ee370 chapter1 slides
DESCRIPTION
a brief introduction to DIGITAL ELECTRONICSTRANSCRIPT
-
7/31/2015
1
EE370Digital Electronics
Yogesh S. ChauhanDepartment of Electrical Engineering
IIT KanpurEmail: [email protected]
Office: WL125, Phone: 7244
Computer
An electronic device which is capable of receiving information (data) in a particular form and of performing a sequence of operations in accordance with a predetermined but variable set of procedural instructions (program) to produce a result in the form of information or signals.
Minimize human efforts with speed and efficiency.
Y. S. Chauhan, IIT Kanpur 2
Why Digital?
Noise Margin Can be reduced to the limit we want
Fidelity/Regeneration Robustness Scalability Data Storage Ease of data handling
Y. S. Chauhan, IIT Kanpur 3
Digital System
Y. S. Chauhan, IIT Kanpur 4
adminHighlight
-
7/31/2015
2
The First Computer
Y. S. Chauhan, IIT Kanpur 5
The BabbageDifference Engine(1832)
25,000 partscost: 17,470
ENIAC - The first electronic computer (1946)
17,468 vacuum tubes
7200 crystal diodes
1500 relays 70,000 resistors 10,000 capacitors Weight > 27 Ton 1800 sq. ft. 150 kW of
electricityY. S. Chauhan, IIT Kanpur 6
The Transistor Revolution
Y. S. Chauhan, IIT Kanpur 7
First transistorBell Labs, 1948
The First Integrated Circuits
Y. S. Chauhan, IIT Kanpur 8
Bipolar logic1960s
ECL 3-input GateMotorola 1966
-
7/31/2015
3
Intel 4004 Micro-Processor
Y. S. Chauhan, IIT Kanpur 9
19711000transistors1MHzoperation
Intel Pentium (IV) microprocessor
Y. S. Chauhan, IIT Kanpur 10
Moores Law In 1965, Gordon Moore noted that the number of
transistors on a chip doubled every 18 to 24 months. He made a prediction that semiconductor technology
will double its effectiveness every 18 months
Y. S. Chauhan, IIT Kanpur 11
16151413121110
9876543210
1
9
5
9
1
9
6
0
1
9
6
1
1
9
6
2
1
9
6
3
1
9
6
4
1
9
6
5
1
9
6
6
1
9
6
7
1
9
6
8
1
9
6
9
1
9
7
0
1
9
7
1
1
9
7
2
1
9
7
3
1
9
7
4
1
9
7
5
L
O
G
2
O
F
T
H
E
N
U
M
B
E
R
O
F
C
O
M
P
O
N
E
N
T
S
P
E
R
I
N
T
E
G
R
A
T
E
D
F
U
N
C
T
I
O
N
Transistor Counts
Y. S. Chauhan, IIT Kanpur 12
adminHighlight
-
7/31/2015
4
Moores law in Microprocessors
Y. S. Chauhan, IIT Kanpur 13
40048008
80808085 8086
286386
486Pentium proc
P6
0.001
0.01
0.1
1
10
100
1000
1970 1980 1990 2000 2010Year
T
r
a
n
s
i
s
t
o
r
s
(
M
T
)
2Xgrowthin1.96years!
TransistorsonLeadMicroprocessorsdoubleevery2years
Die Size Growth
Y. S. Chauhan, IIT Kanpur 14
40048008
80808085
8086286
386486 Pentiumproc
P6
1
10
100
1970 1980 1990 2000 2010Year
D
i
e
s
i
z
e
(
m
m
)
~7%growthperyear~2Xgrowthin10years
Diesizegrowsby14%tosatisfyMooresLaw
Frequency
Y. S. Chauhan, IIT Kanpur 15
P6Pentiumproc
48638628680868085
8080800840040.1
1
10
100
1000
10000
1970 1980 1990 2000 2010Year
F
r
e
q
u
e
n
c
y
(
M
h
z
)
LeadMicroprocessorsfrequencydoublesevery2years
Doublesevery2years
Power Dissipation
Y. S. Chauhan, IIT Kanpur 16
P6Pentiumproc
486386
2868086
80858080
80084004
0.1
1
10
100
1971 1974 1978 1985 1992 2000Year
P
o
w
e
r
(
W
a
t
t
s
)
LeadMicroprocessorspowercontinuestoincrease
Courtesy,Intel
adminHighlight
-
7/31/2015
5
Power density
Y. S. Chauhan, IIT Kanpur 17
400480088080
8085
8086
286 386 486Pentium proc
P6
1
10
100
1000
10000
1970 1980 1990 2000 2010Year
P
o
w
e
r
D
e
n
s
i
t
y
(
W
/
c
m
2
)
HotPlate
NuclearReactor
RocketNozzle
Powerdensitytoohightokeepjunctionsatlowtemp
Power is a major problem
Y. S. Chauhan, IIT Kanpur 18
5KW18KW
1.5KW500W
40048008
80808085
8086286
386486
Pentium proc
0.1
1
10
100
1000
10000
100000
1971 1974 1978 1985 1992 2000 2004 2008Year
P
o
w
e
r
(
W
a
t
t
s
)
Powerdeliveryanddissipationwillbeprohibitive
Not Only Microprocessors
Y. S. Chauhan, IIT Kanpur 19
AnalogBaseband
DigitalBaseband
(DSP+MCU)
PowerManagement
SmallSignalRF
PowerRF
CellPhone
Challenges in Digital Design
Y. S. Chauhan, IIT Kanpur 20
Microscopic Problems Ultra-high speed design Interconnect Noise, Crosstalk Reliability, Manufacturability Power Dissipation Clock distribution.
Everything Looks a Little Different
Macroscopic Issues Time-to-Market Millions of Gates High-Level Abstractions Reuse & IP: Portability Predictability etc.
and Theres a Lot of Them!
adminHighlight
-
7/31/2015
6
Why Scaling?
Technology shrinks by 0.7/generation With every generation can integrate 2X more
functions per chip; chip cost does not increase significantly
Cost of a function decreases by 2x But
How to design chips with more and more functions? Design engineering population does not double every
two years Hence, a need for more efficient design methods
Exploit different levels of abstraction
Y. S. Chauhan, IIT Kanpur 21
Design Abstraction Levels
Y. S. Chauhan, IIT Kanpur 22
n+n+S
GD
+
DEVICE
CIRCUIT
GATE
MODULE
SYSTEM
DesignAutomationisthekey.
Design Metrics
How to evaluate performance of a digital circuit (gate, block, )? Cost Reliability Scalability Speed (delay, operating frequency) Power dissipation Energy to perform a function
Y. S. Chauhan, IIT Kanpur 23
Cost of Integrated Circuits
NRE (non-recurring engineering) costs Design time and effort, mask generation One-time cost factor
Recurring costs Silicon processing, packaging, test proportional to volume proportional to chip area
Y. S. Chauhan, IIT Kanpur 24
adminHighlight
adminHighlight
adminHighlight
-
7/31/2015
7
Cost of Integrated Circuits
Cost per IC = Variable cost per IC + (fixed cost/volume)
The impact of fixed cost is more pronounced for small volume products.
Y. S. Chauhan, IIT Kanpur 25
NRE Cost is Increasing
Y. S. Chauhan, IIT Kanpur 26
Die Cost
Y. S. Chauhan, IIT Kanpur 27
Single die
Wafer
From http://www.amd.com
Going up to 12 (30cm)
Cost per Transistor
Y. S. Chauhan, IIT Kanpur 28
0.0000001
0.000001
0.00001
0.0001
0.001
0.01
0.11
1982 1985 1988 1991 1994 1997 2000 2003 2006 2009 2012
cost: -per-transistor
Fabrication capital cost per transistor (Moores law)
-
7/31/2015
8
Yield
Y. S. Chauhan, IIT Kanpur 29
%100per wafer chips ofnumber Total
per wafer chips good of No. Y
yield Dieper wafer Dies
costWafer cost Die
area die2
diameterwafer
area die
diameter/2wafer per wafer Dies
2
Yield
Y. S. Chauhan, IIT Kanpur 30
%100per wafer chips ofnumber Total
per wafer chips good of No. Y
yield Dieper wafer Dies
costWafer cost Die
Defects
Y. S. Chauhan, IIT Kanpur 31
area dieareaunit per defects1yield die is approximately 3
4area) (die cost die f
Some Examples (1994)
Y. S. Chauhan, IIT Kanpur 32
adminHighlight
-
7/31/2015
9
ReliabilityNoise in Digital Integrated Circuits
Noise unwanted variations of voltages and currents at the logic nodes
Y. S. Chauhan, IIT Kanpur 33
i(t)
Inductive coupling Capacitive coupling Power and groundnoise
v(t) VDD
Voltagechangeononewirecaninfluencesignalontheneighboringwire
crosstalk
Currentchangeononewirecaninfluencesignalontheneighboringwire
Example of Capacitive Coupling
Signal wire glitches as large as 80% of the supply voltage will be common due to crosstalk between neighboring wires as feature sizes continue to scale
Y. S. Chauhan, IIT Kanpur 34
Crosstalkvs.Technology
0.16mCMOS0.12mCMOS
0.35mCMOS
0.25mCMOS
PulsedSignal
BlacklinequietRedlinespulsedGlitchesstrengthvstechnology
FromDunlop,Lucent,2000
Lets start Digital Electronics
Digital Electronics uses binary number system Logic 0 == Logic Low Logic 1 == Logic High
Real world is not digital
Y. S. Chauhan, IIT Kanpur 35
Analog to Digital conversion
Y. S. Chauhan, IIT Kanpur 36
adminHighlight
adminHighlight
-
7/31/2015
10
Binary digital signal
Y. S. Chauhan, IIT Kanpur 37
Digital Logic Inverter
The logic inverter is the most basic element in digital circuit design.
It plays a role parallel to that of the amplifier in analog circuits. Inverter inverts the logic value of its input signal. Thus, for a logic-0 input, the output will be a logic 1, and vice
versa.
Y. S. Chauhan, IIT Kanpur 38
Input-Output characteristics of an Inverter
Lets draw.
Y. S. Chauhan, IIT Kanpur 39
Amplifier vs. Inverter
Y. S. Chauhan, IIT Kanpur 40
-
7/31/2015
11
DC OperationVoltage Transfer Characteristic (VTC)
VIL Max. value vI can have while being interpreted as logic 0.
VIH Min. value vI can have while being interpreted as logic 1.
Y. S. Chauhan, IIT Kanpur 41
IsitInverter?
Exactvalueofinputdoesntmatter.
Noise Margins
Y. S. Chauhan, IIT Kanpur 42
Insensitivity of the inverter output to the exact value of vI within allowed regions is a great advantage that digital circuits have over analog circuits.
Consider an inverter driving inverter.
Noise Margins
Four parameters, VOH, VOL, VIH, and VIL, define the VTC of an inverter and determine its noise margins Changes in vI within the noise margins are rejected by the inverter. Noise is not allowed to propagate further through the system, a definite
advantage of digital over analog circuits. You can think of the inverter as restoring the signal levels to
standard values (VOL and VOH) even when it is presented with corrupted input signal levels (within the noise margins).
Y. S. Chauhan, IIT Kanpur 43
ImportantParametersoftheVTCoftheLogicInverter
Voltage Transfer Characteristic
Switching Threshold Voltage
Y. S. Chauhan, IIT Kanpur 44
Vin
Vout
VOH
VOL
VM
VOHVOL
Vout=Vin
SwitchingThreshold
NominalVoltageLevels
adminHighlight
adminHighlight
adminHighlight
adminHighlight
adminHighlight
adminHighlight
-
7/31/2015
12
Voltage Transfer Characteristic
Switching Threshold Voltage (VM) Output is short circuited with input.
Y. S. Chauhan, IIT Kanpur 45V(x)
V(y)
f
V(y)V(x)
VOH = f (VIL)
VIL VIH
V(y)=V(x)
Switching ThresholdVM
VOL = f (VIH)
Vout=f(Vin)
Mapping between analog and digital signals The regions of acceptable high and low voltages are delimited
by VIH and VIL that represent the points on the VTC curve where the gain = -1
Y. S. Chauhan, IIT Kanpur 46V IL V IH V in
Slope = -1
Slope = -1
V OL
V OH
Vout
0 VOL
VIL
VIH
VOH
UndefinedRegion
1
Noise Margins For robust circuits, want the 0 and 1 intervals to be as
large as possible
Y. S. Chauhan, IIT Kanpur 47
UndefinedRegion
"1"
"0"
Gate Output Gate Input
VOH
VIL
VOL
VIHNoise Margin High
Noise Margin Low
NMH = VOH - VIH
NML = VIL - VOL
Gnd
VDD VDD
Gnd
Large noise margins are desirable, but not sufficient
The Regenerative Property A gate with regenerative property ensure that a disturbed
signal converges back to a nominal voltage level
Y. S. Chauhan, IIT Kanpur 48
v0 v1 v2 v3 v4 v5 v6
-1
1
3
5
0 2 4 6 8 10
t (nsec)
V
(
v
o
l
t
s
) v0
v2
v1
adminHighlight
adminHighlight
adminHighlight
adminHighlight
adminHighlight
-
7/31/2015
13
Conditions for Regeneration
To be regenerative, the VTC must have a transient region with a gain greater than 1 (in absolute value) bordered by two valid zones where the gain is smallerthan 1. Such a gate has two stable operating points.
Y. S. Chauhan, IIT Kanpur 49
v1 = f(v0) v1 = finv(v2)v0 v1 v2 v3 v4 v5 v6
v0
v1
v2
v3 f(v)
finv(v)
Regenerative Gate
v0
v1
v2
v3
f(v)
finv(v)
Nonregenerative Gate
in
out
in
out
Noise Immunity
Noise margin expresses the ability of a circuit to overpower a noise source noise sources: supply noise, cross talk, interference, offset
Absolute noise margin values are deceptive a floating node is more easily disturbed than a node driven by a
low impedance (in terms of voltage) Noise immunity expresses the ability of the system to
process and transmit information correctly in the presence of noise Transfer function between noise source and signal node is
-
7/31/2015
14
VTC of an ideal inverter
Y. S. Chauhan, IIT Kanpur 53
VTC of an ideal inverter The ideal gate should have
Infinite gain in the transition region Switching threshold located in the middle of the logic swing High and low noise margins equal to half the swing Input and output impedances of infinity and zero, respectively
Y. S. Chauhan, IIT Kanpur 54
Ri = Ro = 0Fanout = NMH = NML = VDD/2
g =
V in
V out
An Old-time Inverter
VOH=3.5V VIH=2.35V VM=1.64V NMH=1.15V
VOL=0.45V VIL=0.66V NML=0.21V
Y. S. Chauhan, IIT Kanpur 55
NM H
Vin (V)
V
o
u
t
(
V
)
NM L
VM
0.0
1.0
2.0
3.0
4.0
5.0
1.0 2.0 3.0 4.0 5.0
VOL
VOL
VOH
VOH
VIL VIH
Note Asymmetric NML is low Swing=3.05V
-
7/31/2015
15
Delay Definitions
Y. S. Chauhan, IIT Kanpur 57
t
Vout
Vin
inputwaveform
outputwaveform
tp =(tpHL +tpLH)/2
Propagationdelay
t
50%
tpHL
50%
tpLH
tf
90%
10%tr
signalslopes
Vin Vout
How fast we can toggle?
tpLH and tpHL can be asymmetric
Time period T> max(tpLH, tpHL)x2 If symmetric, T> max(tp)x2 Duty cycle
time for which it is ON/ (Time period) Usually its 50%: Rise and fall should occur within T/2 time
Y. S. Chauhan, IIT Kanpur 58
Vin Vout
How to compare delay in different technologies?
Y. S. Chauhan, IIT Kanpur 59
v0 v1 v5
v1 v2v0 v3 v4 v5
T = 2 tp N
RingOscillator
N=numberofinverters
Modeling Propagation Delay
Model circuit as first-order RC network
Y. S. Chauhan, IIT Kanpur 60
R
C
vin
vout
vout (t) = (1 et/)V
where = RCTime to reach 50% point is
0V
adminHighlight
adminHighlight
adminHighlight
-
7/31/2015
16
Modeling Propagation Delay
Model circuit as first-order RC network
Y. S. Chauhan, IIT Kanpur 61
R
C
vin
vout
vout (t) = (1 et/)V
where = RCTime to reach 90% from 10% point is
0V
Modeling Propagation Delay
Model circuit as first-order RC network
Matches the delay of an inverter gateY. S. Chauhan, IIT Kanpur 62
R
C
vin
vout
vout (t) = (1 et/)V
where = RC
Time to reach 50% point ist = ln(2) = 0.69
Time to reach 90% from 10% point ist = ln(9) = 2.2
0V
Power and Energy Dissipation
Power consumption: how much energy is consumed per operation and how much heat the circuit dissipates Supply line sizing (determined by peak power)
Ppeak = Vdd.ipeak=max(p(t)) Battery lifetime (determined by average power dissipation)
.
1
Packaging and cooling requirements
Y. S. Chauhan, IIT Kanpur 63
Static and Dynamic power
Two important components: static and dynamic Static power dissipation:
Caused by static conductive paths between the supply rails or by leakage currents.
Always present, even when the circuit is in stand-by. Minimization of this consumption source is a worthwhile goal.
Dynamic power dissipation: Occurs only during transients, when the gate is switching.
Charging of capacitors & temporary current paths between supply rails. Proportional to the switching frequency: the higher the number
of switching events, the higher the dynamic power consumption.
Y. S. Chauhan, IIT Kanpur 64
adminHighlight
adminHighlight
adminHighlight
adminHighlight
-
7/31/2015
17
A First-Order RC Network
Y. S. Chauhan, IIT Kanpur 65
vout
vin CL
R
E0 1 P t dt0
T
=
Energydissipatedfromsource
Energytransferredtocapacitor
Ecap Pcap t dt0
T
=
V(t)=VDD
V(t)=Vout
A First-Order RC Network
Y. S. Chauhan, IIT Kanpur 66
vout
vin CL
R
E0 1 P t dt0
T
Vdd isupply t dt0
T
Vdd CLdVout0
Vdd
CL Vdd 2= = = =
Energydissipatedfromsource
Energytransferredtocapacitor
Whereisotherhalf?
Ecap Pcap t dt0
T
Vouticap t dt0
T
CLVoutdVout0
Vdd
12---CL Vdd2
= = = =
Energy and Power
Y. S. Chauhan, IIT Kanpur 67
E (joules) = CL Vdd2 P01 + tsc Vdd Ipeak P01 + Vdd Ileakage
P (watts) = CL Vdd2 f01 + tscVdd Ipeak f01 + Vdd Ileakage
f01 = P01 * fclock
Energy and Energy-Delay
Propagation delay and power consumption of a gate are related Propagation delay is mostly determined by the speed at which a given
amount of energy can be stored on the gate capacitors. The faster the energy transfer (or the higher the power consumption),
the faster the gate. Product of power consumption and propagation delay is generally a
constant for a given technology and gate topology.
An ideal gate is one that is fast, and consumes little energy. The energy-delay product is a combined metric that brings those two
elements together, and is often used as the ultimate quality metric.
Y. S. Chauhan, IIT Kanpur 68
Power-Delay Product (PDP) =
E = Energy per operation = Pav tp
Energy-Delay Product (EDP) =
quality metric of gate = E tp
adminHighlight
adminHighlight
adminHighlight
adminHighlight
adminHighlight
adminHighlight
-
7/31/2015
18
Analog vs. DigitalAnalog Digital
1. Computewithcontinuousvaluesofphysicalvariablesinsomerange
Computewithdiscretevaluesofphysicalvariables
2. Primitivesofcomputationarisefromthephysicsofthecomputingdevices:physicalrelationsoftransistors,capacitors,resistors,floatinggatedevices,Kirchoffs currentandvoltagelawsandsoforth.Theuseoftheseprimitivesisanartformanddoesnotlenditselfeasilytoautomation. Theamountofcomputationsqueezedoutofasingletransistorishigh.
PrimitivesofcomputationarisefromthemathematicsofBooleanlogic:logicalrelationslikeAND,OR,NOT,NAND,andXOR.Theuseoftheseprimitivesisascienceandlendsitselfeasilytoautomation.Thetransistorisusedasaswitch,andtheamountofcomputationsqueezedoutofasingletransistorislow.
3. Computationisoffsetpronesinceitissensitivetomismatchesintheparametersofthephysicaldevices.Thedegradationinperformanceisgraceful.
Computationisnotoffsetpronesinceitisinsensitivetomismatchesintheparametersofthephysicaldevices.However,asinglebiterrorcanresultincatastrophicfailure.
Y. S. Chauhan, IIT Kanpur 69
Analog vs. Digital
Y. S. Chauhan, IIT Kanpur 70
Analog Digital4. Noiseisduetothermalfluctuationsin
physicaldevices.Noiseisduetoroundofferror.
5. Signalisnotrestoredateachstageofthecomputation.
Signalisrestoredto1or0ateachstageofthecomputation.
6. Inacascadeofanalogstages,noisestartstoaccumulate.Thus,complexsystemswithmanystagesaredifficulttobuild.
Roundofferrordoesnotaccumulatesignificantlyformanycomputations.Thus,complexsystemswithmanystagesareeasytobuild.
7. StaticpowerdissipationPA=N.VDD.I
Static powerisduetoleakagecurrentsonly.
8. Littleornodynamicpowerdissipation DynamicpowerconsumptionPD=N.f.C.VDD2
Inside chip
Y. S. Chauhan, IIT Kanpur 71
Acknowledgement
Jan M. Rabaey book Irwin & Vijays slides, Penn State University Sedra and Smith book S.S.K. Iyer, IIT Kanpur Course TAs
Y. S. Chauhan, IIT Kanpur 72
adminHighlight