Sie sind auf Seite 1von 8

ENERGY-CONSUMPTION IN RC CIRCUITS: A BRIEF

INTRODUCTION TO ENERGY RECOVERY

Prof. Kaushik Roy


@ Purdue Univ.
Revisit Dynamic Energy Comsumption

Simple R-C model of Pass pMOS transistor


Consider a pMOS pass transistor
with a capacitive load C at the out
put. The voltage at the power
terminal swings from 0 to Vdd to
charge the node capacitance
through the transistor channel. The
channel is modeled by a normal
resistance R in fig (a).

Let us compute energy dissipated


while charging capacitance C from
0 to Vdd in time T with a linear
supply voltage as shown in fig (b).
Source: Intel

Prof. Kaushik Roy


@ Purdue Univ.
RC Circuit: Energy Dissipation

The voltage relations are dVC


shown here. RC ( )  VC  
dt

The supply voltage can be


expressed as
 0, t0
 Vdd
VC  ( ), 0  t  T
 T
 Vdd , t  T
Source: Intel

Prof. Kaushik Roy


@ Purdue Univ.
RC Circuits ( contd.)

Solving the voltage equation, we’ve


 0,
 t t0
 RC 
VC     ( )Vdd (1  e RC ), 0  t  T
 T
t ( t T ) t T
 RC   ,
  ( )Vdd (1  e RC )e RC
T

The energy dissipation in the charging process can be calculated as


T 
Elinear   iVR dt   iVR dt
0 T
Source: Intel

Prof. Kaushik Roy


@ Purdue Univ.
RC Circuits (contd.)

Term wise expanding, the first term can be expressed as


T T (  V ) 2
 iVR dt  C
dt
0 0 R
t
Vdd
T ( )
 [ RC (1  e RC ] / Rdt
0 T
T
t
RC 2  t
2 RC
 ( ) CVdd  (1  e RC ) 2 d ( )
T 0 RC
RC  3 RC RC ( 
T
) 1 RC ( 
2T
)
 ( )CVdd 1  ( )  2( )e RC  ( )e RC 
2
T  2 T T 2 T 
Source: Intel

Prof. Kaushik Roy


@ Purdue Univ.
RC Circuits (contd.)
The second term of the energy dissipation

  (  V ) 2
 iVR dt  C
dt
T T R
T ( t T )
RC   2
 CVdd2 (1  e RC ) 2 e
 RC dt
T T
T
RC 2 1 
( ) CVdd2 [ (1  e RC ) 2 ]
T 2
Final energy expression
T
RC RC RC  RC
Source: Intel Elinear  ( )CVdd [1 
2
 e ]
T T T

Prof. Kaushik Roy


@ Purdue Univ.
Energy Dissipation: 2 Cases

Let us consider the two extreme cases, when T >> RC

RC
Elinear  ( )CVdd2
T

And when T << RC, as in normal CMOS

RC RC RC T 1 T 2
Elinear  ( )CVdd2 [1   (1   ( ) )]
T T T RC 2 RC
1
 CVdd2
2

Source: Intel

Prof. Kaushik Roy


@ Purdue Univ.
Adiabatic Switching Concept

Source: Intel

Prof. Kaushik Roy


@ Purdue Univ.

Das könnte Ihnen auch gefallen