What is the maximum bits-per-symbol possible using FSK on telephone devices?
in [DSP]
|
From: Clay on 18 Dec 2009 11:24 On Dec 18, 11:11 am, "steveu" <ste...(a)coppice.org> wrote: > >Flex - it is a Motorola propriatary paging standard. It uses 4 level > >FSK. > > >It supports 4 mixes of rate and number of levels. Essentially the > >rates are 1600 or 3200 bauds and each baud can be either 1 or 2 bits. > >The header is always 1600 - 2 level and then may change to a higher > >data rate based upon the header's content. The 4 level data is grey > >encoded and FSK modulated on an RF carrier. The spec calls for a 10th > >order Bessel filter (IIRC cutoff = 3.8kHz) to be used to slew limit > >the modulation. The RF lives in a 25kHz wide channel, although the > >actuall spectral occupancy is much less. But the RF may be viewed as a > >4 FSK. > > >There's your example. > > Ermes was 4 level FSK, too. The paging people seemed to like 4 level FSK. > > Steve- Hide quoted text - > > - Show quoted text - In fact Flex was designed to be able to use Ermes transmitters. The 10th order Bessel requirement came from the Ermes spec. Back when paging was still being used a lot, effort was expended on being able to wring more throuput out of FSK systems since your in situ paging was FM and digital paging used 2 level FSK. So we saw APOC, Ermes, and Flex come on to the scene. We still have some Flex and ReFlex(2 way paging) here in the US. There is a small push for dedicated paging systems on campi - such as factories (not many of those left) and hospitals. Pager users often need access to phones and payphones are almost all extinct. I had a Flex pager watch I used up until about 2000. Payphones were becoming hard to find then. Clay
From: Clay on 18 Dec 2009 11:33 On Dec 18, 11:16 am, Vladimir Vassilevsky <nos...(a)nowhere.com> wrote: > Clay wrote: > > > Flex - it is a Motorola propriatary paging standard. It uses 4 level > > FSK. > It supports 4 mixes of rate and number of levels. Essentially the > > rates are 1600 or 3200 bauds and each baud can be either 1 or 2 bits. > > ERMES is similar European standard. Also 4-FSK with slightly different > parameters. > > > The header is always 1600 - 2 level and then may change to a higher > > data rate based upon the header's content. The 4 level data is grey > > encoded and FSK modulated on an RF carrier. The spec calls for a 10th > > order Bessel filter (IIRC cutoff = 3.8kHz) to be used to slew limit > > the modulation. > > Nitpick: 9th order Bessel is suggested pulse shaping. For that matter, I > actually used FIR filters. > > > The RF lives in a 25kHz wide channel, although the actuall spectral occupancy is much less. > > They coexist in the same band with analog narrowband FM, so they want > their spectrum to be similar to that of FM. > > > But the RF may be viewed as a 4 FSK. > > I don't know if any commertial paging system actually used 4-level > modulation and/or higher baud rate. There is significant tradeoff in > reliable coverage. > > Vladimir Vassilevsky > DSP and Mixed Signal Design Consultanthttp://www.abvolt.com Pagenet and Skytel are two examples of commercial carriers who both used Flex at 3200 baud and with 4 levels. Around 1995 I recall at least 15 channels of high speed flex being used. I designed intercept devices for paging in a prior life and saw a lot of over the air Flex. Sure it could be 9th order instead of 10th order for the filter, it has been almost 13 years since I worked in paging so my memory will show some failings. I do still have the specs though but I'm too lazy to look it up. Clay
From: Clay on 18 Dec 2009 13:59 On Dec 18, 12:06 pm, Vladimir Vassilevsky <nos...(a)nowhere.com> wrote: > Clay wrote: > > On Dec 18, 11:16 am, Vladimir Vassilevsky <nos...(a)nowhere.com> wrote: > > >>I don't know if any commertial paging system actually used 4-level > >>modulation and/or higher baud rate. There is significant tradeoff in > >>reliable coverage. > > > Pagenet and Skytel are two examples of commercial carriers who both > > used Flex at 3200 baud and with 4 levels. Around 1995 I recall at > > least 15 channels of high speed flex being used. I designed intercept > > devices for paging in a prior life and saw a lot of over the air Flex. > > Sure it could be 9th order instead of 10th order for the filter, it > > has been almost 13 years since I worked in paging so my memory will > > show some failings. I do still have the specs though but I'm too lazy > > to look it up. > > In one of the prior lifes, I designed paging equipment for operators. > I may be mistaken, however AFAIK for paging in US they use 900MHz with > very powerful (~kW) transmitters. Back in Russia, they used mainly VHF > band, transmitters on high towers, with typical power ~100W. There was > significant difference in message delivery probability depending on baud > rate and modulation; 2-FSK 1200/1600 was considered to be the best > compromise. > Besides, there is no good way to achieve simulcast coverage with 4-FSK. > > Vladimir Vassilevsky > DSP and Mixed Signal Design Consultanthttp://www.abvolt.com- Hide quoted text - > > - Show quoted text - Here (the US) I had worked for BBL (bought by Glenayre) and TGA. Paging here was in the 150 MHz band and the 929MHz band. There was some 450MHz stuff as well. Two way paging was in one of the PCS bands - IIRC was about 940MHz. And yes the 6400 bps version of Flex required a lot more power and the missed page rate was high. The timing was provided by GPS at the transmitters. Basically data packets are built with a time to transmit tag in them at the paging terminal and the packets are distributed to the transmitters. The smart transmitter emits the scheduled paging data at the right time. Needless to say such networks presented interesting challenges. The encoder software was smart enough to slow down the data rates during low traffic periods. So really only the peak hour or so saw the higher bit error rates. At one paging equipment manufacturer I worked on both the encoders and then the protocol monitors and pager intercept devices. I used Moto56309 DSPs for those projects. The company was already using that processor and its predessor (56001 and 56002) in its product line, so were very comfortable with manufacturing. I certainly enjoyed the available horsepower and 24 bit depth! The specialized equipment was not very price sensitive, so $30 or more for a DSP was not an issue. Clay
From: Green Xenon on 18 Dec 2009 19:15 >Green Xenon wrote: >>> Green Xenon wrote: >>>>> Green Xenon wrote: >>>>>>> Green Xenon wrote: >>>>>>>>> Green Xenon <green_xenon1(a)yahoo.com> wrote: >>>>>>>>> >>>>>>>>>> [glen writes] >>>>>>>>>>> All the FSK I know of are one bit/symbol. >>>>>>>>>> Isn't MFSK [Multiple Frequency Shift Keying] a form of FSK that >>>> uses >>>>>>>> more >>>>>>>>>> than 1-bit-per-symbol? >>>>>>>>> Does the M stand for Multiple or does it mean Minimal? >>>>>>>> M stands for multiple >>>>>>> Can you cite an example? >>>>>> Quote from >> http://en.wikipedia.org/wiki/Multiple_frequency-shift_keying >>>> : >>>>>> "Multiple frequency-shift keying (MFSK) is a variation of >>>> frequency-shift >>>>>> keying (FSK) that uses more than two frequencies." >>>>> Thanks. >>>>> >>>>> You wrote in another post that 1 symbol/sec satisfies you. Since >>>>> sampling for a second theoretically yields 1 Hz resolution, that's a >> lot >>>>> of frequencies in the telephone voice band. Using a spacing of 2 Hz to >> >>>>> allow for less-than-ideal conditions, and taking the band to be 300 to >> >>>>> 8,000 Hz, you can see that it is trivially simple to get 3,850 bits in >> a >>>>> symbol provided you can synchronize properly. Now, 2^3850 is a big >>>>> number, nearly 10^1159. What will you do with an alphabet of that many >> >>>>> letters? >>>>> >>>> If 3,580-bits-per-symbol can be done in FSK using telephone systems, >> then >>>> why isn't it already being done? You said it is "trivially simple". >>> Green, >>> you are using the term 'symbol' in a non-standard way. In >>> Data Comms, 'symbol' refers to the information being >>> conveyed e.g. the ASCII character set is made up of 128 >>> symbols and it takes 7 bits to represent each symbol. >>> The term you need instead of 'symbol' is 'state' which is a >>> combination of amplitude and either frequency and phase. >>> The related term 'Baud' is the number of state changes per >>> second. >>> >>> Regards, >>> John >>> >> >> Isn't baud the same thing symbol-per-second? 1 baud being >> 1-symbol-per-second? >> >> As for the term "symbol", QAM and PSK devices use it, so I thought FSK >> also uses it in a similar manner. > >No, the signalling rate in Baud is the number of changes per >second in the state of the carrier. For example, in QAM >modulation schemes that have 128 points in their >constellation (128 possible amplitude / phase combinations >available) it so happens that you can transmit one 7-bit >ASCII character, or one symbol, with each change in the >carrier so in that case the signalling rate in Baud does >correspond to symbols per second. > >People designing these schemes want to send as many bits as >possible with each signal change, but that is not the same >as wanting to increase the bits per symbol. >In a hypothetical QAM scheme with 16K amplitude / phase >combinations you could send 14 bits per signal change. That >is two ASCII symbols per signal change but is still works >out at seven bits per symbol. > >Regards, >John > Thanks for clarifying this. This brings up another question. What is the maximum amount of bits-per-baud of FSK possible using a telephone system [including the phone lines and any devices from start to finish of the phone's signal chain]? Let's assume a baud rate of only 1 baud.
From: glen herrmannsfeldt on 18 Dec 2009 19:40
John Monro <johnmonro(a)optusnet.com.au> wrote: (snip) > No, the signalling rate in Baud is the number of changes per > second in the state of the carrier. For example, in QAM > modulation schemes that have 128 points in their > constellation (128 possible amplitude / phase combinations > available) it so happens that you can transmit one 7-bit > ASCII character, or one symbol, with each change in the > carrier so in that case the signalling rate in Baud does > correspond to symbols per second. I believe that the term 'symbol' is commonly used in the case of digital QAM. It is not related to the ASCII symbol, but the same word is used. -- glen |