摘要:
The present invention tolerates the decay of a dynamic logic circuit by preserving the logic state of the output before the decay. A slow clock detector is configured to detect a slow clock condition of the clock pertaining to the dynamic logic circuit. A tolerant storage device is configured to preserve the data output by command of the slow clock detector upon a detection of the slow clock condition.
摘要:
A clocking system and method are provided for logic blocks having cascaded self-timed dynamic logic gates. The dynamic logic gates are precharged in parallel and collectively perform self-timed logic evaluation on vector inputs to derive a vector output. An evaluation done detector monitors the output of the logic block and determines when the vector output is valid. An edge detector detects the rising and falling edges of an arbitrary periodic timing signal. Finally, a logic block clock generator is set by the edge detector and reset by the evaluation done detector so as to provide precharging signals to the logic block, thereby defining respective precharge periods, and to provide evaluation periods for the self-timed logic evaluations in the logic block. In a specific implementation, the speed of logic evaluations is twice the speed of the system clock.
摘要:
The operation of a pipeline is observed by launching two or more sets of data into the pipeline on consecutive clock cycles. The clock free-runs for as many cycles as it takes the data to propagate through the stages of the pipeline. The output latches of each stage of the pipeline are only sampled when the data of interest is held in each output latch, respectively. Observation may be completely controlled through a standard test access port (TAP). Observation may be accomplished by halting the clock to scan new data in and results out, or with the clock free-running. The inputs to the pipeline may come from test registers or from circuitry which feeds the pipeline during normal operation.
摘要:
A system and method for improving alpha-particle induced soft error rates in integrated circuits is provided. Logic isolation circuits implemented using a substantially fewer number of pn-junctions are situated at the outputs of fast logic portions containing a substantially greater number of pn-junctions. The present invention reduces the vulnerability of a dynamic logic circuit of incurring alpha soft errors by effectively trading the probability of an isolation circuit composed of only a few pn-junctions incurring alpha-particle strikes with the probability of a fast logic circuit having substantially more pn-junctions incurring alpha-particle strikes. By reducing the number of pn-junctions susceptible to alpha-particle strikes, the present invention significantly lowers the potential alpha-particle induced soft error rate. In one embodiment, isolation circuits used in the present invention are implemented using self-timed logic, to reduce the window in which a circuit is logically vulnerable to alpha strikes, in which a loss of state can occur.
摘要:
A system and method for quiescent current testing of dynamic logic circuitry. Nodes shorted to ground are detected during a dynamic pre-charge state. Nodes shorted to a power supply potential are detected by driving all nodes of interest to ground during a dynamic evaluation phase. Nodes of interest are driven to ground directly by one additional transistor per node or indirectly by logical propagation from upstream nodes. As a result, only two current measurements are needed for all shorted node faults, even for pipelined systems with multiple clocks. There is no need for input test signal sequences and no need for signal propagation to outputs for detection. Specific embodiments are provided for single-rail logic, single-rail pipelined systems, dual-rail logic and dual-rail pipelined systems. For single-rail pipelined systems, optional transistors between stages enable preservation of logical states during testing. For dual-rail logic, storage nodes and static nodes are forced to a logical state that is not possible during normal operation. For pipelined dual-rail logic, testing of alternate stages inherently preserves the logical state of the system during testing.
摘要:
A clock buffer circuit achieves insensitivity to the particular voltage levels and drift therein of input signals used to generate the clock, by use of a differential common gate amplifier incorporating an internally generated threshold voltage. Four separate gain paths couple the differential common gate amplifier to an output stage. Two of the gain paths are used to propagate edges that cause respective abrupt transitions in each direction for a first of two complementary clock signals. The other two accomplish the same for the other complementary clock signal. Each gain path is optimized to propagate a leading edge of a particular direction (relative to its point of origin, the direction of the edge inverts as it goes from stage to stage). Of these four gain paths, a first pair are used to create a high level of drive for the low to high transitions in the clock signal and its complement. Because of the optimization, this drive cannot be removed as abruptly as it can be applied. A latch-like circuit in the gain paths cause early removal of the high level, or hard, drive, leaving in place a maintenance, or holding, level of drive. Each holding drive is abruptly removed by an associated gain path in the remaining pair of gain paths. The hard drive is left in place only long enough to ensure that the capacitance of the clock line is adequately charged. An anti-glitch mechanism bullet proofs the entire circuit against drive fights caused by ambiguities arising from slow transitions that might arise from the differential common gate amplifier.