Добавил:
Опубликованный материал нарушает ваши авторские права? Сообщите нам.
Вуз: Предмет: Файл:

Cadence / DSD 4 / EE560_DynLog

.pdf
Скачиваний:
21
Добавлен:
16.04.2013
Размер:
309.15 Кб
Скачать

41

REDUCE CHARGE SHARING DEGRADATION OF Vx

WEAK PULLUP

 

 

 

 

 

 

 

 

 

 

 

VDD

 

 

weak pullup pMOS

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

V

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

V

 

 

 

 

 

x

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

out

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

nMOS

 

 

 

 

 

 

 

 

 

Pushes Vx to VDD unless there

inputs

 

 

 

 

Logic

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

is a strong pull-down path

 

 

 

 

 

 

 

 

 

 

 

 

 

Block

 

 

 

 

 

 

 

 

 

between Vout and ground.

CK

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Kenneth R. Laker, University of Pennsylvania

42

SEPERATE pMOS TRANSISTORS TO PRECHARGE INTERMEDIATE HIGH CAPACITANCE NODES

VDD

 

 

 

CK

 

Vx1

 

 

 

Vout1

 

 

 

inputs

nMOS

C1

 

 

Logic

 

 

 

Sub-Block

Vx2

Vout2

inputs

nMOS

 

C2

 

Logic

 

 

 

 

 

Sub-Block

 

 

 

nMOS Logic Block

EFFECTIVELY ELIMINATES ALL CHARGE SHARING PROBLEMS

DURING EVALUATION

Kenneth R. Laker, University of Pennsylvania

 

 

VDD

43

 

 

 

 

 

 

 

 

 

 

 

 

CK

 

C4

P4

G4

P3

C3

 

G3

C2

P2

 

G2

C1

P1

 

G1

 

C0

 

C1 = G1 + P1*C0

C2 = G2 + P2*G1 + P2*P1*C0

C3 = G3 + P3*G2 + P3*P2*G1 + P3*P2*P1*C0

C4 = G4 + P4*G3 + P4*P3*G2 + P4*P3*P2*G1 + P4*P3*P2*P1*C0

Kenneth R. Laker, University of Pennsylvania

VDD

VA

VB = 0

CK

44

Vx1 = 0 at t = 0

Vout

C1

Let C1 = C2 = 0.05 pF

Vx2 = 0 at t = 0

C2

nMOS Logic Block

WITHOUT

 

 

 

 

 

 

VDDC1

 

VDD

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

V

¹ V

EVALUATION:V =

=

 

PRECHARGE:

 

 

 

 

x1

x 2

x1

C1 + C2

2

 

WITH

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

PRECHARGE: Vx1 = Vx2 EVALUATION: Vx1 = VDD

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Kenneth R. Laker, University of Pennsylvania

45

NP DOMINO LOGIC (NORA or ZIPPER CMOS)

 

V

VDD

 

 

DD

 

 

CK

 

CK

CK

inputs

nMOS

pMOS

 

logic

inputs logic

inputs

 

 

block

block

 

To other

To other

pMOS

nMOS

blocks

blocks

VDD To other

pMOS blocks

nMOS logic block

NOTE: INVERTERS ARE NOT REQUIRED AT OUTPUTS OF STAGES

ALL inputs stable when CK = 1

CK

nMOS stages

all stages precharge all stages evaluate pMOS stagesevaluate

pre-discharge

Kenneth R. Laker, University of Pennsylvania

46

NP DOMINO LOGIC (NORA or ZIPPER CMOS) EXAMPLE

 

V

VDD

VDD

 

DD

 

 

CK

CK

 

CK

Kenneth R. Laker, University of Pennsylvania

47

SINGLE PHASE CLOCK PIPELINED DYNAMIC CMOS STRUCTURE

 

V

VDD

VDD

VDD

 

DD

 

 

 

CK

 

CK

 

 

 

 

 

 

inputs

nMOS

 

pMOS

CK

 

logic

inputs

logic

 

 

block

CK

block

To Next

 

 

 

 

 

 

 

 

nMOS

 

 

 

 

block

N-BLOCK

P-BLOCK

USING TRISTATE INVERTERS BETWEEN STAGES DECOUPLES THE STAGES AND ENABLES PIPILINED OPERATION.

CK - LOW: nMOS Blocks Precharge to VDD

pMOS Blocks Evaluate by selective pullup to VDD CK-HIGH: pMOS Blocks Pre-discharge to 0V

nMOS Blocks Evaluate by selective pull down to 0V Since CK-Inverse is not used, no clock skew problem can arise. PROVIDES SIMILAR PERFROMANCE TO NORA STRUCTURE.

Kenneth R. Laker, University of Pennsylvania

48

COMMON ADVANTAGES OF DYNAMIC LOGIC STYLES

1.Smaller area than fully static gates.

2.Smaller parasitic capacitances, hence higher speed.

3.Glitch free operation if DESIGNED CARFULLY.

Kenneth R. Laker, University of Pennsylvania

SUMMARY - GUIDELINES

49

 

1.Full complementary static logic is best option in the majority of CMOS

circuits. Noise-immunity not sensitive to kn/kp; does not involve pre charging of nodes; dissipates no DC power; layout can be automated. Large fan-in gates lead to complex circuit structures (2N transistors); larger parasitics; slower and higher dynamic power dissipation than alternatives; no clock.

2.Pseudo-nMOS static logic finds widest utility in large fan-in NOR gates. Requires only N+1 transistors for N fan-in; smaller parasitics; faster and lower dynamic power dissipation than full COS. Noise-immunity sensitive

to kn/kp; dissipates DC power when pulled down; and not well suited for automated layout; no clock.

4. CMOS domino logic should be used for low-power, high speed applications. Requires N+k transistors for N fan-in, size advantages of psuedo-nMOS; dissipates no DC power; noise immunity not sensitive to kn/kp; use of clocks enables synchronous operation. Relies on storage on soft nodes; will require exhaustive simulation at all the process corners to insure proper operation; some of the speed advantage over static gates is diminished by the required pre-charge (pre-discharge) time.

Kenneth R. Laker, University of Pennsylvania

Соседние файлы в папке DSD 4