# thx for clay and julius, 'Still mathetical problem, but for AM demodulation with FIR filter"

Started by May 26, 2007
my problem is the mathematical equation for how the FIR filter to process
demodulation for AM.
oya, bout the windowing method, u tell me...i found the equation for
bandpass filter ideal.
i read a book, and i got the equation for unideal filter. is it h=w.hd?
h:transfer function unideal
w:windowing coef.
hd: transfer function ideal filter

and do i have to make an ADAPTIVE filter for AM demodulation with the
windowing method?

_____________________________________
Do you know a company who employs DSP engineers?
Is it already listed at http://dsprelated.com/employers.php ?
c1910 wrote:
> Clay, thanks for your opinion... > my problem is the mathematical equation for how the FIR filter to process > demodulation for AM. > oya, bout the windowing method, u tell me...i found the equation for > bandpass filter ideal. > i read a book, and i got the equation for unideal filter. is it h=w.hd? > h:transfer function unideal > w:windowing coef. > hd: transfer function ideal filter > > and do i have to make an ADAPTIVE filter for AM demodulation with the > windowing method?
I just jumped in here, so maybe I'm missing something. As far as I know, a filter won't demodulate AM. Generally, AM is demodulated with rectifiers (half wave, full wave, synchronous) followed by low-pass filtering to remove the carrier (or double-frequency carrier) components. Because the carrier is ordinarily so much higher than the modulation, that filter can be trivial. Jerry -- Engineering is the art of making what you want from things you can get. &macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;
On May 26, 11:24 am, "c1910" <c_19...@hotmail.com> wrote:
> Clay, thanks for your opinion... > my problem is the mathematical equation for how the FIR filter to process > demodulation for AM. > oya, bout the windowing method, u tell me...i found the equation for > bandpass filter ideal. > i read a book, and i got the equation for unideal filter. is it h=w.hd? > h:transfer function unideal > w:windowing coef. > hd: transfer function ideal filter > > and do i have to make an ADAPTIVE filter for AM demodulation with the > windowing method? >
You are throwing around too many buzzwords here. If you want to know how to implement a finite impulse response (FIR) filter, you should look in any text in digital signal processing. Some books are even tailored towards particular implementation, such as "Numerical Recipes in C" or other text that are suited for FPGA, DSP systems. Why don't you say more clearly what you have? A C compiler? Verilog? You are way overthinking the problem. Julius
>I just jumped in here, so maybe I'm missing something. As far as I know,
> a filter won't demodulate AM. Generally, AM is demodulated with >rectifiers (half wave, full wave, synchronous) followed by low-pass >filtering to remove the carrier (or double-frequency carrier) >components. Because the carrier is ordinarily so much higher than the >modulation, that filter can be trivial. > >Jerry >-- >Engineering is the art of making what you want from things you can get. >&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr; >
thx... hmm, my problem is how to get the FIR filter transfer function... like u said, i want to remove the carrier with filter. I use DSP TMS320C54, and my problem is how to get the filter transfer function with mathtematical equation. oya, can u give me the Block Diagram of your AM demodulation?please...or u can give me the process...thx so much -chris(c1910)- _____________________________________ Do you know a company who employs DSP engineers? Is it already listed at http://dsprelated.com/employers.php ?
c1910 wrote:
>> I just jumped in here, so maybe I'm missing something. As far as I know, > >> a filter won't demodulate AM. Generally, AM is demodulated with >> rectifiers (half wave, full wave, synchronous) followed by low-pass >> filtering to remove the carrier (or double-frequency carrier) >> components. Because the carrier is ordinarily so much higher than the >> modulation, that filter can be trivial.
...
> thx... > hmm, my problem is how to get the FIR filter transfer function... > like u said, i want to remove the carrier with filter. I use DSP > TMS320C54, and my problem is how to get the filter transfer function with > mathtematical equation. > oya, can u give me the Block Diagram of your AM demodulation?please...or u > can give me the process...thx so much
Removing the carrier will leave you with the upper and lower sidebands that surround it, but no baseband. I did not write that you want to do that. An AM signal has nothing audible in its spectrum. Removing some of it doesn't make the rest audible. You are on the wrong track. You can look up what AM is more quickly than I can write a treatise on it for you, but I'll be happy to answer any questions. http://en.wikipedia.org/wiki/Amplitude_modulation One way to demodulate AM digitally: Assure that there is no DC component. Take the absolute magnitude of the signal. (This leaves a slowly varying part -- the recovered modulation -- and even harmonics of the carrier and sidebands.) Low-pass filter the result, leaving only the recovered modulation. Since the recovered modulation is far in frequency from the carrier with its sidebands, the transition band of the filter can be very broad. A single R-C suffices for analog receivers. I-Q demodulation is called for in the rare cases in which the carrier and modulating frequencies are not separated by a factor of 20 or more. That an entirely different subject. Jerry -- Engineering is the art of making what you want from things you can get. &macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;
On May 26, 12:24 pm, "c1910" <c_19...@hotmail.com> wrote:
> Clay, thanks for your opinion... > my problem is the mathematical equation for how the FIR filter to process > demodulation for AM. > oya, bout the windowing method, u tell me...i found the equation for > bandpass filter ideal. > i read a book, and i got the equation for unideal filter. is it h=w.hd? > h:transfer function unideal > w:windowing coef. > hd: transfer function ideal filter > > and do i have to make an ADAPTIVE filter for AM demodulation with the > windowing method? > > _____________________________________ > Do you know a company who employs DSP engineers? > Is it already listed athttp://dsprelated.com/employers.php?
Hello, Sorry I didn't answeer earlier - I was on a trip. For an example of a bandpass filter defined mathematically, you may look at my paper of raised cosine stuff. The formula is on page 3. http://www.claysturner.com/dsp/Raised%20Cosine%20and%20Root%20Raised%20Cosine%20Formulae.pdf Now with that being done, I don't think this will help you much. To extract the modulation from an AM signal, you will need more than just a filter. While bandpass filters are present in pretty much all receivers, the filters themselves do not do the demodulation. Jerry gave a common (in hardware for broadcast AM) method for demodulation. There are also many other ways to demodulate AM. One of the 1st questions is your AM single sideband, double sideband, vestigal sidband etc? Let us know more about your signal and what it is coming from, then we can point you towards a solution. Clay
Clay wrote:

> ... To > extract the modulation from an AM signal, you will need more than just > a filter. While bandpass filters are present in pretty much all > receivers, the filters themselves do not do the demodulation. Jerry > gave a common (in hardware for broadcast AM) method for demodulation. > There are also many other ways to demodulate AM. One of the 1st > questions is your AM single sideband, double sideband, vestigal > sidband etc?
Just a comment. When "AM" is mentioned, the plain double-sideband with carrier is usually assumed. Generally, SSB implies suppressed carrier, but SSSC makes that explicit.
> Let us know more about your signal and what it is coming from, then we > can point you towards a solution.
Indeed. Assumptions can be wrong, especially when made vy someone who doesn't know how to demodulate AM. Thanks for asking to make it explicit. Jerry -- Engineering is the art of making what you want from things you can get. &macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;
Jerry Avins wrote:

> ... Assumptions can be wrong, especially when made vy someone who > doesn't know how to demodulate AM. ...
Or by me. Jerry -- Engineering is the art of making what you want from things you can get. &macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;
On May 29, 5:38 pm, Jerry Avins <j...@ieee.org> wrote:
> Clay wrote: > > ... To > > extract the modulation from an AM signal, you will need more than just > > a filter. While bandpass filters are present in pretty much all > > receivers, the filters themselves do not do the demodulation. Jerry > > gave a common (in hardware for broadcast AM) method for demodulation. > > There are also many other ways to demodulate AM. One of the 1st > > questions is your AM single sideband, double sideband, vestigal > > sidband etc? > > Just a comment. When "AM" is mentioned, the plain double-sideband with > carrier is usually assumed. Generally, SSB implies suppressed carrier, > but SSSC makes that explicit. >
Hello Jerry, I recall back when I was in high school and had heard that AM was effected by multiplication and you get an upper sideband, lower sideband, and carrier. Of course when I looked at the standard trig identities, I saw how the upper and lower sidebands were created, but I couldn't see how some carrier was left over. I took me a a while to realize that the early radio designs were unable to multiply by negative numbers. So the modulating signal was offset so that the whole of the signal was now positive. And this gives "standard AM." Certainly for a long time AM referred to this offset signal - positive number only multiplication. Although TV has long been an exception to this rule. But the last 10 to 20 years have made more complicated forms of AM quite practical. Clay
Clay wrote:
> On May 29, 5:38 pm, Jerry Avins <j...@ieee.org> wrote: >> Clay wrote: >>> ... To >>> extract the modulation from an AM signal, you will need more than just >>> a filter. While bandpass filters are present in pretty much all >>> receivers, the filters themselves do not do the demodulation. Jerry >>> gave a common (in hardware for broadcast AM) method for demodulation. >>> There are also many other ways to demodulate AM. One of the 1st >>> questions is your AM single sideband, double sideband, vestigal >>> sidband etc? >> Just a comment. When "AM" is mentioned, the plain double-sideband with >> carrier is usually assumed. Generally, SSB implies suppressed carrier, >> but SSSC makes that explicit. >> > > Hello Jerry, > > I recall back when I was in high school and had heard that AM was > effected by multiplication and you get an upper sideband, lower > sideband, and carrier. Of course when I looked at the standard trig > identities, I saw how the upper and lower sidebands were created, but > I couldn't see how some carrier was left over. I took me a a while to > realize that the early radio designs were unable to multiply by > negative numbers. So the modulating signal was offset so that the > whole of the signal was now positive. And this gives "standard AM." > Certainly for a long time AM referred to this offset signal - positive > number only multiplication. Although TV has long been an exception to > this rule. But the last 10 to 20 years have made more complicated > forms of AM quite practical.
Clay, Sure they're practical. They were doable 50 years ago or more. (I once built an AM system using a balanced modulator to suppress the carrier just to demonstrate carrier phase reversal when the modulating signal went negative. Synchronous demodulation was easy because the original carrier was available.) Has the generally assumed meaning of uncharacterized "AM" changed while I wasn't looking? Or is "regular AM" a new and necessary retronym? Jerry -- Engineering is the art of making what you want from things you can get. &macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;