M-ary Modulation

Contents:
y M-ary ASK
y M-ary PSK
y M-ary FSK
y M-ary
M QAM
 Bandwidth in M
M--ary Modulation
Multi symbols signals are often called M-ary signals
Bi
Binary Sequence → 0 1 0 0 1 0 1 1 1 0 1 1 0 0 0 1
S

Symbol Sequence → 0 1 0 0 1 0 1 1 1 0 1 1 0 0 0 1
S1 S2 S3 S4 S3 S4 S2 S1

Bandwidth
B d idth Bandwidth
B d idth
Requirement Requirement
= 2/Tb =2/(2Tb)
Tb Ts=2T
2Tb
Time period for M-ary signal = nTb , n= No. bits to form a symbol
IIn case off 4-ary
4 PSK n=2,
PSK, 2 BBandwidth=2/nT
d idth 2/ Tb = fb
In case of 16-ary PSK, n=4, Bandwidth=2/nTb = fb /2
 M-ary ASK
Number of symbol M=2n, where n=no of successive
binary digits

Magnitude

11
10
01
00
 M-ary PSK
In case of M-ary PSK the waveforms used to identify the symbols are

vm (t ) = 2 Ps cos(ω0t + φm )[∴ m = 0,1,2.....M − 1]

The symbol phase angle is given by φm = (2m + 1) π 2
v3 v1
2π M
v0
Es = PsTs π M
u1 (t ) = 2 Ts cos ω0t
vM −1

Es = Ps × nTb = PsTs
u2 (t ) = 2 Ts sin ω0t
 M-aryy PSK
Distance between two symbols d = 4 Es sin 2 (π M )
Es = Ps × nTb = PsTs

Eb = PsTb

d 2 = rθ If π/M is vey small, θ=sin π/M

d = 2r sin (π M ) d = 2 P T
s s sin 2
(π M ) r = PsTs

d = 4 PsTs sin 2 (π M ) d = 4nEb sin 2 (π M )

Ts = nTb PsTs = PsTb n = nEb

 M-aryy PSK
vm (t ) = 2 Ps cos(ω0t + φm )[∴ m = 0,1,2.....M − 1]
vm (t ) = ( 2 Ps cos φm ) cos ω0t − ( 2 Ps sin φm ) sin ω0t

Pe = 2 Ps cos φm
vm (t ) = Pe cos ω0t − P0 sin ω0t
P0 = 2 Ps sin φm

v( S m ) Sinusoidal Output
Serial to Digital
g to
b(t) Source Phase
Parallel Analog
Controlled by
Converter Converter
v(Sm)

M-ary PSK Transmitter

 M-ary PSK Receiver Ts
vm (t ) = Pe cos ω0t − P0 sin ω0t
∫ ( )dt
0
Raised
input
to Mth PeTs
power
p vm (t ) sin ω0t 0
vmM (t ) 1
A/D 2
BF N-1
N 1
filter
P0Ts
Mf0 Reconstructed
Ts Symbol
cos Mω0t
Freque
ncy cos ω0t
vm (t ) cos ω0t ∫ ( )dt
0
devider
÷M sin ω0t
 M-ary QAM
In M-ary PSK magnitude is same but phase is different. However, in
case of M-ary QAM both magnitude and phase of the symbols are
changed. If n=4, possible number of symbols are 16
u2 (t )
2a (a,3a) (3a,3a)

d = 2a = 0.4 Es
(a, a) (3a, a)
u1 (t )
 M-ary QAM
We can determine the average energy associated with a signal from the
four signals in the first quadrant. The average normalized energy of a signal
1 2
[
Es = (a + a 2 ) + (9a 2 + a 2 ) + (a 2 + 9a 2 ) + (9a 2 + 9a 2 )
4
]
Es = 10a 2 a = 0.1Es d = 2a = 0.4 Es
In the present case since each symbol represents 4 bits, the normalized
symbol energy is Es=4Eb where Eb is the normalized bit energy. So the
distance between two symbols is

d16−QAM = 0.4 Es = 4 × 0.4 Eb = 2 0.4 Eb

In case of 16-ary
16 ary PSK,
PSK the distance between two symbols

d16− PSK = 4nEb sin 2 (π M ) = 16Eb sin 2 (π 16) = 2 0.15Eb

In case of QPSK, the distance between two symbolsd QPSK = 2 Eb
Thus the probability of error QPSK < M − QAM < M − PSK
 M-ary QAM
vQAM (t ) = K1au1 (t ) + K 2 au2 (t )

Where K1 and K2 are each equal to ±1 or ±3

u1 (t ) = 2 Ts cos ω0t , u2 (t ) = 2 Ts sin ω0t , a = 0.1Es

0.2 Es 0.2 Es
vQAM (t ) = K1 cos ω0t + K 2 sin ω0t
Ts Ts
Since Es Ts = Ps

vQAM (t ) = K1 0.2 Ps cos ω0t + K 2 0.2 Ps sin ω0t

 M-ary QAM Transmitter
vQAM (t ) = Ae (t ) Ps cos ω0t + A0 (t ) Ps sin ω0t Ae , A0 = ± 0.2or ± 3 0.2
Since all four value of Ae and A0 are equally likely, we readily verify that Ae2 = A02 = 1
Ps cos ω0t
bk D Q

A/D

For Evaluation Only.

Copyright(C) by Foxit Corporation,2005-2009
Edited by Foxit Reader
bk+1 D Q Ae ((t )
Adder vQAM (t )
bk+2 D Q A0 (t )
A/D

bk+3 D Q
Ps sin ω0t
Clock
period =Ts
 M-ary QAM Receiver T s
vQAM (t ) = Ae (t ) Ps cos ω0t + A0 (t ) Ps sin ω0t
∫ ( )dt
0
Raised b0
A/D
input b1
to 4th Ae (t )
power
p vQAM (t ) cos ω0t
vmM (t ) Ts

BF
filter
∫ ( )dt
0 b2
4f0 A/D
A0 (t ) b3
cos Mω0t
Freque
sin ω0t vQAM (t ) sin ω0t
ncy
devider
÷4 cos ω0t
 M-ary QAM Receiver
v 4
QAM (t ) = P ( Ae (t ) cos ω0t + A0 (t ) sin ω0t )
s
2 4

Neglecting all terms not at the frequency 4f0 we left with

v 4 QAM (t )
Ps2
= A[
e
4
(t ) + A0
4
(t ) − 6 Ae
2
(t ) A 2
]
0 (t ) cos 4ω0t

⎡ Ae2 (t ) − A02 (t ) ⎤
+ Ae (t ) A0 (t ) ⎢ ⎥ sin 4ω0t
⎣ 2 ⎦

The average value of coefficient of cos4ω0t is not zero. While

the average value of coefficient of sin4ω0t is zero. Thus at the
output of the filter we recover signal at frequency 4f0
 M-ary FSK Modulation
Bandwidth requirements for M-ary FSK system:

f0=kfs f1=(k+2fs) f2=(k+4fs) f3=(k+6fs)

Carrier frequency can be successive even harmonics of the
symbol frequency fs=1/Ts. Say f0=kfs while f1=(k+2)fs and so
on. To pass M-ary FSK spectrum, the required frequency
range is
B = 2Mf s Where M=2n. Since fs=fb/n

B = (2 × 2n f b ) n = 2n +1 f b n
 Probability of error in M-
M-ary FSK
2
u3 = cos ω3t
Ts
d = 2 Es
Es
2
u2 = cos ω2t
Es Ts
Es

2
u1 = cos ω1t
Ts
M-ary FSK requires a considerable increased bandwidth in
comparison with M-ary PSK. The probability of error for M-
ary FSK decreases as M increases while for M-ary PSK, the
probability of error increases with M.
 M-ary FSK Tx and Rx
d0
d1 vM − FSK (t )
b(t) S/P N-bit D/A Frequency
converter converter M d l
Modulator

dN-1

ED d0

vM − FSK (t ) f0 d1
Select
ED N-bit A/D
largest
g Converter
f1 output

ED
dN-1
fM-1
 Minimum Shift Keying (MSK)
b(t) b1=1 b2=1 b6=-1 b7=-1 b8=-1
b3=-1 b4=-1 b5=-1
Tb

b0(t) b1=1 b7=-1

b3=-1 b5=-1

be(t) b2=1 b6=-1 b8=-1

2π
cos t b4= 1
b4=-1 2π
4Tb sin t
4Tb

2π 2π
b0 (t ) cos t be (t ) sin t
4Tb 4Tb
 Minimum Shift Keying (MSK)
vMSK (t ) = 2 Ps [be (t ) sin 2π (t 4Tb )]cos ω0t
+ 2 Ps [b0 (t ) cos 2π (t 4Tb )]sin ω0t
⎡ be (t ) + b0 (t ) ⎤
vMSK (t ) = 2 Ps ⎢ ⎥ sin(ω0 + Ω)t
⎣ 2 ⎦
Ω = 2π 4Tb = 2π f b 4
⎡ e
b (t ) − b (t ) ⎤
+ 2 Ps ⎢ 0
⎥ sin(ω0 − Ω)t
⎣ 2 ⎦
C H = (b0 + be ) 2, C L = (b0 − be ) 2 ω H = ω0 + Ω, ω L = ω0 − Ω

vMSK (t ) = 2 Ps C H (t ) sin ω H t + 2 Ps C L (t ) sin ω L t

b0 = ±1, be = ±1 If b0=be, CL=0 while CH=b0=±1

If b0=-be, CH=0 while CL=b0=±1
 Minimum Shift Keying
y g (MSK)
( )
In MSK, the two frequencies fH and fL are chosen to ensure
that the two p
possible signals
g are orthogonal
g over the bit
interval Tb Tb

∫ sin ω
0
H t sin ω L tdt = 0

sin(a − b) x sin(a + b) x
∫ sin ax sin bxdx = 2(a − b) − 2(a + b)
∴ 2π ( f H − f L )Tb = nπ , 2π ( f H + f L )Tb = mπ
fb fb fb
∴ f H = f0 + , f L = f0 − ∴ω H = ω0 + Ω = 2πf 0 + 2π
4 4 4
Putting fH and fL in above equation we have n = 1 and f 0 = mf b 4
 Minimum Shift Keying (MSK)
Since n=1, fH and fL are as close together as possible for
orthogonality to prevail. It is for this reason the present
system
t is
i called
ll d minimum
i i shift
hift kkeying
i

mff b f b fb mf b f b fb
∴ fH = + = (m + 1) and f L = − = (m − 1)
4 4 4 4 4 4

fb fb fb
∴ f H − f L = (m + 1) − (m − 1) =
4 4 2
fb 2
fb
∴ fH − fL =
2
fb