Pulse transmission
A pulse or data train can be considered repetitive and analyzed using the Fourier series. In case of square wave, the frequency spectrum consists of only odd
harmonics.
Transmitting a square
wave or a pulse through a transmission medium
requires transmission of all the frequency components of the wave for a perfect reconstruction.
This implies that a channel must have infinite
bandwidth.
Pulse transmission
As the amplitudes of the harmonics decreaseexponentially, it is adequate that the channel has a
capability to transmit the fundamental and a few
more harmonic components that would permit
unambiguous determination of the pulse levels at the
receiving end.
Besides, most of the pulse trains are not square waves and have a d.c. component. Hence, the channel must be capable of transmitting d.c. components as well or
techniques may be adopted to remove d.c. components from the waveforms before transmission.
Pulse transmission
Apart from the adequate bandwidth, the channel must offer
equal attenuation and equal delay for all the frequencies within the bandwidth.
Measurable transmission line parameters corresponding to these two preceding requirements are attenuation constant Ac and phase velocity Vp.
Another source of distortion is the reflection of the transmitted wave by the transmission line. The related parameter for this is the characteristic impedance Z0 of the line.
When the Vp is not constant for all frequencies, certain
frequencies may be delayed so much that they interfere with frequencies corresponding to later pulses. This is known as intersymbol interference.
A constant Ac value for all frequencies would avoid amplitude distortion.
If the source and load impedances are equal to Z0, there is no reflection of the wave from the load, and hence there is no
Pulse transmission
There are two basic transmission techniques for establishing atime reference between the transmitter and the receiver:
Asynchronous transmission Synchronous transmission
In AT, the start bit at the
beginning of the transmission synchronizes the phase of the receive clock with that of the transmit clock.
The end of transmission is signalled by stop bits.
In AT, the bit stream length is usually eight bits corresponding to a character & hence it is often termed as character mode
transmission.
As there is 25% transmission
overhead, it is not economical to transmit large volume data using
Pulse transmission
In ST, a single clock is used forboth transmission and reception.
Transmitter clock is sent to the
receiver which uses the same clock to sample the incoming data.
Simple but an expensive way of
sending the clocking information to the receiver is to use a
separate channel for this purpose.
Alternatively, synchronous
receivers can be equipped with special tuned circuits or PLL circuits that are capable of
deriving the clocking information from the data itself.
The key to extracting clock from
the data is in the signal
transitions that occur in the data.
Line Coding
Binary digital information representation/encoding usually fall into three
broad categories:
Unipolar
Polar
Bipolar
Unipolar encoding is so named because it uses only one polarity. This
polarity is assigned to one of the two binary states, usually the 1. The other state, usually the 0, is represented by zero voltage.
Two problems that make it less desirable: a DC component and
synchronization
Line coding
Most transmission lines do not pass dc signals as they are ac coupled using transformers or capacitors which results in gradual decay of amplitudes of NRZ waveforms. This
phenomenon known as dc wander causes amplitude
reference to be lost for optimally discriminating between a one level and a zero level.
It is possible to nullify the effect of dc wander by using techniques that restore the dc value to zero level after
every pulse or line codes can be designed such that they do
Line coding
Polar encoding uses two voltage levels: one +ve & one –ve. By using both levels, in most polar encoding methods the average voltage level on the line is reduced and the dc component problem of unipolar encoding is alleviated. In Manchester (used by ethernet LANs) and differential
Manchester encoding ( used by Token Ring LANs), each bit consists of both positive and negative voltages, so the dc component is totally eliminated.
Line coding
In NRZ-L, the level of the signal is dependent upon the state of the bit; a positive voltage usually means the bit is a 0 & a negative voltage means the bit is a 1(or vice versa).
In NRZ-I, an inversion of the voltage level represents a 1 bit. A 0 bit is represented by no change. NRZ-I is superior to
NRZ-L due to the synchronization provided by the signal change each time a 1 bit is encountered. A string of 0s can still cause problems, but because 0s are not as likely, they are less of a problem.
Line coding
RZ encoding uses three values: positive, negative, and zero.
In RZ, the signal changes not between bits but during each bit. Like NRZ-L, a positive voltage means 1 and a negative voltage means 0. But, unlike NRZ-L, halfway through each bit interval, the signal returns to zero. A 1 bit is actually represented by positive-to-zero and a 0 bit by negative-to-zero, rather than by positive and negative alone.
Main disadvantage is that it requires two signal changes to encode one bit and therefore occupies more bandwidth.
Line coding
Probably the best existing solution to the problem of synchronization is
biphase encoding where the signal changes at the middle of the bit interval but does not return to zero. Instead, it continues to the opposite pole.
Manchester encoding uses the inversion at the middle of each bit interval for
both synchronization and bit representation. A n-to-p transition represents 1 and vice versa.
In differential Manchester, the inversion at the middle of the bit interval is
used for sync, but the presence or absence of an additional transition at the beginning of the interval is used to identify the bit. A transition means 0, no transition means 1. It requires two signal changes to represent 0 but only one to represent 1.
Line coding
Bipolar encoding, like RZ, uses three voltage levels: positive, negative, and zero. Unlike RZ, however, the zero level in
bipolar encoding is used to represent binary 0.
The 1s are represented by alternating positive and negative voltages. If the first 1 bit is represented by the positive
amplitude, the second will be represented by the negative amplitude, the third by the positive amplitude, and so on. This alternation occurs even when the 1 bits are not
consecutive.
Line coding
Bipolar alternate mark inversion (AMI) is the simplest type of bipolar encoding. A neutral zero voltage represents 0 and 1s are represented by alternating positive and negative
voltages.
A variation of bipolar AMI is called pseudoternary, in which 0 alternates between positive & negative voltages.
By inverting on each occurrence of a 1, bipolar AMI
accomplishes two things: first, the dc component is zero, and second, a long sequence of 1s stays synchronized. There is no mechanism to ensure the synchronization of a long string of 0s.
Line coding
Bipolar 8-Zero Substitution (B8ZS) is the convention adopted in North America to provide synchronization of long strings of 0s.
The solution provided by B8ZS is to force artificial signal
changes, called violations, within the 0 string. Anytime eight 0s occur in succession, B8ZS introduces changes in the
pattern based on the polarity of the previous 1 ( the 1 occurring just before the 0s).
Line coding
Using B8ZS, encode the bit stream
10000000000100. Assume that the polarity of the
first 1 is positive.
Line coding
The High-density Bipolar 3 (HDB3), used in Europe and
Japan introduces changes into the bipolar AMI pattern every time four consecutive 0s are encountered.
As in B8ZS, the pattern of violations in HDB3 is based on the polarity of the previous 1 bit. But unlike B8ZS, HDB3 also
looks at the number of 1s that have occurred in the bit stream since the last substitution.
Between two violation pulses, the odd number of bipolar pulses should be ensured.
Line coding
Using HDB3, encode the bit stream 10000000000100. Assume that the polarity of the first 1 is positive and the number of 1s so far is odd.
