I've read that fiber optic cable has an almost nonexistent rate of data loss that is much less than that of cable or DSL.

I also learned in physics that when a ray of light passes through a tube with a bend in it, some of the light will be lost through refraction. This would seem to me to prevent fiber optic cables from being practical, whereas electrical signals traveling through wires don't experience this same phenomena.

What am I not thinking of? Does the insulation on fiber optic cables not allow any light to leave the cable?

 

I'm no expert, but the optic cable reflections are internal and not lost.

 

The process of making the cable from glass tubing creates a really clear inner core and a really good mirror sheathing it.

The optics has also improved radically over the decades. We now have fiber amplifiers that will amplify the signal +30 db or more, to the point the output requires laser filter glasses.

The spectral purity as also gone through roof. The early transmitter could do a fraction of a Meg (they were LEDs), now (because each laser is very pure with no sidebands) we can stack 40 channels at 40Gig/sec on one fiber. I used to be a tech that manufactured fiber optic gear (telephony before that). And this was 15 years ago, I suspect it has only gotten faster and better.

 

Wendy is correct. Total internal reflection is an essentially lossless phenomenon, and there is no refraction out of the glass (or core) when the angle of light rays are shallow enough (below the critical angle). The primary losses in silica glass fiber are OH absorption from very tiny trace amount of water molecules, Raleigh scattering (bouncing off SiO2 and other dopant molecules basically) and and a tail from a far infrared absorption band for the glass molecules.

The Raleigh scattering is higher in the blue-ultraviolet region and decreases (wavelength to the -4 power) as we go into the infrared communication band. The infrared absorption is high in the far infrared and decreases we go up in wavelength th the near infrared communications band. The lowest loss occurs at 1550 nm, right where Er-doped fiber amplifiers naturally amplify light (how fortunate), and the OH-absorption peak around 1450 nm creates another local minimum loss wavelength at 1310 nm (still higher loss than at 1550 nm), right where the glass's chromatic dispersion is zero (how fortunate). Also, with waveguide design, the zero dispersion wavelength can be shifted to the 1550 nm band (again, how fortunate).

There are no words that can adequately describe the amazing progress that man has made in making silica glass fiber just about as perfect as it can be. There are no words that can adequately describe the amazing aspects of nature that can provide us with a practically lossless and dispersionless transmission medium.

 

whereas electrical signals traveling through wires don't experience this same phenomena.

Upon further thought, I guess that's not true, given the inductive properties of wire coils?

 

Upon further thought, I guess that's not true, given the inductive properties of wire coils?

Actually, it is not the inductance, although you are thinking on the right track now. Losses like this would be radiation losses. Also, resistive losses in wires dissipate energy.

 

Also, resistive losses in wires dissipate energy.

Right, I forgot about that. I knew that the energy loss caused by resistance would weaken the signals, but not that it could cause some of them to get lost. But I suppose that the second logically follows from the first?

 

If there were such a thing as room temperature superconductors I suspect fiber would have competition.

 

Yes, only weaken the signals, and we can use the fiber amplifier to enhance the signals.

 

If there were such a thing as room temperature superconductors I suspect fiber would have competition.

That would be a great thing for power transmission, but for communications transmission, it would only mitigate one of the many limitations of wire based transmission. Bandwidth limits due to modulation limits and dispersion are important factors. Further, optical fiber diameter is much smaller than even the smallest coax cable designs.

 

Yes, only weaken the signals, and we can use the fiber amplifier to enhance the signals.

Yes, but that's not what I was asking. Since we know that energy loss due to resistance weakens the electrical signals traveling through a wire, would this mean that an extremely small percentage of the signals become so weak that they get lost completely? I'm trying to figure out what causes data loss in electrical wires (not really talking about fiber optic cables anymore).

 

Yes, but that's not what I was asking. Since we know that energy loss due to resistance weakens the electrical signals traveling through a wire, would this mean that an extremely small percentage of the signals become so weak that they get lost completely? I'm trying to figure out what causes data loss in electrical wires (not really talking about fiber optic cables anymore).

Your original question was about data loss, now you are talking about power or signal loss.
Data loss can happen from a weak signal but more likely because of extra bits introduced into a data stream that screws up the transmission. Power fluctuations/spikes/radio interference/ .... Can effect electrical signals whereas fiber is shrouded and suffers no similar issues. That is, stray photons do not jump into a glass fiber like capacitive or inductive effects can cause electrical fluctuations into a copper cable.

 

The following link tells fiber optic cable signal loss, attenuation, and dispersion.

http://www.juniper.net/techpubs/en_...oss-attenuation-dispersion-understanding.html

 

I saw tags attached to a cable in the underground parking lot in a building that said "Caution - Lightwire" . Took me a few moments to realize it was a bundle of fiber optic cable.