After Google searching "Do Fibre Optic Cables attract any noise", most results return that they attract virtually no noise. Is this the case or are there some exceptions?
-
1\$\begingroup\$ Well, in the context of data communications, pretty much no noticable noise. Just the channel effects that @dll mentioned in his answer. But you don't actually mention you're doing communications and not e.g. quantum physical experiments, where the number of things somehow ending up in the optical fibre path might be relevant. Just out of clarity: You're asking because you want to build a network or a S/PDIF thing or so, right? \$\endgroup\$– Marcus MüllerMay 29, 2016 at 8:46
-
\$\begingroup\$ I'm actually doing a report on cabling for networking and the potential for noise difference between copper and fibre cabling was asked. \$\endgroup\$– Oliver KuchiesMay 29, 2016 at 8:50
-
1\$\begingroup\$ Ah well. But "noise in cables" is practically something that is technologically solved in even the cheapest networking equipment, and boils down to: a) Cat 6 copper is cheap, and for Gigabit Ethernet, the bit errors are negligible up to 100m easily. After that, it still reliably works outside very bad environmental settings b) for large distance, fiber optics are the cheapest and most reliable solution c) for 10GigE < 3m: direct attach copper. >3m of 10GE: fibre optics, or 10GBase-T, which has, again, Cat6 cabling,but transceivers are more expensive,so 10GBase-T only for legacy cabling. \$\endgroup\$– Marcus MüllerMay 29, 2016 at 9:30
-
\$\begingroup\$ But since you're doing a report, maybe there's another question that refers to that report and might be worth being asked as a separate electronics.SE question :) \$\endgroup\$– Marcus MüllerMay 29, 2016 at 9:31
6 Answers
You are right, this is the case but fiber optics can still have problems that can be perceived as noise that lead to incorrect data:
Intersymbol interference: This is a kind of noise because the previous symbol that was sent will interfere with the actual symbol that is being sent. Thus the previous symbol will act as noise. Well known techniques to help it are called Orthogonal frequency-division multiplexing (OFDM) and Orthogonal Frequency Division Multiple Access (OFDMA). You can find whole books about intersymbol interference.
Chromatic dispersion [ps/(nm·km)]: The refractive index of fibers varies slightly with the frequency of light, and light sources are not perfectly monochromatic. This has the effect that, over long distances and at high modulation speeds, the different frequencies of light can take different times to arrive at the receiver, ultimately making the signal impossible to discern, and requiring extra repeaters or special cables with adjusted indexes for every wavelength (so they arrive at the same time).
-
2\$\begingroup\$ These effects could be classed as distortion \$\endgroup\$ May 29, 2016 at 11:35
-
\$\begingroup\$ @Jasen Isn't what could be referred to as "noise" one way distortion could manifest itself, though? I know, it's kind of splitting hairs, but I suspect the intended meaning of "noise" is something like "less than the ideal, theoretically achievable useful signal amount at the other end of the cable". \$\endgroup\$– userMay 29, 2016 at 11:38
-
\$\begingroup\$ I mean that as an addition to the answer, not as a criticusm. \$\endgroup\$ May 29, 2016 at 11:41
-
1\$\begingroup\$ @MichaelKjörling er... no, noise has nothing to do with getting less than ideal, theoretically achievable useful signal. It has to do with getting more than ideal, theoretically achievable non-useful signal. In particular, noise (at least in my book) is by definition stochastic (unlike other inteference such a mains hum) and hence can only be removed after the fact by blunt-force wideband filtering. That's not quite true for noise-like effects that are actually artifacts from the signal itself, hence I would agree that such effects should better be termed distortions rather than noise. \$\endgroup\$ May 29, 2016 at 20:56
To make proper comparisons between fibre and cable you have to consider the photodiode at the end of the fibre to be part of the fibre and this is the weak link in terms of noise. Typically the Hamamatsu S5973 photodiode produces a noise equivalent power (NEP) of \$1.5 \times 10^{-15}\$ watts per Hz and given that the device is good for 1 GHz the noise power is going to be about 1.5 uW.
This photodiode converts watts to amps at approximately 2:1 therefore the noise current is about 0.75 uA RMS. You then have to ask yourself how much "signal" current is the photodiode producing and how this compares to the noise current.
I'm just trying to point out that you need to compare apples with apples.
Noise in optical communication typicaly refers to the deviation from an ideal signal, and is usually assocoated with random processes.
In general, the noise sources in a fiber optic link include noise from the RF amplifiers in the transmitter, the laser diode, the photodiode and RF amplifiers in the receiver. Of which laser noise arises from random fluctuations in the intensity of the optical signal.
The two main noise contributors are fluctuations in light intensity, which comes from the laser diode, and interferometric noise, which arises because of multiple light reflections in the optical fiber.
-
2\$\begingroup\$ A link on its own usually is not that helpful (because the link destination can vanish). Its better to include a brief summary of what the pdf contains, e.g. images of diagrams (use the upload button of stackoverflow). \$\endgroup\$– 0x6d64Jul 21, 2016 at 10:12
Assuming you are talking about DIGITAL optical cabling, there is virtually no possibility of "noise". There are potentials for OTHER kinds of artifacts which may be mistaken for "noise".
Fibre optic communication can be said that it does not attract any noise. But this is due to the fact that the sensitivity of the receiver is on lower side [for the noises]. Optical fibres are highly sensitive to different kinds of disturbances like temperature vibrations chemical changes, bending, etc. These can be refereed from fibre based sensors. Fibre sensors has been developed and some are being developed for direct detection of above parameters or dependent parameters. In reality the receivers for these sensors is much complex as detectors or receiver module has to be highly sensitive and it is a complex signal which needs to be decoded/analysed. In the communication systems optical loses, dispersion, ISI etc has more effect than induced noise so the power-budgeting and other considerations keeps communication data unaffected from other noises.
Fibre optic cables can be suitable for longer distance communication.Since the electrical energy is transmitted in the form of light,it may not be affected by noise from its environment.