Abstract:
An integrated circuit includes a programmable logic device, a dedicated device, and an interface circuit between the two. The interface circuit can be easily modified to accommodate the different interface I/O demands of various dedicated devices that may be embedded into the integrated circuit. In one embodiment, the interface circuit may be implemented using a plurality of mask programmable uni-directional interface buffer circuits. The direction of any desired number of the interface buffer circuits can be reversed based on the needs of a desired dedicated device by re-routing the conductors in the interface buffer circuits in a single metal layer of the integrated circuit. In another embodiment, the interface circuit may be implemented using a hardware configurable bi-directional interface buffer circuit.
Abstract:
A differential charge pump includes a transient reducing circuit that provides multiple switching current paths to reduce transients caused by the charge transfer as the charge pump is switched. The differential charge pump includes separate current sources in the transient reducing circuit that are switchably coupled to the non-active current source in the charge pump. In one embodiment, each current sources include a static current source and a variable current source that is controlled by a common mode feedback circuit. The variable current source may produce a current with less magnitude than the current produced by the static current source.
Abstract:
To protect logic module output devices from high voltages, logic modules are not powered during antifuse programming. In some embodiments, two separate power input terminals VCC1 and VCC2 are provided: power input terminal VCC1 being coupled to power the logic modules, and power input terminal VCC2 being coupled to power the programming control circuitry. Power terminal VCC1 is left floating or is grounded during antifuse programming such that the logic modules are not powered but such that the programming circuitry is powered during antifuse programming via the second power terminal VCC2. Logic module output protection transistors are not required nor is the associated charge pump. Because the logic module input devices are not powered, a current surge through the input devices on power up does not occur and an internal disable signal and associated circuitry is not required. In one embodiment, the field programmable gate array is made smaller because it has no internal disable signal and associated circuitry, no logic module output protection transistors, and no charge pump that operates during normal circuit operation. In embodiments, power input terminal VCC2 is a high voltage compatible power input terminal.
Abstract:
A programmable logic device having redundant sets of logic blocks which are capable of being enabled or disabled. The programmable logic device includes a plurality of sets of logic blocks, a plurality of routing resources and a programming circuit. Good logic blocks are enabled and fully operational when programmed. Nonfunctional logic blocks are disabled, powered off and invisible to the programming software. Each set of logic blocks has a corresponding routing resource. The routing resource corresponding to an enabled set of logic blocks is capable of being configured to provide input and output data paths for the enabled set of logic blocks. The routing resource corresponding to a disabled set of logic blocks is capable of being configured to bypass the disabled set of the logic blocks. The programming circuit stores the configuration data for the routing resources and is capable of providing the configuration data to a routing resource that corresponds to an enabled set of logic blocks.
Abstract:
A metal-to-metal conductive plug-type antifuise has a conductive plug disposed in an opening in an insulating layer. A programmable material feature (for example, amorphous silicon) overlies the conductive plug. A conductor involving a metal (for example, aluminum or copper) that migrates in the programmable material overlies the programmable material. To prevent migration of metal from the conductor into the programmable material when the antifuse is not programmed, the conductor has a layer of barrier metal between the metal that migrates and the programmable material. In some embodiments, there are two layers of barrier metal. An airbreak after formation of the first barrier metal layer improves the ability of the barrier metal to prevent diffusion between the programmable material and the overlying conductor. The airbreak may stuff grain boundaries in the upper surface of the first barrier metal and/or may cause the first barrier metal layer to have different grains and/or a different grain orientation than the overlaying second barrier metal layer. In some embodiments, a capping layer over the top surface of the programmable material protects the underlying programmable material during an ashing step when a mask used to etch the programmable material is removed. The capping layer and the programmable material form a capping layer/programmable material layer stack within the antifuse underneath the two barrier metal layers. The capping layer may also be made of a barrier metal and constitute an additional barrier.
Abstract:
A programmable integrated circuit (see FIG. 10) includes a routing conductor, i.e., "express wire," that extends substantially across the array of the integrated circuit. Because of the metal resistance through the long express wire, the express wire is simultaneously supplied with programming current from two different programming voltage terminals. Thus, programming current may be supplied to an electrode of an antifuse being programmed using programming current flowing through two separate programming voltage terminals. One programming voltage terminal supplies programming current via a first programming transistor and a first programming conductor to the express wire near one end of the programming conductor whereas another programming voltage terminal supplies programming current via a second programming transistor and a second programming conductor to the express wire near an opposite end of the express wire. The programming drivers that drive the first and second programming conductors are disposed adjacent opposite sides of the integrated circuit.
Abstract:
A programmable integrated circuit (see FIG. 5) has a plurality of linearly extending wire segments with antifuses disposed between each wire segment and a plurality of linearly extending programming conductors that are perpendicular to the wire segments. A plurality of programming transistors are disposed between a corresponding respective one of the wire segments and a corresponding respective one of the programming conductors. A programming control conductor extending from a programming control driver is coupled to the gate electrode of each of the programming transistors as well as the gate electrode of a test transistor. A test antifuse is coupled in series with the test transistor. When the programming control conductor can drive the test transistor with an adequately high voltage to program the test antifuse, it is assumed that the programming control conductor can drive the programming transistor with an adequately high voltage to program the antifuses. The test transistor may be disposed on the programming control conductor at the opposite end from the programming control driver.
Abstract:
A field programmable gate array has a security antifuse which when programmed prevents readout of data indicative of how the interconnect structure is programmed but which does not prevent readout of data indicative of which other antifuses are programmed. In some embodiments, the programming control shift registers adjacent the left and right sides are the field programmable gate array are disabled when the security antifuse is programmed but the programming control shift registers adjacent the top and bottom sides of the field programmable gate array are not disabled. A second security antifuse is also provided which when programmed disables a JTAG boundary scan register but does not disable a JTAG bypass register. Information can therefore be shifted through the JTAG test circuitry without allowing the JTAG circuitry to be used to extract information indicative of how the interconnect structure is programmed. Logic module and interface cell scan paths are provided and special test instructions are supported which allow test vectors to be loaded into the logic module and interface cell scan paths.
Abstract:
A clock network of a field programmable gate array has a first clock bus extending across the chip in a first dimension. A clock pad can be coupled to the first clock bus if the clock network is to be driven from the clock pad. An output of a selected logic cell can be coupled to the first clock bus if the clock network is to be driven from a logic cell. To increase speed of the clock network, the first clock bus is segmented (in one embodiment, collinearly extending segments can be selectively coupled together via selectively programmable antifuses) so that only a short piece of the first clock bus is used to couple either the pad or the logic cell to the clock network in high speed applications.
Abstract:
Select sets of a logic function corresponding to an output of a first logic circuit are determined. These select sets are used to obtain a second logic circuit, the logic function corresponding to the output of which is the same as the logic function corresponding to the output of the first logic circuit. A propagation delay through the second logic circuit may be smaller than a corresponding delay through the first logic circuit. Sometimes, such a smaller propagation delay through the second logic circuit results in the second logic circuit having a smaller critical path delay. The second logic circuit may therefore have a greater maximum operating speed than the first logic circuit.