摘要:
A method of reliability evaluation and system fail warning using on chip parametric monitors. The method includes determining impact of parametric variation on reliability by identifying key parametric questions to be answered by stress, identifying parametric macros for each parameter, and identifying layout sensitive areas of evaluation. The process can also include a set of parametric macros in one of a test site or a product to be stressed, testing the set of parametric macros prior to start of stress and at each stress read out, and setting life time parameter profile for technology.
摘要:
A method of reliability evaluation and system fail warning using on chip parametric monitors. The method includes determining impact of parametric variation on reliability by identifying key parametric questions to be answered by stress, identifying parametric macros for each parameter, and identifying layout sensitive areas of evaluation. The process can also include a set of parametric macros in one of a test site or a product to be stressed, testing the set of parametric macros prior to start of stress and at each stress read out, and setting life time parameter profile for technology.
摘要:
A method of reliability evaluation and system fail warning using on chip parametric monitors. The method includes determining impact of parametric variation on reliability by identifying key parametric questions to be answered by stress, identifying parametric macros for each parameter, and identifying layout sensitive areas of evaluation. The process can also include a set of parametric macros in one of a test site or a product to be stressed, testing the set of parametric macros prior to start of stress and at each stress read out, and setting life time parameter profile for technology.
摘要:
A method of reliability evaluation and system fail warning using on chip parametric monitors. The method includes determining impact of parametric variation on reliability by identifying key parametric questions to be answered by stress, identifying parametric macros for each parameter, and identifying layout sensitive areas of evaluation. The process can also include a set of parametric macros in one of a test site or a product to be stressed, testing the set of parametric macros prior to start of stress and at each stress read out, and setting life time parameter profile for technology.
摘要:
Disclosed are testing method embodiments in which, during post-manufacture testing, parametric measurements are taken from on-chip parametric measurement elements and used to optimize manufacturing in-line parametric control learning and/or to optimize product screening processes. Specifically, these post-manufacture parametric measurements can be used to disposition chips without shipping out non-conforming products, without discarding conforming products, and without requiring high cost functional tests. They can also be used to identify yield sensitivities to parametric variations from design and to provide feedback for manufacturing line improvements based on the yield sensitivities. Additionally, a historical database regarding the key parameters that are monitored at both the fabrication and post-fabrication levels can be used to predict future yield and, thereby, to preemptively improve the manufacturing line and/or also to update supply chain forecasts.
摘要:
Disclosed are testing method embodiments in which, during post-manufacture testing, parametric measurements are taken from on-chip parametric measurement elements and used to optimize manufacturing in-line parametric control learning and/or to optimize product screening processes. Specifically, these post-manufacture parametric measurements can be used to disposition chips without shipping out non-conforming products, without discarding conforming products, and without requiring high cost functional tests. They can also be used to identify yield sensitivities to parametric variations from design and to provide feedback for manufacturing line improvements based on the yield sensitivities. Additionally, a historical database regarding the key parameters that are monitored at both the fabrication and post-fabrication levels can be used to predict future yield and, thereby, to preemptively improve the manufacturing line and/or also to update supply chain forecasts.
摘要:
A plurality of digital circuits are manufactured from an identical circuit design. A power controller is operatively connect to the digital circuits, and a non-volatile storage medium is operatively connected to the power controller. The digital circuits are classified into different voltage bins, and each of the voltage bins has a current leakage limit. Each of the digital circuits has been previously tested to operate within a corresponding current leakage limit of a corresponding voltage bin into which each of the digital circuits has been classified. The non-volatile storage medium stores boundaries of the voltage bins as speed-binning test data. The power controller controls power-supply signals applied differently for each of the digital circuits based on which bin each of the digital circuit has been classified and the speed-binning test data.
摘要:
A plurality of digital circuits are manufactured from an identical circuit design. A power controller is operatively connect to the digital circuits, and a non-volatile storage medium is operatively connected to the power controller. The digital circuits are classified into different voltage bins, and each of the voltage bins has a current leakage limit. Each of the digital circuits has been previously tested to operate within a corresponding current leakage limit of a corresponding voltage bin into which each of the digital circuits has been classified. The non-volatile storage medium stores boundaries of the voltage bins as speed-binning test data. The power controller controls power-supply signals applied differently for each of the digital circuits based on which bin each of the digital circuit has been classified and the speed-binning test data.
摘要:
Methods of modeling across reticle variations and a related reticle are disclosed. One embodiment of the method includes defining a test for determination across a multiple chip wafer; identifying a measurement structure for performing the test; implementing the measurement structure on the multiple chip wafer using a reticle including the measurement structure between copies of the multiple chips on the reticle, wherein no one of the multiple chips covers an entirety of the reticle; performing the test on the multiple chip wafer using the measurement structure to acquire data across the reticle; using data from the performing to establish an across reticle variation model; and using the across reticle variation model to predict across chip variation for at least one of the multiple chips.
摘要:
Methods of modeling across reticle variations and a related reticle are disclosed. One embodiment of the method includes defining a test for determination across a multiple chip wafer; identifying a measurement structure for performing the test; implementing the measurement structure on the multiple chip wafer using a reticle including the measurement structure between copies of the multiple chips on the reticle, wherein no one of the multiple chips covers an entirety of the reticle; performing the test on the multiple chip wafer using the measurement structure to acquire data across the reticle; using data from the performing to establish an across reticle variation model; and using the across reticle variation model to predict across chip variation for at least one of the multiple chips.