Browse Prior Art Database

Substrate Programming of Partial Good Modules

IP.com Disclosure Number: IPCOM000080646D
Original Publication Date: 1974-Jan-01
Included in the Prior Art Database: 2005-Feb-27
Document File: 3 page(s) / 23K

Publishing Venue

IBM

Related People

Fredricks, RE: AUTHOR [+3]

Abstract

Described is a method of fabricating modules so that no matter which segment is defective, all modules can be made to appear to have the same segment bad. In this way, only one module and card is required regardless of the segment size. Further, in the memory card design, using this method minimizes the total number of substrates required.

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 3

Substrate Programming of Partial Good Modules

Described is a method of fabricating modules so that no matter which segment is defective, all modules can be made to appear to have the same segment bad. In this way, only one module and card is required regardless of the segment size. Further, in the memory card design, using this method minimizes the total number of substrates required.

Defective segments are locatable by a fixed subset of the total chip address field. In any partial good program, three bits are required to identify one of possibly eight bad octants. By making the address of each of the failing octants look like one normalized address, only one module and card are required to package the partial good product. This normalization is accomplished by providing current-switching circuits as input powering for the array chip.

If the signal is brought into the base of transistor T1 with a reference voltage brought into the base of transistor T2, the collector of transistor T2 is out-of- phase while that of transistor T2 is in phase. However, if the input was exchanged with the reference voltage so that it appeared on the base of transistor T2 and the Vref were imposed upon the base of transistor T2, the outputs would reverse, effectively complementing the input. This switching of addresses is accomplished by providing individual chip pads for both inputs to the circuits powering the addresses which select the defective octant. Module substrates are wired to provide all possible wiring input combinations to the chips.

The following would be the result. If octant 000 were defective, it would be packaged in the same module as all good chips and at the module level octant 000 would, of course, appear to be failing. If any other octant failed, for example, 101, the chip could be packaged on a substrate which reversed the inputs to the high and low-order bits which select octants. This means that when 101 was addressed, the address distributed on the chip would be 000, a good octant, but when 000 were addressed, 101 would be distributed on the chip. Thus it would appear that 000 were the failing octant and not quadrant 101. In this manner, all failing octants can be made t...