Approach to demodulating AM-signals?

Hello Jerry,

>> WDF    Washington Department of Fisheries    *****
>> WDF    Web Design Factory    ****
>> WDF    Web Design Forums    ****
>> WDF    Washington Department of Fish    ****
>> WDF    World Diabetes Foundation    ***
>> WDF    World Development Federation    ***
>> WDF    Wave Digital Filters    ***
>>

And then there is the World Dart Federation.


>> I think Wave Digital Filter sounds likely.
> 

Bingo!

> 
> What distinguishes a wave filter from other kinds of signal filters?
> Is there a wave analog filter? I'm unfamiliar with the jargon.
> 

I don't know why it's called digital, probably because it is the digital 
rendition of an analog filter architecture. The only design method I 
know of is to first design the analog equivalent, add one order in case 
this wasn't odd order and then calculate the adapter coefficients. There 
may be some newer CAD SW out there but I haven't found any, except for 
Prof.Mildenbergers DOS routines which I have. There used to be another 
bigger package called "Falcon" but I am still trying to find out whether 
it is at all available to the public. The research group around 
Prof.Fettweis appears to be gone.

Regards, Joerg

http://www.analogconsultants.com

Hello Tim,

>>
> Let's try that again.  What in the world does 'WDF' stand for?
> 

As Allan said, "wave digital filter".


> Forget the aliasing from sampling, I was talking about the 30kHz carrier 
> plus 10kHz modulation intemodding in the absolute value thingie.  You 
> start with energy at 20, 30 and 40 kHz, then run it through a whopping 
> big nonlinearity.  You end up with 20 mixing with 30 down to 10 (good), 
> 30 mixing with 40 down to 10 (also good), but you also end up with the 
> 40 from the bottom end of the second harmonic mixing with the 30kHz 
> carrier (bad), etc., etc.  I think you'd have to simulate it to really 
> see if it's going to have issues.
> 

It won't be non-linear. That's the way classical AM detectors work. If 
it would produce lots of unwanted products even our TVs wouldn't produce 
any decent picture quality. They are basically AM, with a suppressed 
sideband to economize on the spectrum and reduced carrier to curb power 
consumption of the transmitters. Of course, now the gvt has decided to 
do away with all that in a few years. I am already seeing a ton of grief 
flying at the representatives but that's a whole other topic.

Another method would be IQ demodulation but that kind of effort is 
usually only justified if you have to distinuish between the sidebands.

Regards, Joerg

http://www.analogconsultants.com

Hello Al,


> Analog guys might think we're all crazy and just implement the whole 
> thing with a more traditional analog approach. 
> 

Guilty. That's what I sometimes thought (but never said aloud). I have 
reverted a lot of designs to analog and when the BOM cost page was 
flipped on during the design review it generally floored the managers. 
Many didn't know that you can have a jelly bean transistor that is very 
RF capable for a mere 2 cents including assembly. Some modules went to 
less than 20% of the cost of their digital brethren. But that picture 
can change and has in some areas.


> I brought up the DSP choice to broaden the possibilities. DSPs have all 
> gotten both cheaper and faster over the years. Design tradeoffs are 
> always a changing dynamic. I can certainly attest to this in my career, 
> because I used to be an analog engineer before I started using DSP.
> 

Yes, and then there is stuff such as wideband Hilbert transforms that 
can be done analog but will be less expensive and more accurate on a 
DSP. There, analog just doesn't make sense anymore. The AD2105 was kind 
of the breakthrough because it's price was right. I was hoping this 
could become the 8051 of the DSP world but now much of that range is gone.

Regards, Joerg

http://www.analogconsultants.com

Joerg <notthisjoergsch@removethispacbell.net> wrote in
news:mXnOf.65779$PL5.6156@newssvr11.news.prodigy.com: 

> Hello Al,
> 
> 
>> Analog guys might think we're all crazy and just implement the whole 
>> thing with a more traditional analog approach. 
>> 
> 
> Guilty. That's what I sometimes thought (but never said aloud). I have
> reverted a lot of designs to analog and when the BOM cost page was 
> flipped on during the design review it generally floored the managers.
> Many didn't know that you can have a jelly bean transistor that is
> very RF capable for a mere 2 cents including assembly. Some modules
> went to less than 20% of the cost of their digital brethren. But that
> picture can change and has in some areas.
> 
> 
>> I brought up the DSP choice to broaden the possibilities. DSPs have
>> all gotten both cheaper and faster over the years. Design tradeoffs
>> are always a changing dynamic. I can certainly attest to this in my
>> career, because I used to be an analog engineer before I started
>> using DSP. 
>> 
> 




> Yes, and then there is stuff such as wideband Hilbert transforms that 
> can be done analog but will be less expensive and more accurate on a 
> DSP. There, analog just doesn't make sense anymore. The AD2105 was
> kind of the breakthrough because it's price was right. I was hoping
> this could become the 8051 of the DSP world but now much of that range
> is gone. 

I have probably written more 2105 code than anyone you know, but I can't 
say I miss it. You can do so much more with just about any DSP today. I 
think some of the Blackfins are less expensive than the 2105 was at 
$10.00. 

The SHARCs I use today cost more but I can do so much more with them. I 
can program faster with the SHARC than the older fixed point processors 
as well.  The other thing about modern DSPs is they are no longer idiot 
savants. Most DSPs make for pretty good general purpose micros. The ADSP-
21369 for example has two UARTS, I2C, Dual SPI, Timers, PWM, SDRAM 
controller as well as all the normal DSP stuff.

When I was exclusively an analog guy, I used to talk about a DSP core 
with an ADC and DAC that could replace most of my analog functions. Where 
DSP solutions were possible, they were usually way to expensive to 
consider. I'm certainly not anti-analog but I think many solutions are 
better suited to a DSP approach that were once strictly analog 
implementations.

Al

> 
> Regards, Joerg
> 
> http://www.analogconsultants.com
> 



-- 
Al Clark
Danville Signal Processing, Inc.
--------------------------------------------------------------------
Purveyors of Fine DSP Hardware and other Cool Stuff
Available at http://www.danvillesignal.com

More info about the project:



We want to make a digital demodulator because:



1.)    we want to be able to make modifications to the signal processing in
the future

2.)    we want small size

3.)    we want to get rid of a lot of analog tolerances



The unit creates a carrier wave that is sent through a medium. The medium
AM-modulates the signal. So the receiver end in the unit receives an
AM-modulated signal. The carrier wave can be selected from about 30 kHz up
to 100 kHz. And the AM-bandwidth we are interested in is 10 kHz. So if the
carrier is 50 kHz the received AM-signal is between 45 kHz to 55 kHz.



The medium contains interference so we want to be able to move our carrier
frequency to the "quietest" region of this band (30 kHz  to 100 kHz).



The transmitter is in the same unit as the receiver. We could maybe make a
synchronous detector?



And thank you all for the answers I have received so far.

Hello Al,

> 
> I have probably written more 2105 code than anyone you know, but I can't 
> say I miss it. You can do so much more with just about any DSP today. I 
> think some of the Blackfins are less expensive than the 2105 was at 
> $10.00. 
> 

Hey, we got them for under $5. For the truck load :-)

I am no expert in code writing. I worked on algorithms but the code was 
done by folks who knew more about that than I do.

> 
> When I was exclusively an analog guy, I used to talk about a DSP core 
> with an ADC and DAC that could replace most of my analog functions. Where 
> DSP solutions were possible, they were usually way to expensive to 
> consider. I'm certainly not anti-analog but I think many solutions are 
> better suited to a DSP approach that were once strictly analog 
> implementations.
> 

True. Nowadays there is so much that can be done digitally. However, it 
almost always (except for some 8051 version) comes with a caveat: Single 
source. One hickup in the production or delivery chain and your own 
production could be toast. I learned to hate that dreaded phrase "It's 
on allocation", meaning only the well-connected will get any this quarter.

And think about curb appeal: A nice clear-and-red MELF diode plus a 
shiny metallic-green film cap versus a bland 144TQFP with some 
diarrhea-brown 0.1uF decoupling caps. Ok, now I am getting carried away 
here...

Regards, Joerg

http://www.analogconsultants.com

"Mr M" <blobb@bredband.net> wrote in
news:440b34f7$0$15795$14726298@news.sunsite.dk: 

> More info about the project:
> 
> 
> 
> We want to make a digital demodulator because:
> 
> 
> 
> 1.)    we want to be able to make modifications to the signal
> processing in the future
> 
> 2.)    we want small size
> 
> 3.)    we want to get rid of a lot of analog tolerances
> 
> 
> 
> The unit creates a carrier wave that is sent through a medium. The
> medium AM-modulates the signal. So the receiver end in the unit
> receives an AM-modulated signal. The carrier wave can be selected from
> about 30 kHz up to 100 kHz. And the AM-bandwidth we are interested in
> is 10 kHz. So if the carrier is 50 kHz the received AM-signal is
> between 45 kHz to 55 kHz. 
> 
> 
> 
> The medium contains interference so we want to be able to move our
> carrier frequency to the "quietest" region of this band (30 kHz  to
> 100 kHz). 
> 
> 
> 
> The transmitter is in the same unit as the receiver. We could maybe
> make a synchronous detector?
> 
> 
> 
> And thank you all for the answers I have received so far.
> 
> 
> 

I'm not sure if this has been mentioned before but there is a very good 
book called Digital Signal Processing in Communication Systems by Marvin 
Frerking. It is a hard book to find but worth the search (and relatively 
high price)

-- 
Al Clark
Danville Signal Processing, Inc.
--------------------------------------------------------------------
Purveyors of Fine DSP Hardware and other Cool Stuff
Available at http://www.danvillesignal.com

Hello Mr.M,

> 
> The medium contains interference so we want to be able to move our carrier
> frequency to the "quietest" region of this band (30 kHz  to 100 kHz).
> 

You could have it auto-tune. First sweep the frequency while the 
transmitter is off and have the uC or DSP detect the most quiet range. 
Then turn the transmitter on. Repeat regularly if the noise pattern 
might change over time.

> 
> The transmitter is in the same unit as the receiver. We could maybe make a
> synchronous detector?
> 

You could but if you don't need accurate phase information it might be 
too much. Regular AM detection and possibly an AGC amp up front may suffice.

There are plenty of demod chips (Analog Devices etc.) that almost take 
out any tolerance issue. But if you have enough horsepower in the DSP 
you may not need that. However, mind the dynamic range. If you need more 
than 10 to 12 bits the AD conversion tends to become expensive and the 
data volume at more than 250ksps is quite massive.

Regards, Joerg

http://www.analogconsultants.com

Mr M wrote:

> More info about the project:
> 
> 
> 
> We want to make a digital demodulator because:
> 
> 
> 
> 1.)    we want to be able to make modifications to the signal processing in
> the future
> 
> 2.)    we want small size
> 
> 3.)    we want to get rid of a lot of analog tolerances
> 
> 
> 
> The unit creates a carrier wave that is sent through a medium. The medium
> AM-modulates the signal. So the receiver end in the unit receives an
> AM-modulated signal. The carrier wave can be selected from about 30 kHz up
> to 100 kHz. And the AM-bandwidth we are interested in is 10 kHz. So if the
> carrier is 50 kHz the received AM-signal is between 45 kHz to 55 kHz.
> 
> 
> 
> The medium contains interference so we want to be able to move our carrier
> frequency to the "quietest" region of this band (30 kHz  to 100 kHz).
> 
> 
> 
> The transmitter is in the same unit as the receiver. We could maybe make a
> synchronous detector?
> 
> 
> 
> And thank you all for the answers I have received so far.
> 
> 
If you know the phase shift then synchronous demod will probably work 
well.  I would have suggested it except that it's a pain to lock onto 
the carrier -- but hey, you have that already.

-- 

Tim Wescott
Wescott Design Services
http://www.wescottdesign.com

Posting from Google?  See http://cfaj.freeshell.org/google/

Hello Tim,

>>
> If you know the phase shift then synchronous demod will probably work 
> well.  I would have suggested it except that it's a pain to lock onto 
> the carrier -- but hey, you have that already.
> 

Or do the full I/Q scheme and it'll tell you the phase.

Regards, Joerg

http://www.analogconsultants.com