CIRF Circuit IntégréRadio Fréquencehassan/cirf10_lna_lecture.pdfVin Vout M1 M2 Lout M3 R Rref...
Transcript of CIRF Circuit IntégréRadio Fréquencehassan/cirf10_lna_lecture.pdfVin Vout M1 M2 Lout M3 R Rref...
CIRFCircuit Intégré Radio Fréquence
Low Noise Amplifier
Hassan AboushadyUniversité Paris VI
Multidisciplinarity of radio design
H. Aboushady University of Paris VI
• M. Perrott, “High Speed Communication Circuits and
Systems”, M.I.T.OpenCourseWare, http://ocw.mit.edu/,
Massachusetts Institute of Technology, 2005.
• D. Leenaerts, J. van der Tang, and C. Vaucher, “Circuit design for RF transceivers”, Kluwer academic publishers, 2001.
References
H. Aboushady University of Paris VI
Heterodyne
Homodyne(Direct conversion)
Low-IF
4
LNA : 1st block of an RF Receiver
Noise Factor :SNRin
SNRoutF =
NF = 10 log10(F)
F1
G1
F2
G2
F3
G3
Si So
5
University of Paris VI
LNA Requirements: Noise
Noise Figure :
( ) ( ) ( )
12121
3
1
21
...
1...
11)1(1
−
−++
−+
−+−+=
n
n
GGG
F
GG
F
G
FFF
Friis Equation :
LNA Requirements to reduce the RF receiver overall Noise Factor:
• Low Noise Factor, F1• High Gain, G1
x(t) y(t)
Vi Vo
α
1,3IPA
...11
2
3,3
2
1
2
1
2
2,3
2
1
2
1,3
2
3
+++≈IPIPIPIP AAAA
βαα
ωω1ω2ωω1ω2
2ω1-ω2 2 ω2 -ω1
∆P
dBmindB
dBm PP
IIP +∆
=2
3
6
β
2,3IPA
γ
3,3IPA
University of Paris VI
LNA Requirements: Linearity
LNA Requirements to improve the RF receiver overall linearity:
• High AIP3,1• Low αααα1
Vin
Rout
Vout
M1
Cgd
- Parasitic capacitance Cgd between input and output nodes. - Low Gain- Complex Load Impedance
outmv ZgA .=
)1
( 0
par
outoutCj
rRZω
=
17
Common Source Amplifier
D. Haghighitalab, M. Vasilevski & H. Aboushady University of Paris VI
Vin
Vout
M1
M2
Lout- Isolation between input and output nodes- Higher Gain
)1
( 0
0
0 out
par
cascodeout LjCj
rZ ωω
=
18
D. Haghighitalab, M. Vasilevski & H. Aboushady University of Paris VI
Common Source Cascode with Inductive Load
Benefits of Cascode
- Real Load Impedance can be obtained
Purpose of Inductive Load
Vin
Vout
M1
M2
Lout
M3
R
Rref
Current Mirror
- The current in M1 is fixed by Rref and the ratio (W1/L1)/(W3/L3).
- R is to increase the impedance seen by the input signal.
19
D. Haghighitalab, M. Vasilevski & H. Aboushady University of Paris VI
Biasing
Current Mirror
Vin
Vout
M1
M2
Lout
M3
R
Rref
Ldeg
Lg
- Lg compensates the complex impedance due to Cgs1 .
- Ldeg adjusts the real part of the input impedance to 50 Ω .
20
D. Haghighitalab, M. Vasilevski & H. Aboushady University of Paris VI
Matching Input Impedance
Inductive Degeneration
21
Input Impedance
Matching Zin to 50Ω
For a Real Impedance @ ω0
degdeg )(1
)( LC
gLLs
sCsZ
gs
mg
gs
in +++=
degLC
gR
gs
ms =
gsg CLL )(
1
deg
0+
=ω
D. Haghighitalab, M. Vasilevski & H. Aboushady University of Paris VI
Ldeg and Lg (1)
22
For different
Ldeg values
ReZ
in
Lg
⇒ReZin
independent of Lg
D. Haghighitalab, M. Vasilevski & H. Aboushady University of Paris VI
Ldeg and Lg (2)
23
ReZ
in
ImZ
in
Lg
Lg
⇒Ldeg for:
ReZin=50
Lg for:
ImZin=0
D. Haghighitalab, M. Vasilevski & H. Aboushady University of Paris VI
For different
Ldeg values
⇒ReZin
independent of Lg
Ldeg and Lg (2)
24
Techno. 0.35µm
c= -0.4j
γ = 2/3
D. Haghighitalab, M. Vasilevski & H. Aboushady University of Paris VI
NF of a MOS Transistor
)()()(
)(
)(
)(
)(
)(
)(
.).(
addedosourceototalo
sourcei
totalo
totaloi
sourceii
totaloo
sourceii
o
i
NNN
NG
N
NGS
NS
NS
NS
SNR
SNRF
+=
====
)(
)(
)(
)()(
)(
)(1
sourceo
addedo
sourceo
addedosourceo
sourceo
totalo
N
N
N
NN
N
NF +=
+==⇒
1) No(added) : Output Reffered Noise due to the LNA.
2) No(source): Output Reffered Noise due to the source.
25
D. Haghighitalab, M. Vasilevski & H. Aboushady University of Paris VI
NF Characterization
R
RkTi R 42 =
R
Résistance
D. Haghighitalab, M. Vasilevski & H. Aboushady University of Paris VI
Noise Sources
27
R
RkTi R 42 =
R
fgkTi
fgkTi
gsg
dd
∆=
∆=
δ
γ
4
4
2
0
2
Résistance Transistor
D. Haghighitalab, M. Vasilevski & H. Aboushady University of Paris VI
Noise Sources
28
R
RkTi R 42 =
R Ls
Rs
Csub
RsubRsub Rs
fgkTi
fgkTi
gsg
dd
∆=
∆=
δ
γ
4
4
2
0
2
Résistance Transistor Inductance
D. Haghighitalab, M. Vasilevski & H. Aboushady University of Paris VI
Noise Sources
R
RkTi R 42 =
R Ls
Rs
Csub
RsubRsub Rs
29
fgkTi
fgkTi
gsg
dd
∆=
∆=
δ
γ
4
4
2
0
2
),,,,,.( 22222
RsubRsubRsRsRRggddi HiHiHiHiHiNGfNF =
)(
)(1
sourceo
addedo
N
NF +=
)(addedoN
Resistance Transistor Inductance
)(sourceoN
D. Haghighitalab, M. Vasilevski & H. Aboushady University of Paris VI
Noise Sources
I
inI
)(
)()(
sI
sIsH
in
outd =
30
D. Haghighitalab, M. Vasilevski & H. Aboushady University of Paris VI
Noise Sources
out
out
out
Vin
Vin
Circuit
Circuit
Circuit
Iin
in
Gain :
Input Impedance:
Noise Factor:
)(
)()(
sVin
sVoutsGain =
)(
)()(
sIin
sVinsZin =
31
)(
)(1
sourceo
addedo
N
NF +=
D. Haghighitalab, M. Vasilevski & H. Aboushady University of Paris VI
Characterization
out
1)()2
)()(.)1
=
−=
sV
R
sVsVVsC
in
outinout
=
+−
1
0
)(
)(
01
/1/1
sV
sVsCRR
out
in
+−
1
0
01
/1/1 sCRR
+ sRC11
1
10
01
Gauss-Jordan
sRCsV
sVsGain
in
out
+==
1
1
)(
)()(
32
D. Haghighitalab, M. Vasilevski & H. Aboushady University of Paris VI
Characterization Example
– Substitution
– Numerical solution of
the matrix
Solve Matrix
Maxima
GiNaC (C++)
33
– Gain Matrix
– Zin Matrix
– Matrices of Noise
Transfer Functions of
each Noise Source
Create Matrices
Netlist Description
Small Signal Parameters
Zin , Gain , No(added)
Cairo+
D. Haghighitalab, M. Vasilevski & H. Aboushady University of Paris VI
Automatic Characterization Procedure
34
Length: Minimum L to reduce parasitic capacitances.
Width: Ids calculated for NFmin.
Layout: Maximum number of fingers to reduce Gate resistance.
Ldeg : Adjusted for Re Zin = 50 ΩΩΩΩ .
Lg : Adjusted for Im Zin = 0 Ω Ω Ω Ω .
Lout : Adjusted for the desired Zout and Gain.
D. Haghighitalab, M. Vasilevski & H. Aboushady University of Paris VI
Design Procedure
Transistors
Inductors
⇒ ⇒ ⇒
Circuit Design
Parameters
VIN,i
VOUT,i
Veg,i
IB,i
Li
Temp
Transistors’
Sizes
And Biasing
W1
Vgs1
Wn
.
Vds1
.
Small Signal
Parameters
gm,1
gds,1
gm,n
Cgs,1
Linear Circuit
Performance
Ad0
FT
.
φm
.
Sizing
Biasing
Small Signal
Characterization
Performance
Modeling
[Ramy ISKANDER “Knowledge-aware synthesis for analog integrated circuit
design and reuse” Ph.D. Thesis UPMC,LIP6 2008]
2
di
?2
gi
35
CAIRO+
Techno
Transistors Sizing
NF
min?
Rs f0 Vdd
Ids-min (Wmin,Lmin)
Caracterization of NF , Gain, Zin
No
Yes
Gain Topology
Choose Inductances
Calculate NF
Impedance Matching
Gain
OK?Yes
NoGain Adjustment
Ids++
36
University of Paris VI
Automatic Sizing Procedure
Techno
Transistors Sizing
NF
min?
Rs f0 Vdd
Ids-min (Wmin,Lmin)
Caractérisation de NF , Gain, Zin
No
Yes
Gain Topology
Choose Inductances
Calculate NF
Impedance Matching
Gain
OK?Yes
NoGain Adjustment
Ids++
37
University of Paris VI
Automatic Sizing Procedure
Techno
Transistors Sizing
NF
min?
Rs f0 Vdd
Ids-min (Wmin,Lmin)
Caracterization of NF , Gain, Zin
No
Yes
Gain Topology
Choose Inductances
Calculate NF
Impedance Matching
Gain
OK?Yes
NoGain Adjustment
Ids++
50Re
:
deg ≈in
gout
ZpourL
initialesvaleursLetL
38
University of Paris VI
Automatic Sizing Procedure
Techno
Transistors Sizing
NF
min?
Rs f0 Vdd
Ids-min (Wmin,Lmin)
Caracterization of NF , Gain, Zin
No
Yes
Gain Topology
Choose Inductances
Calculate NF
Impedance Matching
Gain
OK?Yes
NoGain Adjustment
Ids++
39
University of Paris VI
Automatic Sizing Procedure
Techno
Transistors Sizing
NF
min?
Rs f0 Vdd
Ids-min (Wmin,Lmin)
Caracterization of NF , Gain, Zin
Non
Yes
Gain Topology
Choose Inductances
Calculate NF
Impedance Matching
Gain
OK?Yes
NoGain Adjustment
I++
40
University of Paris VI
Automatic Sizing Procedure
Techno
Transistors Sizing
NF
min?
Rs f0 Vdd
Ids-min (Wmin,Lmin)
Caracterization of NF , Gain, Zin
No
Yes
Gain Topology
Choose Inductances
Calculate NF
Impedance Matching
Gain
OK?Yes
NoGain Adjustment
Ids++
41
University of Paris VI
Automatic Sizing Procedure
Techno
Transistors Sizing
NF
min?
Rs f0 Vdd
Ids-min (Wmin,Lmin)
Caracterization of NF , Gain, Zin
No
Yes
Gain Topology
Choose Inductances
Calculate NF
Impedance Matching
Gain
OK?Yes
NoGain Adjustment
Ids++
42
University of Paris VI
Automatic Sizing Procedure
43
University of Paris VI
Automatic Sizing Procedure
Techno
Transistors Sizing
NF
min?
Rs f0 Vdd
Ids-min (Wmin,Lmin)
Caracterization of NF , Gain, Zin
No
Yes
Gain Topology
Choose Inductances
Calculate NF
Impedance Matching
Gain
OK?Yes
NoGain Adjustment
Ids++
Ldeg for ReZin=50
Lg for ImZin=0
Gain Adjustment
44
University of Paris VI
Automatic Sizing Procedure
Techno
Transistors Sizing
NF
min?
Rs f0 Vdd
Ids-min (Wmin,Lmin)
Caracterization of NF , Gain, Zin
No
Yes
Gain Topology
Choose Inductances
Calculate NF
Impedance Matching
Gain
OK?Yes
No
Ids++
Gain Adjustment
45
University of Paris VI
Automatic Sizing Procedure
Techno
Transistors Sizing
NF
min?
Rs f0 Vdd
Ids-min (Wmin,Lmin)
Caracterization of NF , Gain, Zin
No
Yes
Gain Topology
Choose Inductances
Calculate NF
Impedance Matching
Gain
OK?Yes
No
Ids++
Choice of Lout :
)(
.
outout
outmv
LfZ
ZGA
=
−=
Gain Adjustment
46
University of Paris VI
Automatic Sizing Procedure
Techno
Transistors Sizing
NF
min?
Rs f0 Vdd
Ids-min (Wmin,Lmin)
Caracterization of NF , Gain, Zin
No
Yes
Gain Topology
Choose Inductances
Calculate NF
Impedance Matching
Gain
OK?Yes
No
Ids++
Cairo+
GiNaC
GiNaC
GiNaC
Introduction
Les paramètres de conception
Circuit proposé
Méthode de conception
Caractérisation automatique
Conception automatique
Exemple de conception
47
D. Haghighitalab, M. Vasilevski & H. Aboushady University of Paris VI
Matching Input Impedance
Technologie 0.13 um
Vdd 1.2 V
Résistance de source
(Rs)
50 Ω
Fréquence centrale
(f0)
2.4 GHz
Gain ≥ 15 dB
W (um) L (um) Vgs (V) Vds (V) Ids(mA
)M1 48.96 0.13 0.6 0.6 3.3122
M2 48.96 0.13 0.6 0.6 3.3122
M3 5.08 0.13 0.6 0.6 0.33122
R= 100 KΩ , Rref = 1.811 KΩ , Ldeg = 0.3709 nH , Lg = 48.65 nH , Lout = 3 nH48
Spécifications
Résultats de la procédure proposée
D. Haghighitalab, M. Vasilevski & H. Aboushady University of Paris VI
Design Example I: A 2.4GHz LNA in 130 nm CMOS
Zin
(Cairo+)
(Ω)
Zin (ELDO)
(Ω)
50.00 49.5749
D. Haghighitalab, M. Vasilevski & H. Aboushady University of Paris VI
Estimated and Simulated Input Impedance
Gain (dB) NF (dB)
Cairo+ 19.14 0.774099
ELDO 19.04 0.727277
50
D. Haghighitalab, M. Vasilevski & H. Aboushady University of Paris VI
Estimated and Simulated Gain and NF
S11 (
dB
)S
21 (
dB
)
F (Hz)
51
S11 @ 2.4 GHz =-24.24 dB
S21 @ 2.4 GHz = 17.028 dB
D. Haghighitalab, M. Vasilevski & H. Aboushady University of Paris VI
Simulated S Parameters
IIP3= - 0.44 dBm
52
Fondamentale
IM3
D. Haghighitalab, M. Vasilevski & H. Aboushady University of Paris VI
Simulated IIP3
)))1(
(100log(10)(
2
0)(21
mWdc
lineaire
PF
fSFOM
−=
53
(*) résultats de simulation
[V. Chandrasekhar, C.M. Hung, Y.C. Ho, and K. Mayaram. “A Packaged 2.4GHz LNA in a 0.15umCMOS Process
with 2kV HBM ESD Protection” , ESSCIRC, 2002]
[1]
[2] *
[3]
[4]
[5]
[6]
[Chandrasekhar 2002]
*
D. Haghighitalab, M. Vasilevski & H. Aboushady University of Paris VI
Comparison with state of the art