International Business Machines Corporation

Steven Michael Eustis

Michael Richard Ouellette

Peter Francis Croce

Robert L Barry

Endre P Thoma

Yabin Wang

Erik L Hedberg

Kevin William Gorman

David E Lackey

Dale E Pontius

Michael Thomas Fragano

John B DeForge

Jeremy Paul Rowland

David E Douse

Susan M Litten

Steven Frederick Oakland

Krishnendu Mondal

Thomas Gregory Sopchak

Ciaran J Brennan

Jeffery Howard Oppold

Ronald A Piro

Thomas R Bednar

David J Wager

Leah Marie Pfeifer Pastel

Neelesh Govindaraya Pai

Cheryl Jean Herdey

James A Monzel

Kevin M Tompsett

Eric Stephen Machat

A method is provided for operating an interface between a first unit and a second unit supplying its data. The method includes switching control between LSSD_B and LSSD_C clocks and system clock (CLK) to provide a test mode of operation and a functional mode of operation to optimize setup and hold times depending on conditions under which the unit is operating. In the test mode, data is launched by the LSSD_C clock. In the functional mode, the data is launched by the system clock (CLK) to RAM. A method is also provided to determine which memory inputs should use a circuit that provides adequate setup and hold margins.

Method and apparatus for a robust embedded interface

A design structure is embodied in a machine readable medium for designing, manufacturing, or testing an integrated circuit. The design structure includes an input register coupled to a data processing unit input and a test operation mode and functional operation mode. In the test mode operation, the register operates in a clocked mode such that, during the test operation mode, the register propagates data to the data processing unit in response to a clock signal. In the functional operation mode, the register operates in a data flush mode such that the register propagates data to the data processing unit in response to the data. The functional mode is enabled by a flush enable signal and the test mode is enabled by an opposite state of the flush enable signal.

Design structure and apparatus for a robust embedded interface

A memory structure configured for supporting multiple test methodologies includes a first plurality of multiplexers configured for selectively coupling at least one data input path and at least one address path between an external customer connection and a corresponding internal memory connection associated therewith. A second multiplexer is configured for selectively coupling an input of a test latch between a functional memory array connection and a memory logic connection coupled to the at least one data input path, with an output of the test latch defining a data out customer connection. Flush logic is further configured to direct data from the memory logic connection to the data out customer connection during a test of logic associated with a customer chip, thereby facilitating observation of the memory logic connection at the customer chip, wherein test elements of the memory structure comprise a scan architecture of a first type, and test elements of the customer chip comprise a scan architecture of a second type.

Compilable memory structure and test methodology for both ASIC and foundry test environments

A method of manufacturing a device having embedded memory including a plurality of memory cells. During manufacturing test, a first test stress is applied to selected cells of the plurality of memory cells with a built-in self test. At least one weak memory cell is identified. The at least one weak memory cell is repaired. A second test stress is applied to the selected cells and the repaired cells with the built-in self test.

Method and apparatus for test and repair of marginally functional SRAM cells

A method and system for locating connector defects in a defective scan chain that has a parallel non-defective scan chain on a different wiring level, with both scan chains being laid out in a regular array pattern. A predetermined bit sequence is scanned into the defective scan chain. The contents of the defective scan chain are then parallel shifted into the non-defective scan chain. The contents of the non-defective scan chain is then scanned out and compared with the predetermined bit sequence. The comparison of the scanned out bits with the predetermined bit sequence facilitates locating both physically and logically where a connector defect has occurred in the defective scan chain.

Diagnosable scan chain

A method of assigning bits to redundant regions for variable bit redundancy region boundaries in a compliable memory such as a 1-port SRAM is provided. Methods include allocating bits between the redundant regions in nearly equal proportions while minimizing the amount of chip real estate consumed by the memory. Methods also includes allocating bits in equal portions between redundant regions while occupying slightly more memory chip real estate. Methods also allocate bits into redundant regions with a simplified procedure which may or may not allocate bits into the redundant regions in equal proportions. All of the methods allow the total number of memory bits in the complied memory to be re-defined while maintaining the same allocation characteristics for each method. Accordingly, the methods allow efficient use of redundant memory bits while also conserving chip real estate or offering simplified allocation steps.

Variable column redundancy region boundaries in SRAM

Self-test architectures are provided to implement data column and row redundancy with a totally integrated self-test and repair capability in a Random Access Memory (RAM), either a Dynamic RAM (DRAM) or a Static Ram (SRAM), and are particularly applicable to compileable memories and to embedded RAM within microprocessor or logic chips. The invention uses two passes of self-test of a memory. The first pass of self-test determines the worst failing column, the column with the largest number of unique failing row addresses. After completion of the first pass of self-test, the spare column is allocated to replace the worst failing column. In the second pass of self-test, the BIST (Built In Self-Test) collects unique failing row addresses as it does today for memories with spare rows only. At the completion of the second pass of self-test, the spare rows are then allocated. Once the second pass of self-test is completed, the column and unique failing row addresses are transported to the e-fuse macros and permanently stored in the chip.

Self-test architecture to implement data column redundancy in a RAM

A method and structure for a read only memory (ROM) cell array has the first drain of a first transistor connected to a true bitline and a second drain of a second transistor connected to a complement bitline. The first transistor also includes a first source, and the second transistor includes a second source. The connection of the first source or the second source to ground programs the ROM cell. With the invention, only the first source or the second source is connected to the ground and the other is insulated from electrical connections. Further, the connection of the source to ground comprises an electrical connection formed during manufacturing of the first transistor and the second transistor.

Complementary two transistor ROM cell

A method of designing an integrated circuit chip includes preparing a first macro to have a first power consumption rate, preparing a second macro to have a second power consumption rate different than the first power consumption rate, designing the circuit, measuring performance characteristics of the circuit, and substituting the second macro for the first macro to improve the performance characteristics. The first macro and the second macro have the same function, devices, surface area size, external wiring pattern, and timing characteristics.

Substituting high performance and low power macros in integrated circuit chips

A memory circuitry is designed to efficiently obtain a predictable array output when an invalid address is requested. The memory circuit comprises an invalid word line path in addition to the standard valid word line path. In order to provide correct output, a dummy word line output of a first decode logic is delayed and the delayed dummy word line output is ANDed with a word line output to update the data out latch. Further, the invalid word line output of a second decode logic is also delayed, and the delayed invalid word line output is ORed with the delayed dummy word line output to reset the control logic. ORing the delayed signals allows the predictable output to be provided at a same clock time, irrespective of whether a valid address or an invalid address is decoded.

Memory circuitry with auxiliary word line to obtain predictable array output when an invalid address is requested

A method and structure for a pair of read only memory (ROM) cells having a first latch and a second latch connected to the first latch. The first latch and the second latch behave as master and slave latches to one another. The first latch and the second latch include a write bitline connection that is permanently connected to a fixed voltage source to permanently program the first latch and the second latch to permanent ROM values.

Compilable writeable read only memory (ROM) built with register arrays

Electrical mask identification of memory modules

A Random Access Memory including a redundancy scheme wherein redundant memory elements are organized in a mixture of redundancy patches of various sizes, i.e., various number of word/ bit lines in each patch. The number of lines, e.g., 1, 2, 4 or 8 word or bit lines, in each of the patches is selected as appropriate with many different sized patches existing within the same redundancy reservoir. The size the particular patch selected depends on the size of the replaced defect detected during programming.

Growing community of inventors

Essex Junction, VT, United States of America

Steven Michael Eustis