Lecture 4 Differential Amplifiers

Dr. Cuong HuynhTelecommunications DepartmentHCMUT

CMOS ANALOG IC DESIGN Spring 2013

1

Dr. Cuong Huynh [email protected]

Department of Telecommunications

Faculty of Electrical and Electronics Engineering

Ho Chi Minh city University of Technology

mailto:[email protected]

Dr. Cuong HuynhTelecommunications DepartmentHCMUT

CMOS ANALOG IC DESIGN Spring 2013

2

Dr. Cuong Huynh [email protected]

Department of Telecommunications

Faculty of Electrical and Electronics Engineering

Ho Chi Minh city University of Technology

Lecture 4: Differential Amplifiers

mailto:[email protected]

• A single-ended signal is measured with respect to a fixed potential (ground)

• A differential signal is measured between two equal and opposite signals which swing around a fixed potential (common-mode level)

• You can decompose differential signals into a differential mode (difference) and a common-mode (average

Single-Ended & Differential Signals

32 outout

CMoutoutDM

VVVVVV

Differential SignalSingle-Ended Signal

Single-Ended & Differential Amplifiers

4

• Differential signaling advantages

• Common-mode noise rejection

• Higher (ideally double) potential output swing

• Simpler biasing

• Improved linearity

• Main disadvantage is area, which is roughly double

• Although, to get the same performance in single-ended designs, we often have to increase the area dramatically

TnGSDDVVV

SwingOutput Max

TnGSDDVVV2

SwingOutput Max

Common-Mode Level Sensitivity

5

• A design which uses two single-ended amplifiers to realize a differential amplifier is very sensitive to the common-mode input level

• The transistors’ bias current and transconductance can vary dramatically with the common-mode input

• Impacts small-signal gain

• Changes the output common-mode, which impacts the maximum output swing

Differential Pair

6

• An improved differential amplifier topology utilizes a “tail” current source to keep the transistor bias current ideally constant over the common-mode input range

• Allows for a constant small-signal gain and output common-mode level

• Note, you still have to have keep the input pair and tail current source transistors in saturation

D

SS

DDCMO

DDDmmm

Dm

inin

outout

DM

RI

VV

RRRggg

RgVV

VVA

2

andwhere

,

2121

21

21

Differential Pair Input-Output Characteristics

7

• For large-signal differential inputs, the maximum output levels are well defined and ideally independent of the input common-mode

• For small-signal differential inputs, the small-signal gain is maximum at low-input signal levels

• As the differential input level increases, the circuit becomes more nonlinear and the gain decreases

Differential Pair I-V Characteristics

8

2

21

4

21

2

2

21

2

21

22

21

2

2121

21

2

21

21

2

21

21

2121

212121

4

1

4usingandsidesboth Squaring

22

1

22

usingandsidesboth Squaring

22

:DifferenceVoltageInput

ininoxnSSininoxnDD

DDssDDDDDD

DDSSininoxn

DDSS

oxn

inin

SSDD

oxn

D

oxn

D

inin

TGSTGSGSGSinin

VVL

WCIVV

L

WCII

IIIIIIIII

IIIVVL

WC

III

L

WC

VV

III

L

WC

I

L

WC

IVV

VVVVVVVV

2

212121

4

2

1inin

oxn

ss

ininoxnDDVV

L

WC

IVV

L

WCII


9

2

212121

4

2

1inin

oxn

ss

ininoxnDDVV

L

WC

IVV

L

WCII

• The differential current is an odd function of the differential input voltage which increases linearly for small inputs

• For large differential input voltages, the output differential current compresses due to the sqrt term

• The differential output current maxes out when all the current flows through one transistor at Vin1

L

WC

IV

oxn

ss

in

21


10

• The differential output current will saturate if the differential input voltage exceeds sqrt(2) times the equilibrium input overdrive voltage

2:OverdriveInput alDifferentiZero

overdriveinput aldifferentizero the to thisrelatecan We

2:InputalDifferentiMaximum

ofallsupport must M1,At this

ideally ,0For

0

M1 through flowscurrent all when caseheconsider trangecurrent maximumFor the

1

2,1

11

1

1121

22

1121

in

oxn

SS

TGS

oxn

SS

TGSin

SSin

inTGSGSGSin

TGSD

SSDSSDDD

V

L

WC

IVV

L

WC

IVVV

IV

VVVVVV

VVI

IIIIII

Differential Pair Transconductance

11

• The differential pair transconductance and gain is maximum near zero input differential voltage

DssoxnDmv

D

ssoxnm

in

in

oxn

ss

in

oxn

ss

oxn

in

D

m

inininDDD

inin

oxn

ss

ininoxnDD

RIL

WCRGA

R

IL

WCG

V

V

L

WC

I

V

L

WC

I

L

WC

V

IG

VVVIII

VV

L

WC

IVV

L

WCII

isgain signal-small the,resistorsload thegConsiderin

is0atuctance transcondsignal-smallThe

4

24

2

1

andDefine

4

2

1

2

2

2121

2

212121

Differential Pair Small-Signal AnalysisMethod 1 - Superposition

12

• The X output from Vin1 is modeled as a degenerated CS amplifier

Find Vout(Vin1)

21

and that Note

2

1

11

1

2121

Dm

m

m

Dm

in

X

DDDmmm

Rg

g

g

Rg

V

V

RRRggg


13

• The Y output from Vin1 is modeled as a Thevenin equivalent driving a CG amplifier

Find Vout(Vin1)

21

and that Note

1

2

12

1

2121

Dm

m

m

Dm

in

Y

DDDmmm

Rg

g

g

Rg

V

V

RRRggg


14

Dm

inin

ininDm

inin

totYX

inDmVYXVout

VoutVout

inDmin

DmDm

in

m

m

Dm

m

m

Dm

VYXVout

Vout

RgVV

VVRg

VV

VV

VRgVVV

VV

VRgVRgRg

V

g

g

Rg

g

g

RgVVV

V

inin

inin

inin

in

21

21

21

2

111

1

2

12

2

1

11

:GainalDifferenti

symmetry,circuit theFrom

2211

total thefindTo

22

12

11

1

Differential Pair Small-Signal AnalysisMethod 2 – Half Circuit

• The symmetric differential pair can be modeled as a Thevenin equivalent to observe how the tail node P changes with the differential input signal

• If RT1=RT2 and the input is a truly differential signal, node P remains constant

• This allows the tail node to be treated as a “virtual ground”

15

Differential Pair Small-Signal AnalysisMethod 2 – Half Circuit

16

Dm

in

inDm

inin

YX

Dm

in

Y

Dm

in

X

RgV

VRg

VV

VV

RgV

V

RgV

V

1

1

11

1

1

2

2:GainalDifferenti

• Applying the virtual ground concept allows modeling as two “half circuits”

Differential Pair Common-Mode Response

• Ideally, a differential amplifier completely rejects common-mode signals, i.e. Av,CM=0

• In reality, the finite tail current source impedance results in a finite common-mode gain

17

SSm

Dm

SSm

Dm

CMin

out

CMvRg

Rg

Rg

Rg

V

VA

2121

22

,

,

Differential Pair with Diode Loads

• While the gain of this amplifier is relatively small, it is somewhat predictable, as it is defined by the ratio of the transistor sizes and the n/p mobility

18

3

1

3

1

3

1

3

1

331

1

0Assuming

L

WC

L

WC

IL

WC

IL

WC

g

gA

g

g

ggg

gA

oxp

oxn

SSoxp

SSoxn

m

m

v

m

m

omo

m

v

Differential Pair w/ Current-Source Loads

• While the gain of this amplifier is higher, it is somewhat unpredictable, as it is defined by the transistor output resistance, which changes dramatically with process variations

19

31

1

0Assuming

oo

m

vgg

gA

Differential Pair w/ Diode & Parallel Current-Source Loads

• Adding a parallel current source to a diode connected load allows for increase gain which is still somewhat predictable

20

loaddiode thefromsteals""sourcecurrent that thepercentagecurrent theiswhere

1

0Assuming

3

1

3

1

3

1

5331

1

SSoxp

SSoxn

m

m

v

m

m

oomo

m

v

IL

WC

IL

WC

g

gA

g

g

gggg

gA

Cascode Differential Pair

21

5753131

57575313131

0Assuming

oomoommv

oomoooomoomv

rrgrrggA

rrgrrrrgrrgA

• Using a cascode differential pair and cascode current-source loads allows for aconsiderable increase in gain

• However, a relatively large power supply may be required to supply the necessary voltage “headroom” to keep all the transistors in saturation

Lecture 4 Differential Amplifiers

Documents

Transcript of Lecture 4 Differential Amplifiers