources?

Except that by his original description he's not peak-seeking -- he's 
averaging the absolute value.  That _ought_ to work better, but I don't 
know by how much.

-- 
Tim Wescott
Control system and signal processing consulting
www.wescottdesign.com

HardySpicer wrote:


> Synchronous demodulation using a PLL will give you 3dB improvement
> over ordinary envelope detection.

This is wrong.

Synchronous demodulation makes improvement from 0dB to infinity, 
depending on SNR.

> The problem arrises when you want to do synchronous demod and the
> carrier isn't there!
> What I mean by that is that when you have
> double sideband supressed carrier. There is no power at the carrier
> freq then and nothing to lock onto.
> 
> Solution...among otehr things you need to square the received waveform
> and lock into twice the carrier then divide down (missing some other
> crucial steps).

Let the dumb lead the blind.


Vladimir Vassilevsky
DSP and Mixed Signal Design Consultant
http://www.abvolt.com

Clay wrote:


> This is a wonderful example of how a trivial method in analog (diode
> and RC filter) is not really how you want to do it in digital. Now if
> he has an analytic signal ....

Except for the diode detector doesn't work like the simplified textbook 
desccription. An accurate analysis requires involved math.


Vladimir Vassilevsky
DSP and Mixed Signal Design Consultant
http://www.abvolt.com

>Synchronous demodulation using a PLL will give you 3dB improvement
>over ordinary envelope detection.

3dB improvement of what? I don't quite understand how to qualify such a
system. I assume you mean SNR, but how does it get measured?


>What Brent said. Keep in mind that you not only shift the carrier to 
>baseband, you also shift everything else down by a similar amount. Where 
>do the aliases of the out-of-band signals go?

Ok this is the part I don't understand! Can you elaborate a bit more?

Thanks a lot!

On 4/23/2010 4:24 PM, Tim Wescott wrote:
> Jerry Avins wrote:
>> On 4/23/2010 1:52 PM, gretzteam wrote:
>>>>> Use it, but understand it. Understand the implication of in-band
>>>>> interference. Understand the need to exclude out-of-band signals from
>>>>> the demodulation process. (The baseband low-pass filter can't remove
>>>>> aliases.)
>>>>
>>>> I am assuming that he is properly prepping the signal prior to the
>>>> multiplication by sin/cos and will pick appropriate filters at
>>>> baseband.
>>>
>>>
>>> Ok I must admit that I'm more confused than before! Why do you still
>>> need a
>>> bandpass filter for method 2? Isn't multiplying by sin/cos shifting the
>>> carrier frequency to DC?
>>
>> What Brent said. Keep in mind that you not only shift the carrier to
>> baseband, you also shift everything else down by a similar amount.
>> Where do the aliases of the out-of-band signals go?
>>
>>> About method 1 having the problem of peak values not being close to full
>>> scale, can we say that this is not a problem when fs>> carrier?
>>
>> When the carrier is adequately oversampled, method 1 works. I leave it
>> to you to determine what "adequate" means. How many samples per
>> carrier cycle are needed to ensure that one is at least 95% of either
>> peak? Is that a reasonable expenditure of resources?
>
> Except that by his original description he's not peak-seeking -- he's
> averaging the absolute value. That _ought_ to work better, but I don't
> know by how much.

How does the average value of samples of the carrier relate to the 
approximate value of the envelope? Would it help if the "carrier" were 
triangular?

Jerry
-- 
"It does me no injury for my neighbor to say there are 20 gods, or no
God. It neither picks my pocket nor breaks my leg."
          Thomas Jefferson to the Virginia House of Delegates in 1776.
���������������������������������������������������������������������

On 4/23/2010 4:32 PM, Vladimir Vassilevsky wrote:
>
>
> Clay wrote:
>
>
>> This is a wonderful example of how a trivial method in analog (diode
>> and RC filter) is not really how you want to do it in digital. Now if
>> he has an analytic signal ....
>
> Except for the diode detector doesn't work like the simplified textbook
> desccription. An accurate analysis requires involved math.

Except for diagonal clipping with deep modulation and the limited 
charging current, what is missing from the classical analysis?

Whenever it was in my control, I used a full-wave peak detector. That 
suppresses the IF in the output, leaving the weaker second harmonic to 
dominate. You then need only a much smaller capacitor, making the onset 
of diagonal clipping more remote. Besides, 456 KHz can get through many 
Hi-Fi power amps, but the response at 902 is much less.

Jerry
-- 
"It does me no injury for my neighbor to say there are 20 gods, or no
God. It neither picks my pocket nor breaks my leg."
          Thomas Jefferson to the Virginia House of Delegates in 1776.
���������������������������������������������������������������������

>
>
>How does the average value of samples of the carrier relate to the 
>approximate value of the envelope? Would it help if the "carrier" were 
>triangular?
>

Yes you have a point here! All I've proven so far is that when the input
signal contains only a carrier, full scale, then the output of the lowpass
filter is pretty much exactly 0.63 (2/pi), which is the average value of a
full scale sine wave. 

I was pretty happy to see this, but that's probably not AM demodulation
just yet! But isn't this what the Analog version does when using bandpass,
full wave rectifier and capacitor?

Thanks

Jerry Avins wrote:

> On 4/23/2010 4:32 PM, Vladimir Vassilevsky wrote:
> 
>>
>>
>> Clay wrote:
>>
>>
>>> This is a wonderful example of how a trivial method in analog (diode
>>> and RC filter) is not really how you want to do it in digital. Now if
>>> he has an analytic signal ....
>>
>>
>> Except for the diode detector doesn't work like the simplified textbook
>> desccription. An accurate analysis requires involved math.
> 
> Except for diagonal clipping with deep modulation and the limited 
> charging current, what is missing from the classical analysis?

Even with ideal diode and ideal source, the angle of conduction is 
determined by transcendental equation. Now account for impedance, 
nonlinearity and add noise, and it gets really messy.

> Whenever it was in my control, I used a full-wave peak detector. That 
> suppresses the IF in the output, leaving the weaker second harmonic to 
> dominate. You then need only a much smaller capacitor, making the onset 
> of diagonal clipping more remote. Besides, 456 KHz can get through many 
> Hi-Fi power amps, but the response at 902 is much less.

Jerry, do you know what was the rationale for choosing 455kHz vs 465kHz 
standard IF ?

VLV

On 4/23/2010 4:33 PM, gretzteam wrote:

I wrote:
>> What Brent said. Keep in mind that you not only shift the carrier to
>> baseband, you also shift everything else down by a similar amount. Where
>> do the aliases of the out-of-band signals go?
>
> Ok this is the part I don't understand! Can you elaborate a bit more?

Rarely is a carrier to be demodulated sampled at more than twice the 
carrier frequency; that would be a waste. The sampling theorem tells us 
that we have to sample more than twice the frequency corresponding to 
the bandwidth of interest. There are some practical restrictions (the 
second edition of Understanding Digital Signal Processing by Rick Lyons 
has an excellent analysis of them) but in general, sampling 20 KHz wide 
signal on a 356 KHz carrier can be accomplished with a 50 KHz sample 
rate. A bandpass filter assures that the AM signal is not contaminated 
by adjacent channels. With bandpass sampling, we need to exclude signals 
both above and below the band being sampled. Even if the signal were 
sampled at 1 MHz, a low-pass filter would be needed to substantially 
eliminate all signals above 500 KHz.

Incidentally, sampling at 1 MHz provides a little over 2 samples per 
carrier cycle, with little chance that either of them will be near a 
carrier peak and hence representative of the envelope. With bandpass 
sampling at 50 KHz, there will be only one sample for every 9 or so 
carrier cycles. There is then no hope of peak detection.

Jerry
-- 
"It does me no injury for my neighbor to say there are 20 gods, or no
God. It neither picks my pocket nor breaks my leg."
          Thomas Jefferson to the Virginia House of Delegates in 1776.
���������������������������������������������������������������������

On 4/23/2010 5:05 PM, Vladimir Vassilevsky wrote:

   ...

> Jerry, do you know what was the rationale for choosing 455kHz vs 465kHz
> standard IF ?

The AM band is roughly 560 to 1600 KHz, a nearly 3:1 range. A little 
thought reveals that the LO must be above the signal in order avoid a 
band switch in the receiver. Someone did an image analysis to pick the 
IF, and 455 came out. I don't know the calculation or if it made sense. 
I remember 455, 456, even 460. Standard IF transformers locked in 45x. 
It couldn't be very different. It needs to be below the AM band itself 
and not so low that the selectivity becomes too high. Also, one doesn't 
want many harmonics that fall into the band.

If my cousin Jack were still alive, I could ask him. Among his patents 
are the VTVM and the ratio detector. http://tinyurl.com/2g8koyp

Jerry
-- 
"It does me no injury for my neighbor to say there are 20 gods, or no
God. It neither picks my pocket nor breaks my leg."
          Thomas Jefferson to the Virginia House of Delegates in 1776.
���������������������������������������������������������������������

On Apr 24, 8:24=A0am, Vladimir Vassilevsky <nos...@nowhere.com> wrote:
> HardySpicer wrote:
> > Synchronous demodulation using a PLL will give you 3dB improvement
> > over ordinary envelope detection.
>
> This is wrong.
>
It's in the textbooks...read it!

On 4/23/2010 5:03 PM, gretzteam wrote:
>>
>>
>> How does the average value of samples of the carrier relate to the
>> approximate value of the envelope? Would it help if the "carrier" were
>> triangular?
>>
>
> Yes you have a point here! All I've proven so far is that when the input
> signal contains only a carrier, full scale, then the output of the lowpass
> filter is pretty much exactly 0.63 (2/pi), which is the average value of a
> full scale sine wave.

How many samples per carrier cycle do you have? How many carrier cycles 
do you average over? How long does that take, and what does that imply 
about the highest envelope frequency you can demodulate without attenuation?

> I was pretty happy to see this, but that's probably not AM demodulation
> just yet! But isn't this what the Analog version does when using bandpass,
> full wave rectifier and capacitor?

The analog version is essentially the same as the digital version, the 
salient difference being an effectively infinite sample rate. Peak 
detection works just fine with that degree of oversampling!

Jerry
-- 
"I view the progress of science as ... the slow erosion of the tendency
  to dichotomize." --Barbara Smuts, U. Mich.
�����������������������������������������������������������������������

On 4/23/2010 5:33 PM, HardySpicer wrote:
> On Apr 24, 8:24 am, Vladimir Vassilevsky<nos...@nowhere.com>  wrote:
>> HardySpicer wrote:
>>> Synchronous demodulation using a PLL will give you 3dB improvement
>>> over ordinary envelope detection.
>>
>> This is wrong.
>>
> It's in the textbooks...read it!

What is ordinary envelope detection? Peak detection?

Jerry
-- 
"I view the progress of science as ... the slow erosion of the tendency
  to dichotomize." --Barbara Smuts, U. Mich.
�����������������������������������������������������������������������

On Apr 23, 4:33=A0pm, "gretzteam" <gretzteam@n_o_s_p_a_m.yahoo.com>
wrote:
> >Synchronous demodulation using a PLL will give you 3dB improvement
> >over ordinary envelope detection.
>
> 3dB improvement of what? I don't quite understand how to qualify such a
> system. I assume you mean SNR, but how does it get measured?
>
> >What Brent said. Keep in mind that you not only shift the carrier to
> >baseband, you also shift everything else down by a similar amount. Where
> >do the aliases of the out-of-band signals go?
>
> Ok this is the part I don't understand! Can you elaborate a bit more?
>
> Thanks a lot!

I will explain, as long as you take your assigned role as "blind". I
will take my assigned role as "dumb".

I was going to try to explain a little further, but I suggest you
actually try to mess around a bit with scilab or matlab if you have it
and see what you can generate.

If your signals do not have any noise or outside interference, then
just do the mixing process and don't worry about it.  If you are
collecting data from a noisy environment and there is embedded energy
in your signal that is outside the frequency of your carrier, then you
need to think about a Band Pass Filter and think about aliasing.

HardySpicer <gyansorova@gmail.com> wrote:
(snip)
 
> The problem arrises when you want to do synchronous demod and the
> carrier isn't there! What I mean by that is that when you have
> double sideband supressed carrier. There is no power at the carrier
> freq then and nothing to lock onto.
 
> Solution...among otehr things you need to square the received waveform
> and lock into twice the carrier then divide down (missing some other
> crucial steps).

As I understand it, commonly used by many modems doing anything
except FSK.  To make sure that there are enough transitions for
carrier recovery, scramblers are commonly used on the bit stream.

-- glen

HardySpicer wrote:

> On Apr 24, 8:24 am, Vladimir Vassilevsky <nos...@nowhere.com> wrote:
> 
>>HardySpicer wrote:
>>
>>>Synchronous demodulation using a PLL will give you 3dB improvement
>>>over ordinary envelope detection.
>>
>>This is wrong.
>>
> 
> It's in the textbooks...read it!

"Making errors is a way of a man. Insisting on errors is a way of an idiot."

Who said that?


Vladimir Vassilevsky
DSP and Mixed Signal Design Consultant
http://www.abvolt.com

Jerry Avins wrote:

> On 4/23/2010 5:33 PM, HardySpicer wrote:
> 
>> On Apr 24, 8:24 am, Vladimir Vassilevsky<nos...@nowhere.com>  wrote:
>>
>>> HardySpicer wrote:
>>>
>>>> Synchronous demodulation using a PLL will give you 3dB improvement
>>>> over ordinary envelope detection.
>>>
>>>
>>> This is wrong.
>>>
>> It's in the textbooks...read it!
> 
> 
> What is ordinary envelope detection? Peak detection?

Doesn't matter; It is very simple. Think of |I| vs sqrt(I^2 + Q^2)

VLV

>On 4/23/2010 5:03 PM, gretzteam wrote:
>>>
>>>
>>> How does the average value of samples of the carrier relate to the
>>> approximate value of the envelope? Would it help if the "carrier" were
>>> triangular?
>>>
>>
>> Yes you have a point here! All I've proven so far is that when the
input
>> signal contains only a carrier, full scale, then the output of the
lowpass
>> filter is pretty much exactly 0.63 (2/pi), which is the average value of
a
>> full scale sine wave.
>
>How many samples per carrier cycle do you have? How many carrier cycles 
>do you average over? How long does that take, and what does that imply 
>about the highest envelope frequency you can demodulate without
attenuation?

Ok I'm way oversampled. I'm doing this to learn about it so I don't want to
have the added difficulty of sample rate (just yet). Here is the current
system - I should have posted this FIRST! 

parameters:
fs = 4MHz
carrier: 99kHz

Currently, there is no noise, and no 'information' being modulated. Just a
carrier sine wave:) One gotta start somewhere!

I then bandpass using a 2nd order bandpass filter centered at the carrier.

Then take the absolute value.
Then lowpass filter using a 2nd order CIC filter all the way down to
something ridiculous like 10-50Hz. 

The output matches surprisingly well the 2*A/pi formula depending on the A
of the carrier.

Now if I do a frequency sweep, using a full scale sine wave from 0 to 2MHz
(fs/2), and plot the obtained average value after it settled, I get the
shape of the bandpass filter! I guess this was to be expected, which is why
I asked if 'method-1' was only dependent on the performance of the bandpass
filter.

I guess so far my 'information' is only at DC, but it works well. 

Now, ff I modulate a 2Hz signal, I also see it at the output, with a DC
offset. 

And this is where I decided to post, since I didn't know how to measure
performance of the system when there IS information.

Does this make sense?

Thanks!

gretzteam wrote:
>>
>> How does the average value of samples of the carrier relate to the 
>> approximate value of the envelope? Would it help if the "carrier" were 
>> triangular?
>>
> 
> Yes you have a point here! All I've proven so far is that when the input
> signal contains only a carrier, full scale, then the output of the lowpass
> filter is pretty much exactly 0.63 (2/pi), which is the average value of a
> full scale sine wave. 
> 
> I was pretty happy to see this, but that's probably not AM demodulation
> just yet! But isn't this what the Analog version does when using bandpass,
> full wave rectifier and capacitor?

Some analog implementations rectify and average, which is what you're doing.

Some (the ones that have a series diode to a shunt cap) detect the peak 
of the waveform, then have the charge drained out of the cap by other 
circuit elements.  I.e. a peak detector.

Different things, but close enough when the carrier frequency is way 
higher than the audio.

-- 
Tim Wescott
Control system and signal processing consulting
www.wescottdesign.com

Jerry Avins wrote:
> On 4/23/2010 4:24 PM, Tim Wescott wrote:
>> Jerry Avins wrote:
>>> On 4/23/2010 1:52 PM, gretzteam wrote:
>>>>>> Use it, but understand it. Understand the implication of in-band
>>>>>> interference. Understand the need to exclude out-of-band signals from
>>>>>> the demodulation process. (The baseband low-pass filter can't remove
>>>>>> aliases.)
>>>>>
>>>>> I am assuming that he is properly prepping the signal prior to the
>>>>> multiplication by sin/cos and will pick appropriate filters at
>>>>> baseband.
>>>>
>>>>
>>>> Ok I must admit that I'm more confused than before! Why do you still
>>>> need a
>>>> bandpass filter for method 2? Isn't multiplying by sin/cos shifting the
>>>> carrier frequency to DC?
>>>
>>> What Brent said. Keep in mind that you not only shift the carrier to
>>> baseband, you also shift everything else down by a similar amount.
>>> Where do the aliases of the out-of-band signals go?
>>>
>>>> About method 1 having the problem of peak values not being close to 
>>>> full
>>>> scale, can we say that this is not a problem when fs>> carrier?
>>>
>>> When the carrier is adequately oversampled, method 1 works. I leave it
>>> to you to determine what "adequate" means. How many samples per
>>> carrier cycle are needed to ensure that one is at least 95% of either
>>> peak? Is that a reasonable expenditure of resources?
>>
>> Except that by his original description he's not peak-seeking -- he's
>> averaging the absolute value. That _ought_ to work better, but I don't
>> know by how much.
> 
> How does the average value of samples of the carrier relate to the 
> approximate value of the envelope? Would it help if the "carrier" were 
> triangular?

Well, the RF signal (not the carrier) is carrier * (audio signal + 
offset) -- so you can find a scaled value of the audio signal either by 
finding the peaks (as in traditional AM receivers) or by rectifying and 
averaging.

I suspect (but would have to play with it to find out) that the rectify 
and average is not as harshly nonlinear, and therefore would stand a 
lower sampling (or carrier) rate.

-- 
Tim Wescott
Control system and signal processing consulting
www.wescottdesign.com

Vladimir Vassilevsky <nospam@nowhere.com> wrote:
(snip)
 
> Even with ideal diode and ideal source, the angle of conduction is 
> determined by transcendental equation. Now account for impedance, 
> nonlinearity and add noise, and it gets really messy.
 
(snip)
> Jerry, do you know what was the rationale for choosing 455kHz 
> vs 465kHz standard IF ?

Interesting question.  I don't see anything in wikipedia that
hints as an answer.  It seems obvious that it should be a multiple
of 5kHz, and presumably set such that image frequencies don't come
out in places you don't want them.  

Otherwise, my guess is that once one was chosen, and the appropriate
parts started to get popular, there would be incentive to keep
using the same frequency.  That might be enough reason.

I am interested to see what Jerry has to say.

-- glen

Jerry Avins <jya@ieee.org> wrote:
> On 4/23/2010 5:05 PM, Vladimir Vassilevsky wrote:
 
>> Jerry, do you know what was the rationale for choosing 
>> 455kHz vs 465kHz standard IF ?
> 
> The AM band is roughly 560 to 1600 KHz, a nearly 3:1 range. A little 
> thought reveals that the LO must be above the signal in order avoid a 
> band switch in the receiver. Someone did an image analysis to pick the 
> IF, and 455 came out. I don't know the calculation or if it made sense. 
> I remember 455, 456, even 460. 

I was thinking a few minutes ago that it was obvious it should
be an odd multiple of 5kHz, but maybe not so obvious.  It is
convenient in that the image frequency is also an odd multiple
of 5kHz, and so between stations that are multiples of 10kHz.

I used to know about the rules for TV channel assignment, partly 
based on where the image frequencies of other channels come out.
(Especially for UHF.)  

It does seem that there are AM radio stations at 910kHz.

> Standard IF transformers locked in 45x. 
> It couldn't be very different. It needs to be below the AM band itself 
> and not so low that the selectivity becomes too high. Also, one doesn't 
> want many harmonics that fall into the band.
 
> If my cousin Jack were still alive, I could ask him. Among his patents 
> are the VTVM and the ratio detector. http://tinyurl.com/2g8koyp

I still have my Heathkit VTVM that I built many years ago.
(Christmas present from my dad.)  I don't use it so much,
though, as I have a nice Fluke handheld.  Sometimes that analog
readout is useful, though.

-- glen

On Apr 23, 7:59=A0pm, glen herrmannsfeldt <g...@ugcs.caltech.edu> wrote:
> Vladimir Vassilevsky <nos...@nowhere.com> wrote:
>
> (snip)
>
> > Even with ideal diode and ideal source, the angle of conduction is
> > determined by transcendental equation. Now account for impedance,
> > nonlinearity and add noise, and it gets really messy.
>
> (snip)
>
> > Jerry, do you know what was the rationale for choosing 455kHz
> > vs 465kHz standard IF ?
>
> Interesting question. =A0I don't see anything in wikipedia that
> hints as an answer. =A0It seems obvious that it should be a multiple
> of 5kHz, and presumably set such that image frequencies don't come
> out in places you don't want them. =A0
>
> Otherwise, my guess is that once one was chosen, and the appropriate
> parts started to get popular, there would be incentive to keep
> using the same frequency. =A0That might be enough reason.
>
> I am interested to see what Jerry has to say.
>
> -- glen

I am not sure that a multiple of 5KHz was required.  the AM band is 10
KHz wide and when AM was defined it was in the context of a continuous
tunable oscillator.  Also, If I am not mistaken, Europe has a 9 KHz
channel spacing.

Jerry Avins <jya@ieee.org> wrote:
(snip)
 
> Rarely is a carrier to be demodulated sampled at more than twice the 
> carrier frequency; that would be a waste. The sampling theorem tells us 
> that we have to sample more than twice the frequency corresponding to 
> the bandwidth of interest. There are some practical restrictions (the 
> second edition of Understanding Digital Signal Processing by Rick Lyons 
> has an excellent analysis of them) but in general, sampling 20 KHz wide 
> signal on a 356 KHz carrier can be accomplished with a 50 KHz sample 
> rate. A bandpass filter assures that the AM signal is not contaminated 
> by adjacent channels. 

OK, but say one wants to minimize the analog circuitry, and fast
digital circuitry is available, including a fast ADC.  That would
seem to go against the analog bandpass filter, but a lot of digital
filtering after the ADC could be provided.

> With bandpass sampling, we need to exclude signals 
> both above and below the band being sampled. Even if the signal were 
> sampled at 1 MHz, a low-pass filter would be needed to substantially 
> eliminate all signals above 500 KHz.

But 1MHz is pretty slow for an ADC by now, isn't it?
 
> Incidentally, sampling at 1 MHz provides a little over 2 samples per 
> carrier cycle, with little chance that either of them will be near a 
> carrier peak and hence representative of the envelope. With bandpass 
> sampling at 50 KHz, there will be only one sample for every 9 or so 
> carrier cycles. There is then no hope of peak detection.

Even a small FPGA should allow for a lot of digital filtering
that can run at 100MHz or so.  

-- glen

On 4/23/2010 6:09 PM, gretzteam wrote:
>> On 4/23/2010 5:03 PM, gretzteam wrote:
>>>>
>>>>
>>>> How does the average value of samples of the carrier relate to the
>>>> approximate value of the envelope? Would it help if the "carrier" were
>>>> triangular?
>>>>
>>>
>>> Yes you have a point here! All I've proven so far is that when the
> input
>>> signal contains only a carrier, full scale, then the output of the
> lowpass
>>> filter is pretty much exactly 0.63 (2/pi), which is the average value of
> a
>>> full scale sine wave.
>>
>> How many samples per carrier cycle do you have? How many carrier cycles
>> do you average over? How long does that take, and what does that imply
>> about the highest envelope frequency you can demodulate without
> attenuation?
>
> Ok I'm way oversampled. I'm doing this to learn about it so I don't want to
> have the added difficulty of sample rate (just yet). Here is the current
> system - I should have posted this FIRST!
>
> parameters:
> fs = 4MHz
> carrier: 99kHz
>
> Currently, there is no noise, and no 'information' being modulated. Just a
> carrier sine wave:) One gotta start somewhere!
>
> I then bandpass using a 2nd order bandpass filter centered at the carrier.

You realize that there's nothing for the bandpass filter to remove.

> Then take the absolute value.
> Then lowpass filter using a 2nd order CIC filter all the way down to
> something ridiculous like 10-50Hz.

You learned that the numbers are correctly manipulated, but little more.

> The output matches surprisingly well the 2*A/pi formula depending on the A
> of the carrier.

It doesn't surprise me. The effective sample rate is even higher than it 
might seem. Successive carrier cycles have their sampled points in 
slightly different places along the curve. You get information about 
every point along the curve (well, almost) by overlaying enough cycles.

> Now if I do a frequency sweep, using a full scale sine wave from 0 to 2MHz
> (fs/2), and plot the obtained average value after it settled, I get the
> shape of the bandpass filter! I guess this was to be expected, which is why
> I asked if 'method-1' was only dependent on the performance of the bandpass
> filter.

The sweep probes the bandpass filter. Overall output falls when the 
filter attenuates the signal. How could it be otherwise?

> I guess so far my 'information' is only at DC, but it works well.
>
> Now, ff I modulate a 2Hz signal, I also see it at the output, with a DC
> offset.
>
> And this is where I decided to post, since I didn't know how to measure
> performance of the system when there IS information.
>
> Does this make sense?

It makes sense in that it fits together in an understandable way. It 
won't get you to a practical demodulator without a lot of tweaking.

Jerry
-- 
"I view the progress of science as ... the slow erosion of the tendency
  to dichotomize." --Barbara Smuts, U. Mich.
�����������������������������������������������������������������������

On 4/23/2010 7:23 PM, Tim Wescott wrote:
> Jerry Avins wrote:
>> On 4/23/2010 4:24 PM, Tim Wescott wrote:
>>> Jerry Avins wrote:
>>>> On 4/23/2010 1:52 PM, gretzteam wrote:
>>>>>>> Use it, but understand it. Understand the implication of in-band
>>>>>>> interference. Understand the need to exclude out-of-band signals
>>>>>>> from
>>>>>>> the demodulation process. (The baseband low-pass filter can't remove
>>>>>>> aliases.)
>>>>>>
>>>>>> I am assuming that he is properly prepping the signal prior to the
>>>>>> multiplication by sin/cos and will pick appropriate filters at
>>>>>> baseband.
>>>>>
>>>>>
>>>>> Ok I must admit that I'm more confused than before! Why do you still
>>>>> need a
>>>>> bandpass filter for method 2? Isn't multiplying by sin/cos shifting
>>>>> the
>>>>> carrier frequency to DC?
>>>>
>>>> What Brent said. Keep in mind that you not only shift the carrier to
>>>> baseband, you also shift everything else down by a similar amount.
>>>> Where do the aliases of the out-of-band signals go?
>>>>
>>>>> About method 1 having the problem of peak values not being close to
>>>>> full
>>>>> scale, can we say that this is not a problem when fs>> carrier?
>>>>
>>>> When the carrier is adequately oversampled, method 1 works. I leave it
>>>> to you to determine what "adequate" means. How many samples per
>>>> carrier cycle are needed to ensure that one is at least 95% of either
>>>> peak? Is that a reasonable expenditure of resources?
>>>
>>> Except that by his original description he's not peak-seeking -- he's
>>> averaging the absolute value. That _ought_ to work better, but I don't
>>> know by how much.
>>
>> How does the average value of samples of the carrier relate to the
>> approximate value of the envelope? Would it help if the "carrier" were
>> triangular?
>
> Well, the RF signal (not the carrier) is carrier * (audio signal +
> offset) -- so you can find a scaled value of the audio signal either by
> finding the peaks (as in traditional AM receivers) or by rectifying and
> averaging.
>
> I suspect (but would have to play with it to find out) that the rectify
> and average is not as harshly nonlinear, and therefore would stand a
> lower sampling (or carrier) rate.

Even so, it takes more oomph to adequately upsample for averaging than 
to generate an analytic signal. Unless there is a frequency lock between 
the carrier and the sample clock, rectifying effectively doubles the 
chance of finding a point near the carrier peak and also doubles the 
number of points averaged over. That's a wash.

Jerry
-- 
"I view the progress of science as ... the slow erosion of the tendency
  to dichotomize." --Barbara Smuts, U. Mich.
�����������������������������������������������������������������������

On 4/23/2010 5:56 PM, Vladimir Vassilevsky wrote:
>
>
> Jerry Avins wrote:
>
>> On 4/23/2010 5:33 PM, HardySpicer wrote:
>>
>>> On Apr 24, 8:24 am, Vladimir Vassilevsky<nos...@nowhere.com> wrote:
>>>
>>>> HardySpicer wrote:
>>>>
>>>>> Synchronous demodulation using a PLL will give you 3dB improvement
>>>>> over ordinary envelope detection.
>>>>
>>>>
>>>> This is wrong.
>>>>
>>> It's in the textbooks...read it!
>>
>>
>> What is ordinary envelope detection? Peak detection?
>
> Doesn't matter; It is very simple. Think of |I| vs sqrt(I^2 + Q^2)

Still, when someone claims "3dB improvement", I want to know what is 
improved upon.

Jerry
-- 
"I view the progress of science as ... the slow erosion of the tendency
  to dichotomize." --Barbara Smuts, U. Mich.
�����������������������������������������������������������������������

On 4/23/2010 8:27 PM, glen herrmannsfeldt wrote:
> Jerry Avins<jya@ieee.org>  wrote:
> (snip)
>
>> Rarely is a carrier to be demodulated sampled at more than twice the
>> carrier frequency; that would be a waste. The sampling theorem tells us
>> that we have to sample more than twice the frequency corresponding to
>> the bandwidth of interest. There are some practical restrictions (the
>> second edition of Understanding Digital Signal Processing by Rick Lyons
>> has an excellent analysis of them) but in general, sampling 20 KHz wide
>> signal on a 356 KHz carrier can be accomplished with a 50 KHz sample
>> rate. A bandpass filter assures that the AM signal is not contaminated
>> by adjacent channels.
>
> OK, but say one wants to minimize the analog circuitry, and fast
> digital circuitry is available, including a fast ADC.  That would
> seem to go against the analog bandpass filter, but a lot of digital
> filtering after the ADC could be provided.

I'll rephrase this:
>> With bandpass sampling, we need to exclude signals
>> both above and below the band being sampled. Even if the signal were
>> sampled at 1 MHz, a low-pass filter would be needed to substantially
>> eliminate all signals above 500 KHz.

With baseband sampling, we need an anti-alias lowpass filter. With 
bandpass sampling, we need an anti-alias bandpass filter.

> But 1MHz is pretty slow for an ADC by now, isn't it?

Sure, but needing all those samples means that you need the MIPS (and 
the watts) to process them. Why do that?

>> Incidentally, sampling at 1 MHz provides a little over 2 samples per
>> carrier cycle, with little chance that either of them will be near a
>> carrier peak and hence representative of the envelope. With bandpass
>> sampling at 50 KHz, there will be only one sample for every 9 or so
>> carrier cycles. There is then no hope of peak detection.
>
> Even a small FPGA should allow for a lot of digital filtering
> that can run at 100MHz or so.

I think the trade-offs between analog and digital selectivity keep 
changing as technology advances. What sample rate would be needed to 
digitize the whole AM band? Should we build receivers that way?

Jerry
-- 
"I view the progress of science as ... the slow erosion of the tendency
  to dichotomize." --Barbara Smuts, U. Mich.
�����������������������������������������������������������������������

On Fri, 23 Apr 2010 16:03:32 -0500, gretzteam <gretzteam@n_o_s_p_a_m.yahoo.com> wrote:
>>
>>
>>How does the average value of samples of the carrier relate to the 
>>approximate value of the envelope? Would it help if the "carrier" were 
>>triangular?
>>
> 
> Yes you have a point here! All I've proven so far is that when the input
> signal contains only a carrier, full scale, then the output of the lowpass
> filter is pretty much exactly 0.63 (2/pi), which is the average value of a
> full scale sine wave. 
> 
> I was pretty happy to see this, but that's probably not AM demodulation
> just yet! But isn't this what the Analog version does when using bandpass,
> full wave rectifier and capacitor?

The size of the capacitor (a.k.a. lowpass filter) matters.

Put a 100pF capacitor across your signal and you filter out stuff you
can't hear (and presumably, in this context, don't care about). It
won't have a noticeable effect on audio frequencies. (But you should
check to see what impedance 100pF represents at your carrier and
modulating frequencies.)

Put a 1000uF capacitor across your signal hand and you filter out
everything down nearly to DC; in other words, you've reinvented the
wall-wart (power supply, AC->DC converter, etc.). (Calculate the
impedance of 1000uF at your carrier and modulating frequencies.)

So... design your trailing LPF to eliminate signals above (say)
20kHz, feed it a carrier modulated by a 2000Hz signal, and see what
you get out the far end.


Hope this helps...


Frank McKenney
-- 
    Liberty not only means that the individual has both the
    opportunity and the burden of choice; it also means that he
    must bear the consequences of his actions... Liberty and
    responsibility are inseparable. 
        -- Friedrich von Hayek, The Constitution of Liberty 1960
-- 
Frank McKenney, McKenney Associates
Richmond, Virginia / (804) 320-4887
Munged E-mail: frank uscore mckenney ayut mined spring dawt cahm (y'all)

Jerry Avins <jya@ieee.org> wrote:
(snip, I wrote)

>> OK, but say one wants to minimize the analog circuitry, and fast
>> digital circuitry is available, including a fast ADC.  That would
>> seem to go against the analog bandpass filter, but a lot of digital
>> filtering after the ADC could be provided.
 
(snip)
 
> With baseband sampling, we need an anti-alias lowpass filter. With 
> bandpass sampling, we need an anti-alias bandpass filter.
 
>> But 1MHz is pretty slow for an ADC by now, isn't it?
 
> Sure, but needing all those samples means that you need the MIPS 
> (and the watts) to process them. Why do that?

Microprocessors are an inefficient way to do digital logic
operations, but they are convenient to program.  If you do
it in an FPGA, as a systolic array, it should take a fairly
small amount of logic and, hopefully, not so much power.
 
>>> Incidentally, sampling at 1 MHz provides a little over 2 samples per
>>> carrier cycle, with little chance that either of them will be near a
>>> carrier peak and hence representative of the envelope. With bandpass
>>> sampling at 50 KHz, there will be only one sample for every 9 or so
>>> carrier cycles. There is then no hope of peak detection.

>> Even a small FPGA should allow for a lot of digital filtering
>> that can run at 100MHz or so.
 
> I think the trade-offs between analog and digital selectivity keep 
> changing as technology advances. What sample rate would be needed to 
> digitize the whole AM band? Should we build receivers that way?

I haven't followed it so closely, but I think that they now have
most of an analog AM radio on a single chip.  It would seem,
though, that pretty soon AM radios could easily be built mostly
digital, and maybe not so much longer for FM.

Well, I had the idea not so long ago in an FPGA newsgroup about
an FPGA based development board for college level digital logic
classes.  It would have the parts for a digital clock that could
be used for a freshman level class, and then in later years one
could convert to a clock radio.  As above, the amount of analog
circuitry would be minimized, such that much of the work goes 
into the digital part in the FPGA.

As I understand it, many college level digital design classes
are taught entirely with simulations.  Students never see any
actual hardware!

-- glen

>The size of the capacitor (a.k.a. lowpass filter) matters.
>
>Put a 100pF capacitor across your signal and you filter out stuff you
>can't hear (and presumably, in this context, don't care about). It
>won't have a noticeable effect on audio frequencies. (But you should
>check to see what impedance 100pF represents at your carrier and
>modulating frequencies.)
>
>Put a 1000uF capacitor across your signal hand and you filter out
>everything down nearly to DC; in other words, you've reinvented the
>wall-wart (power supply, AC->DC converter, etc.). (Calculate the
>impedance of 1000uF at your carrier and modulating frequencies.)
>
>So... design your trailing LPF to eliminate signals above (say)
>20kHz, feed it a carrier modulated by a 2000Hz signal, and see what
>you get out the far end.
>
>Hope this helps...

Hi,
Thanks, this helps!

Now, I'm not building an AM radio, and I'm mostly interested in very low
frequency (up to 20Hz), but I need quite good performance. My sampling rate
and carrier are fixed and won't change. This is not under my control and
were decided because of other stuff that this system is doing.  

So another way to ask the original question:
When taking the absolute value of a digital signal, what really happens?
I'm trying to see this from a frequency domain perspective. 

My original experiment - reinventing the wall-wart I guess - showed that
the carrier is shifted to DC pretty well. I'm trying to get a feel for what
happens to the sidebands, how much distortion is introduced when I move out
of DC? I understand abs() is nonlinear, but there might be some analysis
possible?

Now, since multiplying by sin/cos is a perfect shift, I know the sidebands
won't get affected, and my problem becomes designing a good lowpass filter.


Thanks! 

On Apr 25, 12:39=A0am, Jerry Avins <j...@ieee.org> wrote:
> On 4/23/2010 5:56 PM, Vladimir Vassilevsky wrote:
>
>
>
>
>
> > Jerry Avins wrote:
>
> >> On 4/23/2010 5:33 PM, HardySpicer wrote:
>
> >>> On Apr 24, 8:24 am, Vladimir Vassilevsky<nos...@nowhere.com> wrote:
>
> >>>> HardySpicer wrote:
>
> >>>>> Synchronous demodulation using a PLL will give you 3dB improvement
> >>>>> over ordinary envelope detection.
>
> >>>> This is wrong.
>
> >>> It's in the textbooks...read it!
>
> >> What is ordinary envelope detection? Peak detection?
>
> > Doesn't matter; It is very simple. Think of |I| vs sqrt(I^2 + Q^2)
>
> Still, when someone claims "3dB improvement", I want to know what is
> improved upon.
>
> Jerry
> --
> "I view the progress of science as ... the slow erosion of the tendency
> =A0 to dichotomize." --Barbara Smuts, U. Mich.
> =AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=
=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=
=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF

If I can remember that far back I believe it is in received SNR
(baseband). Look up synchronous demodulation vs
 envelope detection. Of course you digital guys have I and Q and make
things even more complicated. Do you remember when we just used a cats
whisker!
The I and Q method vs a PLL - I have no idea. How do you get I and Q -
do you need a PLL to get I and Q? If so then I expect it is the same
result.

It was Taub and Shilling or some such that had the details but most
older coms books will have the info. People are so locked into digital
nowadays that they forget the basics
which all comes from analogue. You cannot understand digital without
understanding analogue first. I don't mean you Jerry of course but
many of the new breed of engineer that gets taught exclusively
digital.
(does happen!!) For example, what is an exclusive OR in a PLL? To me
it is a multiplier!!

Hardy

On 4/24/2010 3:30 PM, gretzteam wrote:
>> The size of the capacitor (a.k.a. lowpass filter) matters.
>>
>> Put a 100pF capacitor across your signal and you filter out stuff you
>> can't hear (and presumably, in this context, don't care about). It
>> won't have a noticeable effect on audio frequencies. (But you should
>> check to see what impedance 100pF represents at your carrier and
>> modulating frequencies.)
>>
>> Put a 1000uF capacitor across your signal hand and you filter out
>> everything down nearly to DC; in other words, you've reinvented the
>> wall-wart (power supply, AC->DC converter, etc.). (Calculate the
>> impedance of 1000uF at your carrier and modulating frequencies.)
>>
>> So... design your trailing LPF to eliminate signals above (say)
>> 20kHz, feed it a carrier modulated by a 2000Hz signal, and see what
>> you get out the far end.
>>
>> Hope this helps...
>
> Hi,
> Thanks, this helps!
>
> Now, I'm not building an AM radio, and I'm mostly interested in very low
> frequency (up to 20Hz), but I need quite good performance. My sampling rate
> and carrier are fixed and won't change. This is not under my control and
> were decided because of other stuff that this system is doing.
>
> So another way to ask the original question:
> When taking the absolute value of a digital signal, what really happens?
> I'm trying to see this from a frequency domain perspective.

I'd have to do the math -- a Laplace transform -- to give a definitive 
answer even for a continuous signal, and the discrete-time case is more 
involved. There seems to be a conflict among authorities. In the one 
hand, the waveform is what one gets from a push-push doubler, and should 
contain no odd harmonics, not even the fundamental. On the other hand, 
the ITT Reference Data for Radio Engineers (4th edition; 1949) gives a 
formula that I don't believe involving finite values of coefficients for 
all harmonics, including the fundamental.

> My original experiment - reinventing the wall-wart I guess - showed that
> the carrier is shifted to DC pretty well.

I wouldn't say that. There's a strong second harmonic, so you could also 
say that the carrier is shifted up. What we know is that the average 
value is extracted and that harmonics are produced. Each of those 
harmonics will exhibit sidebands.

>                                         I'm trying to get a feel for what
> happens to the sidebands, how much distortion is introduced when I move out
> of DC? I understand abs() is nonlinear, but there might be some analysis
> possible?
>
> Now, since multiplying by sin/cos is a perfect shift, I know the sidebands
> won't get affected, and my problem becomes designing a good lowpass filter.

What do you mean by "not affected"? they get shifted along with the carrier.

Jerry
-- 
"I view the progress of science as ... the slow erosion of the tendency
  to dichotomize." --Barbara Smuts, U. Mich.
�����������������������������������������������������������������������

On 4/24/2010 8:04 PM, HardySpicer wrote:
> On Apr 25, 12:39 am, Jerry Avins<j...@ieee.org>  wrote:
>> On 4/23/2010 5:56 PM, Vladimir Vassilevsky wrote:
>>
>>
>>
>>
>>
>>> Jerry Avins wrote:
>>
>>>> On 4/23/2010 5:33 PM, HardySpicer wrote:
>>
>>>>> On Apr 24, 8:24 am, Vladimir Vassilevsky<nos...@nowhere.com>  wrote:
>>
>>>>>> HardySpicer wrote:
>>
>>>>>>> Synchronous demodulation using a PLL will give you 3dB improvement
>>>>>>> over ordinary envelope detection.
>>
>>>>>> This is wrong.
>>
>>>>> It's in the textbooks...read it!
>>
>>>> What is ordinary envelope detection? Peak detection?
>>
>>> Doesn't matter; It is very simple. Think of |I| vs sqrt(I^2 + Q^2)
>>
>> Still, when someone claims "3dB improvement", I want to know what is
>> improved upon.
>>
>> Jerry
>> --
>> "I view the progress of science as ... the slow erosion of the tendency
>>    to dichotomize." --Barbara Smuts, U. Mich.
>> �����������������������������������������������������������������������
>
> If I can remember that far back I believe it is in received SNR
> (baseband). Look up synchronous demodulation vs
>   envelope detection. Of course you digital guys have I and Q and make
> things even more complicated. Do you remember when we just used a cats
> whisker!
> The I and Q method vs a PLL - I have no idea. How do you get I and Q -
> do you need a PLL to get I and Q? If so then I expect it is the same
> result.
>
> It was Taub and Shilling or some such that had the details but most
> older coms books will have the info. People are so locked into digital
> nowadays that they forget the basics
> which all comes from analogue. You cannot understand digital without
> understanding analogue first. I don't mean you Jerry of course but
> many of the new breed of engineer that gets taught exclusively
> digital.
> (does happen!!) For example, what is an exclusive OR in a PLL? To me
> it is a multiplier!!

I/Q demodulation is pure envelope detection, even for carriers not much 
higher than the highest modulating frequency. What you seem to refer to 
as envelope detection is peak detection, which is merely a damn good 
approximation under typical conditions of operation.

There are a number of ways to make an analytic (I-Q) signal. A Hilbert 
transformer is one, Clay Turner's Tips & Tricks article is better. Often 
the best produces both I and Q in the sampling process. Synchronous 
demodulation does improve on peak detection, but not on sqrt(I^2+Q^2). 
So does exalted carrier, for the same fundamental reason.

Jerry
-- 
"I view the progress of science as ... the slow erosion of the tendency
  to dichotomize." --Barbara Smuts, U. Mich.
�����������������������������������������������������������������������

On 4/24/2010 1:25 PM, glen herrmannsfeldt wrote:

   ...

> As I understand it, many college level digital design classes
> are taught entirely with simulations.  Students never see any
> actual hardware!

The nature of many of the questions we see here bear that out.

Jerry
-- 
"I view the progress of science as ... the slow erosion of the tendency
  to dichotomize." --Barbara Smuts, U. Mich.
�����������������������������������������������������������������������

>As I understand it, many college level digital design classes
>are taught entirely with simulations.  Students never see any
>actual hardware!

I am told that it is becoming quite normal for entire electronics degree
courses to be free of any lab work. Therefore, analogue, power and other
topics are in the same position as digital. Cost, health and safety issues,
and the narrow inclinations of lecturers were cited to me recently by a
lecturer. One still runs his power electronics courses to include things
that can actually go bang.

Regards,
Steve

On Apr 25, 1:03=A0am, Jerry Avins <j...@ieee.org> wrote:

>Synchronous demodulation does improve on peak detection, but not on sqrt(I=
^2+Q^2).

I think you are wrong on this point.

I am in this group primarily because I have responsibility to
"maintain " a software radio product that we bought from another
company. So I have had to try to learn DSP to maintain understand and
maintain this radio.  I am not yet capable of designing a radio like
this, but I am getting closer as time goes by.

Anyhow,  In their design they use a digital PLL. They run a FFT to
analyze the carrier and then build an PLL around the signal to get
synchronous detection.  If the carrier gets squirrely, the PLL will
break lock, and if it does they revert to I/Q detection so as to not
lose the signal.
My point is, I know they did not go to great lengths to build thisFFT/
PLL detection scheme if I/Q demodulation would give the same result.
I cannot tell you right now what the math is, but I know synchronous
detection is better in some respects.

On 4/25/2010 7:07 AM, brent wrote:
> On Apr 25, 1:03 am, Jerry Avins<j...@ieee.org>  wrote:
>
>> Synchronous demodulation does improve on peak detection, but not on sqrt(I^2+Q^2).
>
> I think you are wrong on this point.

That would hardly be novel!

> I am in this group primarily because I have responsibility to
> "maintain " a software radio product that we bought from another
> company. So I have had to try to learn DSP to maintain understand and
> maintain this radio.  I am not yet capable of designing a radio like
> this, but I am getting closer as time goes by.
>
> Anyhow,  In their design they use a digital PLL. They run a FFT to
> analyze the carrier and then build an PLL around the signal to get
> synchronous detection.  If the carrier gets squirrely, the PLL will
> break lock, and if it does they revert to I/Q detection so as to not
> lose the signal.
> My point is, I know they did not go to great lengths to build thisFFT/
> PLL detection scheme if I/Q demodulation would give the same result.
> I cannot tell you right now what the math is, but I know synchronous
> detection is better in some respects.

Without judging the merits of this particular scheme, I have to keep in 
mind other instances where designers -- Even Edwin Armstrong, the father 
of FM -- acted on superstition rather than fact. The scheme you describe 
leaves me with a question: If synchronous demodulation loses the signal 
even when I/Q demodulation tracks it, in what way is synchronous 
demodulation superior? Have you had a chance to compare the outputs of 
the two detectors when both are working?

Jerry
-- 
"I view the progress of science as ... the slow erosion of the tendency
  to dichotomize." --Barbara Smuts, U. Mich.
�����������������������������������������������������������������������

>> Now, I'm not building an AM radio, and I'm mostly interested in very
low
>> frequency (up to 20Hz), but I need quite good performance. My sampling
rate
>> and carrier are fixed and won't change. This is not under my control
and
>> were decided because of other stuff that this system is doing.
>>
>> So another way to ask the original question:
>> When taking the absolute value of a digital signal, what really
happens?
>> I'm trying to see this from a frequency domain perspective.
>
>I'd have to do the math -- a Laplace transform -- to give a definitive 
>answer even for a continuous signal, and the discrete-time case is more 
>involved. There seems to be a conflict among authorities. In the one 
>hand, the waveform is what one gets from a push-push doubler, and should 
>contain no odd harmonics, not even the fundamental. On the other hand, 
>the ITT Reference Data for Radio Engineers (4th edition; 1949) gives a 
>formula that I don't believe involving finite values of coefficients for 
>all harmonics, including the fundamental.
>
>> My original experiment - reinventing the wall-wart I guess - showed
that
>> the carrier is shifted to DC pretty well.
>
>I wouldn't say that. There's a strong second harmonic, so you could also 
>say that the carrier is shifted up. What we know is that the average 
>value is extracted and that harmonics are produced. Each of those 
>harmonics will exhibit sidebands.
>
>>                                         I'm trying to get a feel for
what
>> happens to the sidebands, how much distortion is introduced when I move
out
>> of DC? I understand abs() is nonlinear, but there might be some
analysis
>> possible?
>>
>> Now, since multiplying by sin/cos is a perfect shift, I know the
sidebands
>> won't get affected, and my problem becomes designing a good lowpass
filter.
>
>What do you mean by "not affected"? they get shifted along with the
carrier.
>
>Jerry

Yes but the sidebands don't get distorted at all during the sin/cos
multiplication, not matter what other stuff is in the spectrum from 0 to
fs/2, right? Everything just gets shifted, and performance will depend on
the lowpass filter.
However, can we say that for the abs() case? Can we say that the sidebands
are shifted along with the carrier back to DC without distortion? I don't
care if everything else becomes all distorted, cause it will get lowpass
filtered.

Thanks.
 

>My point is, I know they did not go to great lengths to build thisFFT/
>PLL detection scheme if I/Q demodulation would give the same result.
>I cannot tell you right now what the math is, but I know synchronous
>detection is better in some respects.


Maybe they try to avoid I/Q demodulation because it involves two
multipliers running 'fast', and twice the lowpass filtering? This is kind
of what I'm trying to avoid, if I can get decent performance with another
scheme.

However, a bandpass filter is not free either, so maybe it all boils down
to the same thing!

Thanks,

Jerry Avins wrote:

> If synchronous demodulation loses the signal 
> even when I/Q demodulation tracks it, in what way is synchronous 
> demodulation superior? Have you had a chance to compare the outputs of 
> the two detectors when both are working?

BTW, in the modern ICs, they demodulate AM in a quasi-synchronous way: 
the bandpass signal is digitized by comparator and the output of the 
comparator drives the analog switch which does synchronous 
rectification. The advantage is much better linearity compared to diode 
detector. Obviously this could be improved by adding a narrowband filter 
in front of the comparator; however it is not worth extra cost for 
consumer applications.


Vladimir Vassilevsky
DSP and Mixed Signal Design Consultant
http://www.abvolt.com

On 4/25/2010 11:59 AM, Vladimir Vassilevsky wrote:
>
>
> Jerry Avins wrote:
>
>> If synchronous demodulation loses the signal even when I/Q
>> demodulation tracks it, in what way is synchronous demodulation
>> superior? Have you had a chance to compare the outputs of the two
>> detectors when both are working?
>
> BTW, in the modern ICs, they demodulate AM in a quasi-synchronous way:
> the bandpass signal is digitized by comparator and the output of the
> comparator drives the analog switch which does synchronous
> rectification. The advantage is much better linearity compared to diode
> detector. Obviously this could be improved by adding a narrowband filter
> in front of the comparator; however it is not worth extra cost for
> consumer applications.

That's the digital equivalent of exalted carrier operation that I 
mentioned earlier in this thread.

Jerry
-- 
"I view the progress of science as ... the slow erosion of the tendency
  to dichotomize." --Barbara Smuts, U. Mich.
�����������������������������������������������������������������������

On 4/25/2010 11:31 AM, gretzteam wrote:
>>> Now, I'm not building an AM radio, and I'm mostly interested in very
> low
>>> frequency (up to 20Hz), but I need quite good performance. My sampling
> rate
>>> and carrier are fixed and won't change. This is not under my control
> and
>>> were decided because of other stuff that this system is doing.
>>>
>>> So another way to ask the original question:
>>> When taking the absolute value of a digital signal, what really
> happens?
>>> I'm trying to see this from a frequency domain perspective.
>>
>> I'd have to do the math -- a Laplace transform -- to give a definitive
>> answer even for a continuous signal, and the discrete-time case is more
>> involved. There seems to be a conflict among authorities. In the one
>> hand, the waveform is what one gets from a push-push doubler, and should
>> contain no odd harmonics, not even the fundamental. On the other hand,
>> the ITT Reference Data for Radio Engineers (4th edition; 1949) gives a
>> formula that I don't believe involving finite values of coefficients for
>> all harmonics, including the fundamental.
>>
>>> My original experiment - reinventing the wall-wart I guess - showed
> that
>>> the carrier is shifted to DC pretty well.
>>
>> I wouldn't say that. There's a strong second harmonic, so you could also
>> say that the carrier is shifted up. What we know is that the average
>> value is extracted and that harmonics are produced. Each of those
>> harmonics will exhibit sidebands.
>>
>>>                                          I'm trying to get a feel for
> what
>>> happens to the sidebands, how much distortion is introduced when I move
> out
>>> of DC? I understand abs() is nonlinear, but there might be some
> analysis
>>> possible?
>>>
>>> Now, since multiplying by sin/cos is a perfect shift, I know the
> sidebands
>>> won't get affected, and my problem becomes designing a good lowpass
> filter.
>>
>> What do you mean by "not affected"? they get shifted along with the
> carrier.
>>
>> Jerry
>
> Yes but the sidebands don't get distorted at all during the sin/cos
> multiplication, not matter what other stuff is in the spectrum from 0 to
> fs/2, right? Everything just gets shifted, and performance will depend on
> the lowpass filter.
> However, can we say that for the abs() case? Can we say that the sidebands
> are shifted along with the carrier back to DC without distortion? I don't
> care if everything else becomes all distorted, cause it will get lowpass
> filtered.

Some of the harmonics caused by taking the magnitude exceed fs/2 and 
alias. The higher the sample rate, the less of a problem that is.

Jerry
-- 
"I view the progress of science as ... the slow erosion of the tendency
  to dichotomize." --Barbara Smuts, U. Mich.
�����������������������������������������������������������������������

On Apr 25, 5:03=A0pm, Jerry Avins <j...@ieee.org> wrote:
> On 4/24/2010 8:04 PM, HardySpicer wrote:
>
>
>
> > On Apr 25, 12:39 am, Jerry Avins<j...@ieee.org> =A0wrote:
> >> On 4/23/2010 5:56 PM, Vladimir Vassilevsky wrote:
>
> >>> Jerry Avins wrote:
>
> >>>> On 4/23/2010 5:33 PM, HardySpicer wrote:
>
> >>>>> On Apr 24, 8:24 am, Vladimir Vassilevsky<nos...@nowhere.com> =A0wro=
te:
>
> >>>>>> HardySpicer wrote:
>
> >>>>>>> Synchronous demodulation using a PLL will give you 3dB improvemen=
t
> >>>>>>> over ordinary envelope detection.
>
> >>>>>> This is wrong.
>
> >>>>> It's in the textbooks...read it!
>
> >>>> What is ordinary envelope detection? Peak detection?
>
> >>> Doesn't matter; It is very simple. Think of |I| vs sqrt(I^2 + Q^2)
>
> >> Still, when someone claims "3dB improvement", I want to know what is
> >> improved upon.
>
> >> Jerry
> >> --
> >> "I view the progress of science as ... the slow erosion of the tendenc=
y
> >> =A0 =A0to dichotomize." --Barbara Smuts, U. Mich.
> >> =AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=
=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=
=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF
>
> > If I can remember that far back I believe it is in received SNR
> > (baseband). Look up synchronous demodulation vs
> > =A0 envelope detection. Of course you digital guys have I and Q and mak=
e
> > things even more complicated. Do you remember when we just used a cats
> > whisker!
> > The I and Q method vs a PLL - I have no idea. How do you get I and Q -
> > do you need a PLL to get I and Q? If so then I expect it is the same
> > result.
>
> > It was Taub and Shilling or some such that had the details but most
> > older coms books will have the info. People are so locked into digital
> > nowadays that they forget the basics
> > which all comes from analogue. You cannot understand digital without
> > understanding analogue first. I don't mean you Jerry of course but
> > many of the new breed of engineer that gets taught exclusively
> > digital.
> > (does happen!!) For example, what is an exclusive OR in a PLL? To me
> > it is a multiplier!!
>
> I/Q demodulation is pure envelope detection, even for carriers not much
> higher than the highest modulating frequency. What you seem to refer to
> as envelope detection is peak detection, which is merely a damn good
> approximation under typical conditions of operation.
>
> There are a number of ways to make an analytic (I-Q) signal. A Hilbert
> transformer is one, Clay Turner's Tips & Tricks article is better. Often
> the best produces both I and Q in the sampling process. Synchronous
> demodulation does improve on peak detection, but not on sqrt(I^2+Q^2).
> So does exalted carrier, for the same fundamental reason.
>
> Jerry
> --
> "I view the progress of science as ... the slow erosion of the tendency
> =A0 to dichotomize." --Barbara Smuts, U. Mich.
> =AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=
=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=
=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF

Ok so we are comparing apples with oranges.

Hardy

On Apr 26, 3:59=A0am, Vladimir Vassilevsky <nos...@nowhere.com> wrote:
> Jerry Avins wrote:
> > If synchronous demodulation loses the signal
> > even when I/Q demodulation tracks it, in what way is synchronous
> > demodulation superior? Have you had a chance to compare the outputs of
> > the two detectors when both are working?
>
> BTW, in the modern ICs, they demodulate AM in a quasi-synchronous way:
> the bandpass signal is digitized by comparator and the output of the
> comparator drives the analog switch which does synchronous
> rectification. The advantage is much better linearity compared to diode
> detector. Obviously this could be improved by adding a narrowband filter
> in front of the comparator; however it is not worth extra cost for
> consumer applications.
>
> Vladimir Vassilevsky
> DSP and Mixed Signal Design Consultanthttp://www.abvolt.com

Can you hear the difference though!

HardySpicer wrote:
> On Apr 26, 3:59 am, Vladimir Vassilevsky <nos...@nowhere.com> wrote:
> 
>>Jerry Avins wrote:
>>
>>>If synchronous demodulation loses the signal
>>>even when I/Q demodulation tracks it, in what way is synchronous
>>>demodulation superior? Have you had a chance to compare the outputs of
>>>the two detectors when both are working?
>>
>>BTW, in the modern ICs, they demodulate AM in a quasi-synchronous way:
>>the bandpass signal is digitized by comparator and the output of the
>>comparator drives the analog switch which does synchronous
>>rectification. The advantage is much better linearity compared to diode
>>detector. Obviously this could be improved by adding a narrowband filter
>>in front of the comparator; however it is not worth extra cost for
>>consumer applications.
> 
> Can you hear the difference though!

You could see the difference. It used to be done in that way in analog TV.


Vladimir Vassilevsky
DSP and Mixed Signal Design Consultant
http://www.abvolt.com

On Apr 26, 3:59=A0am, Vladimir Vassilevsky <nos...@nowhere.com> wrote:
> Jerry Avins wrote:
> > If synchronous demodulation loses the signal
> > even when I/Q demodulation tracks it, in what way is synchronous
> > demodulation superior? Have you had a chance to compare the outputs of
> > the two detectors when both are working?
>
> BTW, in the modern ICs, they demodulate AM in a quasi-synchronous way:
> the bandpass signal is digitized by comparator and the output of the
> comparator drives the analog switch which does synchronous
> rectification. The advantage is much better linearity compared to diode
> detector. Obviously this could be improved by adding a narrowband filter
> in front of the comparator; however it is not worth extra cost for
> consumer applications.
>
> Vladimir Vassilevsky
> DSP and Mixed Signal Design Consultanthttp://www.abvolt.com

Could all be done with an op-amp! What nonsense. How to make radio
1000 times more complicated than it need be.

Hardy

>On Apr 26, 3:59=A0am, Vladimir Vassilevsky <nos...@nowhere.com> wrote:
>> Jerry Avins wrote:
>> > If synchronous demodulation loses the signal
>> > even when I/Q demodulation tracks it, in what way is synchronous
>> > demodulation superior? Have you had a chance to compare the outputs
of
>> > the two detectors when both are working?
>>
>> BTW, in the modern ICs, they demodulate AM in a quasi-synchronous way:
>> the bandpass signal is digitized by comparator and the output of the
>> comparator drives the analog switch which does synchronous
>> rectification. The advantage is much better linearity compared to diode
>> detector. Obviously this could be improved by adding a narrowband
filter
>> in front of the comparator; however it is not worth extra cost for
>> consumer applications.
>>
>> Vladimir Vassilevsky
>> DSP and Mixed Signal Design Consultanthttp://www.abvolt.com
>
>Could all be done with an op-amp! What nonsense. How to make radio
>1000 times more complicated than it need be.

Vlad described a solution consisting of a comparator (i.e. an op-amp with
op-amp crossed out, and comparator written across it in crayon) and an
analogue switch (i.e. one of the simplest structures in mixed signal CMOS).
Your solution requires an op-amp and a diode. Can you explain the 1000
times complexity ratio?

Steve