位 = 1300 ... 1550 nm
How much data in bits per second (bps) can be transferred through such cable WITHOUT ERRORS? Lets assume dismal signal to noise ratio SNR = 1 = 0dB.|||I'll take a guess.
You need a sufficient length so there is a 1 wavelength difference between your upper and your lower wavelength signals. This would allow you to send a data packet from destructive to constructive to destructive interference.
(You should be able to trick out your signal processing to get down to 1/2 wavelength difference because all you are looking for is a change from on to off or off to on. Further, you could probably trick out the signal processing to handle 1/4 wavelength difference. But I'm just going for the single wavepacket)
1300nm * x = 1550 *x-1
1300/1550 = (x-1)/x = 1-1/x
x = 6.2
d = x * 1300 nm
d = 3844 nm/bit
According to wiki, the index of refraction is ~ 1.48 for fibre optical glass. Therefore the time to transmit one bit is:
t = c/n *1/d
and the bits/sec is 1/t
data rate = 3e8m/s *1/1.48 * 1/(3844e-9m/bit)
= 5.27e13 bits/sec
..........
Of course if 位 is for a vacuum, not in glass, then you can ignore the index of refraction and the data rate is
d.r. = 7.8 e13 bits/sec
And if you could simply use 位/2, your data rate is pushed up to
d.r. = 15.6 e13 bits/sec
And if you could push it up to 位/4, your data rate is pushed up to
d.r. = 31.2 e13 bits/sec|||Come on Track. Don't be angry. I just ran out the Shannon鈥揌artley theorem and got 3.72e13 bits/sec which is a factor of 2 less than my vacuum derived information rate. Not bad for an independent derivation.
Report Abuse
|||Very good for independent derivation, I absoluely agree.
The thing is that strict and exact and non-trivial answer exists and I assumed that the answer was very widely known.
Report Abuse
|||I agree that someone should have known about the Shannon鈥揌artley theorem and its importance.
Report Abuse
|||Many, I Applied for Patent On a Design Which was Mil-Std-1773 Compliant, Long Ago.
EDIT: Hey el_santuario, Does it Use Single-Mode Fiber?|||As of right now, you can have 4xOC-192 (SONET) or 4xSTM-64 (SDH) in ONE wavelength for a total of 40 Gbps! And if you take into consideration that most SONET/SDH networks can use DWDM to put between 48 to 96 wavelengths in a single fiber, that would give us ... let me see... for 10G optical signals... a 960G total bandwidth!! And if you use the 40G signals (let's use 48 wavelengths for this case as I haven't seen the first DWDM system using 96 wavelengths with 40G optical signals) a total of 1920Gigs/sec!!!! And I know they are already working in the new 100G standard for DWDM!! So make your math!
No comments:
Post a Comment