Naebad wrote:> ... Why go digital when there is a perfectly cheap analogue solution?Digital is sexy. Jerry -- Engineering is the art of making what you want from things you can get. �����������������������������������������������������������������������
Better FM demodulator
Started by ●January 14, 2006
Reply by ●January 15, 20062006-01-15
Reply by ●January 15, 20062006-01-15
"Leon" <leon_heller@hotmail.com> wrote in message news:1137333930.818314.193610@g47g2000cwa.googlegroups.com...> > > That's what the DAB radios getting popular here in the UK do. My BT > Aviator radio has a Blackfin DSP and can process analogue FM as well as > the digital broadcasts. > > Leon >Execpt for one thing - ye canna re-write the laws of Physics captain! Making it digital does not remove all the fundamental problems.I have heard that DAB in cars still suffers from multipath so it's no better than analogue FM. What would be good is Internet radio in your car so I could listen to the BBC in any country on the move. Most FM radio here in NZ is wall to wall adverts and rap music.Hardly any good for the thinking mind! Naebad
Reply by ●January 16, 20062006-01-16
"Vladimir Vassilevsky" <antispam_bogus@hotmail.com> wrote in message news:fMuyf.270$F01.184@newssvr27.news.prodigy.net...> > > Steve Underwood wrote: > > > >>>> Currently, I have a project that uses the Blackfin DSP. A 8-bit A/D > >>>> will > >>>> do undersampling (at 1.6384MHz) of an IF signal which isdown-converter> >>>> from standard stereo FM broadcast. > >>> > >>> This is pointless because the complete analog FM receiver IC costsabout> >>> 50c and the DSP solution is not going to be any better. > >> > >> I know you're a jerk but perhaps "project" means something to do for > >> school, for learning, for fun? > > > > Perhaps project means for professional use too. A digital radio already > > has a powerful DSP. When the radio is switched to receiving analogue FM > > it makes perfect sense to use the DSP. It is cheaper, technically as > > good, and "fully digital receiver" has immense marketing power. > > > > Instead of continuation of bullshit, here is the formula: > > Vfm = (I*dQ/dt - Q*dI/dt)/(I^2 + Q^2) > > Behold the FM discriminator. It is a bit more complicated then 2+2 = 4.Yes - it's the differential of Q/I. High level stuff eh? However, it doesn't improve on analogue demodulation - it's just the equivalent so why bother? This sort of receiver is still suseptable to Multipath,Thresholding etc - all the usual fundamental problems that a PLL has and it needs much more hardware. Naebad
Reply by ●January 16, 20062006-01-16
This is a multi-part message in MIME format. --------------020608030004020908000500 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Naebad wrote:>"Vladimir Vassilevsky" <antispam_bogus@hotmail.com> wrote in message >news:fMuyf.270$F01.184@newssvr27.news.prodigy.net... > > >>Steve Underwood wrote: >> >> >> >> >>>>>>Currently, I have a project that uses the Blackfin DSP. A 8-bit A/D >>>>>>will >>>>>>do undersampling (at 1.6384MHz) of an IF signal which is >>>>>> >>>>>> >down-converter > > >>>>>>from standard stereo FM broadcast. >>>>>> >>>>>> >>>>>This is pointless because the complete analog FM receiver IC costs >>>>> >>>>> >about > > >>>>>50c and the DSP solution is not going to be any better. >>>>> >>>>> >>>>I know you're a jerk but perhaps "project" means something to do for >>>>school, for learning, for fun? >>>> >>>> >>>Perhaps project means for professional use too. A digital radio already >>>has a powerful DSP. When the radio is switched to receiving analogue FM >>>it makes perfect sense to use the DSP. It is cheaper, technically as >>>good, and "fully digital receiver" has immense marketing power. >>> >>> >>> >>Instead of continuation of bullshit, here is the formula: >> >>Vfm = (I*dQ/dt - Q*dI/dt)/(I^2 + Q^2) >> >>Behold the FM discriminator. It is a bit more complicated then 2+2 = 4. >> >> > >Yes - it's the differential of Q/I. High level stuff eh? However, it doesn't >improve on analogue demodulation - it's just the equivalent so why bother? >This sort of receiver is still suseptable to Multipath,Thresholding etc - >all the usual fundamental problems that a PLL has and it needs much more >hardware. > >Naebad > >Answer 1 - the sane engineering answer ---------------------------------------- Because it can be cheaper. If you are building a DAB radio you will implement the bulk in DSP. When it switches to the analogue FM channels it can continue to use a DSP method at zero cost. As you and Vladimir point out, demodulating FM in DSP is so simple, it is highly unlikely space cannot be made for the relevant code in some odd corner of flash. It is literally a zero cost solution, trumping even the cheapest analogue FM chip. Answer 2 - the marketing answer ---------------------------------- Lets say you have some DSP capacity available, but need to do a little to plumb the FM front end to that capacity (e.g. a cell phone). If it is not too expensive, the DSP solution is technically better. That's because technically the whole point is to make meney, and the fully digital solution will attract a higher price at the retail outlet. :-) Regards, Steve --------------020608030004020908000500 Content-Type: text/html; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit <!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN"> <html> <head> <meta content="text/html;charset=ISO-8859-1" http-equiv="Content-Type"> <title></title> </head> <body bgcolor="#ffffff" text="#000000"> Naebad wrote: <blockquote cite="midrQHyf.15292$vH5.806810@news.xtra.co.nz" type="cite"> <pre wrap="">"Vladimir Vassilevsky" <a class="moz-txt-link-rfc2396E" href="mailto:antispam_bogus@hotmail.com"><antispam_bogus@hotmail.com></a> wrote in message <a class="moz-txt-link-freetext" href="news:fMuyf.270$F01.184@newssvr27.news.prodigy.net">news:fMuyf.270$F01.184@newssvr27.news.prodigy.net</a>... </pre> <blockquote type="cite"> <pre wrap=""> Steve Underwood wrote: </pre> <blockquote type="cite"> <blockquote type="cite"> <blockquote type="cite"> <blockquote type="cite"> <pre wrap="">Currently, I have a project that uses the Blackfin DSP. A 8-bit A/D will do undersampling (at 1.6384MHz) of an IF signal which is </pre> </blockquote> </blockquote> </blockquote> </blockquote> </blockquote> <pre wrap=""><!---->down-converter </pre> <blockquote type="cite"> <blockquote type="cite"> <blockquote type="cite"> <blockquote type="cite"> <blockquote type="cite"> <pre wrap="">from standard stereo FM broadcast. </pre> </blockquote> <pre wrap="">This is pointless because the complete analog FM receiver IC costs </pre> </blockquote> </blockquote> </blockquote> </blockquote> <pre wrap=""><!---->about </pre> <blockquote type="cite"> <blockquote type="cite"> <blockquote type="cite"> <blockquote type="cite"> <pre wrap="">50c and the DSP solution is not going to be any better. </pre> </blockquote> <pre wrap="">I know you're a jerk but perhaps "project" means something to do for school, for learning, for fun? </pre> </blockquote> <pre wrap="">Perhaps project means for professional use too. A digital radio already has a powerful DSP. When the radio is switched to receiving analogue FM it makes perfect sense to use the DSP. It is cheaper, technically as good, and "fully digital receiver" has immense marketing power. </pre> </blockquote> <pre wrap="">Instead of continuation of bullshit, here is the formula: Vfm = (I*dQ/dt - Q*dI/dt)/(I^2 + Q^2) Behold the FM discriminator. It is a bit more complicated then 2+2 = 4. </pre> </blockquote> <pre wrap=""><!----> Yes - it's the differential of Q/I. High level stuff eh? However, it doesn't improve on analogue demodulation - it's just the equivalent so why bother? This sort of receiver is still suseptable to Multipath,Thresholding etc - all the usual fundamental problems that a PLL has and it needs much more hardware. Naebad </pre> </blockquote> Answer 1 - the sane engineering answer<br> ----------------------------------------<br> Because it can be cheaper. If you are building a DAB radio you will implement the bulk in DSP. When it switches to the analogue FM channels it can continue to use a DSP method at zero cost. As you and Vladimir point out, demodulating FM in DSP is so simple, it is highly unlikely space cannot be made for the relevant code in some odd corner of flash. It is literally a zero cost solution, trumping even the cheapest analogue FM chip.<br> <br> Answer 2 - the marketing answer<br> ----------------------------------<br> Lets say you have some DSP capacity available, but need to do a little to plumb the FM front end to that capacity (e.g. a cell phone). If it is not too expensive, the DSP solution is technically better. That's because technically the whole point is to make meney, and the fully digital solution will attract a higher price at the retail outlet. :-)<br> <br> Regards,<br> Steve<br> <br> </body> </html> --------------020608030004020908000500--
Reply by ●January 16, 20062006-01-16
This is a multi-part message in MIME format. ------=_NextPart_000_001F_01C61AF7.4045FA50 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable "Steve Underwood" <steveu@dis.org> wrote in message = news:dqfidk$jrv$1@home.itg.ti.com... Naebad wrote:=20 "Vladimir Vassilevsky" <antispam_bogus@hotmail.com> wrote in message news:fMuyf.270$F01.184@newssvr27.news.prodigy.net... Steve Underwood wrote: Currently, I have a project that uses the Blackfin DSP. A 8-bit A/D will do undersampling (at 1.6384MHz) of an IF signal which is down-converter from standard stereo FM broadcast. This is pointless because the complete analog FM receiver IC = costs about 50c and the DSP solution is not going to be any better. I know you're a jerk but perhaps "project" means something to = do for school, for learning, for fun? Perhaps project means for professional use too. A digital radio = already has a powerful DSP. When the radio is switched to receiving analogue FM it makes perfect sense to use the DSP. It is cheaper, technically as good, and "fully digital receiver" has immense marketing power. Instead of continuation of bullshit, here is the formula: Vfm =3D (I*dQ/dt - Q*dI/dt)/(I^2 + Q^2) Behold the FM discriminator. It is a bit more complicated then 2+2 =3D = 4. =20 Yes - it's the differential of Q/I. High level stuff eh? However, it = doesn't improve on analogue demodulation - it's just the equivalent so why = bother? This sort of receiver is still suseptable to Multipath,Thresholding etc = - all the usual fundamental problems that a PLL has and it needs much more hardware. Naebad Answer 1 - the sane engineering answer ---------------------------------------- Because it can be cheaper. If you are building a DAB radio you will = implement the bulk in DSP. When it switches to the analogue FM channels = it can continue to use a DSP method at zero cost. As you and Vladimir = point out, demodulating FM in DSP is so simple, it is highly unlikely = space cannot be made for the relevant code in some odd corner of flash. = It is literally a zero cost solution, trumping even the cheapest = analogue FM chip. Answer 2 - the marketing answer ---------------------------------- Lets say you have some DSP capacity available, but need to do a little = to plumb the FM front end to that capacity (e.g. a cell phone). If it is = not too expensive, the DSP solution is technically better. That's = because technically the whole point is to make meney, and the fully = digital solution will attract a higher price at the retail outlet. :-) Regards, Steve But such a chip will be mixed signal - it still needs an analogue = front end including possibly AGC. I am all for DSP but not re-inventing = the wheel. Yes if you have DAB in a home then FM demod is no overhead = but I have yet to see a detailed dB comparison between ordinary analogue = PLLs and I/Q demodulation to see which if any is better. Naebad ------=_NextPart_000_001F_01C61AF7.4045FA50 Content-Type: text/html; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN"> <HTML><HEAD><TITLE></TITLE> <META http-equiv=3DContent-Type = content=3Dtext/html;charset=3DISO-8859-1> <META content=3D"MSHTML 6.00.2800.1106" name=3DGENERATOR> <STYLE></STYLE> </HEAD> <BODY text=3D#000000 bgColor=3D#ffffff> <DIV><FONT face=3DArial size=3D2></FONT> </DIV> <BLOCKQUOTE=20 style=3D"PADDING-RIGHT: 0px; PADDING-LEFT: 5px; MARGIN-LEFT: 5px; = BORDER-LEFT: #000000 2px solid; MARGIN-RIGHT: 0px"> <DIV>"Steve Underwood" <<A=20 href=3D"mailto:steveu@dis.org">steveu@dis.org</A>> wrote in message = <A=20 = href=3D"news:dqfidk$jrv$1@home.itg.ti.com">news:dqfidk$jrv$1@home.itg.ti.= com</A>...</DIV>Naebad=20 wrote:=20 <BLOCKQUOTE cite=3DmidrQHyf.15292$vH5.806810@news.xtra.co.nz = type=3D"cite"><PRE wrap=3D"">"Vladimir Vassilevsky" <A = class=3Dmoz-txt-link-rfc2396E = href=3D"mailto:antispam_bogus@hotmail.com"><antispam_bogus@hotmail.com= ></A> wrote in message <A class=3Dmoz-txt-link-freetext = href=3D"news:fMuyf.270$F01.184@newssvr27.news.prodigy.net">news:fMuyf.270= $F01.184@newssvr27.news.prodigy.net</A>... </PRE> <BLOCKQUOTE type=3D"cite"><PRE wrap=3D"">Steve Underwood wrote: </PRE> <BLOCKQUOTE type=3D"cite"> <BLOCKQUOTE type=3D"cite"> <BLOCKQUOTE type=3D"cite"> <BLOCKQUOTE type=3D"cite"><PRE wrap=3D"">Currently, I have a = project that uses the Blackfin DSP. A 8-bit A/D will do undersampling (at 1.6384MHz) of an IF signal which is = </PRE></BLOCKQUOTE></BLOCKQUOTE></BLOCKQUOTE></BLOCKQUOTE></BLOCKQUOTE><P= RE wrap=3D""><!---->down-converter </PRE> <BLOCKQUOTE type=3D"cite"> <BLOCKQUOTE type=3D"cite"> <BLOCKQUOTE type=3D"cite"> <BLOCKQUOTE type=3D"cite"> <BLOCKQUOTE type=3D"cite"><PRE wrap=3D"">from standard = stereo FM broadcast. </PRE></BLOCKQUOTE><PRE wrap=3D"">This is pointless because = the complete analog FM receiver IC costs </PRE></BLOCKQUOTE></BLOCKQUOTE></BLOCKQUOTE></BLOCKQUOTE><PRE = wrap=3D""><!---->about </PRE> <BLOCKQUOTE type=3D"cite"> <BLOCKQUOTE type=3D"cite"> <BLOCKQUOTE type=3D"cite"> <BLOCKQUOTE type=3D"cite"><PRE wrap=3D"">50c and the DSP = solution is not going to be any better. </PRE></BLOCKQUOTE><PRE wrap=3D"">I know you're a jerk but = perhaps "project" means something to do for school, for learning, for fun? </PRE></BLOCKQUOTE><PRE wrap=3D"">Perhaps project means for = professional use too. A digital radio already has a powerful DSP. When the radio is switched to receiving analogue FM it makes perfect sense to use the DSP. It is cheaper, technically as good, and "fully digital receiver" has immense marketing power. </PRE></BLOCKQUOTE><PRE wrap=3D"">Instead of continuation of = bullshit, here is the formula: Vfm =3D (I*dQ/dt - Q*dI/dt)/(I^2 + Q^2) Behold the FM discriminator. It is a bit more complicated then 2+2 =3D = 4. </PRE></BLOCKQUOTE><PRE wrap=3D""><!----> Yes - it's the differential of Q/I. High level stuff eh? However, it = doesn't improve on analogue demodulation - it's just the equivalent so why = bother? This sort of receiver is still suseptable to Multipath,Thresholding etc = - all the usual fundamental problems that a PLL has and it needs much more hardware. Naebad </PRE></BLOCKQUOTE> <DIV>Answer 1 - the sane engineering=20 answer<BR>----------------------------------------<BR>Because it can = be=20 cheaper. If you are building a DAB radio you will implement the bulk = in DSP.=20 When it switches to the analogue FM channels it can continue to use a = DSP=20 method at zero cost. As you and Vladimir point out, demodulating FM in = DSP is=20 so simple, it is highly unlikely space cannot be made for the relevant = code in=20 some odd corner of flash. It is literally a zero cost solution, = trumping even=20 the cheapest analogue FM chip.<BR><BR>Answer 2 - the marketing=20 answer<BR>----------------------------------<BR>Lets say you have some = DSP=20 capacity available, but need to do a little to plumb the FM front end = to that=20 capacity (e.g. a cell phone). If it is not too expensive, the DSP = solution is=20 technically better. That's because technically the whole point is to = make=20 meney, and the fully digital solution will attract a higher price at = the=20 retail outlet. :-)<BR><BR>Regards,<BR>Steve<BR><BR></DIV> <DIV><FONT face=3DArial size=3D2>But such a chip will be mixed signal = - it still=20 needs an analogue front end including possibly AGC. I am all = for DSP=20 but not re-inventing the wheel. Yes if you have DAB in a home then FM = demod is=20 no overhead but I have yet to see a detailed dB comparison between = ordinary=20 analogue PLLs and I/Q demodulation to see which if any is = better.</FONT></DIV> <DIV><FONT face=3DArial size=3D2></FONT> </DIV> <DIV><FONT face=3DArial size=3D2>Naebad</FONT></DIV> <DIV><FONT face=3DArial = size=3D2></FONT> </DIV></BLOCKQUOTE></BODY></HTML> ------=_NextPart_000_001F_01C61AF7.4045FA50--
Reply by ●January 16, 20062006-01-16
Vladimir Vassilevsky wrote:> This is pointless because the complete analog FM receiver IC costs about > 50c and the DSP solution is not going to be any better.I have to disagree. An analog PLL demodulator is limited by the linearity of the VCO. Nonlinearity in the freq->voltage response here causes harmonics, intermodulation and all sorts of other bad things in the recovered output. When L+R and L-R start interfering with each other due to these effects in stereo mode, things go bad really quickly. Then comes stereo demux, which requires precise 19KHz recovery and 38KHz synthesis, and sharp filtering of the resulting L+R/L-R. Maintaining a flat frequency response up to 15KHz and good pilot rejection is easy with a good DSP FIR design, but is a nightmare with cascaded analog filters - and I can't imagine anything better than a 4th order lowpass in a cheap FM demod chip. Overall, I'll need some strong convincing that a 50 cent chip will do an equal or better job than an RF/IF-input, fully digital demodulator. And sure, an incredibly high quality FM demodulator is still limited by the quality of the FM that's going into it, there's not much you can do about multipath, lame FM morning shows or jessica simpson. But "radio sucks" is no reason to attempt to build a better, high quality receiver. And there's certainly a market for high quality radio receivers - companies such as Belar (who make super high quality AM/FM demod equipment for broadcast use) haven't exactly been run out of business by $10 radios from Walmart. I don't know what chris1911's motives are for attempting FM demodulation in the Blackfin, but what's the harm? Best case, he might build a digital FM demodulator with outstanding performance. Worst case, it might not work out and he'll only learn all sorts of DSP theory and practice which could lead him into a great career doing DSP. So hold off on the discouragement for a while. -GM
Reply by ●January 16, 20062006-01-16
"Gary Marsh" <dont@spam.me.plz> wrote in message news:imRyf.94952$6K2.92898@edtnps90...> Vladimir Vassilevsky wrote: > > This is pointless because the complete analog FM receiver IC costs about > > 50c and the DSP solution is not going to be any better. > > I have to disagree. > > An analog PLL demodulator is limited by the linearity of the VCO. > Nonlinearity in the freq->voltage response here causes harmonics, > intermodulation and all sorts of other bad things in the recovered > output. When L+R and L-R start interfering with each other due to these > effects in stereo mode, things go bad really quickly. > > Then comes stereo demux, which requires precise 19KHz recovery and 38KHz > synthesis, and sharp filtering of the resulting L+R/L-R. Maintaining a > flat frequency response up to 15KHz and good pilot rejection is easy > with a good DSP FIR design, but is a nightmare with cascaded analog > filters - and I can't imagine anything better than a 4th order lowpass > in a cheap FM demod chip. > > Overall, I'll need some strong convincing that a 50 cent chip will do an > equal or better job than an RF/IF-input, fully digital demodulator. > > And sure, an incredibly high quality FM demodulator is still limited by > the quality of the FM that's going into it, there's not much you can do > about multipath, lame FM morning shows or jessica simpson. But "radio > sucks" is no reason to attempt to build a better, high quality receiver. > And there's certainly a market for high quality radio receivers - > companies such as Belar (who make super high quality AM/FM demod > equipment for broadcast use) haven't exactly been run out of business by > $10 radios from Walmart. > > I don't know what chris1911's motives are for attempting FM demodulation > in the Blackfin, but what's the harm? Best case, he might build a > digital FM demodulator with outstanding performance. Worst case, it > might not work out and he'll only learn all sorts of DSP theory and > practice which could lead him into a great career doing DSP. > > So hold off on the discouragement for a while. > > -GMSome good points there but they need to be verified with listening tests. I have a Universal Software Radio and it costs a fortune! Still, I have seen software radio ICs that work at IF. For both cases you ened a good analogue front-end and you cannot get away from that. It would be good if you could stick an arial up an ADC but it's not that easy. Naebad