EE 5323 - VLSI Design I
Fall Semester, 2002
Homework Assignment 3
Due: Wednesday, November 6
-
A finite state machine (FSM) has two inputs, X and Y, one output, Z,
and two state variables, A and B. The next-state and output equations
for this FSM are as follows:
Anext = X'A + Y'B , Bnext = XA' + YB , Z = XA' + YB'
(' indicates Boolean complement.)
Use two-phase clocking (i.e., where phase 2 is the complement of phase 1)
and assume that the input signals are latched on phase 1. Draw the
transistor-level circuit diagram for this FSM using a domino CMOS PLA
(without any evaluate transistors in the OR-plane) to generate the
next-state and output functions. Use the drawing style for the PLA that
was discussed in class. Include dynamic D-type latches (i.e., built using
a transmission gate and an inverter) for the input, output and state
signals. You may use macro symbols, but be sure to completely specify the
transistor-level circuitry within each such symbol. You do not have
to specify the widths of the transistors in this circuit.
-
A proposed edge-triggered D-type flip-flop is shown in
Figure 1. (The weak inverter has a weak PMOS
and a weak NMOS.)
(a) Is this a rising-edge triggered or falling-edge triggered flip-flop?
(b) Is Z the true or the complement Q output?
(c) For the 4 possible combinations (current value of Z, next D to be stored),
state what will happen at node X and node Y during the evaluate phase.
(d) Dynamic power dissipation is due to
the charging and discharging of capacitive nodes. Based on your
answer to part (c), briefly explain why this circuit might be useful for
low-power applications.
-
Consider the function f = BC + XOR(A, C), where XOR is the exclusive-OR
function. Assume that true and complement forms of A, B and C are
available as inputs.
(a) Write the nested if-then-else logic statements corresponding
to this function and its complement. Let A be the outer (i.e., the first)
binary decision block, B be the middle binary decision block and C
be the inner binary decision block. Indicate which, if any, of these binary
decision blocks may be shared or eliminated, and give the minimum
number of A-blocks, B-blocks and C-blocks that are required.
(b) Draw the dynamic DCVS (i.e., dual-rail domino) circuit corresponding
to the if-then-else description of part (a), using the minimum number
of NMOS transistors in the merged pull-down network. Be sure that your
circuit is robust against charge sharing and noise. You do not have
to specify transistor widths.
-
In this problem, you will
perform circuit simulations for the circuit of
Figure 2.
The dynamic DCVS full adder is the design discussed in class,
the NOR gates are static CMOS NOR gates, the C-element is
the circuit in Figure 1 of Homework 1 and the inverters are
static CMOS inverters.
Assume that
the true and complement forms of the input signals
x1, x2 and x3 can switch
to new values at the midpoint of each precharge
phase, and let the values for
x1, x2 and x3
during 8 successive clock cycles be
as follows: (0, 0, 0), (0, 0, 1), (0, 1, 0), (0, 1, 1),
(1, 0, 0), (1, 0, 1), (1, 1, 0), (1, 1, 1).
(The complement forms of the inputs have values opposite to these.)
Use a 50% duty cycle clock with a clock period that
is long enough
to allow the output f3 to be computed during the evaluate phase.
Use the BSIM 3 models (Star-HSPICE level 49)
that are available from the MOSIS web site. The fabrication technology
is a 0.5 micron CMOS process called
HP AMOS14TB,
and you will find NMOS and PMOS device
model parameters for several fabrication runs at the following web page:
http://www.mosis.org/Technical/Testdata/hp-amos14tb-prm.html.
The most recent fabrication run is listed first, and the others are
listed in reverse-chronological order. Select the parameters from
any of the 10 most recent runs. (Note: Occasionally, the parameters
in these files give strange results. If you find this
to be the case, then use the parameters from another file.)
Within any of these files, cut and paste the lines starting with
".MODEL CMOSN NMOS (" so that you will have the NMOS and PMOS device models
for that fabrication run.
Specify L = 0.6U for all NMOS and PMOS transistors. (Note: The
transistor model parameter XL will subtract 0.1 micron from this value so
that the actual channel length used for the simulations will
be the design rule minimum value of 0.5 micron.) For simplicity,
use the following width, area and perimeter values:
For all NMOS transistors in this circuit,
use W = 10.8U, AS = AD = 19.44P, PS = PD = 14.4U.
For all PMOS transistors in this circuit (except for keepers),
use W = 43.2U, AS = AD = 77.76P, PS = PD = 46.8U.
For all PMOS keepers in this circuit,
use W = 0.9U, AS = AD = 1.62P, PS = PD = 4.5U.
Simulate your circuit under nominal conditions,
i.e.
with a VDD of 3.3 volts and at
the default simulation temperature.
VERY IMPORTANT:
Hand-annotate your simulation results by tracing over key waveforms
with colored pens and adding comments explaining what the waveforms
mean. In particular, show that you obtain the correct results for
the sum, carry out, f1, f2 and f3 functions
during the 8 consecutive clock cycles.