Abstract: Various new and non-obvious apparatus and methods for testing embedded memories in an integrated circuit are disclosed. One of the disclosed embodiments is an apparatus for testing an embedded memory in an integrated circuit. This exemplary embodiment comprises input logic that includes one or more memory-input paths coupled to respective memory inputs of the embedded memory, a memory built-in self-test (MBIST) controller, and at least one scan cell coupled between the input logic and the MBIST controller. The scan cell of this embodiment is selectively operable in a memory-test mode and a system mode. In memory-test mode, the scan cell can apply memory-test data to the memory inputs along the memory-input paths of the integrated circuit. Any of the disclosed apparatus can be designed, simulated, and/or verified (and any of the disclosed methods can be performed) in a computer-executed application, such as an electronic-design-automation (“EDA”) software tool.

Abstract: Embodiments of the disclosed technology comprise software-based techniques that can be used to improve scan chain test pattern generation and scan chain failure diagnosis resolution. For example, certain embodiments can be used to generate high quality chain diagnosis test patterns that are able to isolate a scan chain defect to a single scan cell. Such embodiments can be used to generate a “complete” test set—that is, a set of chain diagnosis test patterns that is able to isolate any scan chain defect in a faulty scan chain to a single scan cell.

Abstract: X-masking registers are added in front of a compactor in test data compression environment to remove unknown values. The X-masking registers block out some chains due to unknown values and select other chains to feed the compactor. This X-masking capability is used to select one scan cell to observe at a time after a failure is observed at the compactor output.

Abstract: Profiling-based scan chain diagnosis techniques are disclosed. With various implementations of the invention, unloading masking information for each of scan patterns is first determined. A tester then applies the scan patterns to a circuit under test and collects test response data according to the unloading masking information. A profiling-based analysis is performed to determine failing scan cell information based on the test response data.

Abstract: Disclosed are representative embodiments of methods, apparatus, and systems for partitioning-based Test Access Mechanisms (TAM). Test response data are captured by scan cells of a plurality scan chains in a circuit under test and are compared with test response data expected for a good CUT to generate check values. Based on the check values, partition pass/fail signals are generated by partitioning scheme generators. Each of the partitioning scheme generators is configured to generate one of the partition pass/fail signals for one of partitioning schemes. A partitioning scheme divides the scan cells into a set of non-overlapping partitions. Based on the partition pass/fail signals, a failure diagnosis process may be performed.

Abstract: A diagnosis technique to improve scan cell internal defect diagnostic resolution using scan cell internal fault models.

Abstract: Techniques are disclosed for reducing the set of initial candidates in signature based diagnosis methodology. These techniques are based on a unique way of making optimum use of information from logic back-cone tracing along with equations that describe the test response compactor.

Abstract: Various new and non-obvious apparatus and methods for testing embedded memories in an integrated circuit are disclosed. One of the disclosed embodiments is an apparatus for testing an embedded memory in an integrated circuit. This exemplary embodiment comprises input logic that includes one or more memory-input paths coupled to respective memory inputs of the embedded memory, a memory built-in self-test (MBIST) controller, and at least one scan cell coupled between the input logic and the MBIST controller. The scan cell of this embodiment is selectively operable in a memory-test mode and a system mode. In memory-test mode, the scan cell can apply memory-test data to the memory inputs along the memory-input paths of the integrated circuit. Any of the disclosed apparatus can be designed, simulated, and/or verified (and any of the disclosed methods can be performed) in a computer-executed application, such as an electronic-design-automation (“EDA”) software tool.

Abstract: Chain or logic diagnosis resolution can be enhanced in the presence of limited failure cycles using embodiments of the various methods, systems, and apparatus described herein. For example, pattern sets can be ordered according to a diagnosis coverage figure, which can be used to measure chain or logic diagnosability of the pattern set. Per-pin based diagnosis techniques can also be used to analyze limited failure data.

Abstract: Methods, apparatus, and systems for performing fault diagnosis are disclosed herein. In one exemplary embodiment, a failure log is received comprising entries indicative of compressed test responses to chain patterns and compressed test responses to scan patterns. A faulty scan chain in the circuit-under-test is identified based at least in part on one or more of the entries indicative of the compressed test responses to chain patterns. One or more faulty scan cell candidates in the faulty scan chain are identified based at least in part on one or more of the entries indicative of the compressed test responses to scan patterns. The one or more identified scan cell candidates can be reported. Computer-readable media comprising computer-executable instructions for causing a computer to perform any of the disclosed methods are also provided. Likewise, computer-readable media storing lists of fault candidates identified by any of the disclosed methods are also provided.

Abstract: Chain or logic diagnosis resolution can be enhanced in the presence of limited failure cycles using embodiments of the various methods, systems, and apparatus described herein. For example, pattern sets can be ordered according to a diagnosis coverage figure, which can be used to measure chain or logic diagnosability of the pattern set. Per-pin based diagnosis techniques can also be used to analyze limited failure data.

Abstract: A built-in self-test (BIST) circuit is disclosed that allows high fault coverage. Additionally, a method is disclosed for implementing the BIST circuit. In one aspect, the BIST circuit includes a plurality of scan chains that receive test patterns used in testing the integrated circuit. A pseudo random pattern generator provides test patterns to the scan chains. Weight select logic is positioned between the scan chains and the pseudo random pattern generator and controls the weightings of the test patterns that are loaded in the scan chains. In another aspect, the weight select logic can switch the weightings of the test patterns on a per-scan-cell basis. Thus, as the scan chains are loading, the weight select logic can effectively switch between test patterns being loaded into the scan chains.

Abstract: Scan chain diagnosis techniques are disclosed. Faulty scan chains are modeled and scan patterns are masked to filter out loading-caused failures. By simulating the masked scan patterns, failing probabilities are determined for cells on a faulty scan chain. One or more defective cells are identified based upon the failing probability information. A noise filtering system such as the one based upon adaptive feedback may be adopted for the identification process.

Abstract: Among the various embodiments described is a method of detecting defects in a cell of an integrated circuit that analyzes exercising conditions applied to an input of the cell during a capture phase of testing with failed test patterns that produce an indication of a fault and that analyzes the exercising conditions that are applied during a capture phase of testing with observable passing patterns that do not provide an indication of a fault. From the analysis, true failing excitation conditions and passing excitation conditions can be determined and used to identify whether a defect is in the cell or on an interconnect wire of the integrated circuit.

Abstract: Various new and non-obvious apparatus and methods for testing embedded memories in an integrated circuit are disclosed. One of the disclosed embodiments is an apparatus for testing an embedded memory in an integrated circuit. This exemplary embodiment comprises input logic that includes one or more memory-input paths coupled to respective memory inputs of the embedded memory, a memory built-in self-test (MBIST) controller, and at least one scan cell coupled between the input logic and the MBIST controller. The scan cell of this embodiment is selectively operable in a memory-test mode and a system mode. In memory-test mode, the scan cell can apply memory-test data to the memory inputs along the memory-input paths of the integrated circuit. Any of the disclosed apparatus can be designed, simulated, and/or verified (and any of the disclosed methods can be performed) in a computer-executed application, such as an electronic-design-automation (“EDA”) software tool.

Abstract: Methods of diagnostic test pattern generation for small delay defects are based on identification and activation of long paths passing through diagnosis suspects. The long paths are determined according to some criteria such as path delay values calculated with SDF (Standard Delay Format) timing information and the number of logic gates on a path. In some embodiments of the invention, the long paths are the longest paths passing through a diagnosis suspect and reaching a corresponding failing observation point selected from the failure log, and N longest paths are identified for each of such pairs.

Abstract: Disclosed herein are exemplary embodiments of a so-called “X-press” test response compactor. Certain embodiments of the disclosed compactor comprise an overdrive section and scan chain selection logic. Certain embodiments of the disclosed technology offer compaction ratios on the order of 1000×. Exemplary embodiments of the disclosed compactor can maintain about the same coverage and about the same diagnostic resolution as that of conventional scan-based test scenarios. Some embodiments of a scan chain selection scheme can significantly reduce or entirely eliminate unknown states occurring in test responses that enter the compactor. Also disclosed herein are embodiments of on-chip comparator circuits and methods for generating control circuitry for masking selection circuits.

Abstract: Technologies disclosed herein can be used to diagnose defects on die having both scan chain and system logic defects, including in situations where the presence of one or more faults in the system logic potentially obscures the detectability of one or more faults in the scan chains (or channels) and vice versa. At least some embodiments employ an iterative approach where at least some scan chain faults are identified, these chain faults are used to identify system logic faults, and then additional chain faults are identified using the system logic faults and vice versa. Failing bits can be partitioned into at least two groups: failing bits determined as being caused by system logic failures, and failing bits determined as being possibly caused by chain defects, system logic defects, or the compound effects of both types of defects.

Abstract: Speed-path debug techniques based on at-speed scan test patterns. Potential speed paths are identified based upon detected at-speed scan pattern failures and unknown X-value simulation. When the number of identified speed paths is large, the suspect speed paths are ranked.

Abstract: Methods, apparatus, and systems for performing fault diagnosis are disclosed herein. In one exemplary embodiment, a failure log is received including entries indicative of compressed test responses to chain patterns and compressed test responses to scan patterns. A faulty scan chain in the circuit-under-test is identified based at least in part on one or more of the entries indicative of the compressed test responses to chain patterns. One or more faulty scan cell candidates in the faulty scan chain are identified based at least in part on one or more of the entries indicative of the compressed test responses to scan patterns. The one or more identified scan cell candidates can be reported. Computer-readable media comprising computer-executable instructions for causing a computer to perform any of the disclosed methods are also provided. Likewise, computer-readable media storing lists of fault candidates identified by any of the disclosed methods are also provided.