How do electrical pulses limit the bandwidth of a standard telephone line?

[Jimmy87]

Hi PF, please could someone help explain how the bandwidth of a telephone line works. I have read that fibre optics uses light which has a high frequency and therefore can offer a high bandwidth versus a standard telephone line. But what I don't understand is the bandwidth of a standard telephone line. I understand that standard telephone lines use electrical pulses instead of light pulses but why are electrical pulses limited in terms of bandwidth? I didn't know electrical pulses had a frequency in the sense that radio and light waves do?

To be a bit clearer. To send a digital pulse of light requires a range of frequencies and since light operates at a high frequency it can deliver large amounts of digital data quickly. If you send an electrical pulse down a telephone then surely this also requires a range of frequencies but a range of frequencies of what exactly? I didn't know electrical signals had a frequency since they are not electromagnetic radiation in the sense that light and radio waves are.

Many thanks for any help offered!

 

[mfb]

I didn't know electrical pulses had a frequency in the sense that radio and light waves do?

Not directly, but you have to switch the voltage with some frequency to generate "voltage pulses". Those spread out over time, and if you make them too short (in time = in space), they merge after a while and you cannot separate them any more.

 

[davenn]

To be a bit clearer. To send a digital pulse of light requires a range of frequencies and since light operates at a high frequency it can deliver large amounts of digital data quickly.

That is also not quite right...
The first and still some digital optical fibre transmissions use monochromatic laser source

Its no different to the answer mfb gave

There's a whole bunch of modulation systems used for wire, fibre or radio data transmissions

from basic CW (OOK) through PCM, TDM to some of the more recent ones like QAM, to name just a few
of the 20 or so schemes

These in their own way vary the data transfer rateDave

 

[Drakkith]

I understand that standard telephone lines use electrical pulses instead of light pulses but why are electrical pulses limited in terms of bandwidth? I didn't know electrical pulses had a frequency in the sense that radio and light waves do?

Indeed. The quicker the pulse, the higher frequency it is. This puts a limit to how fast you can transfer information down a telephone line because at very high frequencies too much of the energy of the pulse is radiated from the lines as EM radiation.

Also, there is a limit to how fast the electrons in a conductor can respond. At infrared frequencies and higher the electrons simply cannot move back and forth fast enough, so the wave is absorbed.

To be a bit clearer. To send a digital pulse of light requires a range of frequencies and since light operates at a high frequency it can deliver large amounts of digital data quickly. If you send an electrical pulse down a telephone then surely this also requires a range of frequencies but a range of frequencies of what exactly? I didn't know electrical signals had a frequency since they are not electromagnetic radiation in the sense that light and radio waves are.

The change in voltage and current propagates as a wave down the lines. Each pulse can be considered to be a range of frequencies just like a light pulse.

 

[Drakkith]

That is also not quite right...
The first and still some digital optical fibre transmissions use monochromatic laser source

Technically the pulse consists of a range of different frequencies that interfere with each other to form the pulse, even in a monochromatic laser. I assumed this was what the OP was talking about.

 

[Jimmy87]

Indeed. The quicker the pulse, the higher frequency it is. This puts a limit to how fast you can transfer information down a telephone line because at very high frequencies too much of the energy of the pulse is radiated from the lines as EM radiation.

Also, there is a limit to how fast the electrons in a conductor can respond. At infrared frequencies and higher the electrons simply cannot move back and forth fast enough, so the wave is absorbed.



The change in voltage and current propagates as a wave down the lines. Each pulse can be considered to be a range of frequencies just like a light pulse.

Thanks to everyone for your answers. So Drakkith are you saying that as you increase the frequency of the voltage pulse to a similar frequency as optical fibres (i.e. infrared) the electrons cannot respond quick enough? So is the bandwidth of a phone line not limited by the voltage frequency per say but rather the electrons in the wire? What is a voltage pulse essentially created by and is it considered alternating hence the mention of EM radiation?

 

[Drakkith]

So is the bandwidth of a phone line not limited by the voltage frequency per say but rather the electrons in the wire?

No, the attenuation of a signal in a phone line is limited by a combination of effects well before the frequency gets to the IR range.

Loss types include: Skin effect, dielectric loss, radiation losses, and more. This is actually a fairly complicated topic. See chapter 19 here for more information: http://www.physics.wisc.edu/undergrads/courses/fall2011/623/sn/1-Transmission_Line_notes.pdf

There are also plenty of articles online if you do a google search for "attenuation in transmission lines".

What is a voltage pulse essentially created by and is it considered alternating hence the mention of EM radiation?

The voltage pulse is typically created by your computer's network card, a modem, or a telephone. Yes, it is alternating since the voltage is alternating in strength over time. The exact characteristics of the pulse depend on the type of modulation and/or line code chosen.

 

[Drakkith]

Here is an excellent resource for learning the basics of transmission lines: http://www.navymars.org/national/training/nmo_courses/NMO1/module10/14182_ch3.pdf

There is a wealth of info at this site about all things related to communications. See the NMO Second Class and NMO First Class links at the following site for this information: http://www.navymars.org/national/training/nmo_courses/

 

[256bits]

I didn't know electrical signals had a frequency since they are not electromagnetic radiation in the sense that light and radio waves are.

Many thanks for any help offered!

Your electrical power company supplies you with a 60Hz or 50 Hz alternating frequency of the electrical current down the power line.

Analog phone systems use a microphone to change the sound waves of your voice, into electrical waves that travel down the telephone line to the receiver where the small speaker changes the waves of electricity into sound for you ear.

Your stereo amplifier works on the electrical waves where it changes a small signal of varying amplitude to one that can be played out of your speakers. For records and tapes the signal was manipulated as an audio signal from the start with the pickup through the system to the speakers.

So, yes, there are such things as electrical waves that move through wires.

 

[sophiecentaur]

I don't think anyone has mentioned the essential difference between analogue and digital signalling and the actual bandwidths involved. (I may have missed this but it is worth stressing). The word 'bandwidth' has two connotations, these days - it can be actual occupied spectrum (as in the FM broadcast radio signal - about 75KHz) or the share of the bit rate, available to a user (say 10Mb/s for a budget 'broadband' service)
Conventional telephone lines carry audio signals in the form of a simple varying voltage. The bandwidth used for ordinary telephony (voice) is limited to about 3.4kHz (this is just enough for voice) in order that many voice signals can be mixed together and carried over trunk lines (incredibly crude system but pretty clever for the time it was first done). From your house to the exchange the limit is /was 3.4kHz. This can be carried on a trunk line, along with thousands of other analogue conversations - each one taking its own 3.4kHz slice of the bandwidth. (All analogue)
When you start to use digital methods (wires or optical fibre), the analogue signal is sampled at a high rate - way above the audio signal bandwidth - and converted to a series of pulses (these may be binary or multi-level, to confuse the issue further). The rate of these pulses may be many hundreds of kHz /MHz. This, on its own, would require a much higher quality of cabling - except for the fact that you can squeeze a string of binary pulses along the line and the receiving equipment can get over the impairment from a rubbish pair of twisted copper wires (what goes to most houses). You can also drastically bit-rate reduce your digital signal (advanced idea) to reduce the occupied bandwidth to something much more reasonable.
Digital signals from A to B always involve large bandwidths and carry many (thousands of) interleaved streams - telephone, video, data etc.. Hence, it is difficult to relate what your copper wire to your house does, to the optical fibre, from the box at the end of the road to the exchange and connects you to the Internet. The actual analogue bandwidth occupied by the optical signal will be many GHz. (In the end, every signal consists of an analogue variation of some quantity so there is a specifiable bottom line bandwidth of many GHz, in some cases.

 

[rcgldr]

Although not used much any more, dial up modems manage get up to 33.6 kilo bits per second when both calling and called modems are on analog lines, and up to 56 kilo bits per second when both calling and called modems are on digital lines.

 

[sophiecentaur]

Looking at the OP again, I see there is another issue and it's quite subtle. The number of 'levels' in an analogue signal is 'infinite' and there is a lot of information (infinite), potentially, in the actual level of even a very low 'bandwidth' signal. The limit of usability is when the random noise level interferes with the quality of the final audio. Digitising, turns all this on its head. If you are prepared to do without the 'infinite' resolution of an analogue transmission (digital sound is not 'perfect'), you can send quantised pulses, which are resolvable even with very high noise levels (it works until the noise amplitude is worse than half the peak to peak signal value). An analogue line will carry a digital signal at a very much higher frequency because it can resolve a 1 from a 0 (putting it crudely), even when the frequency response of the line has sagged to a very low level.

The weird thing is that the only limit to the bit rate that can be transmitted in a given analogue bandwidth, only depends upon the level of noise in the channel. (Look up Shannon information theory, if you are interested - it rapidly gets pretty meaty) You can. for instance, use more and more levels for your digital signal. If you reduce the noise by half then you can halve the gap between levels and this means you can send more information per pulse. This (plus very clever coding) is the basis of how they can transmit dozens of digital TV signals in the spectrum used by just one analogue TV channel.