摘要:
The present invention is generally directed to a an integrated circuit package having a unique lead configuration, wherein the integrated circuit package is constructed from a die containing an integrated circuit. The die has a plurality of leads for carrying electrical signals to and from the integrated circuit, wherein the plurality of leads are disposed over a bottom side of the die. The package further includes a multi-layer substrate having at least two signal layers. The substrate is juxtaposed against the die and has a plurality of contacts disposed along a top side to align with the leads of the die to carry the electrical signals to conductive paths within the at least two signal layers. The multi-layer substrate has a larger adjoining surface area than the die and further has a plurality of leads disposed across a bottom side for connection with a printed circuit board, the on the bottom side being in communication with the leads of the top side by way of the conductive paths disposed within the substrate. The leads of the die are disposed such that at least two high speed rows of leads are disposed in parallel fashion near the center of the die, wherein the high speed rows are for carrying high frequency electrical signals. At least two sets of low speed rows of leads are disposed in parallel fashion near the sides of the die, and spaced apart from the high speed rows.
摘要:
The present invention is generally directed to a method and apparatus for transferring data from an integrated circuit that is capable of bidirectional data communication. In accordance with one aspect of the invention, an apparatus is provided having a circuit for splitting the data into two portions--a high bit portion and a low bit portion. The circuit also includes two data paths. A first data path communicates the high bit portion of the data and a second data path communicates the low bit portion of the data. The apparatus further includes an output circuit that is configured to connect outputs of the first and second data paths to a common, bidirectional data bus. Finally the apparatus includes a hold circuit configured to hold a last data value on the common data bus for at least one clock cycle before allowing circuitry to receive data from the common bus. A method is also provided for transferring data from an integrated circuit capable of bidirectional data communication. The method operates by splitting data into a first group of bits and a second group of bits. The method further includes steps of transmitting the first group of bits along a first data path, and transmitting the second group of bits along a second data path. The method then electrically connects an output of the first data path to an output of the second data path at a common directional data bus, and alternatively transmits data over the first path and the second path. When alternating the data transmissions in this way, the method ensures that the two data path outputs are not driven at the same time. Finally, the method holds a last data value on the common bus for at least one clock cycle before receiving data over the common bus.
摘要:
A method for increasing the layer density uniformity across a conductive layer, which comprises a plurality of functional blocks, of an integrated circuit is presented. Increased uniformity is achieved by tiling a plurality of capacitors in between the functional blocks. The configuration of the capacitor array and number of the capacitor cells in the array is arranged so as to provide approximate uniformity in the conductor-to-non-conductor density across the entire conductive layer. The capacitor array may be used to reduce power supply switching noise by coupling one or more of the capacitor cells making up the capacitor array between a high power rail and a low power rail.
摘要:
The present invention is generally directed to a driver circuit for a high speed transceiver. In accordance with one aspect of the invention, the driver circuit includes a first driver segment disposed to receive a control signal and configured to drive the control signal from a logic zero state to a logic one state and place the driven signal on a first driver segment output. Similarly, the driver circuit includes a second driver segment disposed to receive the control signal and configured to drive the control signal from a logic one state to a logic zero state and place the driven signal on a second driver segment output. In this regard, the control signal is a signal generated internally (i.e., within the chip) to be driven across a bus to another chip. The strength of the control signal must be increased before driving the control signal onto the bus. For this reason, the first driver segment and the second driver segment each include a plurality of drive units that are disposed in a cascaded configuration. As the control signal passes through each successive drive unit, it gains in signal strength. As will be appreciated by persons skilled in the art, this cascaded drive unit configuration provides for an extremely fast overall power build-up of the signals, as opposed to using a single, more powerful drive unit. To balance the timing delay between the two segments, a delay element is serially disposed within the segment have the fewer inversions.
摘要:
A circuit for matching the impedance of a first array of transistors to an external resistor is used to produce a first set of control signals. This first set of control signals is used to control another array of transistors to replicate the impedance of the first array of transistors. This replicated impedance is then used by another circuit for matching impedance to produce a second set of control signals that control an array of transistor of a different type to match the impedance of the first two array. The two sets of control signals may then be used as calibration signals for the pull-up and pull-down transistors of multiple output drivers.