Browse Prior Art Database

Error Detection of Address Line Faults in Memory Array Usage

IP.com Disclosure Number: IPCOM000108385D
Original Publication Date: 1992-May-01
Included in the Prior Art Database: 2005-Mar-22
Document File: 2 page(s) / 70K

Publishing Venue

IBM

Related People

Zucker, AN: AUTHOR

Abstract

Disclosed is a scheme that can be used in any type of memory array where the entire array space is not functionally required. It is particularly applicable for use in arrays whose values are written only occasionally and then read frequently and rapidly (i.e., arrays defining system configurations).

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 62% of the total text.

Error Detection of Address Line Faults in Memory Array Usage

       Disclosed is a scheme that can be used in any type of
memory array where the entire array space is not functionally
required. It is particularly applicable for use in arrays whose
values are written only occasionally and then read frequently and
rapidly (i.e., arrays defining system configurations).

      To illustrate this algorithm, start with a 16-entry (4 address
lines) by 8-data bit memory array, as in Fig. 1.

      In order to detect errors in address lines feeding a memory
array, parity bit (which covers all address lines) is added to the
functional address to create the physical address into the array.  In
the example, address lines A0, A1, and A2 are used as functional
addresses, and A3 is generated as the odd parity of the first three.
This reduces the functional size of the array by 1/2 since, for every
valid address, an additional entry into the array is used to assure
odd parity.  In the example, the legal available addresses would be
1,2,4,7,8,11,13 and 14 (corresponding to functional addresses
0,1,2,3,4,5,6,7 and 8). Any of the address lines could be used as the
parity bit.

      The data within the memory array must, itself, be modified to
allow detection of invalid addresses.  This data would contain, in
addition to any data error detection scheme, a valid bit which would
be '1' for all valid addresses and '0' for all invalid addresses (in
the example, used data line...