Analogue signals are converted to digital using sampling, basically at least two samples are made per cycle, and the more samples the higher quality of the digital output.
This is called quantisation; this maps the continuous range of signals and breaks down the bits into a digital signal.
Figure 2 - Analogue Signal With Sampling
Figure 3 - Digital Signal After Sampling
Data Sizes
A signal may be comprised of a
Bit - (1 Bit)
Byte - (8 Bits)
Nibble - (4 Bits)
Word – (16 Bits)
Packet -
Task 2b – P5
Synchronous & Asynchronous
Synchronous and Asynchronous are two different transmission methods; both of these methods have two clock signals to lock on to the packet when it’s sent and when it is received. Asynchronous is used with low bitrates.
Figure 4 - Asynchronous Packet
With the Asynchronous packet you have one start bit, eight data bits, one parity bit and one stop bit.
The oscillator will lock on to the start bit and then send the data and the stop bit will tell the oscillator that the packet has ended.
One down fall to Asynchronous is there is a lot of overhead, only 67% of the packet is actual data and 33% is wasted due to overhead.
However with larger amounts of data the clock would lose its accurate lock on the packet; this is where synchronous method is introduced.
Figure 5 - Synchronous Packet
With the Synchronous packet you have three synchronisation bits, one start bit, 1500 data bits and one end bit.
With Synchronous the efficiency is considerably higher with a 99% efficiency and 1% head due to overhead.
The clock will lock onto 3 synchronisation bits then it will start to send the data, then it will stop. This method is better when handling larger amounts of data without losing the lock from the oscillator.
Bandwidth
Bandwidth is the amount of data you can send down a signal per second; this was as little as 56k when modems where used, to present where is can usually average home users at 15Mb/s.
The bandwidth has been improved due to the improved mediums such as moving from RS232 to Coaxial and fibre cabling..
Claude Shannon’s Law
This is a law that states that the higher the noise the lower the channel capacity for the bandwidth, random noise is conflicting with the signal.
The formulae shown on Figure 5 shows how the signal to noise ratio is calculated
Figure 6 - Shannon’s Law Formulae ()
Different types of noise can conflict with the signal; there are different types of noise such as:
White Noise
Pink Noise
Johnson Noise
Error Detection & Error Correction
Parity Bit – RS232
There are a few types of error correction types; I will explain what these are and which ones are more efficient.
With RS232 – Serial you had a parity bit; this was part of the packet when you sent data. This worked by calculating the odd or even numbers. For example if the parity bit looks for an even number of 0’s and there is an odd number it will detect that there is an error and flip a bit to make it even.
However this is bad because RS232 has a limitation of one bit detection, when more than one bit flips over and it won’t detect any errors because it will show that it is even. Another downfall is that there is no retransmission on RS232 so the packet can’t be requested if there is an error that cannot be fixed.
Checksum
Checksum is where you have a fixed size of data which can be compared to rule out any accidental errors in transmission, this is done by calculating how many 1 bits they are and it will compare it with the sender, if the amount of 1 bits match you will have a successful transmission.
CRC (Cyclic Redundancy Check)
CRC is another method of error detection; this is hands down more complex than the parity bit in RS232. With CRC you have a long equation which is built in never NIC which does not change.
x32 + x26 + x23 + x22 + x16 + x12 + x11 + x10 + x8 + x7 + x5 + x4 + x2 + x + 1 []
This equation is divided by the sent data and the answer when calculated is the remainder.
Reed-Solomon
This is hands down the best solution for error detection and error correction, what this method does is check every single bit sent from point a to point b. This is used in large companies and this solution can be expensive.
Figure 7 - Reed-Solomon Error Detection
This can check for more than one error and also correct more than one error unlike RS232
Comparison of Mediums
Table 1 - Comparison of Mediums
As you can see from the table above is a comparison of different transmission mediums. This explains the price, modulation and bandwidth limit. As you would expect satellite is very expensive, this is because a satellite is a very expensive piece of equipment which uses the QAM modulation scheme.
A satellite would cost around the region of 100 Million and to send the satellite to space would cost around about 5 Million.
Radio however is a solution that is a lot cheaper; this uses the FM modulation scheme. Radio is used for low bandwidth signals unlike Satellite which is good for high bandwidth signals.
Copper cabling has an average price range; it is cheaper to implement that microwave and satellite because you can just lay the cable instead of dealing with high priced equipment. Satellites have a large range this is also why they are very expensive.
Data Compression
Data compressions is where you take the data and make the file size smaller than what it is, this is used when sending files over the internet and fitting files on small data mediums. There are several different methods of compression.
Huffman Encoding
Instead gggof using an ASCII code, the binary will get separated into segments of three and it well follow the tree diagram and it will show how many bits it takes to get to the letter in the tree. This will output a binary code a lot shorter than using ASCII codes which are normally about 7 bits.
Figure 8 - Huffman Tree Diagram []
The most letters used will use a certain amount of bits for example 00111 instead of an ASCII code using 0011101; this will save at least 2 bits. This is a very complex method. There are two different variations of data compression one is lossy and one is loss-less. Huffman Encoding uses Loss-Less compression.
Lossy
With file formats such as JPEG, MP3, MPEG, and PNG you can afford to drop and lose some bits here and there and quality won’t decrease noticeably.
Loss-Less
This is used with data compression where if you lose one bit your data will be corrupted.
RLL – Run Length Limited
Run length limited is a loss-less compression method, this is a simple method of reducing a file size for use in media such as CDs, DVDs and Flash Drives.
RLL simply counts how many ones and zero’s they are in a binary code. For example if you had this code
“000111001 11110000 1101011 00”
RLL would change it to this:
“3,0s;3,1s;2,0s;1,1s;4,1s;4,0s;2,1s;1,0s;1,1s;1,0s;2,1s;2,0s”
JPEG – Joint Photographic Experts Group
JPEG is a compressed image format that is a lossy form; JPEG can easily be a fifth of the original file format which would usually be a .BMP or .TIF file. You cannot use JPEG in:
CAD-CAM Applications
Images that have had mask and or shadow effects added
Images containing 256 colours or less. []
Figure 9 - JPEG Image []