ece65 course summaryaries.ucsd.edu › ... › 13-w › slides › ece65_w13-course-summary.pdff....
TRANSCRIPT
ECE65 Course Summary
ECE 65, Winter2013, F. Najmabadi
Devices
Diode iv characteristics equation
F. Najmabadi, ECE65, Winter 2013, Course Summary for Final (3/40)
IS : Reverse Saturation Current (10-9 to 10-18 A)
VT : Volt-equivalent temperature (= 26 mV at room temperature)
n: Emission coefficient (1 ≤ n ≤ 2 for Si ICs)
( )1 / −= TD nVvSD eIi
:bias Reverse :bias Forward
3 ||For /
SD
nVvSD
TD
IieIi
nVvTD
−≈≈
≥
For derivation of diode iv equation, see Sedra & Smith Sec. 3
Sensitive to temperature:
IS doubles for every 7oC increase
VT = T(k) /11,600
Diode piecewise-linear model: Diode iv is approximated by two lines
F. Najmabadi, ECE65, Winter 2013, Course Summary for Final (4/40)
Constant Voltage Model
Sifor V 7.06.0 voltage,in"-cut" and 0 :OFF Diode
0 and :ON Diode
0
0
0
−=<=≥=
D
DDD
DDD
VVvi
iVv
Circuit Models: ON: OFF:
Diode ON
Diode OFF
VD0
Zener Diode piecewise-linear model
F. Najmabadi, ECE65, Winter 2013, Course Summary for Final (5/40)
0 and : Zener and 0 :OFF Diode
0 and :ON Diode
0
0
≤−=<<−=
≥=
DZD
DDZD
DDD
iVvVvVi
iVv
Diode ON
Diode OFF
VD0
Zener
Circuit Models: ON: OFF: Zener:
Recipe for solving diode circuits
F. Najmabadi, ECE65, Winter 2013, Course Summary for Final (6/40)
Recipe:
1. Draw a circuit for each state of diode(s).
2. Solve each circuit with its corresponding diode equation.
3. Use the inequality for that diode state (“range of validity”) to check
o if the solution is valid if circuit parameters are all known.
o to find the range of circuit “variable” which leads to that state.
Accuracy of Constant-Voltage Model
F. Najmabadi, ECE65, Winter 2013, Course Summary for Final (7/40)
Constant Voltage Model Diode ON
Diode OFF
VD0
Diode can be in forward bias with vD as small as 0.4 V when iD is small (Lab 4)
In forward bias, “cut-in” voltage (VD0) can vary between 0.6 & 0.8 V (± 0.1 V)
In forward bias, diode voltage changes slightly as current changes (discussed later in small signal model)
BJT iv characteristics includes four parameters
F. Najmabadi, ECE65, Winter 2013, Course Summary for Final (8/40)
Two transistor parameters can be written in terms of the other four:
BJT iv characteristics equations are:
NPN transistor
CEBEBC
BCE
vvviii−=
+= :KVL
:KCL
Cut-off : BE is reverse biased Active: BE is forward biased BC is reverse biased (Deep) Saturation: BE is forward biased BC is foward biased
0 ,0 == CB ii
+=
==
A
CEVvSC
VvSCB
VveIi
eIii
TBE
TBE
1/
/
ββ
BCsatCE
VvSB
iiVv
eIi TBE
,
/
ββ
<≈
=
),( )(
CEBC
BEB
vi givfi
==
Transistor operates like a “valve:” iC & vCE are controlled by iB
F. Najmabadi, ECE65, Winter 2013, Course Summary for Final (9/40)
Controller part: Circuit connected to BE sets iB
Controlled part: iC & vCE are set by transistor state (& outside circuit)
BJT Linear Model
Cut-off (iB = 0, vBE < VD0): Valve Closed iC = 0,
Active (iB > 0, vBE = VD0): Valve partially open iC = β iB, vCE > VD0
Saturation (iB > 0 , vBE = VD0): Valve open iC < β iB, vCE = Vsat
For PNP transistor, replace vBE with vEB and replace vCE with vEC in the above.
V 2.0 ,V 7.0 Si,For 0 == satD VV
* BJT Linear model is based on a diode “constant-voltage” model for the BE junction and ignores Early effect.
Recipe for solving BJT circuits (State of BJT is unknown before solving the circuit)
F. Najmabadi, ECE65, Winter 2013, Course Summary for Final (10/40)
1. Write down BE-KVL and CE-KVL:
2. Assume BJT is OFF, Use BE-KVL to check: a. BJT OFF: Set iC = 0, use CE-KVL to find vCE (Done!)
b. BJT ON: Compute iB
3. Assume BJT in active. Set iC = β iB . Use CE-KVL to find vCE . If vCE ≥ VD0 , Assumption Correct, otherwise in saturation:
4. BJT in Saturation. Set vCE = Vsat . Use CE-KVL to find iC . (Double-check iC < β iB )
NOTE: o For circuits with RE , both BE-KVL & CE-KVL have to be solved
simultaneously.
MOS i-v Characteristics Equations
NMOS (VOV = vGS – Vtn)
o For PMOS set VOV = vSG – |Vtp| & replace vDS with vSD in the above
F. Najmabadi, ECE65, Winter 2013, Course Summary for Final (11/40)
[ ]
[ ]DSOVoxnDOVDSOV
DSDSOVoxnDOVDSOV
DOV
vVL
WCiVvV
vvVL
WCiVvV
iV
λµ
µ
+=≥≥
−=≤≥
=≤
1 5.0 and 0 :Saturation
2 5.0 and 0 :Triode
0 0 :Off-Cut
2
2
PMOS NMOS
MOS operation is “Conceptually” similar to a BJT -- iD & vDS are controlled by vGS
F. Najmabadi, ECE65, Winter 2013, Course Summary for Final (12/40)
Controller part: Circuit connected to GS sets vGS (or VOV )
Controlled part: iD & vDS are set by transistor state (& outside circuit)
A similar solution method to BJT:
o Write down GS-KVL and DS-KVL, assume MOS is in a particular state, solve with the corresponding MOS equation and validate the assumption.
MOS circuits are simpler to solve because iG = 0 ! o However, we get a quadratic equation to solve if MOS in triode (check MOS in
saturation first!)
Foundation of Transistor Amplifiers
F. Najmabadi, ECE65, Winter 2013, Course Summary for Final (13/40)
MOS is always in saturation (BJT in active)
Input to transistor is made of a constant bias part (VGS ) and a signal (vgs): vGS = VGS + vgs
Response (vo = vDS ) is also made of a constant part (VDS ) and a signal response part (vds): vDS = VDS + vds
VDS , is ONLY related to VGS : o i.e., if vgs = 0, then vds = 0
The response to the signal is linear, i.e., vds / vgs = const. But o vGS / vDS is NOT a constant! o VGS / VDS is NOT a constant!
Issues in developing a transistor amplifier:
F. Najmabadi, ECE65, Winter 2013, Course Summary for Final (14/40)
1. Find the iv characteristics of the elements for the signal (which can be different than their characteristics equation for bias). o This will lead to different circuit configurations for bias versus signal
2. Compute circuit response to the signal o Focus on fundamental transistor amplifier configurations
3. How to establish a Bias point (bias is the state of the system when there is no signal). o Stable and robust bias point should be resilient to variations in
µnCox (W/L),Vt (or β for BJT) due to temperature and/or manufacturing variability.
o Bias point details impact small signal response (e.g., gain of the amplifier).
Summary of signal circuit elements
F. Najmabadi, ECE65, Winter 2013, Course Summary for Final (15/40)
Resistors& capacitors: The Same o Capacitor act as open circuit in the bias circuit.
Independent voltage source (e.g., VDD) : Effectively grounded
Independent current source: Effectively open circuit
Dependent sources: The Same
Non-linear Elements: Different! o Diodes & transistors ?
Diode Small Signal Model vd
id rd = nVT/ID
vD
iD
Transistor Small Signal Models
F. Najmabadi, ECE65, Winter 2013, Course Summary for Final (16/40)
1 2 D
oOV
Dm I
rV
Ig⋅
≈⋅
=λ
Comparison of MOS and BJT small-signal circuit models: 1. MOS has an infinite resistor in the input (vgs) while BJT has a finite resistor, rπ
(typically several kΩ).
2. BJT gm is substantially larger than that of a MOS (BJT has a much higher gain).
3. ro values are typically similar (10s of kΩ). gm ro >> 1 for both.
NMOS/PMOS
C
Ao
T
Cm
B
T
IVr
rVIg
IVr ≈===
ππ
β
NPN/PNP BJT
Transistor Biasing
F. Najmabadi, ECE65, Winter 2013, Course Summary for Final (17/40)
To make bias point independent of changes in transistor parameters (e.g. β,) the bias circuit should “set” IC and NOT IB !
Emitter Degeneration (BJT):
Requires a resistor in the emitter circuit. Requirements:
1.
2.
V 1 ≥EE RI
)1( min EB RR +<< β
Current source:
Source Degeneration (MOS):
Requires a resistor in the source circuit. Requirement:
1.
GSDS VIR >
Emitter-degeneration bias circuits
F. Najmabadi, ECE65, Winter 2013, Course Summary for Final (18/40)
Basic Arrangement
Bias with one power supply
(voltage divider)
Bias with two power supplies
MOS source-degeneration bias circuits are identical
To solve circuits with voltage divider bias: 1. BJT: replace voltage divider with its Thevenin equivalent.
2. MOS: Since iG = 0, calculate VG directly.
BJT Current Mirrors are based on two identical transistor with vBE,ref = vBE1
F. Najmabadi, ECE65, Winter 2013, Course Summary for Final (19/40)
Identical BJTs Qref is always in active since
Identical BJTs and vBE,ref = vBE1
o BJTs will have the same iB and the same iC (ignoring Early effect)
0,,
, 0
DrefBErefCE
refC
VVVi
==
>
βC
CBrefCrefiiiiI 22 :KCL , +=+=
/21
1 refref
C II
iI ≈+
==β
For the current mirror to work, Q1 should be in active:
011 DEECCE VVVV ≥+= Since I1 = const. regardless of
VC1 , this is a current source!
MOS Current-Steering Circuit are based on two identical transistor with vov,ref = vov1
F. Najmabadi, ECE65, Winter 2013, Course Summary for Final (20/40)
Qref is always in saturation since
OVOVrefOV
GSGSrefGS
trefGSrefGSrefDS
VVVVVV
VVVV
==
==
−>=
1,
1,
,,,
2111
2,
)/(5.0
)/(5.0
OVoxnD
OVrefoxnrefDref
VLWCiI
VLWCiI
µ
µ
==
==
( )( )refref LW
LWII
// 11 =
For the current steering circuit to work, Q1 should be in saturation:
tGSOVDS VVVV −=>1
Identical MOS: Same µCox and Vt
Since I1 = const. regardless of VD1 , this is a current source!
How to add signal to the bias
F. Najmabadi, ECE65, Winter 2013, Course Summary for Final (21/40)
Bias & Signal vGS = VGS + vgs
Bias & Signal vDS = VDS + vds
1. Direct Coupling
Use bias with 2 voltage supplies o For the first stage, bias such that
VGS = 0 o For follow-up stages, match bias
voltages between stages
Difficult biasing problem
Used in ICs
Amplifies “DC” signals!
2. Capacitive Coupling
Use a capacitor to separate bias voltage from the signal.
Simplified biasing problem.
Used in discrete circuits
Only amplifies “AC” signals
Discrete CE and CS Amplifiers
F. Najmabadi, ECE65, Winter 2013, Course Summary for Final (22/40)
||
)||||(
oDo
Gi
LDomi
o
rRRRR
RRrgvv
==
−=
i
o
sigi
i
sig
o
vv
RRR
vv
×+
=
|| ||
)||||(
oCo
Bi
LComi
o
rRRrRR
RRrgvv
==
−=
π
∞→ πr
Discrete CE and CS Amplifiers with RE / RS
F. Najmabadi, ECE65, Winter 2013, Course Summary for Final (23/40)
[ ] )1( ||
/)||( 1
)||(
SmoDo
Gi
oLDSm
LDm
i
o
RgrRRRR
rRRRgRRg
vv
+==
++−=
i
o
sigi
i
sig
o
vv
RRR
vv
×+
=
+++=
+
++=
+++−=
sigBE
EoCo
oLC
EEBi
EoLCEm
LCm
i
o
RRRrRrRR
rRRRRrRR
rRrRRRgRRg
vv
||1 ||
]/)||[(1 ||
)/1](/)||[(1)||(
π
π
π
β
β
Discrete CB and CG Amplifiers
F. Najmabadi, ECE65, Winter 2013, Course Summary for Final (24/40) i
o
sigi
i
sig
o
vv
RRR
vv
×+
=
)]||||(1[ || 1
)||( ||||
)||||(
sigEmoCo
om
LCoEi
LComi
o
RRrgrRRrgRRrrRR
RRrgvv
π
π
+=
+
+=
+=
)]||(1[ || 1
)||( ||
)||||(
sigSmoDo
om
LDoSi
LDomi
o
RRgrRRrgRRrRR
RRrgvv
+=
+
+=
+=
∞→ πr
Discrete CC and CD Amplifiers
F. Najmabadi, ECE65, Winter 2013, Course Summary for Final (25/40)
1 ||
)||||(1
)||||(
mSo
Gi
LSom
LSom
i
o
gRR
RRRRrg
RRrgvv
=
=+
=
i
o
sigi
i
sig
o
vv
RRR
vv
×+
=
[ ]
||||
||
)||||( || )||||(1
)||||(
osigB
Eo
LEoBi
LEom
LEom
i
o
rRRr
RR
RRrrRRRRrg
RRrgvv
β
β
π
π
+≈
+=+
=
Gain of a multi-stage amplifier
F. Najmabadi, ECE65, Winter 2013, Course Summary for Final (26/40)
Example: A 3-stage amplifier
2,
3,
1,
2,
1,
1,
1,
3,
1,
o
o
o
o
i
o
i
o
i
o
vv
vv
vv
vv
vv
××== ... 321 ××××
+= vvv
sigi
i
sig
o AAARR
Rvv
But we need to know RL,1, RL,2, RL,3, … in order to find AM s.
2,1, io vv = 3,2, io vv = 3,oo vv =1,1,
1,1, , iisigi
i
sig
i
sig
i RRRR
Rvv
vv
=+
==3,oo RR =
3213,
3,
2,
2,
1,
1,
1,
vvvi
o
i
o
i
o
i
o AAAvv
vv
vv
vv
××=××=
21 io vv = 32 io vv =
What are RL and Rsig for each stage?
F. Najmabadi, ECE65, Winter 2013, Course Summary for Final (27/40)
Amp Model for Stage j-1
1,, −= jojsig RR1,, += jijL RR
RL for each stage is the input resistance of the following stage.
Rsig for each stage is the output resistance of the previous stage.
Amp Model for Stage j+1
Procedure for Solving multi-stage Amplifiers
F. Najmabadi, ECE65, Winter 2013, Course Summary for Final (28/40)
Gain & Ri:
1. Start from the load side (nth stage),
o Find the gain Av,n = (vo/vi)n and Ri,n .
2. For (n-1)th stage, set RL,n-1 = Ri,n
o Find the gain Av,n-1 = (vo/vi)n-1 and Ri ,n-1 .
3. Repeat until reaching to the first stage. Then,
Ro:
1. Start from the source side (1st stage). Find Ro,1 .
2. Go to the second stage. Set Rsig,2 = Ro,1 . Find Ro,2
3. Continue to the last stage (nth stage). Then, Ro = Ro,n
... 3211, ××××+
== vvvsigi
i
sig
oii AAA
RRR
vvRR
Cut-off frequency of a multi-stage amplifier
F. Najmabadi, ECE65, Winter 2013, Course Summary for Final (29/40)
Similar to a single-stage amplifier, each capacitor introduces a pole:
1) Coupling capacitor at the input:
2) Coupling capacitor at the output:
3) Coupling capacitor between stages j-1 and j
4) By-pass capacitors for stage j (if exists)
5) Then:
11 )(2
1
csigip CRR
f+
=π
passbypassbybypassp CR
f−−
= 2
1, π
coLopo CRR
f)(2
1+
=π
cjjojipj CRR
f)(2
1
1,, −+=
π
... 21 ++≈ ppp fff
Diode Functional Circuits
Rectifier & Clipper Circuits
F. Najmabadi, ECE65, Winter 2013, Course Summary for Final (31/40)
Half-wave rectifier
OFF) (Diode 0 ,For ON) (Diode ,For
0
00
=<−=≥
oDi
DioDi
vVvVvvVv
Clipper
OFF) Diode( ,For ON) Diode( ,For
0
00
ioDi
DoDi
vvVvVvVv
=<=≥
“Clipping” voltage can be adjusted
F. Najmabadi, ECE65, Winter 2013, Course Summary for Final (32/40)
vo limited to ≤ VD0 + VZ vo limited to ≤ VD0 + VDC
vo limited to ≥ − VD0 − VDC vo limited ≥ − VD0 −VZ
Both top & bottom portions of the signal can be clipped simultaneously
F. Najmabadi, ECE65, Winter 2013, Course Summary for Final (33/40)
vo limited to ≤ VD0 + VDC1 and ≥ − VD0 − VDC2
vo limited to ≤ VD0 + VZ1 and ≥ − VD0 − VZ2
Peak Detector Circuit
Peak Detector & Clamp Circuits
F. Najmabadi, ECE65, Winter 2013, Course Summary for Final (34/40)
Clamp Circuit
vo is equal to vi but shifted “downward” by − (V + − VD0)
“ideal” peak detector: τ/T → ∞
“Good” peak detector: τ/T >> 1
vo shifted “downward”
Voltage shift in a clamp circuit can be adjusted!
F. Najmabadi, ECE65, Winter 2013, Course Summary for Final (35/40)
)( 0DZio VVVvv −−−= + )( 0DDCio VVVvv −−−= +
vo shifted “upward”
)( 0DZio VVVvv −−+= − )( 0DDCio VVVvv −−+= −
BJT as a switch
F. Najmabadi, ECE65, Winter 2013, Course Summary for Final (36/40)
Use: Logic gate can turn loads ON (BJT in saturation) or OFF (BJT in cut-off)
ic is uniquely set by CE circuit (as vce = Vsat)
RB is chosen such that BJT is in deep saturation with a wide margin (e.g., iB = 0.2 ic /β)
*Lab 4 circuit Solved in Lecture notes (problems 12 & 13)
Load is placed in collector circuit
BJT as a Digital Gate
F. Najmabadi, ECE65, Winter 2013, Course Summary for Final (37/40)
Other variants: Diode-transistor logic (DTL) and transistor-transistor logic (TTL)
BJT logic gates are not used anymore except for high-speed emitter-coupled logic circuits because of o Low speed (switching to saturation is quite slow). o Large space and power requirements on ICs
RTL NOT gate (VL = Vsat , VH = VCC)
Resistor-Transistor logic (RTL)
RTL NOR gate* RTL NAND gate*
*Solved in Lecture notes (problems 14 & 15)
NMOS Functional circuits
F. Najmabadi, ECE65, Winter 2013, Course Summary for Final (38/40)
Similar to a BJTs in the active mode, NMOS behaves rather “linearly” in the saturation region (we discuss NMOS amplifiers later)
Transition from cut-off to triode can be used to build NMOS switch circuits. o Voltage at point C (see graph)
depends on NMOS parameters and the circuit (in BJT vo = Vsat)!
We can also built NMOS logic gate similar to a RTL. But there is are much better gates based on CMOS technology!
Complementary MOS (CMOS) is based on NMOS/PMOS pairs
F. Najmabadi, ECE65, Winter 2013, Course Summary for Final (39/40)
Maximum signal swing: Low State: 0, High State: VDD
o Independent of MOS device parameters!
Zero “static” power dissipation (iD = 0 in each state).
o It uses the fact that if a MOS is ON and iD = 0 only if MOS is in triode and vDS = 0
CMOS Inverter CMOS NAND Gate
How to Solve Problems
F. Najmabadi, ECE65, Winter 2013, Course Summary for Final (40/40)
1. What is Known? (comes from problem description.)
2. What is the aim? (Indentify circuit parameters that has to be calculated.) 3. How to get there?
(use recipes!)
First, write down all equations that govern the circuit.
Make sure that you have enough equations to solve for the unknowns.
Make sure that you the solution will give you the answer to the problem.
Only the, start solving the equations.