Xilinx, Inc.

Robert W Wells

Robert D Patrie

Robert O Conn

Shahin Toutounchi

Christopher H Kingsley

Shekhar Bapat

Zhi-min Ling

Jae Cho

Gary R Lawman

Daniel Y Chung

Andrew W Lai

Eric Thorne

Joseph D Linoff

Vincent L Tong

David L Ferguson

Tassanee Payakapan

Gil A Speyer

Steven P Young

Trevor J Bauer

Erik Vaclav Chmelar

Matthew H Klein

Jongheon Jeong

Michael M Matera

Andrew J DeBaets

Anthony P Calderone

Rick W Dudley

Steven M Trimberger

Shelly G Davis

Clay S Johnson

In the disclosed methods, integrated circuit (IC) dice are manufactured from a common specification, and the IC dice are tested for defective circuitry. Respective defect sets are generated to indicate defective circuitry in the IC die. The dice are assigned to bins based on the respective defect sets. For each bin, all IC dice assigned to the bin have equivalent respective defect sets. Product definitions are provided, and each product definition indicates a respective set of circuitry required for a corresponding product. Respective sets of packages are manufactured for each product. In the manufacturing of each package of a respective set of packages for each product, one or more IC dice are selected from a subset of the plurality of bins such that the IC dice have respective defect sets allowed by the product definition of the product. The selected IC dice are then manufactured into the package.

Methods of making integrated circuit products

A method of testing a programmable device begins by programming at least a portion of the programmable device in accordance with at least a portion of an application to produce a programmed circuit, wherein the programmed circuit includes an input sequential element and an output sequential element. The method continues by providing a test input to the programmed circuit. The method continues by triggering the input sequential element to temporarily store the test input based on a first edge of the test clock. The method continues by triggering the output sequential element to temporarily store a test output of the programmed circuit based on a second edge of the test clock. The method continues by capturing the test output of the programmed circuit in accordance with the second edge of the test clock.

Testing of a programmable device

Described are methods for implementing customer designs in programmable logic devices (PLDs). The defect tolerance of these methods makes them particularly useful with the adoption of 'nanotechnology' and molecular-scale technology, or 'molectronics.' Test methods identify alternative physical interconnect resources for each net required in the user design and, as need, reroute certain signal paths using the alternative resources. The test methods additionally limit testing to required resources so devices are not rejected as a result of testing performed on unused resources. The tests limit functional testing of used resources to those functions required in the user designs.

Application-specific methods for testing molectronic or nanoscale devices

Methods of utilizing partially defective PLDs, i.e., PLDs having localized defects. A partially defective PLD is tested for compatibility with a particular configuration bitstream. If the partially defective PLD is compatible with the bitstream (i.e., if the localized defect has no effect on the functionality of the design implemented by the bitstream), a product is made available that includes both the bitstream and the partially defective PLD. In some embodiments, the bitstream is stored in a memory device such as a programmable read-only memory (PROM). In some embodiments, the product is a chip set that includes the partially defective PLD and a separately-packaged PROM in which the bitstream has previously been stored. In some embodiments, the PROM is manufactured as part of the FPGA die.

Methods of utilizing programmable logic devices having localized defects in application-specific products

Disclosed methods for utilizing programmable logic devices that contain at least one localized defect. Such devices are tested to determine their suitability for implementing selected customer designs that may not require the resources impacted by the defect. If the FPGA is found to be unsuitable for one design, additional designs may be tested. The test methods in some embodiments employ test circuits derived from a user design to verify PLD resources required for the design. The test circuits allow PLD vendors to verify the suitability of a PLD for a given customer design without requiring the vendor to understand the design.

Application-specific methods useful for testing look up tables in programmable logic devices

Methods of testing for shorts (e.g., bridging defects) between interconnect lines in an integrated circuit. For example, in a design implemented in a programmable logic device (PLD), some interconnect lines are used and others are unused. To test for shorts between the used and unused interconnect lines, both used and unused interconnect lines are driven to a first logic level, and the leakage current is measured. The used interconnect lines are driven to a second logic level, while the unused lines remain at the first logic level. The current is again measured, and the difference between the two measurements is determined. If the difference exceeds a predetermined threshold, the device design combination is rejected. Some embodiments provide methods of testing for shorts between used and unused interconnect lines for a design targeted to a partially defective PLD.

Methods of testing for shorts in programmable logic devices using relative quiescent current measurements

Disclosed are methods for utilizing programmable logic devices that contain at least one localized defect. Such devices are tested to determine their suitability for implementing selected designs that may not require the resources impacted by the defect. If the FPGA is found to be unsuitable for one design, additional designs may be tested. The test methods in some embodiments employ test circuits derived from a user's design to verify PLD resources required for the design. The test circuits allow PLD vendors to verify the suitability of a PLD for a given user's design without requiring the PLD vendor to understand the user's design.

Application-specific testing methods for programmable logic devices

FPGAs that contain at least one localized defect may be used to implement some designs if the localized defect is not used in the designs. To determine if the FPGA is suitable to implement a design, the design is loaded into the FPGA. The FPGA is tested to determine whether it can execute the design accurately even with the localized defect. The FPGA will be marked as suitable for that design if it passes the test. If the FPGA is found to be unsuitable for one design, additional designs may be tested. Thus, a FPGA manufacturer can sell FPGAs that are normally discarded. As a result, the price of these FPGAs could be set significantly low.

Method of using partially defective programmable logic devices

Fault coverage for the programmable interconnect of a programmable logic device (PLD) is provided. A user's design is modeled, thereby determining the programmable interconnect path in the device. The user's logic design is then modified, thereby facilitating the detection of faults. Specifically, any function generators in the PLD are implemented as predetermined logic gates, thereby forming a logic gate tree design. The synchronous elements in the user's design are preserved and transformed, if necessary, to provide controllability. Then, a vector can be exercised in the new design. A first readback of the PLD can be compared to a second readback of a fault-free model of the design.

Providing fault coverage of interconnect in an FPGA

A circuit measures the signal propagation delay through a selected test circuit. The test circuit is provided with a feedback path so that the test circuit and feedback path together form a free-running oscillator. The oscillator then automatically provides its own test signal that includes alternating rising and falling signal transitions on the test-circuit input node. A phase discriminator samples the output of the oscillator and accumulates data representing the signal propagation delay of either rising or falling signal transitions propagating through the test circuit. The worst-case delay associated with the test circuit can then be expressed as the longer of the two. Knowing the precise worst-case delay allows IC A designers to minimize the guard band and consequently guarantee higher speed performance.

Built-in self test using pulse generators

A new testing method uses a field programmable gate array to emulate faults, instead of using a separate computer to simulate faults. In one embodiment, a few (e.g., two or three) known good FPGAs are selected. A fault is introduced into the design of a FPGA configuration. The configuration is loaded into the FPGAs. A test vector is applied and the result is evaluated. If the result is different from that of a fault-free configuration, the fault is caught. One application of this method is to evaluate fault coverage. A fault model that can be used in the present invention is disclosed.

Fault emulation testing of programmable logic devices

Described are methods and circuits for accurately placing signal transitions, or &ldquo;edges,&rdquo; simultaneously on two or more pins of an integrated circuit (IC). A conventional tester is connected to an integrated circuit, such as a programmable logic device. The integrated circuit is adapted to include a coincidence detector that compares the timing of edges on two input pins of the integrated circuit. The coincidence detector indicates when the two edges are coincident, allowing an operator of the tester to adjust the tester to establish coincidence. The amount of offset necessary to provide coincident edges is stored in a database for later use in deskewing edges used in subsequent tests. The integrated circuit can be a programmable logic device configured to include one or more coincidence detectors with which to place edges relative to one another on different pins.

Methods and circuits for precise edge placement of test signals

Described is a test circuit that can be instantiated on a programmable logic device to perform at-speed functional tests of programmable resources, including internal memory and routing resources. The resources to be tested are configured to instantiate a counter circuit connected to the address terminals of a linear-feedback shift register (LFSR). LFSRs are cyclic, in the sense that when clocked repeatedly they go through a fixed sequence of states. Consequently, an LFSR that starts with a known set of data has a predictable set of data after a given number of clock periods. The LFSR is preset to a known count and clocked a known number of times. The resulting count is then compared with a reference. If the resulting count matches the reference, then all of the resources used to implement the test circuit, including the memory and routing resources used to implement the LFSR, are deemed fully functional at the selected clock speed. A test circuit employing an LFSR can be duplicated many times on a given device under test to consume (and therefore test) as many resources as possible.

Growing community of inventors

Cupertino, CA, United States of America

Robert W Wells