Re: [digitalradio] Re: Bandwidth v Shift in RTTY ?
Dave Bernstein escribió: > I understand, Jose. My question is whether the "inner tones" -- the > ones between the ensemble's highest and lowest tones -- contribute to > the bandwidth if their magnitudes are identical to those of the > lowest and highest tones. I expect little contribution from them to occupied bandwidth. The significant ones are the extreme tones. > Asked another way, is the bandwidth of 300 baud 1 khz 4-tone FSK > greater than the bandwidth of 300 baud 1 khz 2-tone FSK? (where the 1 > khz is the frequency difference between the ensemble's highest and > lowest tones). Based on the superposition approach suggested by an > earlier poster, one would suspect that the inner tones make little > contribution to bandwidth unless the tones are spaced quite closely. Without the backing of any simulations or calculations, this makes sense at first sight. 73, Jose, CO2JA VI Conferencia Internacional de Energía Renovable, Ahorro de Energía y Educación Energética 9 - 12 de Junio 2009, Palacio de las Convenciones ...Por una cultura energética sustentable www.ciercuba.com
[digitalradio] Re: Bandwidth v Shift in RTTY ?
I understand, Jose. My question is whether the "inner tones" -- the ones between the ensemble's highest and lowest tones -- contribute to the bandwidth if their magnitudes are identical to those of the lowest and highest tones. Asked another way, is the bandwidth of 300 baud 1 khz 4-tone FSK greater than the bandwidth of 300 baud 1 khz 2-tone FSK? (where the 1 khz is the frequency difference between the ensemble's highest and lowest tones). Based on the superposition approach suggested by an earlier poster, one would suspect that the inner tones make little contribution to bandwidth unless the tones are spaced quite closely. 73, Dave, AA6YQ --- In digitalradio@yahoogroups.com, "Jose A. Amador" wrote: > > > Not exactly. You must add the upper and lower keying sidebands spacing > to the upper and lower tones to get an aproximate idea of the occupied > bandwidth. The sidebands lie at half the signalling speed around the > carriers, and the keying harmonics, whose level and width depends on the > modulation index, which is quite large with 1 kHz shift. The Carson Rule > gives an approximate answer. The exact answer could be found by Fourier > analysis. A simple way to get an answer may be using PSpice or LTSpice, > for those willing to use a simulation package. > > The simplistic answer is at least 1300 Hz: 150 + 1000 + 150, > disregarding higher order sidebands. With such a large shift to keying > rate, the occupied bandwidth will be larger than the simplistic, on the > fky answer. > > Maybe some people won't bother with Fourier analysis, Bessel > coefficients, simulation software or even simple math and just mimic it > with MixW and a loopback to some PC based spectrum analyzer. I would use > Spectran. Spectrum Lab should be OK too. The carriers should be as high > as possible to avoid the lower sideband spectrum foldover. > > For those that would like to give it a try with a radio, I would use a > SDR and not a transceiver with an IF crystal filter to find a true > answer. Beware of nonlinearities that might broaden the signal. > > It would be interesting to read about some practical replies to that > question. > > > 73, > > Jose, CO2JA > > --- > > Dave Bernstein wrote: > > > In n-ary FSK, if all tones in the ensemble have identical maximum > > magnitudes, then isn't it true that the maximum bandwidth will be > > identical that of binary (2-tone) FSK with a shift whose value is > > difference in frequency between the highest and lowest tones in the > > ensemble? > > > VI Conferencia Internacional de Energía Renovable, Ahorro de Energía y > Educación Energética > 9 - 12 de Junio 2009, Palacio de las Convenciones > ...Por una cultura energética sustentable > www.ciercuba.com >
Re: [digitalradio] Re: Bandwidth v Shift in RTTY ?
Not exactly. You must add the upper and lower keying sidebands spacing to the upper and lower tones to get an aproximate idea of the occupied bandwidth. The sidebands lie at half the signalling speed around the carriers, and the keying harmonics, whose level and width depends on the modulation index, which is quite large with 1 kHz shift. The Carson Rule gives an approximate answer. The exact answer could be found by Fourier analysis. A simple way to get an answer may be using PSpice or LTSpice, for those willing to use a simulation package. The simplistic answer is at least 1300 Hz: 150 + 1000 + 150, disregarding higher order sidebands. With such a large shift to keying rate, the occupied bandwidth will be larger than the simplistic, on the fky answer. Maybe some people won't bother with Fourier analysis, Bessel coefficients, simulation software or even simple math and just mimic it with MixW and a loopback to some PC based spectrum analyzer. I would use Spectran. Spectrum Lab should be OK too. The carriers should be as high as possible to avoid the lower sideband spectrum foldover. For those that would like to give it a try with a radio, I would use a SDR and not a transceiver with an IF crystal filter to find a true answer. Beware of nonlinearities that might broaden the signal. It would be interesting to read about some practical replies to that question. 73, Jose, CO2JA --- Dave Bernstein wrote: > In n-ary FSK, if all tones in the ensemble have identical maximum > magnitudes, then isn't it true that the maximum bandwidth will be > identical that of binary (2-tone) FSK with a shift whose value is > difference in frequency between the highest and lowest tones in the > ensemble? VI Conferencia Internacional de Energía Renovable, Ahorro de Energía y Educación Energética 9 - 12 de Junio 2009, Palacio de las Convenciones ...Por una cultura energética sustentable www.ciercuba.com
[digitalradio] Re: Bandwidth v Shift in RTTY ?
--- In digitalradio@yahoogroups.com, "expeditionradio" wrote: > > Hi Andy, > > There is no simple universal relationship between > the "shift" and the transmitted "signal bandwidth", However, for the particular case of binary FSK where the shift is wide compared to the bit rate, you get a reasonable approximation by treating the signal as a pair of carriers being switched on and off alternately. For each carrier you get pairs of sidebands. For the worst case of a square wave digital signal you'll get sidebands at the dot-cycle frequency and odd multiples thereof. So, depending on signal shaping, the signal at 850 Hz shift will have a bandwidth extending over about 950 Hz with a big hole in the middle.
[digitalradio] Re: Bandwidth v Shift in RTTY ?
In n-ary FSK, if all tones in the ensemble have identical maximum magnitudes, then isn't it true that the maximum bandwidth will be identical that of binary (2-tone) FSK with a shift whose value is difference in frequency between the highest and lowest tones in the ensemble? 73, Dave, AA6YQ --- In digitalradio@yahoogroups.com, "expeditionradio" wrote: > > Hi Andy, > > There is no simple universal relationship between > the "shift" and the transmitted "signal bandwidth", > because there are so many factors other than "shift" > that contribute to the bandwidth of an "FSK" signal: > > 1. Symbol rate > 2. Shape of waveform > 3. Symbol transition point > 4. Filtering > 5. Number of tone frequencies > 6. Transmitter chain > 7. Other factors related to modulation process > 8. Noise > 9. Transmitter oscillator spectral purity > 10. Definition of bandwidth > > This is an especially complex calculation for > multiple frequency FSK signals, commonly > 4-ary FSK, 8-ary FSK, 16-ary FSK, 32-ary FSK etc. > where the number of shift frequencies is greater > than 2, or the number of carriers is greater than 1. > > The FCC rule says "maximum frequency shift of > 1 kilohertz between mark and space." But, > that FCC rule was written in the old days when > common ham RTTY was Frequency Shift Keyed between > only 2 frequencies, technically described by > "mark and space". However, in modern multiple > tone frequency shift techniques, with binary > symbols there is no such thing as "mark and space". > Thus, the rule became inapplicable to the new > multiple frequency shifting keying modes. > > When the FCC was asked to convert from "shift > limit" to "bandwidth limit", the FCC refused, > and at the same time, FCC said it had chosen not to > limit bandwidth because it is important for > ham radio to have the freedom to innovate and > develop new techniques. > > Thus, the mark and space shift limit became a > mere footnote in history that largely does not > affect most modern digital techniques used in > ham radio today. > > If you wish to delve into the finer math points > of relationship between bandwidth and shift, may > I suggest reading Section 6 (starting on page 37) > of this fine document: > "Necessary Bandwidth and Spectral Properties of > Digital Modulation" by David J. Cohen: > http://www.its.bldrdoc.gov/pub/ntia-rpt/84-168/84-168.pdf > > 73 Bonnie KQ6XA > > > Andy K3UK wrote > > -Bonnie, can you explain to this bozo what the > > difference between a "shift" restriction and > > "bandwidth" restriction would be? My brain > > viewed them to be the same, that is that a > > 170Hz shift would be roughly that amount > > of Hz wide at the usual ham speed. > > >
[digitalradio] Re: Bandwidth v Shift in RTTY ?
Hi Andy, There is no simple universal relationship between the "shift" and the transmitted "signal bandwidth", because there are so many factors other than "shift" that contribute to the bandwidth of an "FSK" signal: 1. Symbol rate 2. Shape of waveform 3. Symbol transition point 4. Filtering 5. Number of tone frequencies 6. Transmitter chain 7. Other factors related to modulation process 8. Noise 9. Transmitter oscillator spectral purity 10. Definition of bandwidth This is an especially complex calculation for multiple frequency FSK signals, commonly 4-ary FSK, 8-ary FSK, 16-ary FSK, 32-ary FSK etc. where the number of shift frequencies is greater than 2, or the number of carriers is greater than 1. The FCC rule says "maximum frequency shift of 1 kilohertz between mark and space." But, that FCC rule was written in the old days when common ham RTTY was Frequency Shift Keyed between only 2 frequencies, technically described by "mark and space". However, in modern multiple tone frequency shift techniques, with binary symbols there is no such thing as "mark and space". Thus, the rule became inapplicable to the new multiple frequency shifting keying modes. When the FCC was asked to convert from "shift limit" to "bandwidth limit", the FCC refused, and at the same time, FCC said it had chosen not to limit bandwidth because it is important for ham radio to have the freedom to innovate and develop new techniques. Thus, the mark and space shift limit became a mere footnote in history that largely does not affect most modern digital techniques used in ham radio today. If you wish to delve into the finer math points of relationship between bandwidth and shift, may I suggest reading Section 6 (starting on page 37) of this fine document: "Necessary Bandwidth and Spectral Properties of Digital Modulation" by David J. Cohen: http://www.its.bldrdoc.gov/pub/ntia-rpt/84-168/84-168.pdf 73 Bonnie KQ6XA > Andy K3UK wrote > -Bonnie, can you explain to this bozo what the > difference between a "shift" restriction and > "bandwidth" restriction would be? My brain > viewed them to be the same, that is that a > 170Hz shift would be roughly that amount > of Hz wide at the usual ham speed. >
