Effects Of Noise In A Data Communication

This report will look into different types of noise that are associated with Unshielded Twisted Pair and Radio Waves. The noise that affects these transmission mediums such as thermal noise, crosstalk, multipath interference, intermodulation noise and impulse noise will be explored and the damages that it can cause to data being transmitted will be explained. I will also discuss the different modulation techniques and technologies that can be used to try and reduce the effect of the noise and reduce the risk of data loss through transmission.

Introduction

In 1962 Computer Scientist Joseph Carl Robnett Licklider developed ARPANET, which connected 4 computers across America; these computers were located in University of California Los Angeles, Stanford Research Institute, University of California Santa Barbra and the University of Utah. This network was designed for the purpose of sharing sensitive military data between different locations securely. However the first attempt at sending data over the network was not successful, as the UCLA computer crashed as they attempted to log into the computer at Stanford [1]. The result of these connection problems was the creation of TCP/IP and since then networks have grown in size and data rates and transmission mediums have evolved and new technology has been introduced, Noise has also started to play a part in how networks are built, as specific techniques can be put in place to try and reduce noise.

Guided Media

In a communication system using guided media, the signal is sent in the form of electromagnetic waves along a physical path. This physical path is what guides the signal, and can come in the form of 4 main media types, Unshielded Twisted Pair; Shielded twisted Pair, Coaxial or Fibre-Optic cables. However each of these mediums has several different standards of cables associated with them. This report will cover Unshielded Twisted Pair and the noise that can affect it.

UTP

UTP first originated in the 1970’s, it consists of 8 insulated copper wires, each of these copper wires has a diameter of 0.4mm to 0.8mm, and these copper wires are twisted together into pairs, so there ends up being 4 pairs of 2 wires, then all 4 pairs are wrapped in a protective plastic sheath. However UTP is susceptible to several different types of noise that can lead to signal impairment and even cause the loss of data. UTP uses Manchester Encoding

UTP Noise

When a data transmission is received, the received signal is often modified from the original signal that was transmitted; this modification is caused by noise. Noise is defined as ‘additional unwanted signals that are inserted somewhere between transmission and reception’ [2]. There are 4 different types of noise that will be researched; these are Thermal Noise, Cross talk, and Intermodulation Noise. These sources of noise can be placed into one of two categories, internal noise or external noise. Internal Noise is caused by the used of electrical components found in all communication systems. This internal noise could be produced by changes in current or imperfections on conducting materials. External Noise can be caused by different factors, such as lighting storms, or the use of large electrical machinery. [3]

Thermal Noise

Thermal Noise also known as Johnston or white noise was first observed in 1926 by John B. Johnston in Bell Labs. Thermal Noise is caused by electrons that become agitated at any temperature above 0, at this stage they begin to move in random patterns and bounce off other electrons, however in theory it could be stopped completely if all the components were kept at a temperature of absolute zero which is 0 Kelvin or -273.15°C as this would mean that all the electrons would move at their slowest meaning thermal noise would be as good as eliminated, however to try and achieve absolute zero and maintain it would be extremely difficult . [4]

Thermal noise is found across all the bandwidths typically used in a communication system and currently there is no practical way to completely eliminate it, however you can use different types of modulation to lower the frequency of the signal which in turn will lower the thermal noise, so for example if you had an Ethernet system and used PAM-5 modulation which has a frequency of 125MHz and this would give you a thermal noise value of 4.8015×10-13 WHz-1 at room temperature, where as if MLT-3 was used, you would end up with a thermal value of 1.200375×10-13WHz-1 under the same temperature conditions. [5]

To work this out the equation Pn= k . T. Δ f was used, where k is Boltzman’s constant, T is the temperature plus 273, in this case 18 degrees plus 273 which ends up as 291 for T, and Δ f is the frequency of 125×106 Hz for PAM-5 and 31.25×106 Hz for MLT-3.

Cross Talk

Cross talk is caused by the coupling of the copper cables magnetic and electric fields, which causes some of the signal to become lost or distorted. There are two main types of cross talk, NeXT (Near End Cross Talk) and FeXT (Far End Cross Talk), NeXT is when the coupling of magnetic and electric fields occurs near the source of the signal and FeXT is when it occurs near the receiver end. To try and prevent cross talk in UTP cables, the copper cables are twisted into pairs, the number of twists per foot/meter is defined as the twist ratio, so a cable with a higher twist ratio will be more efficient eliminating cross talk, as the twisting of the copper wires makes it harder for the coupling of cables as the loop area between the wires is reduced. However if you have a cable with a high twist ratio that means that you will be using more copper cable and the signal will have to travel a further distance to the receiver, meaning attenuation could become a factor. [6]

Intermodulation Noise

Intermodulation noise may be present in any communications system that sends signals at different frequencies across the same medium. Intermodulation noise produces signals that are the difference, sum or multiple of the two original frequencies. Intermodulation noise is caused by the transmission medium, transmitter and receiver not being linear systems, meaning that instead of the output matching the input, the output is different from the input. It can be caused by signal strength being too excessive for the device to handle or a problem with one of the components. An example of intermodulation Noise would be if there were two signals, 10Hz and 15Hz sharing the same transmission medium and there was intermodulation noise present, these two signals could become one signal at 35Hz. This would mean that not only have the two original signals been disrupted it could potentially disrupt a third signal if there was another 35Hz signal sent out on the medium. To overcome intermodulation noise, you can use Orthogonal Frequency-Division Multiplexing, which is explained more in the multipath interference section under unguided media. [2]

Unguided Media

When using Unguided Media in a communications system, the signal is sent through the air via an antenna in the form of electromagnetic waves, these waves have no specific path to follow. Unguided media used for several different communications systems like wireless, Bluetooth, infrared and satellite. Each of these systems use different types of unguided media for example satellite uses microwaves, but this report will focus on wireless and the noise that can affect the radio waves wireless uses.

Wireless

The first radio waves were sent by Guglielmo Marconi in Italy in 1895 and in 1899 he sent the first wireless radio signal across the English Channel [7]. Wireless works by an Omni directional antenna sending out a broadcast of radio waves, these radio waves are sent at a specific frequency depending on which standard they comply to, for example if the standard being used is 802.11n then they will be sent at 2.4GHz or 5GHz.

Wireless Noise

Wireless can be affected by many different things. This is because radio waves travel through air meaning it can be affected by different types of weather, like rain or snow causing scattering, or obstacles such as trees or buildings causing reflections. However it can also be affected by other devices transmitting at the same frequency causing signal loss.

Multi Path Interference

Multi Path interference is where a receiver receives multiple copies of the same signal, at delayed times, this mainly affects radio, as satellite or microwaves generally are line of sight so there would be no obstacles present for reflection to take place. However with radio waves it is caused by the antenna sending out broadcast signals, and these signals are then reflected between obstacles, and if these reflections arrive at the receiver it means that it will end up with several different copies of the same signal arriving at varying times, and depending on the different path lengths of the original direct signal and the reflected signals it could create a larger or smaller signal that is eventually received. Multipath Interference can cause a number of problems like data corruption, which occurs if there receiver picks up multiple different reflected signals and is unable to determine the transmission information, it can also cause signal nulling, where the reflected signals are received exactly out of phase with the original signal causing the original signal to be cancelled out. Not only can it cause data loss it can change the amplitude of the signal up or down, so if the reflected signals arrive out of phase with the original signal it will cause a drop in the signal amplitude but if they arrive in phase with the main signal the amplitude will increase.

To try and reduce multipath interference a diversity solution can be used. This works by using two antennas with the same gain, that are separated from one another but within the range of the same transmitter, this means that one of the antenna receive most of the multipath interference allowing the other antenna to receive a normal signal. [8]

Another way to reduce Multi Path Interference is to modulate the signal with Orthogonal Frequency-Division Multiplexing, OFDM works by splitting the signal up into 48 subcarrier signals. These 48 channels each carry a different portion of the data being sent and transmit them in parallel channels.

[9]

These subcarrier signals are modulated with BPSK, QPSK, 16-QAM or 64 QAM, and they have a convolution code rate of ½, 2/3 or ¾. The data rate of the signals is determined by the modulation used and the convolution code rate. Also there is 0.3125MHz frequency spacing between each of the subcarriers. [10]

[11]

OFDM also has a guard interval which means that any data arriving at the receiver will only be sampled once the signal has become stable and no more reflected signals are picked up that would cause changes to the phase or timing of the signal. Also because each subcarrier is on a different frequency any interference caused by reflected signals only affects a small percentage of the subcarriers meaning that the rest are received correctly.

[9]

Impulse Noise

Impulse Noise is an unpredictable problem. It consists of short spikes of high amplitude or short irregular pulses, these spikes and pulses are generated from a variety of different unpredictable causes usually however they relate to some sort of electromagnetic instability for example a lighting storm or any faults present in the communications devices. Impulse noise generally affects digital signals worse than it does analogue signals, for example if voice data was sent as an analogue signal and there was occurrences of impulse noise, the voice data would still be understandable as the impulse noise would create short crackles in the data, however with a digital signal the result of impulse noise could mean that all the bits sent through the duration of the impulse noise could be lost, it can however be recovered by sampling the received digital waveform once per bit time, but it can still result in a few bits being in error. As impulse noise is unpredictable, there is no way to eliminate it, however to reduce the effects of it, Coded OFDM can be used, this is very similar to OFDM, in the way that it splits the signal into multiple subcarriers, however Coded OFDM also has forward error correction that is included with the data. Because this error correction is included with the data it means that any data lost by impulse noise can be corrected at the receiver. [12]

Conclusion

After researching different types of noise and how it effects data communications, it became clear that it is a factor present in all systems and cannot be completely eradicated, as it can be caused by several different external sources made my man and internal sources caused by the data communication equipment. However, different strategies, techniques and error correction systems have enabled us to limit the effect that noise can have on a system and this has enabled technology to advance, meaning the chance of losing any crucial data due to the effects of noise is sufficiently lower now that what it was years ago.

Reflection

Throughout this report I have gained a better grasp of different aspects of data communications, for example, noise is present in all systems as any electronic device creates noise through the movement of electrons, imperfections in conductive surfaces and fluctuations of current. I also increased my knowledge of different types of modulation, and how they work regarding changing the frequencies or sending additional data to help with error correction. I have also gained knowledge on how noise can be caused by different types of weather and how they can affect the electromagnetic field and cause detrimental effects on data communication systems. Not only did this report help me gain more knowledge on data communications, it also increased my knowledge on different aspects of physics, and how closely the two subjects are connected.

I feel I completed this report to a reasonably high standard and found plenty of information available on the subject, however understanding this information was more difficult than expected as maths features highly in several of the sources I found, however this did not put me off, it simply lead me to try and comprehend the more complex maths side of the topic. Once I had completed the report I had to try and remove some parts as I had overshot the word count, this proved difficult as I felt I would be missing parts out if I removed some. Overall I would say I learned a great deal more about the complexity of noise and data communication systems.