What limits the speed of information through a copper wire?

[shpongle]

What limits the speed of information through a copper wire? What is the speed of signal in optic fibres? Is it faster and why?

 

[sophiecentaur]

There are two issues here. There is the speed at which an electromagnetic wave can travel along the wire, which depends upon the diameter of the wire and the materials surrounding it (insulation). The upper limit for this is, of course, c. But in most situations, the wave speed will be more like 2/3c. In nearly every case you will be using a pair of wires (or coax, with inner and outer conductors and not a single wire.
But there is another important factor when you are actually talking about the transfer of information. If you send a short pulse (flick a very fast switch on and off, for instance) then the bandwidth of the system becomes relevant. The short pulse will be dispersed into a wider, 'soggy' pulse and the receiver will need to decide on when the pulse has actually arrived, as its amplitude starts to build up. Over a short distance, where the propagation delay is short, the widening of the pulse will be the factor that limits the delay in actual information transfer. Also, the background noise (always present in all systems) will affect just when the (on / off) information can be received with certainty.

Amazing how such a short and succinct question can open such a can of worms . . . .

 

[jerromyjon]

http://en.wikipedia.org/wiki/Fiber-optic_communication#Bandwidth.E2.80.93distance_product

 

[sophiecentaur]

Which bit of that link were you referring to?

 

[jerromyjon]

The bandwidth distance product... followed by record speeds over several recent years

 

[gleem]

The speed of transfer is also intrinsically related to the capacitance of the transmitting cable, Before the final expected voltage shows up at the end it must charge the cable.

 

[sophiecentaur]

The speed of transfer is also intrinsically related to the capacitance of the transmitting cable, Before the final expected voltage shows up at the end it must charge the cable.

For a long line, it's normal to use the transmission line equation, which involves the characteristic impedance, rather than just considering the Capacitance. Depending on the terminating load, the line may look like a resistance, Capacitance or Inductance or, in the general case, an Impedance.

 

[jerromyjon]

Makes me wonder if the biggest difference between electrical and optic transmission speed or at least density is the proximity effect?

 

[sophiecentaur]

Makes me wonder if the biggest difference between electrical and optic transmission speed or at least density is the proximity effect?

I am not sure whether you are talking about speed or data rate here - and I'm not actually sure which the OP really wants to know about.

 

[shpongle]

I am not sure whether you are talking about speed or data rate here - and I'm not actually sure which the OP really wants to know about.

I should have mentioned : speed of information transfer.

 

[jbriggs444]

A less ambiguous phrasing would be "throughput" versus "latency". You appear to care about latency.

 

[sophiecentaur]

I should have mentioned : speed of information transfer.

I think you must mean 'rate' of transfer. That is Bits per second.
"Speed" is assumed to be Distance per unit time. Two entirely different ideas. Delay and rate can both be of importance in data systems.

 

[PaoloDiM]

Which bit of that link were you referring to?

Haha Good pun!

 

[PaoloDiM]

What limits the speed of information through a copper wire?

Interesting thread! Most here have their engineering hats on. I come from linguistics and A.I., so the limits of the "information" transfer would depend on the local "neural" configurations much more than the medium - unless that medium gave more "information" about its relevance to the content.
Is this relevant?

 

[Delta2]

What limits the speed of information through a copper wire? What is the speed of signal in optic fibres? Is it faster and why?

According to shannon's theorem http://en.wikipedia.org/wiki/Shannon–Hartley_theorem it is bandwidth and noise the main factors that affect the data throughput (bits per second) transmitted through a medium. If we care about latency , then the speed of the electromagnetic waves in the medium also plays a role. Optic fibers are better in all those 3 factors over copper wires or coaxial cables. Optic fibers offer enormous bandwidth, zerolike noise and speed of electromagnetic wave almost at theoretical maximum near c=3*10^8m/s^2. The usual twisted copper pair of wires has limited bandwidth because the signal frequencies attenuate the longer the wire is due to electromagnetic radiation and due to conversion of signal energy to heat. Coaxial cable is much better on these two factors as well as shielding the signal from external noise . Speed of the signal in twisted pair and coaxial cable is about 0.70c. As you see only optic fiber offer the absolut best in all factors.

Now SophieCentaur can you tell me how the diameter of wire affects the speed of electromagnetic wave in it? at least using classical physics i don't see how one can arrive at this result.

 

[sophiecentaur]

According to shannon's theorem http://en.wikipedia.org/wiki/Shannon–Hartley_theorem it is bandwidth and noise the main factors that affect the data throughput (bits per second) transmitted through a medium. If we care about latency , then the speed of the electromagnetic waves in the medium also plays a role. Optic fibers are better in all those 3 factors over copper wires or coaxial cables. Optic fibers offer enormous bandwidth, zerolike noise and speed of electromagnetic wave almost at theoretical maximum near c=3*10^8m/s^2. The usual twisted copper pair of wires has limited bandwidth because the signal frequencies attenuate the longer the wire is due to electromagnetic radiation and due to conversion of signal energy to heat. Coaxial cable is much better on these two factors as well as shielding the signal from external noise . Speed of the signal in twisted pair and coaxial cable is about 0.70c. As you see only optic fiber offer the absolut best in all factors.

Now SophieCentaur can you tell me how the diameter of wire affects the speed of electromagnetic wave in it? at least using classical physics i don't see how one can arrive at this result.

Data rate and signal delay, due to path length are two different things - and you mention both. I am still not sure which the OP was asking, and that demonstrates how important the correct vocabulary is in these matters. Actual delay time between the transmission and reception of the information can sometimes be relevant but, very often, a delay is quite acceptable and can allow suitable coding that will actually increase the useful data rate. The word "bandwidth" is often mis-used when information rate is in fact what is meant (as in "broadband data"). Bandwidth is strictly an analogue term because there is no one-to-one relationship between the two. The channel noise level is equally important.

The wave does not travel "in" the wire but is directed by the wire. Also, there are only a very few systems for signal transmission that use only a single conductor. In a practical transmission line, the wires are supported by or even encased in insulators. That will affect the propagation speed (delay).

 

[Doug Huffman]

I remember Admiral Grace Hooper's 'nanosecond' that she carried.
https://en.wikipedia.org/wiki/Grace_Hopper#Anecdotes

 

[tech99]

Data rate and signal delay, due to path length are two different things - and you mention both. I am still not sure which the OP was asking, and that demonstrates how important the correct vocabulary is in these matters. Actual delay time between the transmission and reception of the information can sometimes be relevant but, very often, a delay is quite acceptable and can allow suitable coding that will actually increase the useful data rate. The word "bandwidth" is often mis-used when information rate is in fact what is meant (as in "broadband data"). Bandwidth is strictly an analogue term because there is no one-to-one relationship between the two. The channel noise level is equally important.

The wave does not travel "in" the wire but is directed by the wire. Also, there are only a very few systems for signal transmission that use only a single conductor. In a practical transmission line, the wires are supported by or even encased in insulators. That will affect the propagation speed (delay).

Regarding waves not traveling in the wire, an alternative view might be that a longitudinal wave of compression propagates through the electron plasma which fills the wire. This then creates the fields surrounding the wire, and in some cases produces EM radiation.

 

[sophiecentaur]

Regarding waves not traveling in the wire, an alternative view might be that a longitudinal wave of compression propagates through the electron plasma which fills the wire. This then creates the fields surrounding the wire, and in some cases produces EM radiation.

I cannot imagine any model based on that would produce a propagation speed of more than a small fraction of c. More like the speed of sound, I should have thought. Do you have any reference to that idea or is it one of your own?

 

[davenn]

This then creates the fields surrounding the wire, and in some cases produces EM radiation.

Are you really serious ??
I'm not familiar with any case where an EM field isn't created by the movement of electrons in a wire/conductor

D

 

[Delta2]

Data rate and signal delay, due to path length are two different things - and you mention both. I am still not sure which the OP was asking, and that demonstrates how important the correct vocabulary is in these matters. Actual delay time between the transmission and reception of the information can sometimes be relevant but, very often, a delay is quite acceptable and can allow suitable coding that will actually increase the useful data rate. The word "bandwidth" is often mis-used when information rate is in fact what is meant (as in "broadband data"). Bandwidth is strictly an analogue term because there is no one-to-one relationship between the two. The channel noise level is equally important.

The wave does not travel "in" the wire but is directed by the wire. Also, there are only a very few systems for signal transmission that use only a single conductor. In a practical transmission line, the wires are supported by or even encased in insulators. That will affect the propagation speed (delay).

\

When i say bandwidth i mean the bandwidth in frequency range in Hertz, not the data transmission rate. To tell you the truth i am not sure how delays allow for better coding schemes with faster data rate, but this is not my interest here.

Yes the insulation affects the speed of the wave (which travels inside the wire as well, only the direction of travel is changing, from a direction say along the z-axis it becomes radial (speaking about a cylindrical wire)) BUT how the diameter of the wire (with or without the insulation) affects the speed of the wave?

 

[Tom_K]

\


Yes the insulation affects the speed of the wave (which travels inside the wire as well, only the direction of travel is changing, from a direction say along the z-axis it becomes radial (speaking about a cylindrical wire)) BUT how the diameter of the wire (with or without the insulation) affects the speed of the wave?

The geometry of the conductors has an effect on the characteristic impedance of the transmission line. The idea here is to match the impedance of source and load for maximum transfer of power. The geometry of the conductors has no direct effect on the velocity of transmission.

It is the insulating material that affects the propagation velocity.

 

[jerromyjon]

I'm still a little unsure about a few things but sticking with the op's question... in electrical systems there are 2 wires in the twisted pair, both emitting unwanted radiation which causes interference to the signal, while only 1 conductor in that pair is carrying information? In optical systems, much like wireless, there is only the EM wave carrying the information?

 

[Delta2]

I'm still a little unsure about a few things but sticking with the op's question... in electrical systems there are 2 wires in the twisted pair, both emitting unwanted radiation which causes interference to the signal, while only 1 conductor in that pair is carrying information? In optical systems, much like wireless, there is only the EM wave carrying the information?

Yes you are right about optical and wireless systems.

In twisted pair though both conductors carry the same information (they have equal but opposite currents). The interference is constructive in the space between the two conductors, meaning that the EM-wave is amplified there, while it is destructive in the region outside the conductors which means the EM wave is nearly zero there. And we want this last thing to happen so we don't have much loss of energy due to radiation.

 

[tech99]

Are you really serious ??
I'm not familiar with any case where an EM field isn't created by the movement of electrons in a wire/conductor

D

Agree on that point.

 

[tech99]

I cannot imagine any model based on that would produce a propagation speed of more than a small fraction of c. More like the speed of sound, I should have thought. Do you have any reference to that idea or is it one of your own?

I am afraid it is my own idea, based on plasmons. I have been using this concept to explain surface waves and radiation.

I cannot imagine any model based on that would produce a propagation speed of more than a small fraction of c. More like the speed of sound, I should have thought. Do you have any reference to that idea or is it one of your own?

I thought that as the guided wave travels along, surrounding the wire, each electron in turn will give a back and forth kick as the wave passes. When viewed as a whole, this would constitute a longitudinal wave traveling in the wire at almost c.

 

[davenn]

am afraid it is my own idea, based on plasmons. I have been using this concept to explain surface waves and radiation.

Then you may want to review the PF rules about expounding one's own theories ...
briefly ... its a no-noregards
Dave

 

[oz93666]

No one has yet addressed the fiber optic part to the question...

Two points spring to mind , we need to know the speed of light in the medium the cable is made of , perhaps around 60% c ...and the light does not take a direct rout, but to some extent bounces around inside the fiber.

 

[jack476]

What limits the speed of information through a copper wire?

The simple answer is in two parts.

First, the speed of propagation of an electromagnetic wave is reduced when it travels through material. All electrical signals are waves, but normally the wave behavior of electricity can be ignored. In copper, the speed of an electromagnetic wave is reduced from 299,792,458 meters per second to 285,102,627. So the speed of the wave is reduced in the first place by the fact that it is traveling through a medium, but only by about 5%.

Second, the copper wire has non-zero impedance. That is to say that the copper wire has some resistance over its length, but more importantly due to self-inductance it also has reactance, the imaginary part of impedance (resistance is the real part) that can be thought of as something like a frequency-dependent resistance. The impedance increases as the frequency of the signal increases, and this means that for higher frequencies the amplitude is reduced and the phase of the signal is shifted, so it becomes harder to protect the signal from being degraded by background noise.

So it's not really the velocity of the propagating wave, ie its "speed", that's important, but rather the way that electrical conductors behave when they're carrying high-frequency signals. Limited frequency means that it's harder to send information as quickly.

These two problems are the basis of transmission line theory.

What is the speed of signal in optic fibres? Is it faster and why?

Optical fibers are made of what is essentially glass, so the speed of the propagating wave is reduced to about 200,000,000 meters per second. This may seem significantly lower than with the copper fiber, but remember that it's the maximum frequency the line can carry that matters much more than the speed of propagation of the signal through the wire in terms of carrying information.

Optical fiber can support a vastly higher frequency than copper wire. Glass is not an electrical conductor, so self-inductance does not occur. Furthermore, in a copper line the current is going to want to induce currents in nearby conductors, forcing it to do work, and thereby expending the energy of the signal. It's also going to be influenced by currents induced by background noise. None of this is a problem in optical fiber, so it can support a higher frequency due to not needing to spend energy protecting the signal from corruption by noise.

 

[sophiecentaur]

\

When i say bandwidth i mean the bandwidth in frequency range in Hertz, not the data transmission rate. To tell you the truth i am not sure how delays allow for better coding schemes with faster data rate, but this is not my interest here.

Yes the insulation affects the speed of the wave (which travels inside the wire as well, only the direction of travel is changing, from a direction say along the z-axis it becomes radial (speaking about a cylindrical wire)) BUT how the diameter of the wire (with or without the insulation) affects the speed of the wave?

This involves two entirely different questions - qhich one do you want to discuss on this thread - the propagation speed or the bandwidth?