Randy Yates wrote:> PaavoJumppanen@iname.com (Paavo Jumppanen) wrote in message news:<1a72ed76.0307221814.5e485578@posting.google.com>... > > Jerry Avins <jya@ieee.org> wrote in message news:<3F1DB3DD.2A5B6CB6@ieee.org>... > > > Tom wrote: > > > > > > > > If I have a digital FM system do you have any background on the max > > > > bit rate you can achieve for a given depth of modulation? eg for > > > > 250kHz depth of modulation I can achieve only 500kb/s. Does this sound > > > > right? This must be pretty basic stuff I am sure and related to > > > > Shannons law somewhere. > > > > > > > > Thanks > > > > > > > > Tom > > > > > > I might be able to help if I knew what you meant by digital FM. Can you > > > explain it? Where is it used? What kind of signals are sent with it? > > > > > > Jerry > > > > If he's talking about binary symbols only then I'd assume he means > > FSK. As far as I remember the maximum rate you can achieve using this > > methodology is basically determined by the degree to which the two > > spectral lines for 0 and 1 state become smeared together. > > > > That is, if your symbol rate it very slow then the spectrum will look > > like two spectral lines 500kHz apart. As you increase the bit rate the > > two peaks begin to widen and at some point they will overlap so much > > that you won't be able to tell them apart. At this point you'll > > probably have so much inter-symbol interference that you wont be able > > to detect what you transmitted. > > > > How much you can tolerate will depend upon how you are decoding the > > signal at the receiver end and how good your signal to noise ratio is. > > > > I'd guess that 500kb/s with 250kHz depth is a bit optimistic but I'm > > not experience in FSK so don't quote me on that. > > > > Cheers, > > > > > > Paavo Jumppanen. > > Author of AtSpec : A 2 channel PC based FFT spectrum analyzer > > http://www.taquis.com > > Hello Paavo, > > I don't believe your analysis is correct. On the other hand, I > could be totally off too. > > Here's how I'm thinking of the situation: If you quadrature > downconverted the signal to baseband using a high enough sample > rate that any spectrum that is lost is negligible, then you can > detect the frequency with a simple d theta / dt, where theta is > the instantaneous phase of the complex I/Q data. Now since Tom > didn't specify, we can assume a noiseless system ( ;) ). Then > the frequency will be perfectly detected with no errors and we > will be able to decode the bits. > > What do you think? As I stated in a parallel response, without > any constraints on bandwidth and/or noise, I don't think there > is a limit on the bit rate, no matter what the frequency spacing > is.What I am requesting is the limit on bit rate for a given bandwidth. the bandwidth here I believe is the frequency deviation. SNR has to come into it somewhere too. Tom
Max Bit rate question
Started by ●July 22, 2003
Reply by ●July 24, 20032003-07-24
Reply by ●July 24, 20032003-07-24
Paavo Jumppanen wrote:> yates@ieee.org (Randy Yates) wrote in message news:<567ce618.0307231455.4ce300f7@posting.google.com>... > > PaavoJumppanen@iname.com (Paavo Jumppanen) wrote in message news:<1a72ed76.0307221814.5e485578@posting.google.com>... > > > Jerry Avins <jya@ieee.org> wrote in message news:<3F1DB3DD.2A5B6CB6@ieee.org>... > > > > Tom wrote: > > > > > > > > > > If I have a digital FM system do you have any background on the max > > > > > bit rate you can achieve for a given depth of modulation? eg for > > > > > 250kHz depth of modulation I can achieve only 500kb/s. Does this sound > > > > > right? This must be pretty basic stuff I am sure and related to > > > > > Shannons law somewhere. > > > > > > > > > > Thanks > > > > > > > > > > Tom > > > > > > > > I might be able to help if I knew what you meant by digital FM. Can you > > > > explain it? Where is it used? What kind of signals are sent with it? > > > > > > > > Jerry > > > > > > If he's talking about binary symbols only then I'd assume he means > > > FSK. As far as I remember the maximum rate you can achieve using this > > > methodology is basically determined by the degree to which the two > > > spectral lines for 0 and 1 state become smeared together. > > > > > > That is, if your symbol rate it very slow then the spectrum will look > > > like two spectral lines 500kHz apart. As you increase the bit rate the > > > two peaks begin to widen and at some point they will overlap so much > > > that you won't be able to tell them apart. At this point you'll > > > probably have so much inter-symbol interference that you wont be able > > > to detect what you transmitted. > > > > > > How much you can tolerate will depend upon how you are decoding the > > > signal at the receiver end and how good your signal to noise ratio is. > > > > > > I'd guess that 500kb/s with 250kHz depth is a bit optimistic but I'm > > > not experience in FSK so don't quote me on that. > > > > > > Cheers, > > > > > > > > > Paavo Jumppanen. > > > Author of AtSpec : A 2 channel PC based FFT spectrum analyzer > > > http://www.taquis.com > > > > Hello Paavo, > > > > I don't believe your analysis is correct. On the other hand, I > > could be totally off too. > > > > Here's how I'm thinking of the situation: If you quadrature > > downconverted the signal to baseband using a high enough sample > > rate that any spectrum that is lost is negligible, then you can > > detect the frequency with a simple d theta / dt, where theta is > > the instantaneous phase of the complex I/Q data. Now since Tom > > didn't specify, we can assume a noiseless system ( ;) ). Then > > the frequency will be perfectly detected with no errors and we > > will be able to decode the bits. > > > > What do you think? As I stated in a parallel response, without > > any constraints on bandwidth and/or noise, I don't think there > > is a limit on the bit rate, no matter what the frequency spacing > > is. > > Hello Randy, > > I did qualify my assertion with the notion that it will depend upon > how you detect it at the receiver end and I was thinking of a real > system rather than a noiseless one. Without any noise you're probably > correct, though I'm not experienced enough to enough to say. > > On the other hand I do think that there are real practical limitations > on the bit rate. Was Toms question purely theoretical, or did it have > a practical stance. I'm assuming that his assertion that he could not > achieve more than 500kb/s means it's a practical question. Any > comments Tom? I'd also be interested in knowing what sort of Rx > topology he's using. > > Cheers, > > Paavo Jumppanen, > Author of AtSpec : A 2 channel PC based FFT spectrum analyzer > http://www.taquis.comWe're using a PLL at the receiver. Tom
Reply by ●July 25, 20032003-07-25
Tom <somebody@nOpam.com> wrote in message news:<3F208E22.525B6449@nOpam.com>...> Randy Yates wrote: >> > What I am requesting is the limit on bit rate for a given bandwidth. the bandwidth here I believe is the frequency > deviation. SNR has to come into it somewhere too. > > TomFor given bandwidth, you may use multilevel fsk to increase data-rate. snr is the constraint. deviation is a term used only with analog fm . For fsk, bandwidth and datarate are dependent only on symbol rate; not on tone-separation. If you want to make the waveforms orthogonal to each other, you need a minimum tone-separation of 1/2T where T = symbol duration. This is known as MSK. 2 cents.. shankar
Reply by ●July 25, 20032003-07-25
Tom wrote:>...> > What I am requesting is the limit on bit rate for a given bandwidth. the bandwidth here I believe is the frequency > deviation. SNR has to come into it somewhere too. > > TomWhatever else holds, the bandwidth exceeds the deviation by an amount that depends on the abruptness of the transitions and how often they happen. No matter how small the deviation, the bandwidth is at least equal to twice that of the baseband signal feeding the modulator, and is likely several times that. Jerry -- Engineering is the art of making what you want from things you can get. �����������������������������������������������������������������������
Reply by ●July 25, 20032003-07-25
kbc wrote:> Tom <somebody@nOpam.com> wrote in message news:<3F208E22.525B6449@nOpam.com>... > > Randy Yates wrote: > > > > > > > What I am requesting is the limit on bit rate for a given bandwidth. the bandwidth here I believe is the frequency > > deviation. SNR has to come into it somewhere too. > > > > Tom > > For given bandwidth, you may use multilevel fsk to increase data-rate. > snr is the constraint. > > deviation is a term used only with analog fm . > > For fsk, bandwidth and datarate are dependent only on symbol rate; not on > tone-separation. > > If you want to make the waveforms orthogonal to each other, you > need a minimum tone-separation of 1/2T where T = symbol duration. This > is known as MSK. > > 2 cents.. > > shankarI'll buy that. Tom
Reply by ●July 26, 20032003-07-26
> > Why isn't it? If I quadrature downconvert to baseband and then > > compute the derivative of the instantaneous phase, isn't that > > demodulating an FM signal? > > Say the carrier was 1Hz, FM modulated at 1MHz. Now you want to quadrature > downconvert to 0Hz? Hmm. Where do you get the reference signal for the > downconverter? How accurate is the reference signal? Does it make sense > to call a 1Hz signal modulated at 1MHz FSK? > > -- glenGlen 1 Hz is not carrier 1 Mhz is a carrier on which the information is carried. 1 Hz is the modulating signal. Reference signal for the down converter can be got from crystal oscillators which has to be accurate and precise and the drifts should be very small which is measured in ppm(parts per million) Normally crystals have a tendency to drift at ~25 ppm. to compensate for this drifts plls are used.
Reply by ●July 26, 20032003-07-26
"Baretto" <suhnas1999@yahoo.co.uk> wrote in message news:5b98244c.0307260032.462426c2@posting.google.com...> > > Why isn't it? If I quadrature downconvert to baseband and then > > > compute the derivative of the instantaneous phase, isn't that > > > demodulating an FM signal? > > > > Say the carrier was 1Hz, FM modulated at 1MHz. Now you want toquadrature> > downconvert to 0Hz? Hmm. Where do you get the reference signal forthe> > downconverter? How accurate is the reference signal? Does it makesense> > to call a 1Hz signal modulated at 1MHz FSK?> 1 Hz is not carrier 1 Mhz is a carrier on which the information is > carried. 1 Hz is the modulating signal. > Reference signal for the down converter can be got from crystal > oscillators which has to be accurate and precise and the drifts should > be very small which is measured in ppm(parts per million) Normally > crystals have a tendency to drift at ~25 ppm. to compensate for this > drifts plls are used.I was answering a claim that there was no limit to the modulating frequency. If you don't like those numbers, how about a 1MHz carrier, and modulating at 1THz. Well, maybe 1GHz is enough. I agree that 1Hz doesn't make a very good carrier. (WWV has a transmitter with a 20KHz carrier trying to cover the whole US with one transmitter. That is the lowest I know for a real radio transmission. It might even be FSK.) -- glen
Reply by ●July 28, 20032003-07-28
glen herrmannsfeldt wrote:> > "Randy Yates" <yates@ieee.org> wrote in message > news:567ce618.0307241135.76c03856@posting.google.com... > > "glen herrmannsfeldt" <gah@ugcs.caltech.edu> wrote in message > news:<6hETa.115923$GL4.30071@rwcrnsc53>... > > > "Randy Yates" <yates@ieee.org> wrote in message > > > news:567ce618.0307231342.4637d16f@posting.google.com... > > > > aberdonian_2000@yahoo.com (Tom) wrote in message > > > news:<e1b1658f.0307221244.575e9f86@posting.google.com>... > > > > > If I have a digital FM system do you have any background on the max > > > > > bit rate you can achieve for a given depth of modulation? eg for > > > > > 250kHz depth of modulation I can achieve only 500kb/s. Does this > sound > > > > > right? This must be pretty basic stuff I am sure and related to > > > > > Shannons law somewhere. > > > > > > > If you are talking about FSK, then there is no limit. The problem > > > > is that the bandwidth just keeps getting bigger and bigger. If > > > > you had asked what the maximum bitrate is for a given bandwidth, > > > > then an answer could be negotiated. > > > > > > Well, I think when you exceed, or even get close to, the carrier > frequency > > > you should not call it FSK anymore. It may be that, technically, you > can > > > still decode the bits from the sidebands, it isn't because you are > > > demodulating the FM signal. > > > > Why isn't it? If I quadrature downconvert to baseband and then > > compute the derivative of the instantaneous phase, isn't that > > demodulating an FM signal? > > Say the carrier was 1Hz, FM modulated at 1MHz. Now you want to quadrature > downconvert to 0Hz? Hmm. Where do you get the reference signal for the > downconverter? How accurate is the reference signal? Does it make sense > to call a 1Hz signal modulated at 1MHz FSK?I was presuming that we were far enough away from DC that this type of problem doesn't happen. -- % Randy Yates % "...the answer lies within your soul %% Fuquay-Varina, NC % 'cause no one knows which side %%% 919-577-9882 % the coin will fall." %%%% <yates@ieee.org> % 'Big Wheels', *Out of the Blue*, ELO http://home.earthlink.net/~yatescr
Reply by ●July 28, 20032003-07-28
> I was answering a claim that there was no limit to the modulating frequency. > If you don't like those numbers, how about a 1MHz carrier, and modulating at > 1THz. Well, maybe 1GHz is enough. I agree that 1Hz doesn't make a very > good carrier. > > (WWV has a transmitter with a 20KHz carrier trying to cover the whole US > with one transmitter. That is the lowest I know for a real radio > transmission. It might even be FSK.) > > -- glenI think there is a confusion here, Always a low signal which is called a modulating signal is up converted using a carrier which is at higher frequency..Now lets say a speech signal at 0-4 khz..fm = 4KHz(max mod freq) when this has to be sent over a channel then it is upconverted using a higher freq carrier lats say 10MHz. when this is done the bandwidth occupied here will be twice the modulating freq..at 100MHz which is from (fc-fm) to (fc+fm) ..Glen carrier is not modulated but modutating signal is modulated to get a modulated signal using a higher frequency carrier...in the example u have given 1MHz is again a modulating signal and 1THz is a carrier..There are cases where more than one leven of up conversion or downconversion is used,in which case it is called as IF levels...where it is said that 1MHz IF is again up converted to 1 1THz...>I was answering a claim that there was no limit to the modulatingfrequency. there is a limit to this also...ur modulation freq cannot exceed the carrier frequency raj
Reply by ●July 28, 20032003-07-28
"Baretto" <suhnas1999@yahoo.co.uk> wrote in message news:5b98244c.0307272144.45c58492@posting.google.com...> I think there is a confusion here, > Always a low signal which is called a modulating signal is up > converted using a carrier which is at higher frequency..Now lets say a > speech signal at 0-4 khz..fm = 4KHz(max mod freq) when this has to be > sent over a channel then it is upconverted using a higher freq carrier > lats say 10MHz. when this is done the bandwidth occupied here will be > twice the modulating freq..at 100MHz which is from (fc-fm) to (fc+fm)I think you are right, but a different question was being asked. The question is how high can the modulation signal be for a given carrier frequency, for FM (FSK) modulation. Now, it is normal for the carrier to be higher than the modulation rate, maybe in some cases equal to the modulation rate. There was a claim that there is no limit to the modulation rate, so I took an extreme example, to test the claim. The (fc-fm) and (fc+fm) works well for AM signals, but FM (FSK) was being discussed. Theoretically, the bandwidth is infinite, but practically, it falls of fast enough that a reasonable bandwidth can be defined. -- glen
