S - Digital Electronic Circuits-V7

Download as pdf or txt
Download as pdf or txt
You are on page 1of 136

Digital Electronic Circuits

Dr. Samer Arandi

Dr. Samer Arandi - An-Najah National University

Course Outline

Bipolar Families: RTL,DTL, TTL,ECL and I2L


NMOS Transistor Analysis
CMOS Inverters and gates
CMOS Fabrication
OP-AMPs
Sensors and Transducers (optical and ultrasonic)
Driver Circuits (motors, relays, Triacs, Diacs)
Digital to Analogue and Analogue to Digital
Regulated Power Supplies

Dr. Samer Arandi - An-Najah National University

Introduction
Integrated Circuits (IC): arbitrary number of
interconnected gates in a silicon
semiconductor crystal (chip)
Contains electronic components for constructing
digital gates
Mounted in a ceramic or plastic container that
exposes a number of pins (tens to thousands)

Complexity is measured by the number of


logic gates in a single chip

Dr. Samer Arandi - An-Najah National University

Levels of Integration
Small scale Integration (SSI): typically fewer than 10
independent gates early 60s
Medium Scale Integration (MSI): 10 to 1000 of gates
(decoders, adders, multiplexers) late 60s
Large Scale Integration (LSI): thousands of gates (1st and 2nd
generation processors, memory chips, PLD) mid 70s
Very Large Scale Integration (VLSI/ULSI): hundreds of
thousands of gates to several billions (complex
microcomputers chips) 80s till present
System-on-a-Chip (SoC): all components needed by a
computer is included on a single chip.
Lower manufacturing cost and reduced power budget. Why?

3 Dimensional Integrated Circuits (3D-ICs): two or more


layers of components integrated both vertically and
horizontally. Reduced power budget and overall wire length
Dr. Samer Arandi - An-Najah National University

Digital Logic Families


The basic circuit in each technology/family is a
NAND, NOR or INVERTER.
Many families have been introduced:
Bipolar Transistors (flow of two carriers)

RTL: Resistor-Transistor Logic


DTL: Diode-Transistor Logic
TTL: Transistor-Transistor Logic
ECL: Emitter-Coupled Logic
I2L: Integrated Injection Logic

Unipolar Transistors (Field Effect Transistors)


MOS: Metal-Oxide Semiconductor high density
CMOS: Complementary MOS lower power consumption,
dominant family, used in VLSI
Dr. Samer Arandi - An-Najah National University

Logic Family Characterization

Fan-Out
Power Dissipation
Propagation Delay
Noise Margin

Dr. Samer Arandi - An-Najah National University

The Standard TTL series Characteristics


1964 TI introduced first line of standard TTL ICs.
54/74 series
one of the most widely used series
known as the 74 series (54 wider temperature range)
Many semiconductor IC manufactures produce TTL ICs
Fortunately all use the same numbering system allowing
interchanging of components
SN74
DM74
S74

Texas Instruments 74 Series


National Semiconductor 74 Series
Signetics 74 Series

Dr. Samer Arandi - An-Najah National University

The Standard TTL series Characteristics


Several other TTL series have been developed since
the introduction of the 74 standard series.
74LS, 74S etc.
Provide a wide choice of speed and power
characteristics
Speed Power (picoJoules): the lower the better
Differences in TTL family are not in Logic Levels, but in
internal construction of the basic NAND gate.

Dr. Samer Arandi - An-Najah National University

The Standard TTL series Characteristics


The propagation delay of a transistor which goes into
saturation depends on two factors:
Saturation delay (storage time delay)
RC time constant

By reducing resistor values


reduces RC time constant
decreases propagation delay
Trade-off is high power dissipation
lower resistance draws more current from power supply
Dr. Samer Arandi - An-Najah National University

Low Power, 74L series


Similar to a standard TTL
But all resistor values have been increased.
low power dissipation
1 mW
greater propagation delay
33 nseconds

good for applications with low frequency, battery


operated circuits
calculators etc.

Dr. Samer Arandi - An-Najah National University

High Speed, 74H series


Similar to a standard TTL
But all resistor values have been reduced.

faster switching propagation delay 6 nseconds


but
increased power dissipation
22 mW

Dr. Samer Arandi - An-Najah National University

How to increase speed ?


Whatever is done to the value of the resistors
Speed is ultimately limited by the time required to pull
the output transistors out of saturation.

74, 74L and 74H series all operate with saturated


switching
many of the transistors, when conducting will be in
a saturated condition
As has been seen this causes a saturation delay
(storage delay), when switching from ON to OFF
limits the circuits switching speed.
Dr. Samer Arandi - An-Najah National University

Schottky TTL, 74S series


Can this be improved ?
In Schottky TTL (STTL)
Transistors kept out of saturation by using Schottky barrier
diodes (SD)
Formed by a junction of a metal and semiconductor
conventional diode with a junction of p-type and n-type
semiconductor material
SD connected between the base and the collector
Do not allow the transistors to go
as deeply into saturation
SD has a forward voltage
drop of 0.4V

Dr. Samer Arandi - An-Najah National University

Schottky TTL, 74S series


When the Collector-Base junction becomes forward
biased at the on-set of saturation
SD will conduct, diverting some input current away
from base.
this has effect of reducing the excess base current.
decreases saturation (storage time) delay at turn-off

74S00 NAND has average propagation delay of 3 nsecs


twice as fast as the 74H00
makes the 74H series redundant nowadays
Dr. Samer Arandi - An-Najah National University

Schottky TTL, 74S series


Circuit also uses smaller resistor values to improve
switching times
Improves the circuit average power dissipation to
20 mW
NOTE
All transistor are Schottky Transistors.
Q4 is not required to be a Schottky, why?

Dr. Samer Arandi - An-Najah National University

Schottky TTL, 74S series


+5
+5VV
55
55
2.8k
2.8k

760
760

Q3
Q3
AA
BB
CC

QQ4 4

QQ1 1

3.5k
3.5k

QQ2 2

Output
Output
QQ5 5

370
370

350
350

QQ6 6

Dr. Samer Arandi - An-Najah National University

Low Power Schottky TTL, 74LS series


74 LS is a low powered version of the 74S series
uses with larger resistor values than 74S
Low circuit power requirements
but at the expense of increase in switching times.
Power Dissipation
2 mW
Propagation Delay
9.5 nseconds

This is the mainstay of the TTL family


Found in nearly all new designs that do not require max
speed.
Dr. Samer Arandi - An-Najah National University

Advanced Schottky TTL, 74AS series


As a result of the recent development in IC design and
manufacturing process
High speed Schottky diodes

Power Dissipation
Propagation Delay

Dr. Samer Arandi - An-Najah National University

10 mW
1.5 nseconds

Advanced Low-Power Schottky TTL, 74ALS series

Improvement in both power and speed.


Power Dissipation
Propagation Delay

1 mW
4 nseconds

This series has


the lowest speed-power product of the TTL series
very close to the lowest gate power dissipation
This will eventually replace 74LS as the most widely
used TTL series.
Dr. Samer Arandi - An-Najah National University

Emitter Coupled Logic

Dr. Samer Arandi - An-Najah National University

Emitter Coupled Logic (ECL)


Non-saturated digital logic family
Since transistors do not saturate, it is possible to
achieve a prop. delay of 1-2 ns (fastest family)
Utilized in systems that require very high speed

Power dissipation and noise immunity is the


worst!
Power dissipation 25 mw

Dr. Samer Arandi - An-Najah National University

Emitter Coupled Logic (ECL)

Dr. Samer Arandi - An-Najah National University

ECL - Simplified
Emitter Follower Output

Differential Input amplifier

-0.8 V High
-1.8 V Low

GND

GND

245
Q8
OR

220
Q7

NOR
Q1

Q2

Q5

VBB=-1.3V

VBB: bias voltage, mid-point of the


logic swing
779
50 K

50 K

Dr. Samer Arandi - An-Najah National University

VEE=-5.2 V

Use of VCC as ground and VEE at -5.2 V


results in best noise immunity

ECL - Simplified
-0.8 V High
-1.8 V Low

GND

GND

245
Q8
OR

220
Q7

NOR
Q1

Q2

Q5

779
50 K

50 K

Dr. Samer Arandi - An-Najah National University

VEE=-5.2 V

VBB=-1.3V

ECL - Simplified
-0.8 V High
-1.8 V Low

GND

GND

245
Q8
OR

220
Q7

NOR
Q1

Q2

Q5

779
50 K

50 K

A = High (-0.8) B
Dr. Samer Arandi - An-Najah National University

VEE=-5.2 V

VBB=-1.3V

ECL - Simplified
-0.8 V High
-1.8 V Low

GND

GND

245
Q8
OR

220
Q7

NOR

ON
Q1

Q2

Q5

-1.6 V
779
50 K

50 K

A = High (-0.8) B
Dr. Samer Arandi - An-Najah National University

VEE=-5.2 V

VBB=-1.3V

ECL - Simplified
-0.8 V High
-1.8 V Low

GND

GND

245
Q8
OR

220
Q7

NOR
Q1

Q2

Q5

-1.6 V
779
50 K

50 K

A = High (-0.8) B
Dr. Samer Arandi - An-Najah National University

VEE=-5.2 V

VBB=-1.3V

Difference between VBB & VE is 0.3 V only


Q5 is cut-off

ECL - Simplified
-0.8 V High
-1.8 V Low

GND

GND

245
Q8
OR

220
Q7

NOR
Q1

Q2

Q5

-1.6 V
779
50 K

50 K

A = High (-0.8) B
Dr. Samer Arandi - An-Najah National University

VEE=-5.2 V

VBB=-1.3V

Difference between VBB & VE is 0.3 V only


Q5 is cut-off

-0.8 V

ECL - Simplified
-0.8 V High
-1.8 V Low

GND

GND

245
Q8

(-1.6 - -5.2)*220/779

1.0 V

220

OR

-0.8 V

NOR

-1.8 V

Q7

Q1

Q2

Q5

-1.6 V
779
50 K

50 K

A = High (-0.8) B
Dr. Samer Arandi - An-Najah National University

VEE=-5.2 V

VBB=-1.3V

Difference between VBB & VE is 0.3 V only


Q5 is cut-off

ECL - Simplified
-0.8 V High
-1.8 V Low

GND

GND

245
Q8
OR

220
Q7

NOR
Q1

Q2

Q5

779
50 K

50 K

Dr. Samer Arandi - An-Najah National University

VEE=-5.2 V

VBB=-1.3V

ECL - Simplified
-0.8 V High
-1.8 V Low

GND

GND

245
Q8
OR

220
Q7

NOR
Q1

Q2

Q5

779
50 K

= Low (-1.8)

50 K

VEE=-5.2 V

= Low (-1.8)

Dr. Samer Arandi - An-Najah National University

VBB=-1.3V

ECL - Simplified
-0.8 V High
-1.8 V Low

GND

GND

245
Q8
OR

220
Q7

cut-off

Q1

NOR

cut-off

Q2

Q5

-2.1 V
779
50 K

= Low (-1.8)

50 K

VEE=-5.2 V

= Low (-1.8)

Dr. Samer Arandi - An-Najah National University

VBB=-1.3V

ECL - Simplified
-0.8 V High
-1.8 V Low

GND

GND

(-2.1 - -5.2) * 245 / 779


245

1.0 V

Q8
OR

220
Q7

cut-off

Q1

NOR

cut-off

Q2

Q5

-2.1 V
779
50 K

= Low (-1.8)

50 K

VEE=-5.2 V

= Low (-1.8)

Dr. Samer Arandi - An-Najah National University

VBB=-1.3V

-1.8 V

ECL - Simplified
-0.8 V High
-1.8 V Low

GND

GND

245

1.0 V

Q8
OR

220

-1.8 V

Q7

cut-off

Q1

NOR

cut-off

Q2

Q5

-2.1 V
779
50 K

= Low (-1.8)

50 K

VEE=-5.2 V

= Low (-1.8)

Dr. Samer Arandi - An-Najah National University

VBB=-1.3V

-0.8 V

Integrated Injection Logic

Dr. Samer Arandi - An-Najah National University

Integrated Injection Logic (I2L)


Saturated bipolar logic family
Utilized in MSI and LSI
Small size
Low power consumption
Cheaper fabrication cost

Merges transistor components - Merged


Transistor Logic (MTL)
One semi-conductor region is part of two or more
devices.
This allows considerable saving in silicon chip area.
Dr. Samer Arandi - An-Najah National University

Integrated Injection Logic (I2L)


A single inverter is implemented as:
- Multi-collector npn transistor
- Multi-collector pnp transistor that is
driving the npn transistor.
- The emitter of the driving transistor is
called the injector

Dr. Samer Arandi - An-Najah National University

Integrated Injection Logic (I2L)


The injector distributes current to multiple units
which form its multi-collectors.
Open-collector output which allows a wired-and
configuration (typically followed by an inverter)

Dr. Samer Arandi - An-Najah National University

Integrated Injection Logic (I2L)


The collector of the current
injector transistor and the
base of each multiple collector
transistor are merged, i.e. one
p region is used for both

The base of the injector


transistor is merged with the
emitter of the multiple collector
transistor.
Dr. Samer Arandi - An-Najah National University

Metal Oxide Semiconductor

Dr. Samer Arandi - An-Najah National University

Metal Oxide Semiconductor


Field Effect Transistor (FET): unipolar transistor
as its operation depends on the flow of one
type of carriers
Two types of FETs:
Junction Field Effect Transistor (JFET) used
mainly in linear circuits
Metal-Oxide Semi-conductor FET (MOSFET)
used in digital circuits
Advantage: fabricated in less area than Bipolar
Transistors

Dr. Samer Arandi - An-Najah National University

MOS Transistors

p-channel
n-type substrate (lightly doped)
Two heavily doped regions with p-type impurities (source and drain)
The region between the source and drain is called the channel
The gate is the metal plate separated from the substrate by an insulated
dielectric of silicon dioxide
A negative voltage at the gate => induced electric field at the channel =>
attracts p-type carriers (holes) => conductivity increases => current flows
from source to drain
Dr. Samer Arandi - An-Najah National University

MOS Transistors
Two modes of operation (depends on the
state of the channel at zero voltage):
Depletion
channel is initially doped with p-type impurities,
i.e. a conducting channel exists at 0 V
Enhancement
region beneath the gate is initially uncharged
a voltage must be applied to the gate to induced a
channel i.e. current is enhanced by gate voltage

Dr. Samer Arandi - An-Najah National University

MOS Transistors
The majority carriers enter the device from the source
terminal to the drain when the voltage applied to the gate
exceeds a threshold VT
p-channel
The source is connected to the substrate and a negative voltage
is applied to the drain
When the gate voltage is sufficiently negative below VT p-type
carriers flow from S to D (positive current flow)

n-channel
The source is connected to the substrate and a positive voltage
is applied to the drain
When the gate voltage is sufficiently positive above VT n-type
carriers flow from S to D (equivalent to positive current flow
from D to S)
Dr. Samer Arandi - An-Najah National University

MOS Transistors

Symbols for MOS Transistors

Enhancement Mode

Dr. Samer Arandi - An-Najah National University

MOS Logic Gates


MOSFET can be used as resistor as well as a transistor
Resistance value can be determined as the ratio between VDS and IDS
Different resistor values may be constructed during manufacturing by
fixing the channel length and width of the MOS device

RQ1: Load Resistor

RQ2<<RQ1

Dr. Samer Arandi - An-Najah National University

MOS Logic Gates


RQ2+RQ3<<RQ1
Q1

Q2

Q3

Dr. Samer Arandi - An-Najah National University

Complementary MOS
Takes advantage of the ability to fabricate
both n-type and p-type on the same substrate
Basic circuit is the inverter:
The n-MOS conducts when
gate-to-source is pos+
The p-MOS conducts when
gate-to-source is neg Both are turned off when
gate-to-source is zero

Dr. Samer Arandi - An-Najah National University

s
d
d
s

CMOS Logic Gates

Dr. Samer Arandi - An-Najah National University

CMOS Models/Symbols

Dr. Samer Arandi - An-Najah National University

MOS Operation and Analysis

Dr. Samer Arandi - An-Najah National University

L: channel length (350nm 90nm ~ 2004)


W: channel width
tox: gate oxide thickness (< 25 ~2004)
bulk(substrate): doping density
xj: junction depth (150 70 nm ~2004)

Dr. Samer Arandi - An-Najah National University

Determine the
electrical
Characteristics of
the transistor

MOS Threshold Voltage


Threshold Voltage (VT): the minimum gate voltage needed to
initiate the forming of a conducting channel
VT mainly depends on the material properties - largely
determined at the time of fabrication
The most important factors affecting VT include:

The gate conductor material (poly vs. metal)


The gate insulation material (SiO2)
Impurities in the oxide
The thickness of the gate material
The channel doping concentration

VT is also depends on:


Temperature: changes by -2mV/degree C for low substrate doping levels
Source-Bulk voltage VSB
Dr. Samer Arandi - An-Najah National University

As VGS moves from 0 to a positive value:


positive charge accumulates on the gate
negative charge (electrons) accumulates in the substrate
under the gate

The mobile holes are pushed down creating a depletion


layer (contains the immobile negative charge)
As the voltage increases the depletion layer thickness
increases and a layer of mobile electrons appear
(weak inversion condition)
Further increase in the voltage
increases the concentration
of the electrons until its
equal to the concentration
of the holes
(strong inversion condition)
Dr. Samer Arandi - An-Najah National University

For gate voltage above this point:


The depletion layer thickness remains approximately
constant
Additional mobile carriers accumulate in the channel
(drawn from S & D)

Application of VDS > 0 causes current to flow in


the channel
The value of the VGS required to produce strong
inversion is called
the threshold voltage

Dr. Samer Arandi - An-Najah National University

Threshold Voltage Components


VT has three main components:

GC: difference in work function between the gate


material (G) and the silicon substrate on the channel
side (C)
undesirable positive charge Qox due to imperfections
in the semi-oxide interface doping
voltage required to change the surface potential to
the strong inversion condition and to offset the
induced depletion layer charge QB
VT = GC Qox/Cox 2F QB/Cox
At VSB = 0, VT and QB are denoted as VT0 and QB0

Dr. Samer Arandi - An-Najah National University

VT0 = GC Qox/Cox 2F QB0/Cox


VFB

VFB
F

: flat band voltage


: equilibrium electrostatic potential

k * t ni
ln
F =
q
NA
Thermal voltage
= 26mV

(the shift in Fermi level due to doping)


|2F| : band bending voltage

ni
NA
K

: intrinsic (non-doped carrier concentration)


depends on temperature (1.45x1010 cm-3)
: carriers density in doped semi-conductor
for p-type material (ND for n-type)
: Boltzman constant, T: temperature, q: electron charge

Dr. Samer Arandi - An-Najah National University

QB0 =

q * NA * Xd

Xd =

QB0 = 2q * Si * NA* | 2F|

VSB

: electronic charge (1.6x10-19)


: depletion layer width
: permittivity of silicon (1.06x10-12)
for p-type material (ND for n-type)
: Source-Bulk voltage (typically zero)

Cox
ox
tox

: gate-oxide capacitance
: permittivity of oxide
: oxide thickness

q
Xd
Si

Dr. Samer Arandi - An-Najah National University

2 * Si* | 2F |
q * NA
Cox =

ox
tox

Signs
VFB is negative in nMOS, positive in pMOS
F is negative in nMOS, positive in pMOS
QB0 and QB are negative in nNMOS, positive in
pMOS
VSB is positive in nMOS, negative in pMOS

Dr. Samer Arandi - An-Najah National University

Current-Voltage Characteristics
MOS has three modes of operation:
Cut-off
Linear
Saturation

VGS < VT (IDS is zero)

VGS > VT, VDS>0 (IDS is non-zero)

The mode depends on the terminal voltages:


VGS
VDS

Dr. Samer Arandi - An-Najah National University

Linear Mode
At VDS = 0, the inverted channel is in equilibrium
and IDS = 0
If a small VDS > 0 is applied, IDS will flow from S
to D, such that IDS is proportional to VDS
The transistor
(the channel) acts
as a voltage
controlled resistor

Dr. Samer Arandi - An-Najah National University

Linear Mode

K
2
IDS [2(VGS VT )VDS VDS ]
2
W
K K'
L
K ' n * Cox

nox
tox

n : carrier mobility

IDS depends on both


VGS and VDS
Dr. Samer Arandi - An-Najah National University

Saturation Mode
As the drain voltage is increased the inversion
layer charge and the channel depth at the drain
decreases
At VDS = VGS-VT0 (called VDSAT) the channel
inversion charge at the drain becomes zero
(pinch-off point)
As VDS > VDSAT the inversion layer near the drain
vanishes -> channel length decreases
Dr. Samer Arandi - An-Najah National University

Saturation Mode
2
K
IDS (VGS VT )
2
2
K
IDS (VGS VT )(1 *VDS )
2
(account for channel modulation)

:channel length
modulation
coefficient
Dr. Samer Arandi - An-Najah National University

Pinch-off point

Saturation Mode

Dr. Samer Arandi - An-Najah National University

Practical Considerations
The I-V characteristics discussed so far neglect
many effects that are important in modern
fabrication processes:
Velocity Saturation
We assumed that carrier drift velocity and hence
current increase linearly with the electric field
between source and drain.
This is only true for weak electric fields, at high
electric fields and short channel lengths carrier drift
velocity saturates (saturation occurs when carriers
reach saturation velocity)
Thus, saturation current is linearly dependent on
voltage rather than quadratically dependent
Dr. Samer Arandi - An-Najah National University

Practical Considerations (cont.)


Mobility Degradation
Strong vertical electric fields resulting from large Vgs
reduce carrier mobility .

Sub-threshold conduction
We assumed that no current flows from source to drain
when Vgs < Vt. In reality this is not true and current drops
off exponentially. This current is termed as leakage current

Tunneling
According to quantum mechanics, there is a finite
probability that carriers will tunnel through the gate oxide.
This results in gate leakage current flowing into the gate.
Probability of tunneling drops off exponentially with oxide
thickness, and so was negligible until recently.

Dr. Samer Arandi - An-Najah National University

Practical Considerations (cont.)


Junction leakage
The p-n junctions between S & D and the substrate form
diodes.
The well to substrate junction is another diode.
Although these diodes are reverse biased they still conduct a
small amount of current

Temperature
Carrier mobility decreases with temperature
The magnitude of the threshold voltage decreases nearly
linearly with temperature
Junction leakage increases with temperature because Is is
strongly temperature dependent.
Net effect: ON current decreases, OFF current increases,
worse performance at high temperature
Dr. Samer Arandi - An-Najah National University

Practical Considerations (cont.)


Geometry Dependence
Layout designers draw transistors with some
width and length, Wdrawn and Ldrawn.
The actual dimensions may differ, due to
Polysilicon over-stretching
Effective dimensions should be used rather than
drawn dimensions for analysis or values can be
significantly off.
Below 0.25m, transistor orientation and amount
of nearby poly affect the effective length

Dr. Samer Arandi - An-Najah National University

MOS Capacitance
The switching speed of MOS circuits is limited
by the time required to charge/discharge the
capacitances at the internal nodes
The capacitances must be calculated from
device dimensions and dielectric constants
The values are usually specified in femtoFarads per meter of width (fF/m)
CG vs. Cg

Dr. Samer Arandi - An-Najah National University

Types of MOS Capacitance


Thin-oxide Capacitance
Cg => (Cgs, Cgd,Cgb)
Junction capacitance - Csb and Cdb
Overlap Capacitance Col

Non-linear
voltage-dependant

linear
voltage-independant

Depletion layer capacitance - under the


channel - Cjc (associated with Cgb)
Dr. Samer Arandi - An-Najah National University

Dr. Samer Arandi - An-Najah National University

Thin-Oxide Capacitance
The most important capacitance in MOS
The two plates of the capacitor are the gate
and the channel
The dielectric is the oxide between the two
plates

CG W * Cg
Cg Cox *L
Dr. Samer Arandi - An-Najah National University

ox
tox

*L

Thin-Oxide Capacitance Calculations


In 5 m technology (L=5 m), tox = 1100
Cg = Cox*L =((4*8.85x10-14)/110)*5 m = 1.6 fF/m
In 0.35 m technology (L=0.35 m), tox = 75
Cg = Cox*L =((4*8.85x10-14)/75)*0.35 m = 1.6 fF/m
In 0.13 m technology (L=0.1 m), tox = 22
Cg = Cox*L =((4*8.85x10-14)/22)*0.1 m = 1.6 fF/m
The factor remained constant for over 25 years
Both L and tox are scaled at the same rate and so their effects
cancel each other
Dr. Samer Arandi - An-Najah National University

Thin-Oxide Capacitance
The gate capacitance is decomposed into
three capacitances:
Cgs: gate-to-source capacitance
Cgd: gate-to-drain capacitance
Cgb: gate-to-bulk capacitance

The composition vary depending on whether


the device is in cut-off, linear or saturation
states

Dr. Samer Arandi - An-Najah National University

Thin-Oxide Capacitance
Capacitance

Cut-off

Linear

Saturation

Cgb
Cgd
Cgs

Cg *

Cg

Cg

2/3 Cg

*assuming a VGS = 0 (in the cut-off region Cgb is in


series with the depletion capacitance Cjc)

Dr. Samer Arandi - An-Najah National University

Junction Capacitance
The S and D regions and the substrate forms
pn junctions that give rise to two additional
capacitances - Csb and Cdb
In addition, a junction capacitance is formed
between the inverted channel and the
substrate Cjc
This capacitance depends on the terminal
voltage
Dr. Samer Arandi - An-Najah National University

Junction Capacitance
Cj 0 * A
CJ
VJ m
(1 )
B

CJ: junction capacitance


Cj0: zero-bias junction capacitance (fF/cm2)

A: area of the junction = W*(xj+Y)


m: the junction grading coefficient
(approx. 1/2 for abrupt junctions)
B: built-in junction potential

(si * q) NA * ND
si * q * NA
Cj 0
*

(2 *B) NA ND
2 *B

k * T NA * ND
B
ln
q
ni 2
Dr. Samer Arandi - An-Najah National University

Junction Capacitance
Since the terminal voltages change during
dynamic operation we need to calculate the
capacitance under such transient conditions
We approximate this capacitance by finding the
large-signal average junction capacitance*:

CJ Keq * Cj 0 * A
2 * B
Keq
* ( B V 2 B V 1)
V 2 V1
Keq: dimensionless coefficient
*calculated for a transition between two known voltages, V1 and V2
Dr. Samer Arandi - An-Najah National University

Overlap Capacitance
Voltage-independent capacitance that exists on both sides of the gate
Composed of two components:
diffusion capacitance
between the gate and
the diffusion extensions
of the source
and drain - Cov
fringing capacitance
between sidewall of
the poly-silicon
and surface of drain
and source - Cf

Col = Cov + Cf

Dr. Samer Arandi - An-Najah National University

Overlap Capacitance

Col Cov Cf
2ox

Tpoly
Cf
ln(1
)

tox

Cov Cox * LD
Dr. Samer Arandi - An-Najah National University

MOS Inverter
Static Characteristics

Dr. Samer Arandi - An-Najah National University

Voltage Transfer Characteristics


In an ideal implementation of an inverter, the two
binary levels are ground and VDD.
The output transition between the 1 and 0 states
occurs when the input is exactly VDD/2.

DC Voltage
Transfer
Characteristics

Vout
Gain =
Vin
Dr. Samer Arandi - An-Najah National University

Voltage Transfer Characteristics


Range: voltage interval over which a signal is considered to be
logic 0 or logic 1. For an ideal inverter:
-The input range is very large
-The output range is small
This is desirable as it implies: the input can vary significantly
with little or no effect on the output (the gate rejects noise at
the input)

Dr. Samer Arandi - An-Najah National University

Practical VTC

VOL may not reach GND


VOH may not reach VDD
Vout doesnt switch from VDD
to GND at VDD/2
Vs: point where Vout = Vin
instead of 3 regions
(0 gain, infinite gain, 0 gain)
we have
(low gain, high gain, low gain)

Dr. Samer Arandi - An-Najah National University

Practical VTC
The input ranges that define 0 and 1 are smaller than the ideal case
- 0 to VIL
- VIH to VDD
The two output ranges are larger than the ideal case
- VOL to VOUL
- VOUH to VOH

Dr. Samer Arandi - An-Najah National University

Noise
When the input range is larger than the output
range, the gate retains the noise rejection property
i.e. output fluctuation are small even with large
input fluctuation
This attenuates the noise when multiple gates are
connected successively

Dr. Samer Arandi - An-Najah National University

Noise Margin Definitions


If the noise amplitude at the input of any logic circuit
is smaller than the noise margin of the circuit the
noise will be attenuated
Noise may be transferred to logic nodes or
interconnecting lines by:
Unwanted capacitive or inductive coupling
Series inductance and resistance in the shared ground
and power supply lines

The robustness of a gate depends on:


How much noise can be applied before the gate fails (gate
properties)
How much noise actually couples into the gate (environ.)
Dr. Samer Arandi - An-Najah National University

Noise Margins
The noise margin specifies the range over
which the circuit will function properly.
Noise margin metrics:
single-source noise margin (SSNM): assumes a
single noise source affecting one logic node (not
realistic)
multiple-source noise margin (MSNM): assumes
multiple noise sources potentially affecting all
nodes (more realistic)

Dr. Samer Arandi - An-Najah National University

Noise Margins (cont.)

Noiseless System

System with
single-stage
noise of
magnitude Vn
System with multi-stage
noise
Dr. Samer Arandi - An-Najah National University

Noise Margins (cont.)


For a noiseless system:
Vout = f(Vin)

When adding noise Vn the output:


Vout = f(Vin+Vn)
Can be eventually simplified into:

Vout = Noiseless_Output + noise*gain


This implies that if the inverter is operating in
the region where:
The gain > 1, noise is amplified (undesired)
The gain < 1, noise is attenuated (desired)
Dr. Samer Arandi - An-Najah National University

Noise Margins (cont.)


We can define the noise margins by using the
VTC points where the gain is 1 to establish the
transition points of the range

Dr. Samer Arandi - An-Najah National University

Noise Margins (cont.)

Inverter

NMH = VOH - VIH


NML = VIL - VOL

Dr. Samer Arandi - An-Najah National University

Vout

Vin

Noise Margins (cont.)

Buffer

NMH = VOH - VIH


NML = VIL - VOL

Dr. Samer Arandi - An-Najah National University

Resistive-Load NMOS Inverter


Vin = VGS

Vout = VDS
VSB = 0
VT = VT0
Since IG is
negligible
I R = ID
Dr. Samer Arandi - An-Najah National University

VTC

To calculate the noise margins we find the five critical


points on the VTC: VOH, VOL, VIH, VIL and Vs
Dr. Samer Arandi - An-Najah National University

Calculating VOH
To calculate VOH we set
the input voltage
below VT
Therefore, no current
flows
And so:
VOH = VDD
Dr. Samer Arandi - An-Najah National University

Calculating VOL
When a logic 1 is applied at the gate input
(represented by VOH of a previous gate) the
transistor is driven into the linear region.
This implies:
IR = IDS (linear)
VDD VOL K
2
[2(VOH VT )VOL VOL ]
RL
2

Dr. Samer Arandi - An-Najah National University

Calculating VOL
VDD
VOL
1 k * RL(VDD VT )
To make VOL small:
increase k (i.e. W/L) [increase area, faster falltime]
increase RL[increase area, slower rise-time]

Decreasing VOL increases the power


marginally, however increasing RL reduces
power
VDD VOL
P I *V (

Dr. Samer Arandi - An-Najah National University

RL

) *VDD

Calculating VIL
At Vin = VIL the output voltage is near VDD and
the transistor is operating in the saturation
region.
This implies:
IR = IDS (saturation)
2
VDD VOut k
(VIL VT )
RL
2

Dr. Samer Arandi - An-Najah National University

Calculating VIL
1
VIL VT
k * RL
To increase VIL:
Decrease k and RL. However, this increases VOL
which is not desirable
Both VIL and VOL shift in the same direction and so
it's difficult to affect the NML significantly
Dr. Samer Arandi - An-Najah National University

Calculating VIH
When Vin=VIH the output voltage is near 0 and
the transistor is operating in the linear region
IR = IDS (linear)
Following a similar procedure we end-up with:

8 *VDD
1
VIH VT

3 * k * RL k * RL
Dr. Samer Arandi - An-Najah National University

Calculating VIH
8 *VDD
1
VIH VT

3 * k * RL k * RL
To decrease VIH we can increase kRL, however,
this increases the rising delay and increases
the power slightly
Changes in k and RL have no effect on VOH so
its possible to increase NMH
Dr. Samer Arandi - An-Najah National University

NMOS as Load Devices


The resistor load (RL) requires large amount of
chip area if realized in a standard MOS process
The area would be more than 100 times that of a
transistor!
It would also result in a large Rise Time

Consequently resistors are rarely used as loads


in MOS digital circuits
Instead an NMOS transistor is used to perform
the function of a pull-up resistor.

Dr. Samer Arandi - An-Najah National University

Saturated Enhancement Load Inverter


G of the pull-up is
connected to VDD
The transistor is
working in the
saturation mode or
cut-off mode
The relative sizes of
the two transistors
determine the output
voltages (must ensure
VOL is < VT)

Dr. Samer Arandi - An-Najah National University

Pull-up
transistor

Pull-down
transistor

Saturated Enhancement Load Inverter


Advantages:
- Single power supply
- Simple Fabrication
Disadvantages:
VOH is limited to
VDD - VTL

Dr. Samer Arandi - An-Najah National University

Pull-up
transistor

Pull-down
transistor

Linear Enhancement Load Inverter


G of the pull-up is
connected to VGG such
that VGG > VDD + VTL
The transistor is
working in the
linear mode for all the
range of VOH
VOH can reach to VDD

Dr. Samer Arandi - An-Najah National University

Linear Enhancement Load Inverter


Advantages:
- VOH can reach to VDD
Disadvantage:
-The extra voltage source is
associated with additional
interconnections
(more chip area)
- DC power dissipation in the
output low state
Dr. Samer Arandi - An-Najah National University

Solution?
Depletion Load Inverter
Better noise margins
Single power supply
Smaller overall layer
More complicated fabrication
process

CMOS Inverter
Steady-state power dissipation negligible (only
due to leakage currents)
VTC Transition is very sharp (close to an ideal
inverter)
Dr. Samer Arandi - An-Najah National University

CMOS Inverter
Vin is at VDD: the NMOS
device is conducting while the
PMOS (VGS=0) is cut-off and so
VOUT is 0 and IDN is almost
zero.
Vin is at GND, the NMOS is
cut-off while the PMOS is
conducting and so Vout is
equal to VDD and IDN is almost
zero.
large noise
VOH = VDD
VOL = 0
margins
Dr. Samer Arandi - An-Najah National University

VTC
5 regions

Dr. Samer Arandi - An-Najah National University

VTC

Region

Vin

Vout

NMOS

PMOS

< VT0,n

VOH

Cut-off

Linear

II

VIL

III

Vs

Vs

IV

VIH

Low VOL

Linear

Saturation

> VDD+VT0,p

VOL

Linear

Cut-off

Dr. Samer Arandi - An-Najah National University

High VOH Saturation

Linear

Saturation Saturation

VTC Symmetry
A completely symmetrical VTC is obtained if
VTP = - VTN and kP = kN

VDD | VTP | VTN


Vs
1

Dr. Samer Arandi - An-Najah National University

WN
ECN * LN
WP
ECP * LP

nWN
pWP

CMOS Inverter Timing

tpHL (propagation delay high to low): time delay between V50% transition of
the rising input voltage and the V50% transition of the falling output voltage.

tpLH (propagation delay low to low): time delay between V50% transition of
the falling input voltage and the V50% transition of the rising output voltage.

Propagation time delay (tP)

Dr. Samer Arandi - An-Najah National University

t PHL t PLH
tP
2

CMOS Inverter Timing


We can model the inverter as an effective onresistance (Reff) driving a load capacitance CL
When the output is falling:
Reff = RN

Vout(t ) VDD * e

-t/RN*CL

When the output is rising:


Reff = RP

Vout(t ) VDD * (1 e

-t/RP*CL

In both cases the 50%


occurs at => t = 0.69*Reff*CL
Dr. Samer Arandi - An-Najah National University

Effective On-Resistance
Reff is inversely proportional to W/L
The NMOS and PMOS have difference Reff.
RN = Reqn*(LN/WN)
RP = Reqp*(LP/WP)
Reqn is defined in units
of L/W

Dr. Samer Arandi - An-Najah National University

Load Capacitance (CL)


CL accounts for (is the sum of) :
the internal capacitance of the inverter
the capacitance of the wiring
the capacitance of the fan-out
extrinsic

Dr. Samer Arandi - An-Najah National University

intrinsic

Internal Capacitance
Comprised of the Capacitances connected to
the output (D):
Cdb1 and Cdb2 are the junction capacitance (D-to-B)
Since we mostly work in cut-off
and saturation Cgd12 (gate
capacitance) is almost
negligible except from
the overlap
capacitance

Dr. Samer Arandi - An-Najah National University

Fan-out Capacitance
The Fan-out capacitance is due to the input of
subsequent gates.
This capacitance can be large:
the sum of each of the driven
gate capacitances

Each gate capacitance


comprises:
gate capacitance
(thin-oxide capacitance)
overlap capacitances
Dr. Samer Arandi - An-Najah National University

Minimizing Propagation Delay


Reducing CL
Careful layout can reduce this capacitance.

Increase W/L ratio of the transistors.


Warning: doing so increases the intrinsic capacitance

Increase VDD (Req depends on VDD)

The delay of a gate can be modulated by modifying the


supply voltage.
This allows the designer to trade off energy dissipation
for performance.
However, rising above a certain level yields only a minor
improvement.
Also, reliability concerns (oxide breakdown, hot-electron
effects) set firm upper bounds.

Dr. Samer Arandi - An-Najah National University

Req vs. VDD

Dr. Samer Arandi - An-Najah National University

tp vs. VDD
5.5
5

4
3.5
3

t (normalized)

4.5

2.5
2
1.5
1
0.8

1.2

Dr. Samer Arandi - An-Najah National University

1.4

1.6

1.8

(V)

DD

2.2

2.4

CMOS Fabrication

Dr. Samer Arandi - An-Najah National University

CMOS Fabrication
CMOS Technology depends on using both N-Type and PType devices on the same chip.
The two main technologies to do this task are:
P-Well
The substrate is N-Type. The N-Channel device is built into a PType well within the parent N-Type substrate. The P-channel
device is built directly on the substrate.
N-Well
The substrate is P-Type. The N-channel device is built directly
on the substrate, while the P-channel device is built into a Ntype well within the parent P-Type substrate.

Dr. Samer Arandi - An-Najah National University

Two more advanced technologies to do this task are


becoming more popular for sub-micron geometries
where device performance and density must be pushed
beyond the limits of the conventional p & n-well CMOS
processes.
Twin Tub
Both an N-Well and a P-Well are manufactured on
a lightly doped N-type substrate.
Silicon-on-Insulator (SOI) CMOS Process
SOI allows the creation of independent,
completely isolated nMOS and pMOS transistors
virtually side-by-side on an insulating substrate.

Dr. Samer Arandi - An-Najah National University

N- & P-Well Method


CMOS Requires that both n-channel (nMOS) and pchannel (pMOS) transistors be built on the same chip
substrate.
To accommodate both nMOS and pMOS devices,
special regions must be created in which the
semiconductor type is opposite to the substrate type.
These regions are called wells or tubs.
A p-well is created in an n-type substrate or,
alternatively, an n- well is created in a p-type substrate.
In the simple n-well CMOS fabrication technology, the
nMOS transistor is created in the p-type substrate, and
the pMOS transistor is created in the n-well, which is
built-in into the p-type substrate.
Dr. Samer Arandi - An-Najah National University

N-Well Fabrication Process


The creation of the n-well regions for pMOS transistors, by
impurity implantation into the substrate.
A thick oxide is grown in the regions surrounding the nMOS and
pMOS active regions.
The thin gate oxide is subsequently grown on the surface
through thermal oxidation.

The creation of n+ and p+ regions (source, drain and channelstop implants).


The metallization is created (creation of metal interconnects).
Dr. Samer Arandi - An-Najah National University

Lithography
Each processing step requires that certain areas
are defined on chip by appropriate masks.
Consequently, the integrated circuit may be
viewed as a set of patterned layers of doped
silicon, polysilicon, metal and insulating silicon
dioxide.
In general, a layer must be patterned before the
next layer of material is applied on chip.
The process used to transfer a pattern to a layer
on the chip is called lithography.
Dr. Samer Arandi - An-Najah National University

Basic Steps

Thermal oxidation of the silicon


surface: an oxide layer of about 1
micrometer thickness is created on the
substrate, see (b).
The entire oxide surface is covered
with a layer of photoresist, which is a
light-sensitive, acid-resistant organic
polymer, initially insoluble in the
developing solution (c).
The photoresist material is exposed to
ultraviolet (UV) light, the exposed
areas become soluble so that the they
are no longer resistant to etching
solvents.
To selectively expose the photoresist,
we have to cover some of the areas on
the surface with a mask during
exposure. Thus, when the structure
with the mask on top is exposed to UV
light:

areas which are covered by the opaque


features on the mask are shielded.
areas where the UV light strikes the
photoresist, it is exposed and
becomes soluble in certain solutes (d).

Dr. Samer Arandi - An-Najah National University

PhotoResist

PhotoResist normally comes in powder form, which is insensitive to light. It


is reconstituted into liquid form by adding a solvent, typically alcohol.
The wafer is mounted on a turntable, spinning slowly, and the photoresist is
discharged into its center. Centrifugal force spreads the resist outward
across the wafer. The thickness that remains on the wafer is a function of the
rate of wafer spin and the viscosity of the photoresist. The thickness is
monitored by light diffraction, which is used to adjust the spin rate to reach
the correct PR thickness.
Phase Interference gives
Photoresist Thickness

After the PR is applied, the wafer is heated (~160C) to evaporate the


solvent, leaving a smooth solid coating.
The type of photoresist which is initially insoluble and becomes soluble after
exposure to UV light is called positive photoresist.

Dr. Samer Arandi - An-Najah National University

Following the UV exposure step, the


unexposed portions of the photoresist can be
removed by a solvent.
The silicon dioxide regions which are not
covered by hardened photoresist can be
etched away either by using a chemical solvent
(HF acid) or by using a dry etch (plasma etch)
process (e).
At the end of this step, we obtain an oxide
window that reaches down to the silicon
surface (f).
The remaining (unexposed) photoresist can
be stripped from the silicon dioxide surface by
using another solvent, leaving the patterned
silicon dioxide feature on the surface, see (g).
The fabrication of semiconductor devices
requires several such pattern transfers to be
performed on silicon dioxide, polysilicon, and metal.
The basic patterning process used in all fabrication
steps, however, is quite similar to the one shown.
Dr. Samer Arandi - An-Najah National University

Summary
9 steps to make this simple
hole:
1. Oxidize silicon surface
2. Deposit photoresist
3. Anneal photoresist
4. Mount mask above silicon
5. Expose to UV light
6. Develop photoresist
7. Etch photoresist exposed
to UV
8. Etch SiO2 through
photoresist hole
9. Remove photoresist

Dr. Samer Arandi - An-Najah National University

Making a CMOS device

The oxidation of the silicon (field oxide) is


created on the surface (b).
The field oxide is selectively etched to expose
the silicon surface on which the MOS transistor
will be created (c).
The surface is covered with a thin, high-quality
oxide layer which will eventually form
the gate oxide of the MOS transistor (d).
A layer of polysilicon (polycrystalline silicon) is
deposited (e). Polysilicon is used
both as gate electrode material for MOS
transistors and also as an interconnect medium
in silicon integrated circuits. The resistivity of
polysilicon is reduced by doping
it with impurity atoms.
The polysilicon layer is patterned and etched to
form the interconnects and the MOS transistor
gates (f).

Dr. Samer Arandi - An-Najah National University

Making a CMOS device

The thin gate oxide not covered by


polysilicon is also etched away, which exposes
the bare silicon surface on which the source
and drain junctions are to be formed (g)

The entire silicon surface is then doped


with a high concentration of impurities, either
through diffusion or ion implantation (in this
case with donor atoms to produce n-type
doping).

(h) shows that the doping penetrates the


exposed areas on the silicon surface to create
two n-type regions (source and drain
junctions) in the p-type substrate. Note that
the polysilicon gate, which is patterned before
doping actually defines the precise location of
the channel region and, hence, the location of
the source and the drain regions. (self-aligned
process)

Once the source and drain regions are


completed, the entire surface is again covered
with an insulating layer of silicon dioxide (i).

The insulating oxide layer is then


patterned in order to provide contact windows
for the drain and source junctions (j).
Dr. Samer Arandi - An-Najah National University

Making a CMOS device


The surface is covered with evaporated aluminum (5000A) which will form
the interconnects (k).
Finally, the metal layer is patterned and etched, completing the
interconnection of the MOS transistors on the surface (l).

Dr. Samer Arandi - An-Najah National University

The N-Well Fabrication Process


We have covered the basic process steps for pattern transfer
through lithography, and gone through the fabrication procedure
of a single n-type MOS transistor.
The n-well CMOS integrated circuits fabrication follows the
same steps, however for creating two devices: N-MOS and PMOS.
The first lithographic mask defines the n-well region. Donor
atoms, usually phosphorus, are implanted through this window
in the oxide. Once the n-well is created
The rest of the steps follows similarly

Dr. Samer Arandi - An-Najah National University

You might also like