Você está na página 1de 11

AIM: To study the adaptive delta modulation.

APPARATUS: 1. Function generator


2. Clock generator

3. Transmitter (Differential amplifier, Comparator, Bistable


circuit, Unipolar Bipolar converter, Integrator, Adaptive control circuit)

4.Receiver (Bistable circuit, Unipolar bipolar convertor, Adaptive


control circuit, Integrator, Low pass filter)

THEORY:
Adaptive delta modulation

In delta modulation there are quantization errors due to slope overload and
granular noise. To overcome these distortions, we use adaptive delta
modulation in which step size is made adaptive to variations in the input signal.

Adaptive delta modulation (ADM) or continuously variable slope delta


modulation (CVSD) is a modulation of DM in which the step size is not fixed.

1. In ADM, the processor has an accumulator and at each active edge of


clock waveform, generate a step.
2. In ADM step size is not fixed. Here to generate a step in response to the
INPUT a clock is generated having clock edge(k-1).
3. If the direction of step is same at the clock edge (k-1), then the processor
increase the magnitude of step by amount. If the direction of step is
opposite at clock edge (k-1) then the processor decreases the magnitude
of step by amount.
4. The output signal So (t) is called e(k), where e(k) represent an error signal.
e(k)=+1 if m(t)>m^(t) immediately before the kth edge
e(k)=-1 if m(t)<m^(t) immediately before the kth edge
the step size
S(k)=|S(k-1)|e(k) + Soe(k-1).
ADAPTIVE DELTA MODULATOR:

ADAPTIVE DELTA DEMODULATOR:


THEORY OF CIRCUIT:

1. The control circuit compares the present data bit from D flip flop with
the previous two data bits.
2. Its output to the counter is high when three bits are identical; the
control circuit output goes low, thus setting the counter advance with
every clock cycle. Each time counter is incremented from 00 to 11 is
reached where it remains in that state.
3. When slope overloading is not occurring, the integrator output always
hunt above and below the analog input even after it is caught up with it.
4. The output from a D flip flop is a constant change 1 to 0 at Tx clock edge.
5. The changing input to the control circuit ensures that output to the
counter is high and hence the counter is resettled at every clock cycle.
6. Thus, the control word from counter is always 00 forcing the integrator
gain at its lowest value thereby reducing quantization noise.
7. Now a fast changing analog signal appears at the input of the modulator
such that slope overloading occurs. The integrator output no longer
follows the analog signal but it spends its time trying to catch up the
analog signal.
8. The output from a D flip flop is a constant change from 1 to 0 for 3
consecutive times.
9. As soon as the third continuous 1/0 is sensed by the control circuit its
output is low.
10. The counter now advances to 01 doubling the integrator gain. This
increases the ramping rate of integrator and it is able to catch the analog
signal faster.
11. In the next clock cycle if the same situation continuous, the counter
advances to 10 forcing the integrator gain to 4 times its standard value.
12. This situation continuous till the counter advances to 11 where it
remains locked until control logic does not detect a change in the bit
level at its input.
13. As soon as the control circuit detects a change in bit level, its output
goes high, thus resetting the counter and normalizing the integrator
gain.
14. The demodulator of the in here spikes. The output from integrator is
passed to a LPF to smooth out the waveforms.
ADM ALGORITHMS

SONG ALGORITHM: -
In the year of 1971, C. L. Song et al. [7] proposed a nice algorithm for step size
adaptation by which the step size of the predicted waveform (delta m(t)) can
nicely be adapted. According to song algorithm, positive slope of the prediction
with respect to time results in the next prediction equal to the previous step size
added with the step size at 0th prediction (delta 0), similarly, for negative slope
of the prediction with respect to time results in the next prediction equal to the
previous step size minus the step size at 0th prediction (delta 0).
If the nth prediction sample is delta (n), the algorithm prescribes the adaptation
as shown in the following two equations.
….(1.1)

…..(1.2)

This addition and subtraction is called as accumulation in other terms. To


combine the Eqs (1.1) and (1.2) let us assume another parameter called as
discrepancy parameter (dn) as
dn =+1 if m(t) > _m(t)
−1 if m(t) < _m(t) ... (1.3)

Now, combining Eqs. (1.1) and (1.2), the nth sample can be calculated using
song algorithm of adaptation as

…..(1.4)

The song algorithm can now be used for digitization of any type of analog
signal shape. Let us take a typical step size analog signal for prediction through
accumulation. The ADM waveform through song algorithm is shown below.
in the Fig., the ADM predicted waveform is showing undue oscillation for
typically step like analog signal. For low slope and high slope signals of other
types can be tracked more efficiently by this method of adaptation. The slope
overload problem is solved totally by the song algorithm. The undue oscillation
can be viewed as a special type of granular noise.
SPACE SHUTTLE ALGORITHM
Space-shuttle algorithm is a modification over song algorithm only. We have
seen that, the negative slope or maintenance of constant levels of the analog
signal cannot be properly tracked by the prediction through adaptive delta
modulation by song algorithm. The step size of the accumulated waveform is
incremented and decremented by same amount (delta 0). But, allowance of this
linear decrement gives rise to an undue oscillation during constant level of m(t).
The new algorithm only changes the rule of decrement. Instead of linear
decrement space shuttle algorithm suggests constant decrement of amount
delta 0.
MATLAB CODE:
clc;
clear all;
close all;
t=0 : 2*pi/100 : 2*pi;
mt = sin(t);
mt = sin(t) +0.1;
quantizationallevels = 16;
s = (max(mt) - min(mt))/quantizationallevels;
totalsamples = length(mt) - 1;
mqt = zeros(1, totalsamples);
dk = zeros(1,totalsamples);
dt = 0;
sk = zeros(1, totalsamples);
for n=2 : totalsamples

dt = mt(n) - mqt(n);
if(dt >= 0)
dk(n) = 1;
else
dk(n) = -1;
end
sk(n) = abs(sk(n-1))*dk(n) + s*dk(n-1);
mqt(n+1) = mqt(n) + sk(n);
end
plot(t, mt,'r','linewidth',2);
hold on;
stairs(t, mqt,'k','linewidth',2);
ylabel('amplitude');
legend('input signal1, m(t)', 'modulated signal,
m_q(t)');
grid on;
WAVEFORM:
How is Adaptive delta modulation superior to Delta modulation?
• Delta modulation provides a staircase approximation of the input sampled
signal where only one bit per sample is transmitted.
• This one bit is sent by comparing the present sample value with the previous
sample value and the result whether the amplitude is to be increased or
decreased is transmitted.
• If the step is reduced, 0 is transmitted and if the step is increased then 1 is
transmitted.
• Delta modulation has a disadvantage of the presence of slope overload
distortion and granular noise.
• Slope overload distortion arises due to large dynamic range of the input
signal which results in large error between the original input signal and the
staircase approximated signal. When the slope of the signal is high, the step
size must be increased to reduce slope overload distortion.
• Granular noise arises when the step size is too large compared to the small
variations in the input signal.

=>To overcome these quantization errors due to slope overload and granular
noise, the step size is made adaptive to the variations in the input signal i.e. the
step size is not fixed and can be increased or decreased depending on the
variations of the input signal.

• Step size is determined by the previous and the present input samples.
• If the input is varied slowly then the step size is decreased. These is then
applied to the accumulator where staircase waveform is built at the
transmitter end and at the receiver low pass filter passes out staircase
waveform to reconstruct the original signal.
• Reduction in the slope overload distortion and granular noise in Adaptive
delta modulation induces improved signal to noise ratio as compared to the
Delta modulation.
• Since the step size is variable, the dynamic range of Adaptive delta
modulation is wider than Delta modulation.
• Bandwidth utilization is better in Adaptive delta modulation as compared to
Delta modulation.
• Thus Adaptive delta modulation superior to Delta modulation.
CONCLUSION:
1. Slope overload distortion and dynamic range of adaptive delta
modulation is better than that of delta modulation.
2. Effect of distortion is more in 1 kHz as compared to 2kHz i.e. as
frequency decreases overload distortion increases.
Q1:-What is adaptive delta modulation?
Ans :-The modulation in which step size is made adaptive to variations in the
input signal is called adaptive delta modulation.

Q2:-What are limitations of delta modulation which are overcome by adaptive


delta modulation?
Ans:-Slope overloading and granular noise.

Q3:-What is granular noise?


Ans:- The noise which arises when output from the circuit is coming “+ve” and
“-ve” alternatively,is called granular noise.

Q3:-What is slope overloading?


Ans:- Slope overload distortion arises due to large dynamic range of the input
signal which results in large error between the original input signal and the
staircase approximated signal. When the slope of the signal is high, the step size
must be increased to reduce slope overload distortion

Q4:-What are advantages of adaptive delta modulation?


Ans:-1.Improved signal to noise ratio.
2. Bandwidth utilization is better.
3.Dynamic range is wider.

Q5:-Give names of adaptive delta modulation algorithms.


Ans:-ADM has 2 algorithms:-
1.Song algorithm
2.Space Shuttle algorithm

Q6:-What is song algorithm?


Ans:- In song algorithm, positive slope of the prediction with respect to time
results in the next prediction equal to the previous step size added with the step
size at 0th prediction(delta 0), similarly, for negative slope of the prediction
with respect to time results in the next prediction equal to the previous step size
minus the step size at 0th prediction (delta 0).

Q7:-What is space shuttle algorihm?


Ans:- In space shuttle algorithm,constant decrement of amount (delta 0) is done
instead of linear decrement.

Q8:-What modifications are done in space shuttle algorithm over song


algorithm?
Ams:- The negative slope or maintenance of constant levels of the analog signal
cannot be properly tracked by the prediction through adaptive delta modulation
by song algorithm. The step size of the accumulated waveform is incremented
and decremented by same amount (delta 0). But, allowance of this linear
decrement gives rise to an undue oscillation during constant level of m(t). The
space shuttle algorithm only changes the rule of decrement. Instead of linear
decrement space shuttle algorithm suggests constant decrement of amount
delta 0.

Q9 :- What is the function of integrator in adaptive delta modulation circuit?


Ans :- Integrator increases or decreases the step size of the signal according to
the requirement.

Q10 :- Give Waveform of adaptive delta modulation circuit.


Ans:-

Você também pode gostar