Hi All I need your help to understand a basic concepts while demodulating FM in digital domain. Do we need a digital PLL to lock the FM while demodulating complex baseband. I am more into DTV demodulator which needs carrier recovery to recover the signal band. Is this the case here or it can withstand carrier offsets? I am demodulating using phase differentiator in digital domain . Please share your thoughts? Thanks and Regards Saurabh
Carrier recovery for FM demodulation
Started by ●December 15, 2010
Reply by ●December 15, 20102010-12-15
On 12/15/2010 07:44 AM, shandilyasaurabh wrote:> Hi All > > I need your help to understand a basic concepts while demodulating FM in > digital domain. > > Do we need a digital PLL to lock the FM while demodulating complex > baseband. I am more into DTV demodulator which needs carrier recovery to > recover the signal band. Is this the case here or it can withstand carrier > offsets? I am demodulating using phase differentiator in digital domain . > > Please share your thoughts?FM doesn't need carrier recovery, and unless your system is esoteric it doesn't need any absolute frequency reference at all. You _can_ demodulate FM with a PLL, but I'm pretty sure that's not the best way to do it in the digital domain. -- Tim Wescott Wescott Design Services http://www.wescottdesign.com Do you need to implement control loops in software? "Applied Control Theory for Embedded Systems" was written for you. See details at http://www.wescottdesign.com/actfes/actfes.html
Reply by ●December 15, 20102010-12-15
On Dec 15, 12:40�pm, Tim Wescott <t...@seemywebsite.com> wrote:> On 12/15/2010 07:44 AM, shandilyasaurabh wrote: > ... > > Do we need a digital PLL to lock the FM while demodulating complex > > baseband. I am more into DTV demodulator which needs carrier recovery to > > recover the signal band. Is this the case here or it can withstand carrier > > offsets? I am demodulating using phase differentiator in digital domain . > > FM doesn't need carrier recovery, and unless your system is esoteric it > doesn't need any absolute frequency reference at all. �You _can_ > demodulate FM with a PLL, but I'm pretty sure that's not the best way to > do it in the digital domain.isn't the normal way is to sample the IF, create an I and Q signal pair, then compute the phase difference (which is proportional to the instantaneous frequency) between adjacent samples? i hadn't done it, but that's what i thought all of this I/Q software radio stuff was about. r b-j
Reply by ●December 15, 20102010-12-15
On 12/15/2010 10:10 AM, robert bristow-johnson wrote:> On Dec 15, 12:40 pm, Tim Wescott<t...@seemywebsite.com> wrote: >> On 12/15/2010 07:44 AM, shandilyasaurabh wrote: >> ... >>> Do we need a digital PLL to lock the FM while demodulating complex >>> baseband. I am more into DTV demodulator which needs carrier recovery to >>> recover the signal band. Is this the case here or it can withstand carrier >>> offsets? I am demodulating using phase differentiator in digital domain . >> >> FM doesn't need carrier recovery, and unless your system is esoteric it >> doesn't need any absolute frequency reference at all. You _can_ >> demodulate FM with a PLL, but I'm pretty sure that's not the best way to >> do it in the digital domain. > > isn't the normal way is to sample the IF, create an I and Q signal > pair, then compute the phase difference (which is proportional to the > instantaneous frequency) between adjacent samples? > > i hadn't done it, but that's what i thought all of this I/Q software > radio stuff was about.I think that's the usual way. Your _could_ do it with a digital PLL, I just don't think you'd _like_ it. -- Tim Wescott Wescott Design Services http://www.wescottdesign.com Do you need to implement control loops in software? "Applied Control Theory for Embedded Systems" was written for you. See details at http://www.wescottdesign.com/actfes/actfes.html
Reply by ●December 15, 20102010-12-15
On Dec 15, 1:49�pm, Tim Wescott <t...@seemywebsite.com> wrote:> On 12/15/2010 10:10 AM, robert bristow-johnson wrote: > > > > > On Dec 15, 12:40 pm, Tim Wescott<t...@seemywebsite.com> �wrote: > >> On 12/15/2010 07:44 AM, shandilyasaurabh wrote: > >> ... > >>> Do we need a digital PLL to lock the FM while demodulating complex > >>> baseband. I am more into DTV demodulator which needs carrier recovery to > >>> recover the signal band. Is this the case here or it can withstand carrier > >>> offsets? I am demodulating using phase differentiator in digital domain . > > >> FM doesn't need carrier recovery, and unless your system is esoteric it > >> doesn't need any absolute frequency reference at all. �You _can_ > >> demodulate FM with a PLL, but I'm pretty sure that's not the best way to > >> do it in the digital domain. > > > isn't the normal way is to sample the IF, create an I and Q signal > > pair, then compute the phase difference (which is proportional to the > > instantaneous frequency) between adjacent samples? > > > i hadn't done it, but that's what i thought all of this I/Q software > > radio stuff was about. > > I think that's the usual way. �Your _could_ do it with a digital PLL, I > just don't think you'd _like_ it.Another way is the quadrature detector. The basic principle works analog or digital. multiply the IF signal by a delayed copy of itself, Ideally a quarter cycle, hence the name. The low-pass filtered output is the demodulated signal. an odd number of quarter cycles increases the sensitivity. I'm sure the ARRL handbook shows this. (The demodulator works whatever the delay. The quarter-cycle delay eliminates DC. Jerry
Reply by ●December 16, 20102010-12-16
>On Dec 15, 1:49=A0pm, Tim Wescott <t...@seemywebsite.com> wrote: >> On 12/15/2010 10:10 AM, robert bristow-johnson wrote: >> >> >> >> > On Dec 15, 12:40 pm, Tim Wescott<t...@seemywebsite.com> =A0wrote: >> >> On 12/15/2010 07:44 AM, shandilyasaurabh wrote: >> >> ... >> >>> Do we need a digital PLL to lock the FM while demodulating complex >> >>> baseband. I am more into DTV demodulator which needs carrierrecovery=> to >> >>> recover the signal band. Is this the case here or it can withstandca=>rrier >> >>> offsets? I am demodulating using phase differentiator in digitaldoma=>in . >> >> >> FM doesn't need carrier recovery, and unless your system is esoterici=>t >> >> doesn't need any absolute frequency reference at all. =A0You _can_ >> >> demodulate FM with a PLL, but I'm pretty sure that's not the best way=>to >> >> do it in the digital domain. >> >> > isn't the normal way is to sample the IF, create an I and Q signal >> > pair, then compute the phase difference (which is proportional to the >> > instantaneous frequency) between adjacent samples? >> >> > i hadn't done it, but that's what i thought all of this I/Q software >> > radio stuff was about. >> >> I think that's the usual way. =A0Your _could_ do it with a digital PLL,I>> just don't think you'd _like_ it. > >Another way is the quadrature detector. The basic principle works >analog or digital. multiply the IF signal by a delayed copy of itself, >Ideally a quarter cycle, hence the name. The low-pass filtered output >is the demodulated signal. an odd number of quarter cycles increases >the sensitivity. I'm sure the ARRL handbook shows this. (The >demodulator works whatever the delay. The quarter-cycle delay >eliminates DC. > >Jerry >Hi All Thanks for the valuable suggestions. Hi Tim Thanks for the information. I understand we don't need carrier recovery but if band is not aligned to 0 Hz it results in DC component after differentiating phase of adjacent samples. Isn't it?. how to avoid it? Can we use a DC removal on demodulated signal? Also if we use low quality Tuner(carrier drift> 200KHz) there is a chance of missing a band itself and demodulating adjacent one? I am not using PLL method which is more popular in analog. But in digital DC component in demodulated signal is directly proportional to the carrier offset from 0 Hz after down conversion. I also tried using DPLL(2nd order PI) to lock to the carrier. Its not working properly as signal is a spread not a tone. Can you suggest something? Thanks and Regards Saurabh
Reply by ●December 16, 20102010-12-16
On 12/15/2010 08:10 PM, shandilyasaurabh wrote:>> On Dec 15, 1:49=A0pm, Tim Wescott<t...@seemywebsite.com> wrote: >>> On 12/15/2010 10:10 AM, robert bristow-johnson wrote: >>> >>> >>> >>>> On Dec 15, 12:40 pm, Tim Wescott<t...@seemywebsite.com> =A0wrote: >>>>> On 12/15/2010 07:44 AM, shandilyasaurabh wrote: >>>>> ... >>>>>> Do we need a digital PLL to lock the FM while demodulating complex >>>>>> baseband. I am more into DTV demodulator which needs carrier > recovery= >> to >>>>>> recover the signal band. Is this the case here or it can withstand > ca= >> rrier >>>>>> offsets? I am demodulating using phase differentiator in digital > doma= >> in . >>> >>>>> FM doesn't need carrier recovery, and unless your system is esoteric > i= >> t >>>>> doesn't need any absolute frequency reference at all. =A0You _can_ >>>>> demodulate FM with a PLL, but I'm pretty sure that's not the best way > = >> to >>>>> do it in the digital domain. >>> >>>> isn't the normal way is to sample the IF, create an I and Q signal >>>> pair, then compute the phase difference (which is proportional to the >>>> instantaneous frequency) between adjacent samples? >>> >>>> i hadn't done it, but that's what i thought all of this I/Q software >>>> radio stuff was about. >>> >>> I think that's the usual way. =A0Your _could_ do it with a digital PLL, > I >>> just don't think you'd _like_ it. >> >> Another way is the quadrature detector. The basic principle works >> analog or digital. multiply the IF signal by a delayed copy of itself, >> Ideally a quarter cycle, hence the name. The low-pass filtered output >> is the demodulated signal. an odd number of quarter cycles increases >> the sensitivity. I'm sure the ARRL handbook shows this. (The >> demodulator works whatever the delay. The quarter-cycle delay >> eliminates DC. >> >> Jerry >> > > Hi All > > Thanks for the valuable suggestions. > > Hi Tim > > Thanks for the information. I understand we don't need carrier recovery but > if band is not aligned to 0 Hz it results in DC component after > differentiating phase of adjacent samples. Isn't it?. how to avoid it? Can > we use a DC removal on demodulated signal?What is the signal that you're trying to receive? If it's audio you can remove the DC bias; as long as you stay in the linear operating region of your demodulation algorithm then you should be OK.> Also if we use low quality Tuner(carrier drift> 200KHz) there is a chance > of missing a band itself and demodulating adjacent one?Always, with any channelized service, if your receiver oscillator drifts you can miss your channel.> I am not using PLL method which is more popular in analog. But in digital > DC component in demodulated signal is directly proportional to the carrier > offset from 0 Hz after down conversion. > > > I also tried using DPLL(2nd order PI) to lock to the carrier. Its not > working properly as signal is a spread not a tone. Can you suggest > something?I'm not sure what you're trying to achieve -- is it something like automatic frequency control? I.e., are you trying to pull the receiver center frequency to match the transmitted signal? If that's the case you can look at the bias of your frequency detector and feed it back to your local oscillator. -- Tim Wescott Wescott Design Services http://www.wescottdesign.com Do you need to implement control loops in software? "Applied Control Theory for Embedded Systems" was written for you. See details at http://www.wescottdesign.com/actfes/actfes.html
Reply by ●December 16, 20102010-12-16
>On 12/15/2010 08:10 PM, shandilyasaurabh wrote: >>> On Dec 15, 1:49=A0pm, Tim Wescott<t...@seemywebsite.com> wrote: >>>> On 12/15/2010 10:10 AM, robert bristow-johnson wrote: >>>> >>>> >>>> >>>>> On Dec 15, 12:40 pm, Tim Wescott<t...@seemywebsite.com> =A0wrote: >>>>>> On 12/15/2010 07:44 AM, shandilyasaurabh wrote: >>>>>> ... >>>>>>> Do we need a digital PLL to lock the FM while demodulating complex >>>>>>> baseband. I am more into DTV demodulator which needs carrier >> recovery= >>> to >>>>>>> recover the signal band. Is this the case here or it can withstand >> ca= >>> rrier >>>>>>> offsets? I am demodulating using phase differentiator in digital >> doma= >>> in . >>>> >>>>>> FM doesn't need carrier recovery, and unless your system isesoteric>> i= >>> t >>>>>> doesn't need any absolute frequency reference at all. =A0You _can_ >>>>>> demodulate FM with a PLL, but I'm pretty sure that's not the bestway>> = >>> to >>>>>> do it in the digital domain. >>>> >>>>> isn't the normal way is to sample the IF, create an I and Q signal >>>>> pair, then compute the phase difference (which is proportional tothe>>>>> instantaneous frequency) between adjacent samples? >>>> >>>>> i hadn't done it, but that's what i thought all of this I/Q software >>>>> radio stuff was about. >>>> >>>> I think that's the usual way. =A0Your _could_ do it with a digitalPLL,>> I >>>> just don't think you'd _like_ it. >>> >>> Another way is the quadrature detector. The basic principle works >>> analog or digital. multiply the IF signal by a delayed copy of itself, >>> Ideally a quarter cycle, hence the name. The low-pass filtered output >>> is the demodulated signal. an odd number of quarter cycles increases >>> the sensitivity. I'm sure the ARRL handbook shows this. (The >>> demodulator works whatever the delay. The quarter-cycle delay >>> eliminates DC. >>> >>> Jerry >>> >> >> Hi All >> >> Thanks for the valuable suggestions. >> >> Hi Tim >> >> Thanks for the information. I understand we don't need carrier recoverybut>> if band is not aligned to 0 Hz it results in DC component after >> differentiating phase of adjacent samples. Isn't it?. how to avoid it?Can>> we use a DC removal on demodulated signal? > >What is the signal that you're trying to receive? If it's audio you can >remove the DC bias; as long as you stay in the linear operating region >of your demodulation algorithm then you should be OK. > >> Also if we use low quality Tuner(carrier drift> 200KHz) there is achance>> of missing a band itself and demodulating adjacent one? > >Always, with any channelized service, if your receiver oscillator drifts >you can miss your channel. > >> I am not using PLL method which is more popular in analog. But indigital>> DC component in demodulated signal is directly proportional to thecarrier>> offset from 0 Hz after down conversion. >> >> >> I also tried using DPLL(2nd order PI) to lock to the carrier. Its not >> working properly as signal is a spread not a tone. Can you suggest >> something? > >I'm not sure what you're trying to achieve -- is it something like >automatic frequency control? I.e., are you trying to pull the receiver >center frequency to match the transmitted signal? If that's the case >you can look at the bias of your frequency detector and feed it back to >your local oscillator. > >-- > >Tim Wescott >Wescott Design Services >http://www.wescottdesign.com > >Do you need to implement control loops in software? >"Applied Control Theory for Embedded Systems" was written for you. >See details at http://www.wescottdesign.com/actfes/actfes.html >Hi Tim Thanks for you reply. I want to improve Quality(SNR) of demodulated FM signal. As explained earlier if FM band is not aligned to 0Hz it results in DC component after demodulation. I thought of using a 2nd order Digital PLL with integral and proportional control. My Idea is to lock to the center i.e carrier of the FM band. It will make sure there are no DC component in the demodulated FM. Digital PLL variance is too much as carrier frequency is changing (Definition of FM). hence I am not able to lock to the FM band. I am using algorithm mentioned in Matlab as below, y = IF Signal Fs = Sampling Freq Fc = FM center Freq t = (0:1/Fs:((size(y,1)-1)/Fs))'; t = t(:,ones(1,size(y,2))); yq = hilbert(y).*exp(-j*2*pi*Fc*t); z = (1/(2*pi*freqdev))*[zeros(1,size(yq,2)); diff(unwrap(angle(yq)))*Fs]; Share your thoughts if I am wrong somewhere? Thanks and Regards Saurabh
Reply by ●December 16, 20102010-12-16
> Hi Tim > > Thanks for you reply. > > I want to improve Quality(SNR) of demodulated FM signal. As explained > earlier if FM band is not aligned to 0Hz it results in DC component after > demodulation. > > I thought of using a 2nd order Digital PLL with integral and proportional > control. My Idea is to lock to the center i.e carrier of the FM band. It > will make sure there are no DC component in the demodulated FM. > > Digital PLL variance is too much as carrier frequency is changing > (Definition of FM). hence I am not able to lock to the FM band. > > I am using algorithm mentioned in Matlab as below, > y =A0=3D IF Signal > Fs =3D Sampling Freq > Fc =3D FM center Freq > t =3D (0:1/Fs:((size(y,1)-1)/Fs))'; > t =3D t(:,ones(1,size(y,2))); > > yq =3D hilbert(y).*exp(-j*2*pi*Fc*t); > z =3D (1/(2*pi*freqdev))*[zeros(1,size(yq,2)); diff(unwrap(angle(yq)))*Fs=];> > Share your thoughts if I am wrong somewhere? > > Thanks and Regards > Saurabh- Hide quoted text - > > - Show quoted text -Saurabh, I'm just curious... are these questions about a project you are working on as a student or hobby or professional for which you are being paid? is this for a consumer product? what country are you in? thanks Mark
Reply by ●December 16, 20102010-12-16
> Hi Tim > > Thanks for you reply. > > I want to improve Quality(SNR) of demodulated FM signal. As explained > earlier if FM band is not aligned to 0Hz it results in DC component after > demodulation. > > I thought of using a 2nd order Digital PLL with integral and proportional > control. My Idea is to lock to the center i.e carrier of the FM band. It > will make sure there are no DC component in the demodulated FM. > > Digital PLL variance is too much as carrier frequency is changing > (Definition of FM). hence I am not able to lock to the FM band. > > I am using algorithm mentioned in Matlab as below, > y �= IF Signal > Fs = Sampling Freq > Fc = FM center Freq > t = (0:1/Fs:((size(y,1)-1)/Fs))'; > t = t(:,ones(1,size(y,2))); > > yq = hilbert(y).*exp(-j*2*pi*Fc*t); > z = (1/(2*pi*freqdev))*[zeros(1,size(yq,2)); diff(unwrap(angle(yq)))*Fs]; > > Share your thoughts if I am wrong somewhere? > > Thanks and Regards > Saurabh- Hide quoted text - > > - Show quoted text -Saurabh, I'm just curious... are these questions about a project you are working on as a student or hobby or professional for which you are being paid? is this for a consumer product? what country are you in? thanks Mark
