Browse Prior Art Database

Computing the Probability of a Delta-I Failure

IP.com Disclosure Number: IPCOM000047136D
Original Publication Date: 1983-Sep-01
Included in the Prior Art Database: 2005-Feb-07
Document File: 2 page(s) / 55K

Publishing Venue

IBM

Related People

Brown, A: AUTHOR [+3]

Abstract

The Delta-I (WI) problem occurs when drivers with fast rise times and high currents switch simultaneously on a module with high inductance. Due to this excessive switching, a noise pulse on the power supply can propagate through quiet drivers to receivers. If the noise is sufficiently large to penetrate threshold, a latch on a chain of gates from the receiver may be falsely set. When designing chips, users must factor in the WI effect. However, in many situations, the user is unaware of conditions where WI exists since there does not exist a sufficiently powerful algorithm that can accurately predict the switching behavior of the logic. This publication discloses a new algorithm which determines the probability that a WI failure exists. Other known algorithms exist which find the input pattern which causes the failure.

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 52% of the total text.

Page 1 of 2

Computing the Probability of a Delta-I Failure

The Delta-I (WI) problem occurs when drivers with fast rise times and high currents switch simultaneously on a module with high inductance. Due to this excessive switching, a noise pulse on the power supply can propagate through quiet drivers to receivers. If the noise is sufficiently large to penetrate threshold, a latch on a chain of gates from the receiver may be falsely set. When designing chips, users must factor in the WI effect. However, in many situations, the user is unaware of conditions where WI exists since there does not exist a sufficiently powerful algorithm that can accurately predict the switching behavior of the logic. This publication discloses a new algorithm which determines the probability that a WI failure exists. Other known algorithms exist which find the input pattern which causes the failure. The algorithm disclosed herein makes use of the cutting algorithm which determines a lower and upper bound for a signal to be equal to one. First a proper WI model is constructed, and then known algorithms are used to solve for the lower and upper bounds which represent the probability of a Delta-I failure. If the lower bound is greater than zero, then there exists an input pattern which will cause many drivers to switch simultaneously. The model is constructed by The box "add number of 1's" is a tree of adders which adds the number of drivers switching. The box ">n" is a combinational network which repres...