DSPRelated.com
Forums

fsk demodulation

Started by harsh17 July 17, 2005
I am a novice in DSP.I am trying to demodulate a FSK signal wherein the
mark and space frequencies are 16MHz and 24MHz respectively.I am thinking
of using delay and multiply method.Can this be implemented using an
ADSP-2181?More specifically will the DSP be able to handle the high
throughput involved?Please correct me if I am conceptually going wrong
somewhere.

		
This message was sent using the Comp.DSP web interface on
www.DSPRelated.com
harsh17 wrote:
> I am a novice in DSP.I am trying to demodulate a FSK signal wherein the > mark and space frequencies are 16MHz and 24MHz respectively.I am thinking > of using delay and multiply method.Can this be implemented using an > ADSP-2181?More specifically will the DSP be able to handle the high > throughput involved?Please correct me if I am conceptually going wrong > somewhere.
Those frequencies, being harmonically related, are a poor choice for digital or any other kind of signal processing. There are other aspects of the choice that seem strange to me. What system requirements led you to choose them? Jerry -- Engineering is the art of making what you want from things you can get. �����������������������������������������������������������������������
"Jerry Avins" <jya@ieee.org> wrote in message 
news:H8ydndzDZYLQGUffRVn-hg@rcn.net...
> harsh17 wrote: >> I am a novice in DSP.I am trying to demodulate a FSK >> signal wherein the >> mark and space frequencies are 16MHz and 24MHz >> respectively.I am thinking >> of using delay and multiply method.Can this be >> implemented using an >> ADSP-2181?More specifically will the DSP be able to >> handle the high >> throughput involved?Please correct me if I am >> conceptually going wrong >> somewhere. > > Those frequencies, being harmonically related, are a poor > choice for digital or any other kind of signal processing.
Would you be willing to elaborate on this, please? I don't doubt that what you say is probably true; I just don't understand the connections.
John E. Hadstate wrote:
> "Jerry Avins" <jya@ieee.org> wrote in message > news:H8ydndzDZYLQGUffRVn-hg@rcn.net... > >>harsh17 wrote: >> >>>I am a novice in DSP.I am trying to demodulate a FSK >>>signal wherein the >>>mark and space frequencies are 16MHz and 24MHz >>>respectively.I am thinking >>>of using delay and multiply method.Can this be >>>implemented using an >>>ADSP-2181?More specifically will the DSP be able to >>>handle the high >>>throughput involved?Please correct me if I am >>>conceptually going wrong >>>somewhere. >> >>Those frequencies, being harmonically related, are a poor >>choice for digital or any other kind of signal processing. > > > Would you be willing to elaborate on this, please? I don't > doubt that what you say is probably true; I just don't > understand the connections.
Both frequencies are multiples of 8 MHz. Various nonlinearities in transmission might make it difficult to separate them. Frequencies used for FSK usually have a small GCD (greatest common divisor). Jerry -- Engineering is the art of making what you want from things you can get. &#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;
"Jerry Avins" <jya@ieee.org> wrote in message 
news:qM2dnQ-TJO_XMkffRVn-pg@rcn.net...
> John E. Hadstate wrote: >> "Jerry Avins" <jya@ieee.org> wrote in message >> news:H8ydndzDZYLQGUffRVn-hg@rcn.net... >> >>>harsh17 wrote: >>> >>>>I am a novice in DSP.I am trying to demodulate a FSK >>>>signal wherein the >>>>mark and space frequencies are 16MHz and 24MHz >>>>respectively.I am thinking >>>>of using delay and multiply method.Can this be >>>>implemented using an >>>>ADSP-2181?More specifically will the DSP be able to >>>>handle the high >>>>throughput involved?Please correct me if I am >>>>conceptually going wrong >>>>somewhere. >>> >>>Those frequencies, being harmonically related, are a poor >>>choice for digital or any other kind of signal >>>processing. >> >> >> Would you be willing to elaborate on this, please? I >> don't doubt that what you say is probably true; I just >> don't understand the connections. > > Both frequencies are multiples of 8 MHz. Various > nonlinearities in transmission might make it difficult to > separate them. Frequencies used for FSK usually have a > small GCD (greatest common divisor). >
Thanks. In fact, I have seen a problem related to exactly the non-linear effects you are describing.
>harsh17 wrote: >> I am a novice in DSP.I am trying to demodulate a FSK signal wherein
the
>> mark and space frequencies are 16MHz and 24MHz respectively.I am
thinking
>> of using delay and multiply method.Can this be implemented using an >> ADSP-2181?More specifically will the DSP be able to handle the high >> throughput involved?Please correct me if I am conceptually going wrong >> somewhere. > >Those frequencies, being harmonically related, are a poor choice for >digital or any other kind of signal processing. There are other aspects >of the choice that seem strange to me. What system requirements led you >to choose them? > >Jerry >-- >Engineering is the art of making what you want from things you can get. >&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295; >
Hi Jerry, The data is to be transmitted at 16MBPS but bandwidth available is just 8MHz.Each bit consists of either 1.5 cycles of 24 MHz or a single cycle of 16MHz.To achieve phase continuity, I have to use two versions of 16MHz each 180 deg out of phase with each other, and two versions of 24MHz each 180 deg out of phase with each other.This enables me to transmit at 16mbps with a bandwidth of 8MHz.Hence the selection of frequencies has to be such that they are multiples of 8 and have a difference=8.Can u suggest an alternative? I do get your point regarding non-linearity. Regards, Harshad This message was sent using the Comp.DSP web interface on www.DSPRelated.com
harsh17 wrote:
>>harsh17 wrote: >> >>>I am a novice in DSP.I am trying to demodulate a FSK signal wherein > > the > >>>mark and space frequencies are 16MHz and 24MHz respectively.I am > > thinking > >>>of using delay and multiply method.Can this be implemented using an >>>ADSP-2181?More specifically will the DSP be able to handle the high >>>throughput involved?Please correct me if I am conceptually going wrong >>>somewhere. >> >>Those frequencies, being harmonically related, are a poor choice for >>digital or any other kind of signal processing. There are other aspects >>of the choice that seem strange to me. What system requirements led you >>to choose them? >> >>Jerry >>-- >>Engineering is the art of making what you want from things you can get. >>&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295; >> > > > Hi Jerry, > The data is to be transmitted at 16MBPS but bandwidth available > is just 8MHz.Each bit consists of either 1.5 cycles of 24 MHz or a single > cycle of 16MHz.To achieve phase continuity, I have to use two versions of > 16MHz each 180 deg out of phase with each other, and two versions of 24MHz > each 180 deg out of phase with each other.This enables me to transmit at > 16mbps with a bandwidth of 8MHz.Hence the selection of frequencies has to > be such that they are multiples of 8 and have a difference=8.Can u suggest > an alternative? I do get your point regarding non-linearity.
Phase continuity is easily achieved by starting each new frequency at the amplitude where the other left off. If you used 25 MHz and 16 MHz, for example, there would still be no need for discontinuity. One way to do that is by using what is called a "phase accumulator". The method uses a look-up table of sine values that is jumped through by a pointer. Frequency is varied by changing the stride of the pointer. I hope you don't expect the frequency deviation to resemble the signal's bandwidth. The actual bandwidth will be much larger than that. FM has sidebands that extend well beyond the deviation, continuous phase notwithstanding. Jerry -- Engineering is the art of making what you want from things you can get. &#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;
Jerry Avins wrote:

(snip)

>>> harsh17 wrote:
>>>> I am a novice in DSP.I am trying to demodulate a FSK signal wherein the >>>> mark and space frequencies are 16MHz and 24MHz respectively.I am >>>> thinking
(snip)
>>> Those frequencies, being harmonically related, are a poor choice for >>> digital or any other kind of signal processing.
(snip)
> Both frequencies are multiples of 8 MHz. Various nonlinearities in > transmission might make it difficult to separate them. Frequencies used > for FSK usually have a small GCD (greatest common divisor).
He didn't say what the modulation rate was, which could be important. I still remember the system used to record data on cassette tapes with early microcomputers which used either one or two cycles of some frequency for each bit. If you synchronously decode it then maybe it is fine, or even better that way. If, say, you wanted 8 MHz or even 16 MHz modulation, though it might be that in that case PSK would be a better choice. -- glen