Analog Integrated Circuits - KNTU · Why CMOS? • Advantages: • CMOS technology is a low-power...

Preview:

Citation preview

Analog Integrated Circuits

Hossein Shamsi

1

Why Analog?

• Processing of Natural Signals

2

Why Analog?

• Digital Communication

3

Why Analog?

• Disk Drive Electronics

4

Why Analog?

• Wireless Receivers

5

Why Analog?

• Optical Receivers

6

Why Analog?

• Sensors

7

Why Analog?• Microprocessors

– Distribution and timing of data and clocks

– Non-idealities in signal and power interconnects

– Issues related to the package parasitics.

• Memories– They needs high-speed “sense amplifiers”

“High-speed digital design is in fact analog design.”8

Why Integrated?• The birthday of Electronics is 1900.

• The main building block vacuum tube• Drawbacks

• Large size• Small life-time

• The birthday of Microelectronics is 1960 (25μm Technology).• Microelectronics: the knowledge of integration of transistors in a small area.• The main building block Transistor• Advantages

• small size• Infinite life-time

• Moore’s law: the number of transistors per chip has continued to double approximately every 1.5 years.

• 25μm CMOS Technology 45nm CMOS Technology • CPU (3cm×3cm) 100-million transistors• Handset 1-million transistors• Memory 1-billion transistors

• The state-of-the-art microelectronics products:• Laptop• Cell phone• Digital camera

9

Why CMOS?• Advantages:

• CMOS technology is a low-power technology.• In order to realize digital circuits in CMOS technology, we need fewer devices than its

Bipolar or GaAs counterparts.• Disadvantages

• Slower• Noisier

• Therefore CMOS technology is a helpful technology for digital circuits.• Since we want to integrate both the analog and digital circuits on a same

substrate, so we must design the analog circuits in CMOS technology.

10

Structure of a NMOS transistor

Ddrawneff LLL 2−=

Typical values:

11

N-Well Process

12

MOS Symbols

13

MOS I/V Characteristics

Formation of the depletion region

Formation of the inversion layer

12 GG VV >

14

MOS I/V Characteristics

15

Derivation of I/V Characteristics

A semiconductor bar carrying a current I

Snapshots of the carriers one second apart.

16

Derivation of I/V Characteristics

( )THGSoxd VVWCQ −−=

( )[ ]THChGoxd VxVWCQ −−= −( )[ ]THchGoxd VxVVWCQ −−−=

( )[ ]vVxVVWCI THGSoxD −−−=17

Derivation of I/V Characteristics

Since ID is constant along the channel

( )[ ]vVxVVWCI THGSoxD −−−=

18

voltageeffectivevoltageoverdriveVV THGS ≡≡−

ratioaspectL

W≡

regiontriodeVVVVVVVif THGSDSTHGDTHGS →−<≡>> ,19

( ) regiontriodedeepVVVif THGSDS →−<< 2

So we have:

Linear operation in deep triode region

MOSFET as a controlled linear resistor20

Saturation of the Drain Current

Pinch-off Behavior21

Saturation of the Drain Current

( )[ ]∫∫−

=

=

−−=THVGSV

VTHGSnOX

L

xD dVVxVVWCdxI

00

µ

22

Saturation Region

For long channel transistor, we have: LL ′≅

So the above formula is simplified as follows:

( )2

21

THGSoxnD VVL

WCI −= µ23

Transconductance of MOSFET

( )2

21

THGSoxnD VVL

WCI −= µ

24

Conceptual Visualization of Saturation and Triode Regions

25

Second Order Effects•Body Effect•Channel-Length Modulation•Subthreshold conduction•Leakage current•Short-channel effects

If VSB>0 the threshold voltage VTH is increased as follows:

26

Second Order Effects•Body Effect•Channel-Length Modulation•Subthreshold conduction•Leakage current•Short-channel effects

The actual length of the inverted channel gradually decreases as the value of VD increases. In other words, L‘ is in fact a function of VDS

(from semiconductor devices course)

27

Second Order Effects

Therefore, for longer channels, λ is smaller.

Channel-length modulation results in a nonzero slop in the ID /VDScharacteristic.

L1

∝λDSVLL λ=∆

28

Second Order Effects

In reality, for VGS<VTH a weak inversion layer still exist and some current flows from D to S. This phenomena is called weak inversion or subthreshold conductance.

•Body Effect•Channel-Length Modulation•Subthreshold conduction•Leakage current•Short-channel effects

5.1,75 =≥ nmVVDS

(David Johns)

29

Second Order Effects

The Leakage current of the PN junctions is important in estimating the maximum time a sample-and-hold circuit or a dynamic memory cell can be left in hold mode.The leakage current doubles for every 11 °C rise in temperature.

•Body Effect•Channel-Length Modulation•Subthreshold conduction•Leakage current•Short-channel effects

30

Second Order Effects

This phenomena will be discussed later (chapter 17, Razavi).

•Body Effect•Channel-Length Modulation•Subthreshold conduction•Leakage current•Short-channel effects

31

MOS Device Layout

The two most important masks are:•Active region mask•Gate polysilicon mask

The intersection of these two masks becomes the channel region of the MOS transistor.

The design rules for laying out transistors are often expressed in terms of λ where λ is ½ the gate length.

32

MOS Device Layout

When we are drawing the layout of a circuit, we must perform both the DRC and LVS tests.DRC Design Rule CheckLVS Layout Versus Schematic 33

MOS Device Layout

Example:

Area:

Peripheral: WλPs 210 +=

WλAJ ×= 51

34

Integrated Resistors

35

Types of integrated capacitors

36

Integrated capacitors

37

Capacitors Features

38

MOS Device Capacitances

C1 or Cch: oxide capacitanceC2 or Cd: depletion capacitanceC3,C4: overlap capacitanceC5,C6: junction capacitance

39

Effect of Layout on Parasitic Capacitances

( ) ( ) jswjSBDB CEWCEWCC ++×== 2

40

MOS Device Capacitances(Saturation Region)

MOS Device Capacitances(Saturation Region)

oxovoxeffGS WLCWCCWLC32

32

≅+=

oxDovGD CWLWCC ≅=

0≅GBC CGB is neglected because inversion layer act as a shield.

( )

perimetersourcePareasourceA

voltageinbuiltmV

CC

V

CC

WLACPCAAC

ss

Bm

B

SB

jswjswm

B

SB

jj

effchjswsjchsSB

::

:,4.03.0

1

,

1

,

00 −≤≤

+

=

+

=

=++=

− ϕ

ϕϕ

m

B

DB

jswjswm

B

DB

jjjswdjdDB

V

CC

V

CCCPCAC

+

=

+

=+= −

ϕϕ1

,

1

, 00

42

MOS Device Capacitances(Deep Triode Region)

MOS Device Capacitances(Deep Triode Region)

ovox

GS WCWLCC +≅2

0≅GBC CGB is neglected because inversion layer act as a shield.

m

B

SB

jswjswm

B

SB

jjjswsj

chsSB

V

CC

V

CCCPCAAC

+

=

+

=+

+= −

ϕϕ1

,

1

,2

00

m

B

DB

jswjswm

B

DB

jjjswdj

chdDB

V

CC

V

CCCPCAAC

+

=

+

=+

+= −

ϕϕ1

,

1

,2

00

ovox

GD WCWLCC +≅2

44

MOS Device Capacitances(Cut-off)

ovGS WCC =

( )( ) oxGBGd

oxd

oxdGB WLCCVC

WLCCWLCCC ≅⇒<<+

≅ 0ife,capacitancdepletion:,

m

B

SB

jjjswsjsSB

V

CCCPCAC

+

=+=

ϕ1

, 0

m

B

DB

jjjswdjdDB

V

CCCPCAC

+

=+=

ϕ1

, 0

ovGD WCC =

45

Beahvior of MOS Device as a Capacitor

NMOS operating in accumulation region

46

MOS Low-Frequency Small-Signal Model(active region)

47

MOS Small-Signal Model(active region)

48

MOS Small-Signal Model(triode region)

( )THGSoxn

ds

VVL

WCr

−=µ

1

49

MOS Small-Signal Model(Turn off region)

50

MOS SPICE Model

TechnologyCMOSμm0.5

51

MOS SPICE Model

52

Comparison between NMOS and PMOS

PMOS devices are quite inferior to NMOS transistors.

For example, due to the lower mobility of holes: oxnoxp CC µµ 5.0≅

nmpm gg <

Besides, for a given current and dimension we have:nopo rr <

53

FinFET

• FinFET utilizes a three-dimensional structure, which results in small channel length about 20nm.

• Equivalent channel width: • Typically, WF=6nm, HF=50nm.• In practice, the designer does not have any control on WF and HF.• Wider transistors can be obtained by increasing the number of fins.

54

Utilizing Multiple Fins

Top View

55

Recommended