Transcript View File

EE210 Digital Electronics
Class Lecture 10
June 19, 2008
Home Work No. 5 (Due June 26, 2008 )
Problems at the End Of Chapter 10.
1.
2.
3.
4.
Problem D10.25
Problem D10.26
Problem D10.46
Problem D10.47
In This Class
We Will Continue to Discuss:
Complete CMOS Gates
10.3.9 Effects of Fan-in and Fanout on Propagation Delay
10.4 Pseudo-NMOS Logic Ckts
10.5 Pass-Transistor Logic Ckts
10.3.9 Effects of Fan-in and Fan-out
• As compared to other MOS logic, CMOS logic
requires two transistors (NMOS and PMOS) for
each additional Input
• This results in Increase in Chip Area and Total
Effective Capacitance per gate hence Increase in
Propagation Delay
• Increase in number of in-put (fan-in) increases
size and increase tp and due to that there is limit
on fan-in of the NAND gate to 4.
• For higher inputs clever logic design is required
10.3.9 Effects of Fan-in and Fan-out
• Increase in gate’s fan-out adds directly to
its load capacitance and increases
propagation delay
• Thus apart from many advantages, CMOS
suffers from increased circuit complexity
when fan-in and fan-out are increased
• In next two sections we shall study some
simplified forms of CMOS logic that reduces
this complexity, however, we will loose
some advantages of CMOS
10.4 Pseudo-NMOS Logic Ckts
10.4.1 Pseudo-NMOS Inverter
Input is applied at QN gate and QP
gate is grounded, thus QP acts as an
active load for QN.
Note that before even going into
more details the advantage over
CMOS is already obvious: One
transistor (NMOS) for each input
10.4.2 Static Characteristics
• The static characteristics of pseudo-NMOS inverter can be
derived in similar manner as we did for CMOS inverter
• So the drain currents of QN and QP are:
1
k n (vI  Vt ) 2 , for vO  vI  Vt (saturation ) (Eq. 10.29)
2
1
iDN  k n [(vI  Vt )vO  vO2 ] for vO  vI  Vt ( triode) (Eq. 10.30)
2
1
iDP  k p (VDD  Vt ) 2 , for vO  Vt (saturation ) (Eq. 10.31)
2
1
iDP  k p [(VDD  Vt )(VDD  vO )  (VDD  vO ) 2 ] for vO  Vt ( triode)
2
iDN 
(Eq. 10.32)
where Vtn = -Vtp = Vt and kn=kn’(W/L)n and kp=kp’(W/L)p
• To Obtain VTC, we superimpose load curve (QP iD-vDS) on
the QN iD-vDS characteristics for various values of vGS=vI
10.4.2 Static Characteristics
NOTE that:
1. QP has much lower saturation
current than QN at vI=VDD. This is
because pseudo-NMOS inverter
usually designed so that kn is 4 to
10 times kp.
In fact, r ≡ kn/kp determines all the
breakpoints of VTC and noise
margins. High value of r reduces
VOL and widens NMs.
2. One tends to think of Qp as
constant-current source, but it
actually operates in saturation for
only a small range of vO ≤ Vt. Rest
is in triode region
10.4.2 Static Characteristics
Now consider two extreme cases
of vI: When vI=0, QN is cut off
and QP is operating in the triode
region with zero current and zero
drain-source voltage. Point A,
where vO=VOH=VDD, Static
current is zero.
When vI=VDD the operating point
is E. Unlike CMOS here VOL is
not zero – a disadvantage.
Another disadvantage is that the
gate conducts Istat in low output
state, thus there will be static
power dissipation: PD=IstatxVDD
10.4.3 Derivation of VTC
• VTC of pseudo-NMOS
inverter has four regions
labeled I thru IV
corresponding to different
modes of operation of QN
and QP.
Region
Segment
QN
QP
Condition
I
AB
Cutoff
Triode
vI < Vt
II
BC
Saturation
Triode
vO ≥ vI-Vt
III
CD
Triode
Triode
Vt ≤ vO ≤ vI-Vt
IV
DE
Triode
Saturation
vO ≤ Vt
10.4.3 Derivation of VTC
Region I (Segment AB)
vO = VOH = VDD
Region II (Segment BC)
Equating iDN and iDP and using kn=rkp
and some manipulation
vO  Vt  (VDD  Vt ) 2  r (vI  Vt ) 2
VIL  Vt 
VDD  Vt
r (r  1)
VM  Vt 
VDD  Vt
r 1
Region
Segment
QN
QP
Condition
I
AB
Cutoff
Triode
vI < Vt
II
BC
Saturation
Triode
vO ≥ vI-Vt
III
CD
Triode
Triode
Vt ≤ vO ≤ vI-Vt
IV
DE
Triode
Saturation
vO ≤ Vt
10.4.3 Derivation of VTC
Region III (Segment CD)
Being short segment is not very
important. Point D is vO = Vt
.
Region
Segment
QN
QP
Condition
I
AB
Cutoff
Triode
vI < Vt
II
BC
Saturation
Triode
vO ≥ vI-Vt
III
CD
Triode
Triode
Vt ≤ vO ≤ vI-Vt
IV
DE
Triode
Saturation
vO ≤ Vt
10.4.3 Derivation of VTC
Region IV (Segment DE)
Equating iDN and iDP and using kn=rkp
results
1
vO  (vI  Vt )  (vI  Vt ) 2  (VDD  Vt ) 2
r
2
VIH  Vt 
(VDD  Vt )
3r

1
VOL  (VDD  Vt ) 1  1  
r

Region
Segment
QN
QP
Condition
I
AB
Cutoff
Triode
vI < Vt
II
BC
Saturation
Triode
vO ≥ vI-Vt
III
CD
Triode
Triode
Vt ≤ vO ≤ vI-Vt
IV
DE
Triode
Saturation
vO ≤ Vt
10.4.3 Derivation of VTC
Finally, we use these equations to determine NML
and NMH

NM L  Vt  (VDD  Vt ) 1 

2

NM H  (VDD  Vt )1 
3r


1
1
1 

r
r (r  1) 



Since VDD and Vt are determined by the process
technology the only parameter for controlling the
values of VOL and the noise margins is the ratio r.
10.4.4 Dynamic Operation
• Similar procedure as CMOS inverter is used to
find tPLH and tPHL for the pseudo-NMOS inverter
and are given as:
1.7C
t PLH 
t PHL 
k pVDD
1.7C
k nVDD
for large value of r
• Although these are identical to CMOS inverter,
the pseudo-NMOS inverter has a special
problem: Since kp is r times smaller than kn, tPLH
will be r times larger than tPHL. Thus ckt exhibits
asymmetrical performance.
10.4.6 Gate Circuits
Except for the load device, pseudo-NMOS
gate circuit is identical to PDN of CMOS gate
Note that NOR and NAND gates each require
five transistors compared to eight in CMOS
10.4.7 Concluding Remark
• The pseudo-NMOS is particularly suited for
applications in which the output remains high
most of the time. Since gate dissipates static
power only in low-output state, the static power
dissipation in this application is low.
• The propagation delay can be made as short as
necessary for the output transition from high-tolow.
• This type of application can be found in ROMs.
Example 10.3
10.5 Pass-Transistor Logic Ckts
• A simple approach to implement logic functions is to use
series and parallel combination of switches controlled by
input logic variables to connect input and output nodes
These switches can be NMOS transistor
or pair of complementary MOS transistor
connected as CMOS transmission
gate configuration
10.5 Pass-Transistor Logic Ckts
• This provides simple form of logic circuit that is
particularly suited for special logic functions and is
frequently used in conjunction with CMOS logic to
implement such functions efficiently
• Because this form of logic utilizes MOS transistors in
series path from input to output, to PASS or block
signal, it is known as Pass-Transistor Logic (PTL)
• As CMOS transmission gates are frequently used to
implement switches in these logic circuits an
alternative name as Transmission-gate logic.
Generally both these names are used irrespective of
actual switches
• Although simple, PTL logic circuits design require care
10.5.1 Essential Design Requirement
• Essential requirement to design PTL circuit is to ensure that every ckt
node has at all times a low-resistance path to VDD or ground.
Switch S1 is used to form AND Function
of its controlling variable B and
variable A at the CMOS inverter output.
Y of PTL is connected to input of another
inverter. When B is Hi S1 closes and Y=A.
Node Y will be connected to either VDD
(if A is Hi) or to ground (if A is Low). But what
happens when B is low and S1 opens? Node Y will now
become high-impedance node. If vY was zero it will remain so. However,
if it was VDD, it will be maintained by the charge on parasitic C, for only
a time. Leakage current will slowly discharge C and vY will be lost.
10.5.1 Essential Design Requirement
Problem can be easily solved by establishing a lowresistance path for node Y that is activated when B
goes low (S1 opens).
In the improved design
here another switch S2
is controlled by B (bar)
And connects Y and ground.
10.5.2 Operation with NMOS as SW
Using NMOS as switch in PTL results in small area and
small node capacitances. These advantages are
achieved at the expense of serious shortcomings in both
static characteristics and the dynamic performance of the
circuit.
10.5.2 Operation with NMOS as SW
When vI goes to Zero
10.5.2 Operation with NMOS as SW
Technique to correct ‘Poor 1’
10.5.3 Use of CMOS Transmission
Gate as SW
10.5.3 Use of CMOS Transmission
Gate as SW
10.5.4 Pass-Transistor Logic Circuit
Examples
10.5.4 Pass-Transistor Logic Circuit
Examples
10.5.4 Pass-Transistor Logic Circuit
Examples