摘要:
A test structure for characterizing a production static random access memory (SRAM) array. The test structure includes a characterization circuit having multiple memory cell columns connected in series to form a ring configuration. The characterization circuit is fabricated on a wafer substrate in common with and proximate to a production SRAM array. The characterization circuit preferably includes SRAM cells having a circuit topology substantially identical to the circuit topology of memory cells within the production SRAM array. In one embodiment, the test structure is utilized for characterizing a multi-port memory array and includes multiple memory cell columns connected in series to form a ring oscillator characterization circuit. Each cell column in the characterization circuit includes multiple SRAM cells each having a latching node and multiple data path access nodes. Selection control circuitry selectively enables the multiple data path access nodes for the SRAM cells within the characterization circuit.
摘要:
A test structure for characterizing a production static random access memory (SRAM) array. The test structure includes a characterization circuit having multiple memory cell columns connected in series to form a ring configuration. The characterization circuit is fabricated on a wafer substrate in common with and proximate to a production SRAM array. The characterization circuit preferably includes SRAM cells having a circuit topology substantially identical to the circuit topology of memory cells within the production SRAM array. In one embodiment, the test structure is utilized for characterizing a multi-port memory array and includes multiple memory cell columns connected in series to form a ring oscillator characterization circuit. Each cell column in the characterization circuit includes multiple SRAM cells each having a latching node and multiple data path access nodes. Selection control circuitry selectively enables the multiple data path access nodes for the SRAM cells within the characterization circuit.
摘要:
A distributed charge pump system uses a delay element and frequency dividers to generate out of phase pump clock signals that drive different charge pumps, to offset peak current clock edges for each charge pump and thereby reduce overall peak power. Clock signal division and phase offset may be extended to multiple levels for further smoothing of the pump clock signal transitions. A dual frequency divider may be used which receives the clock signal and its complement, and generates two divided signals that are 90° out of phase. In an illustrative embodiment the clock generator comprises a variable-frequency clock source, and a voltage regulator senses an output voltage of the charge pumps, generates a reference voltage based on a currently selected frequency of the variable-frequency clock source, and temporarily disables the charge pumps (by turning off local pump clocks) when the output voltage is greater than the reference voltage.
摘要:
A distributed charge pump system uses a delay element and frequency dividers to generate out of phase pump clock signals that drive different charge pumps, to offset peak current clock edges for each charge pump and thereby reduce overall peak power. Clock signal division and phase offset may be extended to multiple levels for further smoothing of the pump clock signal transitions. A dual frequency divider may be used which receives the clock signal and its complement, and generates two divided signals that are 90° out of phase. In an illustrative embodiment the clock generator comprises a variable-frequency clock source, and a voltage regulator senses an output voltage of the charge pumps, generates a reference voltage based on a currently selected frequency of the variable-frequency clock source, and temporarily disables the charge pumps (by turning off local pump clocks) when the output voltage is greater than the reference voltage.
摘要:
Embedded dynamic random access memory (eDRAM) sense amplifier circuitry in which a bit line connected to each of a first plurality of eDRAM cells is controlled by cell control lines tied to each of the cells. During a READ operation the eDRAM cell releases its charge indicating its digital state. The digital charge propagates through the eDRAM sense amplifier circuitry to a mid-rail amplifier inverter circuit which amplifies the charge and provides it to a latch circuit. The latch circuit, in turn, inverts the charge to correctly represent at its output the logical value stored in the eDRAM cell being read, and returns the charge through the eDRAM sense amplifier circuitry to replenish the eDRAM cell.
摘要:
A distributed charge pump system uses a delay element and frequency dividers to generate out of phase pump clock signals that drive different charge pumps, to offset peak current clock edges for each charge pump and thereby reduce overall peak power. Clock signal division and phase offset may be extended to multiple levels for further smoothing of the pump clock signal transitions. A dual frequency divider may be used which receives the clock signal and its complement, and generates two divided signals that are 90° out of phase. In an illustrative embodiment the clock generator comprises a variable-frequency clock source, and a voltage regulator senses an output voltage of the charge pumps, generates a reference voltage based on a currently selected frequency of the variable-frequency clock source, and temporarily disables the charge pumps (by turning off local pump clocks) when the output voltage is greater than the reference voltage.
摘要:
A wordline-to-bitline timing ring oscillator circuit for evaluating storage cell access time provides data on internal bitline access timing, and in particular the total wordline select-to-bitline read output timing. Columns of a storage array are connected in a ring, forming a ring oscillator. The bitline read circuit output of each column is connected to a wordline select input of a next column, with a net inversion around the ring, so that a ring oscillator is formed. The period of oscillation of the ring oscillator is determined by the total wordline select-to-bitline read circuit output timing for a first phase and the pre-charge interval time for the other phase, with the bitline read timing dominating. The circuit may be applied both to small-signal storage arrays, with the sense amplifier timing included within the ring oscillator period, or to large-signal storage arrays, with the read evaluate circuit timing included.
摘要:
A circular edge detector on an integrated circuit including a plurality of edge detector cells, each of the plurality of edge detector cells having an input select block operable to receive a data signal and a previous cell signal and to generate a present cell signal, and a state capture block operably connected to receive the present cell signal. The present cell signal of each of the plurality of edge detector cells is provided to a next of the plurality of edge detector cells as the previous cell signal for the next of the plurality of edge detector cells, and the present cell signal from a last edge detector cell is provided to a first edge detector cell as the previous cell signal for the first edge detector cell.
摘要:
Power-gated circuitry is put in a “sleep mode” that selectively gates both the power supply rails for static power control and the clock distribution for dynamic power control. A time interval M is established following a wake-up signal that includes the time to power-up, perform a computation, and return a result to the following circuitry. Likewise, a time interval N is established that indicates how long to wait after a result is returned before the power-gated circuitry is returned to the sleep mode to assure a desired performance. When a power-gated circuit is going to be needed for a future computation, it is issued a wake-up signal and a predetermined estimated time K for receipt of a next wake-up signal. A decision is made by analyzing the times M, N, and K as to when to return a power-gated circuit to the sleep mode following activation by a wake-up signal.
摘要:
In an exemplary embodiment of the present invention, a local clock buffer (LCB) fabricated in a semiconductor receives a global clock signal as input. The LCB implements a pulse width controller that is operationally coupled to the LCB and an output driver forming a ring oscillator. The output driver outputs a pulse width adjusted signal. The pulse width of the pulse width adjusted signal is adjustable by way of the pulse width controller and is related in frequency to the global clock signal. A second ring oscillator (also referred to as the nclk loop) can also be implemented to server as the global clock signal. The pulse width controller can be used to precisely adjust the pulse width of the pulse width adjusted signal. A pulse width multiplier can be implemented to allow direct observation and measurement of the pulse width of the pulse width adjusted signal.