DSPRelated.com
Forums

What waveform[s] do[es] EDT's FSK use?

Started by GreenXenon February 26, 2009
Hi:

What waveform[s] of FSK do[es] the EDT [European Deaf Telephone]
protocol use?

FSK = Frequency Shift Keying

Quote from http://www.deaflibrary.org/v18/v18.html :

"EDT Operational Mode -- The modulation is frequency shift keyed
modulation (i.e. carrier is 10 ms before a character is transmitted
and removed 1s after the last character) using Recommendation V.21(1)
frequencies.  The data signalling rate is 110 bits/s."

Quote from
http://www.freshpatents.com/Text-telephone-modem-communications-over-packet-networks-dt20050908ptan20050195801.php?type=description:

"European Deaf Telephone (or EDT), which is used in Germany, Austria,
Switzerland, Malta, Spain and Italy, uses ITU-T V.21 Channel 1
modulation at a signaling rate of 110 bps. ITU-T V.21 Channel 1
modulation has a mark frequency at 980 Hz and a space frequency at
1080 Hz. In EDT, the mark frequency is transmitted for 300 ms before a
character is transmitted, and the mark frequency is transmitted for up
to one (1) second after the last character. EDT has a character format
of one (1) start bit, seven (7) data bits, one (1) even parity bit and
two (2) stop bits. The first EDT character can be detected using the
present invention, as described below. For example, at step 506 of
FIG. 5, the mark frequency that precedes the start bit is used for the
detection of the first EDT character. At step 508, tone detector 420
starts looking for the mark frequency, i.e. 980 Hz, and at step 510,
it is determined whether the 980 Hz mark frequency is detected for a
pre-determined amount of time. In one embodiment, detection of 10-20
ms of the mark frequency can be sufficient at step 510."

According to the above links, EDT uses FSK in which the two are
frequencies are 980 Hz [representing a "1"] and 1080 Hz [representing
a "0"].

What are the shapes of the 980 Hz and 1080 Hz tones? Are they sine
waves, square waves, sawtooth waves, triangle waves??

Would an EDT-demodulator recognize only certain wave-shapes while
ignoring others?


Thanks
>Hi: > >What waveform[s] of FSK do[es] the EDT [European Deaf Telephone] >protocol use? > >FSK = Frequency Shift Keying > >Quote from http://www.deaflibrary.org/v18/v18.html : > >"EDT Operational Mode -- The modulation is frequency shift keyed >modulation (i.e. carrier is 10 ms before a character is transmitted >and removed 1s after the last character) using Recommendation V.21(1) >frequencies. The data signalling rate is 110 bits/s." > >Quote from >http://www.freshpatents.com/Text-telephone-modem-communications-over-packet-networks-dt20050908ptan20050195801.php?type=description: > >"European Deaf Telephone (or EDT), which is used in Germany, Austria, >Switzerland, Malta, Spain and Italy, uses ITU-T V.21 Channel 1 >modulation at a signaling rate of 110 bps. ITU-T V.21 Channel 1 >modulation has a mark frequency at 980 Hz and a space frequency at >1080 Hz. In EDT, the mark frequency is transmitted for 300 ms before a >character is transmitted, and the mark frequency is transmitted for up >to one (1) second after the last character. EDT has a character format >of one (1) start bit, seven (7) data bits, one (1) even parity bit and >two (2) stop bits. The first EDT character can be detected using the >present invention, as described below. For example, at step 506 of >FIG. 5, the mark frequency that precedes the start bit is used for the >detection of the first EDT character. At step 508, tone detector 420 >starts looking for the mark frequency, i.e. 980 Hz, and at step 510, >it is determined whether the 980 Hz mark frequency is detected for a >pre-determined amount of time. In one embodiment, detection of 10-20 >ms of the mark frequency can be sufficient at step 510." > >According to the above links, EDT uses FSK in which the two are >frequencies are 980 Hz [representing a "1"] and 1080 Hz [representing >a "0"]. > >What are the shapes of the 980 Hz and 1080 Hz tones? Are they sine >waves, square waves, sawtooth waves, triangle waves?? > >Would an EDT-demodulator recognize only certain wave-shapes while >ignoring others?
You actually referenced the exact spec for what is sent on the wire - ITU V.21 - which is freely downloadable from the ITU web site. What more are you looking for? Steve
On Feb 26, 7:43 pm, "steveu" <ste...@coppice.org> wrote:


> You actually referenced the exact spec for what is sent on the wire - ITU > V.21 - which is freely downloadable from the ITU web site. What more are > you looking for?
What are the shapes of the 980 and 1080 Hz waves? I'm guessing they are sine waves but I could be wrong. That's why I'm asking.
On Feb 26, 11:23 pm, GreenXenon <glucege...@gmail.com> wrote:
> On Feb 26, 7:43 pm, "steveu" <ste...@coppice.org> wrote: > > > You actually referenced the exact spec for what is sent on the wire - ITU > > V.21 - which is freely downloadable from the ITU web site. What more are > > you looking for? > > What are the shapes of the 980 and 1080 Hz waves? I'm guessing they > are sine waves but I could be wrong. That's why I'm asking.
FSK would typically use sinusoids. Square waves, sawtooth waves, etc. have wide bandwidths (lots of harmonics); add frequency modulation onto that and you'll get a lot of spectral junk. Sinusoids are more well-behaved. Jason
steveu wrote:
>> Hi: >> >> What waveform[s] of FSK do[es] the EDT [European Deaf Telephone] >> protocol use? >> >> FSK = Frequency Shift Keying >> >> Quote from http://www.deaflibrary.org/v18/v18.html : >> >> "EDT Operational Mode -- The modulation is frequency shift keyed >> modulation (i.e. carrier is 10 ms before a character is transmitted >> and removed 1s after the last character) using Recommendation V.21(1) >> frequencies. The data signalling rate is 110 bits/s." >> >> Quote from >> http://www.freshpatents.com/Text-telephone-modem-communications-over-packet-networks-dt20050908ptan20050195801.php?type=description: >> >> "European Deaf Telephone (or EDT), which is used in Germany, Austria, >> Switzerland, Malta, Spain and Italy, uses ITU-T V.21 Channel 1 >> modulation at a signaling rate of 110 bps. ITU-T V.21 Channel 1 >> modulation has a mark frequency at 980 Hz and a space frequency at >> 1080 Hz. In EDT, the mark frequency is transmitted for 300 ms before a >> character is transmitted, and the mark frequency is transmitted for up >> to one (1) second after the last character. EDT has a character format >> of one (1) start bit, seven (7) data bits, one (1) even parity bit and >> two (2) stop bits. The first EDT character can be detected using the >> present invention, as described below. For example, at step 506 of >> FIG. 5, the mark frequency that precedes the start bit is used for the >> detection of the first EDT character. At step 508, tone detector 420 >> starts looking for the mark frequency, i.e. 980 Hz, and at step 510, >> it is determined whether the 980 Hz mark frequency is detected for a >> pre-determined amount of time. In one embodiment, detection of 10-20 >> ms of the mark frequency can be sufficient at step 510." >> >> According to the above links, EDT uses FSK in which the two are >> frequencies are 980 Hz [representing a "1"] and 1080 Hz [representing >> a "0"]. >> >> What are the shapes of the 980 Hz and 1080 Hz tones? Are they sine >> waves, square waves, sawtooth waves, triangle waves?? >> >> Would an EDT-demodulator recognize only certain wave-shapes while >> ignoring others? > > You actually referenced the exact spec for what is sent on the wire - ITU > V.21 - which is freely downloadable from the ITU web site. What more are > you looking for?
We went down this path a couple of years ago. He can read the spec, but apparently, he doesn't know what it means. Jerry -- Engineering is the art of making what you want from things you can get. &#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;
GreenXenon wrote:
> On Feb 26, 7:43 pm, "steveu" <ste...@coppice.org> wrote: > > >> You actually referenced the exact spec for what is sent on the wire - ITU >> V.21 - which is freely downloadable from the ITU web site. What more are >> you looking for? > > > What are the shapes of the 980 and 1080 Hz waves? I'm guessing they > are sine waves but I could be wrong. That's why I'm asking.
they are sine waves, probably continuous at the transitions, and filtered to further reduce splatter. Jerry -- Engineering is the art of making what you want from things you can get. &#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;&#4294967295;
On Thu, 26 Feb 2009 18:25:07 -0800 (PST), GreenXenon
<glucegen1x@gmail.com> wrote:

>Hi: > >What waveform[s] of FSK do[es] the EDT [European Deaf Telephone] >protocol use? > >FSK = Frequency Shift Keying > >Quote from http://www.deaflibrary.org/v18/v18.html : > >"EDT Operational Mode -- The modulation is frequency shift keyed >modulation (i.e. carrier is 10 ms before a character is transmitted >and removed 1s after the last character) using Recommendation V.21(1) >frequencies. The data signalling rate is 110 bits/s." > >Quote from >http://www.freshpatents.com/Text-telephone-modem-communications-over-packet-networks-dt20050908ptan20050195801.php?type=description: > >"European Deaf Telephone (or EDT), which is used in Germany, Austria, >Switzerland, Malta, Spain and Italy, uses ITU-T V.21 Channel 1 >modulation at a signaling rate of 110 bps. ITU-T V.21 Channel 1 >modulation has a mark frequency at 980 Hz and a space frequency at >1080 Hz. In EDT, the mark frequency is transmitted for 300 ms before a >character is transmitted, and the mark frequency is transmitted for up >to one (1) second after the last character. EDT has a character format >of one (1) start bit, seven (7) data bits, one (1) even parity bit and >two (2) stop bits. The first EDT character can be detected using the >present invention, as described below. For example, at step 506 of >FIG. 5, the mark frequency that precedes the start bit is used for the >detection of the first EDT character. At step 508, tone detector 420 >starts looking for the mark frequency, i.e. 980 Hz, and at step 510, >it is determined whether the 980 Hz mark frequency is detected for a >pre-determined amount of time. In one embodiment, detection of 10-20 >ms of the mark frequency can be sufficient at step 510." > >According to the above links, EDT uses FSK in which the two are >frequencies are 980 Hz [representing a "1"] and 1080 Hz [representing >a "0"]. > >What are the shapes of the 980 Hz and 1080 Hz tones? Are they sine >waves, square waves, sawtooth waves, triangle waves?? > >Would an EDT-demodulator recognize only certain wave-shapes while >ignoring others? > > >Thanks
It is designed around sine waves. An excess of harmonics may cause bit recognition errors. If you stay properly v.21 compliant (sine waves) you should be OK.