Dismiss
InnovationQ will be updated on Sunday, Oct. 22, from 10am ET - noon. You may experience brief service interruptions during that time.
Browse Prior Art Database

Dynamic Random-Access Memory Error Flag

IP.com Disclosure Number: IPCOM000037952D
Original Publication Date: 1989-Sep-01
Included in the Prior Art Database: 2005-Jan-31
Document File: 1 page(s) / 12K

Publishing Venue

IBM

Related People

Pontius, DE: AUTHOR [+3]

Abstract

A signal (the "error flag") is generated from internal control signals of a dynamic random access memory chip to determine if a major fault has occurred. This signal can be used by an error correction code (ECC) scheme to enhance capability to locate the failing bit(s) within the ECC word and to correct the fails.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 70% of the total text.

Page 1 of 1

Dynamic Random-Access Memory Error Flag

A signal (the "error flag") is generated from internal control signals of a dynamic random access memory chip to determine if a major fault has occurred. This signal can be used by an error correction code (ECC) scheme to enhance capability to locate the failing bit(s) within the ECC word and to correct the fails.

An error flag is generated from a combination of the major timing clock signals, an 'exclusive or' of the word lines, and an 'and' of any critical clocks and signals used in support circuits across the chip. Varying levels of effectiveness are achieved by the number of signal lines in the error flag logic. A simple scheme of checking that the last major clock has fired flags approximately 80% of major chip faults. A sophisticated error flag that checks for one and only one word line to have fired and for other critical signals could catch 99% of major chip faults. If the error flag is not received by the system when access is expected, it is virtually certain that the chip has a major fault, causing all output bits to be suspect.

The error flag can also be timed with the expected access time of the chip. The strobe can then be used as a data latch signal allowing card timing to be optimized to chip performance.

By using the error flag, the ECC logic can locate chip kill fails in the ECC word about 99% of the time since a failing chip flags its error condition to the ECC logic. In the case of a double bit error, on...