Chapter 1 - Noise

Contents
¾ Introduction
¾ Definition of Noise
¾ Sources of High Frequency Noise
¾ Thermal Noise
¾ Shot Noise
¾ Flicker Noise
¾ Propagation of Noise Through a System
¾ Noise Through a Linear Filter
¾ Systems With Several Noise Sources
¾ Frequency Transformation of Noise
¾ Noise Representations
¾ Noise Figure
¾ Noise Temperature
¾ Equivalent Input Noise
¾ Cascade of Noisy Networks
¾ Measurement of Noise
1 2
RF IC Technology - Noise RF IC Technology - Noise
References for Noise Types of Noise
• Selected Papers, Provided by Lecturer

• External Noise
– May be of random or regular nature from outside
• Design of Analog CMOS Integrated Circuits, Behzad Razavi, sources
Chapter 7, pp.201 -245
• Analysis and Design of Analog Integrated Circuits, Gray and – Interaction between the circuit and the outside
Meyer, 4th Ed., Chapter 11, pp. 748 – 807 world
• Design with Operational Amplifiers and Analog Integrated – Interaction between different parts of circuit
Circuits, Franco, Third Ed., Chapter 7, pp. 311 - 346
• Analog Integrated Circuit Design, David Johns, • Internal Noise
Chapter 4, pp. 181 – 220, Ed. 1997 – Random signals due to the natural phenomenon
3 4
Intrinsic Noise RF Signal Along With Noise
Sn(f)
(f)
• Thermal Noise
RX
• Shot Noise RF
Good f
• Flicker Noise (1/f) RF f
• Burst or Popcorn Noise
RX
Bad RF f
5 6
Quantitative Definition of Noise

Definition of Noise
Statistical Models
Probability Density Function (PDF): p(n)

• Noise is a RANDOM PROCESS
• The value of noise can not be predicted at any time p(n) dn = probability of n1 < n < n1 + dn
• The average power for most types of noise is n2

predictable by observing noise over a long time
∫ p(n)dn = P( n < n < n
n1
1 2 )
+∞
−∞
∫ p(n)dn = 1
7 8
Distribution of PDF Average Noise Power
Gaussian Model
Distribution of Probability Density Function (PDF): p(n) For a periodic (T) voltage signal v(t) across
a load RL : 1
+T / 2
v 2 (t )
T −T∫/ 2 RL
Pav = dt
p(n)
1  1n  2
1
+T / 2
x 2 (t )
p(n) = exp−
T −T∫/ 2 RL
2 Pav = lim T →∞ dt
2π σ n  2 σn  2σn
+σ n +T / 2
1
∫ p(n)dn = 0.68 n
Pav = lim T →∞ ∫ x 2 (t )dt
T −T / 2
−σ n - σn + σn
+T / 2
P( - σn < n < + σn ) = 0.68 1
n 2 (t ) = lim T →∞ ∫ n 2 (t )dt
T −T / 2
9 10
Noise in Time Domain Noise in the Frequency Domain

Noise is expressed as a Fourier Series over a period T :
K =+∞
 2π K 
n(t ) = ∑X K exp  j
 T 
t BPF 1 Hz
n f 1 (t ) n 2f 1 (t )
K =−∞
n(t ) ( )2
K =+∞
 2π K  f1
n(t ) = ∑X
K =−∞
K exp  j
 T 
t = X0
K =+∞
n 2 (t ) = X 0 + 2 ∑ X K X K∗
2 Noise average power in a 1 Hz bandwidth around a frequency f1 : S n ( f1 )
K =1
11 12
Power Spectral Density (PSD) Sources of Noise
of Noise
The Amount of mean-squared noise over a finite
bandwidth ∆ f = f1 - f2
f2 • Thermal Noise, Johnson’s Noise
n (t ) = ∫ Sn ( f ) df
2
• Shot Noise
f1
• Flicker Noise (1/f)
n 2 (t ) = Sn ( f ) ∆f
• Burst or Popcorn Noise
n 2 (t )
Sn ( f ) =
∆f Vn2
Sn ( f ) =
∆f
13 14
Thermal Noise Thermal Noise

• A thermally (thermal energy) generated noise due to Power Spectrum
random motion of the charge carriers; electrons Instantaneous
Density (W/Hz)
• The average noise power is proportional to: Noise Voltage
– Temperature
KT
– Frequency bandwidth (spectrum) of the thermal noise
Time
Pn = KT B = KT ∆f (Watt) 0 Frequency (Hz)
Pn
Power Spectrum Density : Sn = = KT (Watt/Hz)
∆f
K = 1.38 x 10 -23 J/oK
15 16
Resistor Thermal Noise Root-Mean-Square Value of Noise
Vn2 Vn
v n (t ) Vn2 Pn = = KTB Vn = 4RKTB
4R
R ≡ R ≡ R
R
Vn2 = 4RKTB I n = 4GKTB

Noisy Resistor Noiseless Resistor
Noiseless Resistor
I = 4GKTB
2
n Valid up to very high frequencies of 10 GHz, f <<
KT
= 6.2 THz
h
Planck’s constant, h = 6.67 x 10 -34 J.S
17 18
Thermal Current Noise PSD of Thermal Noise
SV(f) (V2/Hz)
I n2 = 4GKTB Pav = V = 4RKTB

n
2
(V2)
R in
4RKT
I n = 4GKTB Vn2
Sn ( f ) =
Noiseless ∆f 0 f (Hz)
Resistor
Sn ( f ) = 4KTR (V2/Hz)
19 20
Thermal Noise of Series Resistors Thermal Noise of Parallel Resistors
X nT = X n1 + X n2
2
I nT = I n21 + I n22
V = V +V
n
2 2
n1
2
n2
2
= ( X n1 + X n2 ) 2 I nT = I n21 + I n22
X nT Vn = Vn21 + Vn22
2
I nT = 4(G1 + G2 )KTB
Vn2 = 4(R1 + R2 )KTB
2
X nT = X n21 + X n22 + 2 X n1 X n2
Vn2 = 4RS KTB
2
I nT = 4GT KTB
Vn2 = 4∑ Ri KTB
2
I nT = 4∑Gi KTB
i i
21 22
Effective Bandwidth Effective Bandwidth

Parallel Resistor and Capacitor: 1
| H (ω) |= +
Vno (s) 1 + 1 + (ωRC) 2 R
H (s) = | H (ω) |= R C Vno
Vni (s) 1 + (ωRC) 2
C Vno 1
Sno ( f ) = KT
2
1 + (ωRC) 2 Vni2 -
Sno ( f ) = Sni ( f ) | H (ω) |2 V ni -
~
1 Sno Power Spectrum
Sno ( f ) = KT Density (W/Hz)
1 + (ωRC) 2 KT KT
Pno = = KTBeff
∞ f (Hz) 4RC
KT
Pno = ∫ Sno ( f ) df = = KTBeff 0 Beff =
1
0
4RC 4RC
23 24
MOSFET’s
Noise Voltage at the Output
Channel Thermal Noise
Vno (s) The most significant source of the thermal noise generated
H (s) = +
Vni (s) in the channel is called Drain Noise or Channel Noise;
Noise
R
2 C Vno For long channel devices operating in the saturation region
V KT
Pno = =n
4R 4RC Vni2 - Long channel: γ = 2/3

Submicron : γ = 2.5 for 0.25 micron technology
KT γ : is bias-dependent parameter
Vn2 =
C
KT
Vn,rms = I n2 = 4 KT γ g m ∆f
C
I n2 = 4 KT γ g d 0 ∆f
Ref.: Y. Tsividis
25 26
Induced Gate Noise Induced Gate Noise

ing
ind
VGS _+ G 16
_+ VD 2
ing = KTω 2 C gs2 ∆f
15
S D 2
C gs = CoxWL
Fluctuating channel potential couples
2
ing 3
capacitively into the gate terminal, ∆f Slope=
20 dB/decade
causing a noise gate current “Analysis and Design of Analog Integrated Circuits,” Gray and
4 KT γ g do Meyer, 4th Ed., Chapter 11, pp. 748 – 807
2
 2πf 
2
ing = 4 KTδ g do   ∆f
 ( 5 / α )( g C ) 
 m gs 
5 ft f
α
- δ is gate noise coefficient Typically assumed to be 2γ
- Correlated to drain
27 28
Correlation factor between the induced gate noise
Noise Parameter As a Function of Vd and the channel drain noise
ig id*
B + jC =
ig2id2
Drain Voltage VDS

29 30
Thermal Noise in MOSFETs (Cont.) Thermal Noise in MOSFETs (Cont.)
The ohmic sections also contribute thermal noise. In which R1 =RG /3 Vn2,RD +
_
In a relatively wide transistor:
D
Vn2,R1 RD
R1
RG1 RGn +
_
RG2
We will hereafter neglect
G
the thermal noise due to the RS
ohmic sections of MOS
devices Vn2,RS +
_
S
RG1+ RG2+ … + RGn= RG
31 32
Noise Sources in a CMOS Thermal Noise in MOSFETs (Cont.)
Amplifier
In which R1 =RG /3
Vn2,RD +
_
While the thermal noise caused

in channel is controlled by gm, Vn2,R1 RD
R1
the thermal noise caused by RG +
_
can be reduced in a folded device,
and the thermal noise caused by RS
RS and RD are negligible
due to small resistor values. Vn2,RS +
_
We will hereafter neglect

⇒ the thermal noise due to the
ohmic sections of MOS devices
33 34
Shot Noise Shot Noise

• Associated with discrete packets of charge emission, or the T
charge carriers crossing a boundary; potential barrier region, 1
in2 = (i − I D )2 = limT →∞
T0∫ (i − I D )2 dt
• Depends directly on the direct component of the current

in2 = 2qIDC ∆f (A2)
# of Electrons # of Holes
p n in2
Power Spectrum Density : Si ( f ) = = 2qI DC (A2 /Hz)
∆f
j
It is valid for the fmax (less than or) comparable to 1/τ ,
iJ ID where τ is the junction transit time
t
35 36
Shot Noise Flicker Noise
• Flicker noise or Contact noise occur due to the imperfect contact
p(I n ) “Contamination” between two conducting materials that causes the
1  1i  2
p(i) = exp− 2 conductivity to fluctuate in the presence of a dc current.
2π σ  2σ 
2 σI
Kf Im
σ 2 = 2qI DC ∆f • mean-square flicker noise current in 1 Hz frequency band: i =
2
f
where Kf is the flicker noise coefficient, I is the dc current,
fn
m is the flicker noise exponent, and n~1. (A2 /Hz)
σ = 2qI DC ∆f In
-σI ID +σI
• Flicker noise is modeled by a noise current source in parallel with the
Current amplitude lies between device.
ID ± σI for 68 percent of the time
One-over-f-noise/ low frequency noise/ pink noise
37 38
Flicker Noise, (1/f) Noise Flicker Noise PSD

• Associated with the fluctuation in carrier density due to trapped
electrons “Crystal defects”, for example, the dangling bonds existing Si(f)
Log scale
in the MOS oxide-substrate interface. I2
I α
Si ( f ) = n
= Kf
∆f f
Iα
I n2 = K f ∆f (A2) (A2 /Hz)
fβ Log scale f
Kf : constant for a particular device
1/f Generation
α : constant in the range 0.5 to 2 I n,rms Device
• In devices exhibiting high flicker noise levels,
β : constant of about unity this noise source may dominate the device
noise at frequencies well into mega hertz
Iα • In general, Kf is an unknown constant

I n,rms = K f ∆f • May varies by orders of magnitude from one device type to the next
fβ
• Vary widely for different transistors or integrated circuits from the
same process wafer
39 40
Flicker Noise in MOSFETs Flicker Noise in MOSFETs (Cont.)
The flicker noise in MOSFETs can be easily modeled as a voltage
source in series with the gate and roughly given by: SO WHAT HAPPENS TO TOTAL FLICKER NOISE IF THE LOWER
END OF THE BAND APPROACHES ZERO?
K 1
Vn2 = (V2)
CoxWL f
K is a process-dependent constant
1. Signals in most applications do not contain significant low frequency
K = 10-25 (V2 F), ∆f = 1Hz
components.
(Behzad Razavi)
2. The logarithmic dependence of the flicker noise power upon fL
Flicker noise spectrum: allows some margin for error in selecting fL
• Devices with areas of several thousands square microns in low

noise applications
• PMOS devices exhibit less 1/f noise than NMOS transistors.
41 42
Propagation of Noise Through

Flicker Noise in MOSFETs (Cont.)
a System
For a linear and noiseless system:
Si ( f ) So ( f )
H (ω )
S o ( f ) = S i ( f ) | H ( 2π f ) | 2
2 K' 1 K' 3 R
4 KT ( g m ) =
2
gm ⇒ fc = gm | H (ω ) |
3 CoxWL f c CoxWL 8 KT 1
RC ω
C
Weak dependence for a given L and thus relatively constant “corner
frequency”, leads to fc falling in the vicinity of 500KHz to 1MHz for
submicron transistors.
43 44
Propagation of Noise Through
a System
∞
 2π k 
R n i
= ∑X
k = −∞
k
exp  j
 T
t

vni(t) vno(t)
C ∞
 2π k   2π k 
n o
= ∑ X
k = −∞
k
H
 T 
 exp  j
 T
t

Sni (f) (V2/Hz) Snio(f)
(f) (V2/Hz)
S o ( f ) = [ 2 lim T → ∞ X K X *
K T ][ H ( 2 π f )] 2
4RKT 4RKT
= S i ( f )[ H ( 2 π f )] 2
0 f (Hz) 0 1 f (Hz)
4 RC
45 46
Example
Si ( f ) So ( f )
H1 H2
vin vout = (vin + vn1 + vn 2 )A
+ + A
vin = Vm cos(ω t )
S o ( f ) = S i ( f ) | H 1 ( 2π f ) | 2 | H 2 ( 2π f ) | 2 v n1 v n2
S o ( f ) = Si ( f ) | H 1H |2  2

V m  + A V 
2 2 2 2 2 2
V out
= A  2 
n1
+V n2 
Signal Noise
47 48
Example Example
v A(2πf )
vout
in
+ A(2πf ) V vin + +
+
out
B(2πf )
v n1 v n2 v n1 vn 2
Vm2
V 2
out = δ ( f − f 0 ) | A(2πf ) |2 + Vn21 | A(2πf ) |2 +Vn22
2
Signal Noise
49 50
v Systems with Several Noise Sources

V eq
= v n1
+ n2
A (2 π f )
2
V A (2 π f ) δ  A (2 π f ) B (2 π f )
2 2 2
2
f f  + 2
V out
=
2
o
−
0  v n1
+
...
+ v
2
+ 2 Re {AB (2 π f )}v 2n 2 n
~ n1 ~
n2 j
+
 1  ~ n2
V = v n 2  B (2 π f ) +  Vout + nout
eq
 A (2 π f )  Vin ~ ZL
-
51 52
Procedure for Output Noise
~ n1
~ n2
... ~
n j
+
• Turn off all noise sources, then determine the
Vin ~ output
ZL Vout + nout • Turn off the input voltage (current) source and all
Noisy Network - noise sources except one.
– If the noise source set on is correlated to other noise
sources in the system, turn these sources on as well
nin
~
~ n1 = 0
~ n2 = 0
... ~
n =0
j
+
– Analyze the system for mean-squared noise output
• Repeat the above step until all the noise sources
exhausted
Vin ~ ZL Vout + nout N
Effective
Input Noise
2
Vout Total
= ∑ (Vout
2
)i
Noiseless Network - i =1
53 54
Example Frequency Transformation of Noise

- vd + vo
v d ( t ) = A ( v 2 ( t ) − v1 ( t ))
Power Power
1
A A v 2
ds = A 2 (V12 + V 22 )
2
v1 (t ) + v v + v 2 (t )
2
v dn (
= A 2 v n22 + v n23 ) ⇔
n2 n3 Function of
RF System
(
v d2 = A 2 v n22 + v n23 + ) 1 2 2
2
A (V1 + V 22 )
f
RF
f
+ v n1
1 2
v o2 = A + A 2 (V n21 + V n23 )
2
55 56
[j ω t ]
∞
Mixer
V n
= ∑ X l
exp l
( )
l = −∞
V 1 + vn (t ) × V out
= V 2 V 1 + V n (t )
2π l
ω l
=
T
V 2  
 
 *  ∞ *
* ∞
a a ∑ ∑

A (t )
[e ω ] (V 2 v n (t )) = ∑ Xl X l + X X
k k
V (t )
∞
 
  l = −∞
V = V V +V j t − j ω t
2
l l
out 2 1 2 n
V 1
=
2
c
+ e c
4
 l = −∞  translated
by  translated
by

Wanted Noise
k =1
 ω k
−ω k

∑ a2 {exp [ j ω ] [− j ω ]}
∞
V 2
= a 0
+ k
k
t + exp k
t
+ a
2  ∑ ∞
X X
* 
k =1 0  l = −∞ l l 
V V = a 0
A (t )
{e j ω c
t
+ e
− j ω c
t
}+ *
( f ) = ∑ a k a k  S
1 2
2 ∞
 f f  +  f f   + 2  f 
S − S + a S
+ ∑ a
k =1
∞
k
A
4
{e
(t ) j (ω k
+ ω c
)t
+ e
− j (ω k
− ω c
)t
}+ out
4 k =1 n
 k  n
 k 
 0 n
 
+ ∑ a
k =1
∞
k
4
A (t )
{e j (ω k
− ω c
)t
+ e
− j (ω k
+ ω c
)t
}
57 58
Noise Measures Noise Figure
Si / N
Noise Factor: F = i
So / N o
(S / N )i P Si / P Ni
• Noise Figure; Noise Factor F =
(S / N )o
F =
P So / P No
• Noise Temperature
Si N o
F =
So Ni
Noise Figure: F (dB) = 10 log F
59 60
Noise Figure Noise Figure
Si / N Si / N Si N o
F = i F = i
=
So / N So N i
So / N o
o
N + GAN N
F = a i
= o
GANi G A N i
PNi Noisy Noiseless
Rs Network PNo PNi Network PNo =
RL
+
GA Pne G A PNi + Pn N
Vs GA F = 1+ e
N = G N N
- Pn a A e
i
P
F = 1 + Ne
P Ni
61 62
Noise Temperature Noise Temperature

Maximum available noise power from a P Si / P Ni P
V 2 F = = 1 + Ne
noisy resistor at temperature Ts: P N = n , rms = KT s B P So / P No P Ni
4R
Pn N
Where: P Ne = = a
GA G A
Noisy Pa P Ne = KT e B
Network RL Pa = KTe B
P Ni = KT s B
Pa
Noise Temperature: Te = Te
KB F = 1+
Ts
63 64
Characterization of
Example
Two-Port Noisy Network
i1 e n1 en2 i2
R1 R2
R1 R2
Using z-parameters
en3
R3 Noisy v1 v2
i1 i2 v n1 vn2 i2 Network
i1 R3
Noisy Noiseless
v1 v2 v1 v2 v n1 vn2 i
Network Network i1 R2
R1 2
v1 R3 v2
Q:
v 1 = z 11 i 1 + z 12 i 2 + v n 1
v n1 = v 1 i1 = i 2 = 0 Verify correlation
between vn1 and vn2
v 2 = z 21 i 1 + z 22 i 2 + v n 2 vn2 = v2 i1 = i 2 = 0
Noiseless
65 66
Characterization of Characterization of
Tow-Port Noisy Network Tow-Port Noisy Network
Using y-parameters
Q: Derive in1 and in2
Derive in1 and in2 in terms of vn1 and vn2
in terms of vn1 and vn2
i1 i2 i1 i2 V = ZI + V n
−1
I = Z (V − V n )
−1
Noisy Noiseless In = −Z Vn
v1 v2 v1 i n 1 Network i n 2 v2
Network I n = − YV n
− z 22 z
in1 = v n 1 + 12 v n 2
∆ ∆ i n 1 = − y 11 v n 1 − y 12 v n 2
i n 1 = i1 z − z 11
i 1 = y 11 v 1 + y 12 v 2 + i n 1 v1 = v 2 = 0
in 2 = 21 v n 1 + vn2 i n 2 = − y 21 v n 1 − y 22 v n 2
in 2 = i2 ∆ ∆
i 2 = y 21 v 1 + y 22 v 2 + i n 2 v1 = v 2 = 0
67 68
Characterization of Characterization of
Tow-Port Noisy Network Two-Port Noisy Network
Derive vn1 and vn2 in terms of in1 and in2 Using (ABCD)-parameters
I = YV + I n
−1 i1 i2 i1 vn i2
V = Y (I − In )
−1
Vn = −Y In
V n = − ZI v1 Noisy Noiseless
i n Network
n
v2 v1 v2
Network
− y 22 y
v n1 = i n 1 + 12 i n 2
∆ ∆
z − z 11
vn2 = 21 i n 1 + in 2 Q:
∆ ∆ v 1 = Av 2 + B (− i2 ) + v n Derive in and vn
i 1 = Cv + D (− i2 ) + in in terms of vn1 and vn2
2
69 70
Derive in and vn Derive in and vn

in terms of vn1 and vn2 in terms of in1 and in2
By characterizing the network By characterizing the network
in terms of z parameters: i1 vn i2 in terms of y parameters: i1 vn i2
z 11 −1
v n = v n1 − vn2 vn = in2
z 21 Noiseless y 21 Noiseless
v1 i n Network v2 v1 i n Network v2
−1 y 11
in = vn2 i n = i n1 − in 2
z 21 y 21
71 72
Minimum Noise Figure Input Noise in Terms of
Noise Figure “F”
• Find Noise Figure: F = f(is,Ys,vn,in)
• if in=inu+inc derive Fmin
Hint: assume i nc = Y c v n Y c = G c + jB c P ne Noiseless
F = 1+ PNi Network PNo =
Y s = G s + jB s P Ni
Pne G A PNi + Pn
• Express optimum value of Ys, called Yopt , P Ni , total GA
F =
for Fmin P Ni
i1 vn i2
P Ni , total = FP Ni = F ( KT o B )
P ne = P Ni , total − P Ni = ( F − 1 )( KT o B )
Noiseless
is Ys i n Network v2
Pn = G A Pne = G A ( F − 1 )( KT o B )
73 74
Networks in Cascade Networks in Cascade
PNi Noisy Noisy

Rs Network PNo1 Network PNo 2 PNi Noisy Noisy Noisy
RL PNo 2 Rs Network PNo1 Network PNo 2 Network PNo
+ G A1 G A2 F= RL
Vs G A 2G A1 PNi + G A1 G A2 G A3
- Pn1 F1 Pn 2 F2 Vs
P No = G Pn1 F1 Pn 2 F2 Pn 3 F3
A 2 P Ni 2 , total
2
-
P Ni 2 , total = P No 1 + Pne 2
N
P No 1 = G P Ni 1 , total = G F 1 ( KT 0 B ) F = o
A1 A1 G A1 G A2 G A3 N i
P Ni 2 , total = G A1 F 1 ( KT o B ) + ( F 2 − 1 )( KT o B )
F2 − 1 F −1
P No 2 F −1 F = F1 + + 3
F = = F1 + 2 G1 G 2G1
G A 2 G A 1 P Ni G1
75 76
Minimum Noise Figure Minimum Noise Figure
i sc2
F =
• Find Noise Figure: F = f(is,Ys,vn,in) i s2
• if in=inu+inc derive Fmin
Hint: assume i nc = Y c v n Y c = G c + jB c
Y s = G s + jB s
isc = − is + in + VnYs
• Express optimum value of Ys, called Yopt ,
for Fmin i sc2 = ( − i s + i n + V n Y s ) 2 = i s2 + ( i n + V n Y s ) 2 − 2 i s ( i n + V n Y s )
i1 vn i2
+
is (in + V nY s ) = 0
Noiseless
is Ys in Network v2
(in + V nY s ) 2
_ i 2
= i + (in + V nY s )
2 2 F =1+
sc s
i s2
77 78
(inu + (Yc + Ys )Vn ) 2 Bs = − Bc

in = inu + inc F = 1+
is2 Gu Rn
F Bs = − Bc = 1+ + (Gs + Gc ) 2
inc = YcVn is2 = 4kT0Gs B Gs Gs
in = inu + YcVn dF
Vn2 = 4kT0 Rn B Bs = − Bc
=0
dGs
inVn* = Yc Vn2
2
inu = 4kT0Gu B
dF Bs = − Bc Gu  2Gs (Gs + Gc ) − (Gs + Gc ) 2 
2 = − 2 + Rn   = 0
4kT0Gu B + Gs + jBs + Gc + jBc 4kT0 Rn B dGs Gs  G 2

*
V i F = 1+ s
Yc = n n
4kT0Gs B Gu
Vn2 Gs = Gopt = Gc2 +
G R
[
= 1 + u + n (Gs + Gc ) 2 + ( Bs + Bc ) 2
Gs G s
] Rn
Gu = Rn (Gopt
2
− Gc2 )
79 80
F = Fmin − 2 Rn (Gc + Gopt ) +
Gu Rn
+
Gs G s
[
(Gs + Gc ) 2 + ( Bs − Bopt ) 2 ]
G
Yopt = Gopt + jBopt = G + u − jBc
2
Rn
[ ]
c
Rn F = Fmin + (Gs − Gopt ) 2 + ( Bs − Bopt ) 2
Gs
Gu R Rn 2
Fmin = F Ys =Yopt = 1+ + n (Gopt + Gc ) 2 F = Fmin + Ys − Yopt
Gopt Gopt Gs
rn 2
F = Fmin + ys − yopt
Gc2 R gs
Fmin = 1 + Rn (Gopt − ) + n (Gopt
2
+ 2Gopt Gc + Gc2 )
Gopt Gopt Ys Gs + jBs
ys = = = g s + jbs
Y0 Y0
= 1 + 2 Rn (Gopt + Gc )
Yopt Gopt + jBopt
yopt = = = g opt + jbopt
Y0 Y0
81 82
Minimum Noise Figure
• Find Noise Figure: F = f(vs,Zs,vn,in)

• if vn=vnu+vnc derive Fmin
Hint: assume v nc = Z c i n Z c = R c + jX c
Z s = R s + jX s
• Express optimum value of Zs, called Zopt ,

for Fmin
Zs i1 vn i2
+
vs + Noiseless
- i n Network v2
83
RF IC Technology - Noise

Chapter 1 - Noise

Uploaded by

Document Informationclick to expand document information

Copyright:

Available Formats

Chapter 1 - Noise

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Chapter 1 - Noise

Uploaded by

Copyright:

Available Formats

Contents

References for Noise Types of Noise

• Selected Papers, Provided by Lecturer

• Burst or Popcorn Noise

Quantitative Definition of Noise

Probability Density Function (PDF): p(n)

• The average power for most types of noise is n2

Noise in Time Domain Noise in the Frequency Domain

Thermal Noise Thermal Noise

Pn = KT B = KT ∆f (Watt) 0 Frequency (Hz)

Vn2 = 4RKTB I n = 4GKTB

Thermal Current Noise PSD of Thermal Noise

I n2 = 4GKTB Pav = V = 4RKTB

Effective Bandwidth Effective Bandwidth

4R 4RC Vni2 - Long channel: γ = 2/3

Induced Gate Noise Induced Gate Noise

Drain Voltage VDS

Thermal Noise in MOSFETs (Cont.) Thermal Noise in MOSFETs (Cont.)

RG1+ RG2+ … + RGn= RG

While the thermal noise caused

We will hereafter neglect

Shot Noise Shot Noise

• Depends directly on the direct component of the current

One-over-f-noise/ low frequency noise/ pink noise

Flicker Noise, (1/f) Noise Flicker Noise PSD

Iα • In general, Kf is an unknown constant

• Devices with areas of several thousands square microns in low

Propagation of Noise Through

v Systems with Several Noise Sources

Example Frequency Transformation of Noise

Noise Measures Noise Figure

Noise Figure: F (dB) = 10 log F

Noise Temperature Noise Temperature

Derive in and vn Derive in and vn

Networks in Cascade Networks in Cascade

PNi Noisy Noisy

(inu + (Yc + Ys )Vn ) 2 Bs = − Bc

Minimum Noise Figure

• Find Noise Figure: F = f(vs,Zs,vn,in)

• Express optimum value of Zs, called Zopt ,

You might also like