DOC PREVIEW
Berkeley ELENG 42 - Solution to Problem Set 10

This preview shows page 1 out of 3 pages.

Save
View full document
View full document
Premium Document
Do you want full access? Go Premium and unlock all 3 pages.
Access to all documents
Download any document
Ad free experience
Premium Document
Do you want full access? Go Premium and unlock all 3 pages.
Access to all documents
Download any document
Ad free experience

Unformatted text preview:

04/27/03 Solution to Problem Set # 10 (by Farinaz Koushanfar) 10.1 a) Gate 1: If (E1=0) Æ pull-down: ((A1.B1) + (C1.D1))’ pull-up: (C1’+D1’).(A1’+B1’) If (E1=1) Æ pull-down: ((B1+D1). (A1+C1))’ pull-up: B1’.D1’+A1’.C1’ Æ G1 = E1’. (C1’+D1’).(A1’+B1’)+E1. (B1’.D1’+A1’.C1’) (after simplification)Æ G1 = A1’.C1’+B1’.D1’+E1’. (B1’.C1’+A1’.D1’) Gate 2: If (D2=0) Æ pull-down: (A2.B2)’ pull-up: A2’+B2’ If (D2=1) Æ pull-down: (A2. (B2+C2))’ pull-up: A2’+(B2’.C2’) Æ G2 = D2’. (A2’+B2’) + D2. (A2’+ (B2’.C2’)) (after simplification) Æ G2 = A2’ + B2’. (C2’+D2’) b) As can be seen in the answer to part a), the functions of the pull-up and pull-down parts are complimentary, which makes sure that the pull-up and pull-down paths are not simultaneously on for any given input. Also, the pull-up circuit is all PMOS, and the pull-down is all NMOS. The complimentary feature also says that the pull-up and pull-down oaths are not “on” at the same time, so in the steady state, there is no current flowing from the source to the ground (the output is just connected to the source or to the ground). c) For going from high to low in the output, the pull-down circuit should discharge the current. The pull-down delay of an inverter is: τHL(inv) = 0.69RDC. Gate1: The minimum pull-down time occurs, when (A1=B1=C1=D1=1), Req=2RD||2RD = RD The max pull-down is when E1=1 and we have a path starting at one branch, and ending at the other, for e.g. (A1=E1=C1=1), Req=3RD Thus, the range 1≤ τHL(gate1)/ τHL(inv) ≤3. Gate2: The minimum pull-down time occurs, when (A2=B2=C2=D2=1), Req=5/3RD The max pull-down is when (A2=C2=D2=1, B2=0), Req=3RD Thus, the range 1.66 ≤ τHL(gate2)/ τHL(inv) ≤ 3. d) The pull-up delay of an inverter is: τLH(inv) = 0.69RUC Gate1: The min pull-up happens when both pull-up branches are on (A1=B1=C1=D1=0), Req=2RU||2RU = RU The max pull-up is when there is only on path and that path passes E1 in the pull-up circuit, e.g. (A1=E1=D1=0, B1=C1=1). Req=3RU. Thus, the range is 1≤ τLH(gate1)/ τLH(inv) ≤3. Gate2: The minimum pull-up time occurs, when (A2=B2=C2=D2=0), Req= 0.6 RU The max pull-up is when (B2=C2=0, A2=D2=1), Req=2RU Thus, the range 0.6 ≤ τHL(gate2)/ τHL(inv) ≤ 2. e) The overall worst-case delay for Gate1 is the max(τLH(gate1),τHL(gate1)) = 3τLH(inv)= 3τHL(inv) The overall worst-case delay for Gate2 is the max(τLH(gate2),τHL(gate2)) = 3τLH(inv). 10.2 a) The best (or min) delay happens when the output of gate 2 is independent of the values of A2 (A2 is connected to output of gate 1) that happens when the circuit is pulled up with the other. So, if we assume that D2=C2=1 for a long time, and then the input B2 changes from 1 to 0, the delay for pulling up the output of the gate 2 becomes: τLH(gate2) for (B2=C2=D2=0), regardless of A2, which is 1.5τLH(inv). b) The longest propagation time is: max( delay)/(inv delay)) = Max ( τHL(gate1)/ τHL(inv) + τLH(gate2)/ τLH(inv) ) = 3+3=6 (inverter delays) Assume that (A1=B1=C1=D1=E1=0), then the inputs A1 and B1 change from 0 to 1, causing the worst case HL transition at the output of gate1 (3 inverter delays). Assume that the inputs (B2=D2=C2=1), then EECS 42 –Introduction to Electronics for Computer Science Spring 2003, Prof. A. R. NeureutherDept. EECS, 510 Cory [email protected] 642-4590 UC Berkeley Office Hours (Tentative M, Tu, W, (Th), F 11Course Web Site http://www-inst.eecs.berkeley.edu/~ee42/04/27/03 the change in A2 (that is connected to the output of gate1 and goes to zero), will cause the worst case LH transition at the output of gate 2 (3 inverters delay). 10.3 a) Each of the latches, consist of two stages. The first stage of the latch is transparent when the clock is low (Φ=0) and has the delay of (0.69×2×RDC) or (0.69×2×RUC) depending on if it is going from H to L or vice versa. The second stage of the latch is transparent when clock goes high, and has the delay of (0.69×2×RUC) or (0.69×2×RDC) depending on the transition. Overall, since the outputs of the two stages are compliment, the overall delay of a latch is: (0.69×2×RUC)+(0.69×2×RDC) = 0.69×2×C(RU+RD), which is 4 inverter delays. b) c) 10.4 a) Once the clock goes from low to high, - Time for the output of L0 to change = 2τ(inv) - Worst case (max) for the output of gate 1 to change = 3τ(inv) - Worst case (max) for the output of gate 2 to change = 3τ(inv) When clock goes from high to low, 2 3 1.66 2 2Φ B1 Vout1 Vout(L1) Vout2 Vout(L2) 2 3 1.66 2 2ΦB1 A2 Vout2 VmidVout(L1)04/27/03 - Time for transferring the data in the mid-point of the latch = 2τ(inv) Total (low to high) = 8 τ (inv), total (high to low) = 2τ(inv) “Note that the total delay is dependent on the clock cycle, for all the cases we have here!” b) For the first pipeline stage, once the clock goes from low to high, - Time for the output of L0 to change = 2τ(inv) - Worst case (max) for the output of gate 1 to change = 3τ(inv) - Time for transferring the data in the mid-point of the latch L1= 2τ(inv) Total (low to high) = 5 τ (inv), total (high to low) = 2τ(inv) For the second pipeline stage, once the clock goes from low to high, - Time for the output of L1 to change = 2τ(inv) - Worst case (max) for the output of gate 2 to change = 3τ(inv) - Time for transferring the data in the mid-point of the latch L2= 2τ(inv) Total (low to high) = 5 (inv), total (high to low) = 2τ(inv) c) The latency for the overall logic evaluation, (assuming the clock has τH(clock) and τL(clock)) For lumped logic = τH(clock) + τL(clock) = τH(clock) + τL(clock) For pipelined logic = 2 (τH(clock) + τL(clock)) d) The lumped circuit (using the worst case calculations from 10.4(a)): min {τH(clock)} = 8 τ (inv) min {τL(clock)} = 2 τ (inv) The pipelined circuit (using the worst case calculations from 10.4(a)): min {τH(clock)} = 5 τ (inv) min {τL(clock)} = 2 τ (inv) Assuming minimum clock cycles, the overall delay of the G1 and G2 (for the output to be valid at the midpoint of last latch) is more with the pipelined circuit: 2 (τH(clock) + τL(clock)) = 14τ(inv) as opposed to the lumped circuit: τH(clock) + τL(clock) = 10τ(inv). However, if we just assume the delay of one cycle of the pipelined implementation, it has the minimum τH(clock) +


View Full Document

Berkeley ELENG 42 - Solution to Problem Set 10

Documents in this Course
Lecture 1

Lecture 1

25 pages

Lecture 2

Lecture 2

20 pages

Lecture 3

Lecture 3

21 pages

Midterm 1

Midterm 1

20 pages

Load more
Download Solution to Problem Set 10
Our administrator received your request to download this document. We will send you the file to your email shortly.
Loading Unlocking...
Login

Join to view Solution to Problem Set 10 and access 3M+ class-specific study document.

or
We will never post anything without your permission.
Don't have an account?
Sign Up

Join to view Solution to Problem Set 10 2 2 and access 3M+ class-specific study document.

or

By creating an account you agree to our Privacy Policy and Terms Of Use

Already a member?