Midterm 1 Problems Solution - Digital Integrated Circuit | EECS 312, Exams of Electrical and Electronics Engineering

Material Type: Exam; Class: Digit Integrat Circ; Subject: Electrical Engineering And Computer Science; University: University of Michigan - Ann Arbor; Term: Fall 2007;

Typology: Exams

Pre 2010

Uploaded on 09/02/2009

koofers-user-ryq-1
koofers-user-ryq-1 🇺🇸

10 documents

1 / 8

Toggle sidebar

This page cannot be seen from the preview

Don't miss anything!

bg1
EECS312
Fall 2007
Midterm 1 Solution
Problem 1
Note that capacitances and delays have nothing to do with VTC!
a) No effect.
b) VTC shifts right.
c) VTC shifts left.
d) Can never be the same since tpLH = 0.69RpC, tpHL = 0.69(Rp||Rn)C,
Rp would have to equal RpRn/(Rp+Rn).
Problem 2
a)
b)
pf3
pf4
pf5
pf8

Partial preview of the text

Download Midterm 1 Problems Solution - Digital Integrated Circuit | EECS 312 and more Exams Electrical and Electronics Engineering in PDF only on Docsity!

EECS

Fall 2007

Midterm 1 Solution

Problem 1 Note that capacitances and delays have nothing to do with VTC!

a) No effect. b) VTC shifts right. c) VTC shifts left. d) Can never be the same since tpLH = 0.69RpC, tpHL = 0.69(Rp||Rn)C, Rp would have to equal RpRn/(Rp+Rn).

Problem 2 a)

b)

c) No, only sub-threshold leakage.

I1 has no short circuit current since we assume the rise and fall times are zero. I2 has the following short circuit current:

  • Psc = t (^) sc,rise VDD I (^) peak f + t (^) sc,fall VDD I (^) peak f

t (^) sc,fall = VDD – 2VT tf VDD 0. = (2.5 – 0.8)/2.5 * (82.7e-12 / 0.8) = 70 ps

t (^) sc2,rise = VDD – 2VT tr VDD 0. = (2.5 – 0.8)/2.5 * (100e-12 / 0.8) = 85 ps

*Psc = 155e-12 * 2.5 * 100e-6 * 1.25e = 0.48 mW

tf = 2.2 Reqn C 1 = 1.21 * C 1 * VDD / I (^) DSAT,N = 1.21 * 10e-15 * (2.5 / (731 * 0.5 μA) ) = 82.7 ps

tr = 2.2 Reqp C 1 = 1.21 * C 1 * VDD / I (^) DSAT,P = 1.21 * 10e-15 * (2.5 / (300 * 1.0 μA) ) = 100 ps

b) (5 pts) What happens to my total power dissipation if I reduce Wp , Wn and t (^) ox of both inverters by a factor S (where S > 1) while keeping the same VDD and V (^) T? Be concise and clear in your answer; use equations when necessary.

For dynamic, if C’s don’t scale: Pdyn = (C 1 VDD^2 + C 2 VDD^2 ) f (no scaling)

If C’s scale: Pdyn = ((C 1 / S )VDD2 + (C 2 / S )VDD^2 ) f Pdyn/S

Pstat = 2 (I (^) static,n / S + I (^) static,p / S ) VDD = Pstat / S

Psc = t (^) sc,rise VDD I (^) peak f + t (^) sc,fall VDD I (^) peak f ∝ tf * Ipeak + tr * Ipeak ∝ (S*tf) (Ipeak/S) + (Str) *(Ipeak/S) ∝ 1 (no scaling)

The only trick here was to remember that IpeakI (^) SATW. This scaling situation might be the case when wiring capacitances dominate the gate caps (such that scaling Cg’s doesn’t really affect the power scaling).