On-clip testing circuit and method for improving testing of integrated circuits

[0002] The present invention relates generally to testing of integrated circuits, and more specifically to a method and apparatus that reduces the time and testing resources needed for testing of integrated circuits.

[0003] Integrated circuits are extensively tested both during and after production, and, in some cases, routinely during use after they have been installed in products. For example, memory devices, such as random access memories (“RAMs”) and dynamic random access memories (“DRAMs”), are tested during production at the wafer level and after packaging, and they are also routinely tested each time a computer system using the DRAMs executes a power-up or “boot” routine when power is initially applied to the computer system. As the capacity of DRAMs and other memory devices continues to increase, the time required to test the DRAMs continues to increase, even though memory access times continue to decrease.

[0004] A typical RAM integrated circuit includes at least one array of memory cells arranged in rows and columns. Each memory cell must be tested to ensure that it is operating properly. In a typical prior art test method, data having a first binary value (e.g., a “1”) are written to and read from all memory cells in the arrays, and thereafter data having a second binary value (e.g., a “0”) are typically written to and read from the memory cells. A memory cell is determined to be defective when the data that is read from the memory cell does not equal the data that was written to the memory cell. As understood by one skilled in the art, other test data patterns may be utilized in testing the memory cells, such as an alternating bit pattern, e.g., 101010. . . , written to the memory cells in each row of the arrays.

[0005] One situation requiring testing of memory integrated circuits occurs during fabrication of integrated circuits. Fabrication yields are reduced when fabrication errors occur. Testing of integrated circuits during fabrication allows the sources of some fabrication errors to be promptly identified and corrected. Testing during fabrication may reduce costs by reducing the number of integrated circuits affected by a given fabrication error.

[0006] Another situation requiring testing of integrated circuits also occurs in fabrication of memory integrated circuits. Defective memory cells are identified by testing and are replaced with non-defective memory cells from a set of spare or redundant memory cells. In one conventional method for replacing defective memory cells, fuses on the integrated circuit are blown in a pattern corresponding to the pattern of defective memory cells to select rows or columns of redundant memory cells. The pattern is then read to replace the rows or columns that include the defective memory cells.

[0007] FIG. 1 is a simplified block diagram of several integrated circuits 10 and an automated tester 12 according to the prior art. Separate buses 14 are dedicated to couple each of the integrated circuits 10 to the automated tester 12. The data buses 14 convey stimuli, known as background data, from the automated tester 12 to function circuits 16, such as memory arrays, contained in the integrated circuits 10 that are being tested. Each function circuit 16 generates a response, such as read data, from the background data that are sent to that function circuit 16. The data buses 14 also convey the read data from each function circuit 16 back to the automated tester 12. The automated tester 12 compares the read data from each integrated circuit 10 that is being tested to a corresponding set of expect data. The expect data correspond to read data that would be provided by the integrated circuit 10 if its function circuit 16 was operating properly. When the read data and the corresponding expect data match, the integrated circuit 10 is considered to be functioning normally. When the read data do not match the corresponding expect data, the integrated circuit 10 that provided the read data is considered to be malfunctioning.

[0008] Each bus 14 can only convey data unambiguously from one integrated circuit 10 at a time to the automated tester 12. In turn, the automated tester 12 can only accommodate a finite number of buses 14, limiting the number of integrated circuits 10 that may be tested at one time. The number of memory integrated circuits 10 that may be coupled to the automated tester 12 at one time is known as the “fanout” for the automated tester 12.

[0009] There is a need for an on-chip test circuit to test function circuits in a group of integrated circuits without requiring a separate module or control integrated circuit to read output signals from the function circuits in order to compare the output signals with expected output signals.

[0010] An on-chip test circuit is included in an integrated circuit for testing function circuits in the integrated circuit and for storing failure data from the tests. The on-chip test circuit includes an expect data register and a comparison circuit having a first input coupled to an output of the function circuits and a second input coupled to an output of the expect data register. The on-chip test circuit also includes a fail data register having an input coupled to an output of the comparison circuit. The fail data register stores data describing memory array failures. The combination of the comparison circuit and the fail data register allows many integrated circuits to be tested at one time without waiting for the each integrated circuit to provide read data to a tester and without bus contention. Testing of integrated circuits is thereby facilitated, reducing the time required for testing the integrated circuits and increasing the practical fanout from automated testers.

[0011] FIG. 1 is a simplified block diagram of several integrated circuits and an automated tester according to the prior art.

[0012] FIG. 2 is a simplified block diagram of a portion of an integrated circuit including an on-chip testing circuit in accordance with an embodiment of the present invention.

[0013] FIG. 3 is a simplified block diagram of several integrated circuits and an automated tester in accordance with an embodiment of the present invention.

[0014] FIG. 4 is a flow chart describing an integrated circuit testing method in accordance with an embodiment of the present invention.

[0015] FIG. 5 is a flow chart describing a method of reading stored failure data from the fail data register of an integrated circuit that has been tested in accordance with an embodiment of the present invention.

[0016] FIG. 6 is a flow chart describing an integrated circuit speed testing method in accordance with an embodiment of the present invention.

[0017] FIG. 7 is a simplified block diagram of a computer system including the integrated circuit of FIG. 2 in accordance with an embodiment of the present invention.

[0018] FIG. 2 is a simplified block diagram of a portion of an integrated circuit 20 including an on-chip testing circuit 22 in accordance with an embodiment of the present invention. The integrated circuit 20 also includes I/O pins or pads 24 coupled to a DQ buffer 26, allowing data, address and control signals to be coupled between external circuits (not shown in FIG. 4) and a function circuit 28. The I/O pins or pads 24 used to couple signals to and from the integrated circuit 20 typically include between one and sixty-four pins or pads. In one embodiment, the function circuit 28 includes a microprocessor. In another embodiment, the function circuit 28 includes a memory circuit such as a RAM, DRAM or synchronous DRAM.

[0019] The on-chip testing circuit 22 includes an expect data register 30 having an input bus coupled to the DQ buffer 26. A comparison circuit 32 includes a first input coupled to the expect data register 30 and a second input coupled to an output of the function circuit 28. A test mode enable circuit 33 is responsive to signals provided on a control line TME. A fail data register 34 has a reset input R coupled to the test mode enable circuit 33 and has one or more data inputs coupled to an output of the comparison circuit 32. In one embodiment, the on-chip testing circuit 22 also includes an OR gate 35 having inputs coupled to the output of the comparison circuit 32 and an output coupled to a first input of an AND gate 36. The AND gate 36 has a second input coupled to an output of a clock input buffer 38. The DQ buffer 26, the function circuit 28 and the expect data register 30 all couple data in or out in response to a clock signal, e.g., CLK_IN.

[0020] The on-chip test circuit 22 operates in one of two modes, a test mode and a pass-through mode, as determined by the state of the test mode enable circuit 33. In the pass-through mode, an external circuit (not shown in FIG. 2) provides signals on the control line TME causing the test mode enable circuit 33 to deactivate all address and control signals that it develops, allowing the integrated circuit 20 to function in a normal mode.

[0021] When the external circuit activates the test mode enable circuit 33, the on-chip test circuit 22 operates in the test mode to test the function circuit 28 in the integrated circuit 20. In the test mode of operation, command signals are coupled to the test mode enable circuit 33 from the external circuit in order to (i) put the integrated circuit 20 into the test mode, (ii) clear the fail data register 34 using a signal coupled to a reset input R, (iii) load background data into the function circuit 28 through the DQ buffer 26 and (iv) load expect data into the expect data register 30 through the IO pins or pads 24 of the DQ buffer 26. The comparison circuit 32 then compares read data from the function circuit 28 with corresponding expect data from the expect data register 30 and provides one or more output signals to the fail data register 34 when the read data do not match the corresponding expect data.

[0022] In one embodiment, the expect data register 30 is omitted. In this embodiment, the expect data may be supplied to the comparison circuit 32 through the I/O pins or pads 24 and the DQ buffer 26. Alternatively, the expect data may be stored in, e.g. a ROM (not illustrated) that is also a part of the test circuit 22, or may be stored or generated in the integrated circuit 20.

[0023] In another embodiment, the comparison circuit 32 may be a group of exclusive OR (“XOR”) gates (not illustrated) performing bitwise comparisons between the read data from the function circuit 28 and the corresponding expect data. In this embodiment, the output bits from the comparison circuit 32 will all be logical zeroes unless there is an error in the read data indicating a failure in the function circuit 28.

[0024] The outputs from the comparison circuit 32 may be used to increment a clock signal to the fail data register 34 by coupling each of the output bits from the comparison circuit 32 to a separate input of the OR gate 35 and coupling an output of the OR gate 35 to a first input of the AND gate 36. A clock signal from the clock buffer 38 is coupled to a second input of the AND gate 36. The clock signal from the clock buffer 38 is then passed to the clock input CLK of the fail data register 34 only when one or more bits from the output of the comparison circuit 32 indicate a failure in the function circuit 28. The fail data register 34 then only records data when there is a mismatch between the read data and the corresponding expect data that is indicative of a failure in the function circuit 28. Recording only data relating to failures, rather than all of the results of comparing read and expect data, reduces the amount of data to be stored in the fail data register 34.

[0025] In one embodiment, the fail data register 34 may be a counter. When a single-bit counter is used, the counter can indicate only that at least one failure occurred or that no failures occurred. When a multi-bit counter is used, the fail data register 34 may record how many failures occurred, up to the capacity of the counter employed for the fail data register 34. The addition of an overflow bit provides an indication that a greater number of failures occurred than can be recorded by the counter.

[0026] In some situations, multiple tests of each integrated circuit 20 may require several sets of failure data to be stored in each integrated circuit 20. In one embodiment, the fail data register 34 is segmented into a series of sub-registers FDR₁, FDR₂etc. each dedicated to storing failure data from one of the tests. These kinds of data may be useful in speed grading (determining the maximum clock frequency permitting reliable operation) of the integrated circuits 20, as is explained below in more detail.

[0027] When the integrated circuit 20 includes a memory circuit as the function circuit 28, it may be useful to store additional types of data in the fail data register 34. For example, it may be desirable to store addresses corresponding to defective memory locations in a failed cell address register that is part of the fail data register 34. These addresses may be used to repair the function circuit 28 or to avoid writing data to or reading data from the defective memory locations in the function circuit 28.

[0028] FIG. 3 is a simplified block diagram of N many integrated circuits 20 and an automated tester 50 in accordance with an embodiment of the present invention. A common, i.e., shared, bus 52 couples the I/O pins or pads 24 of the integrated circuits 20 to the automated tester 50. At least one dedicated control line 52_n, where n is an element of the set ranging from 1 to N inclusive, is also coupled between each of the control lines TME of the test mode enable circuits 33 (FIG. 2) in the N many integrated circuits 20 and the automated tester 50.

[0029] A first advantage that the automated tester 50 provides when testing integrated circuits 20 that include the on-chip testing circuit 22 is that the common bus 52 may be employed to send address, data and control signals to all of the integrated circuits 20 simultaneously. In other words, storing the results of the testing in the fail data registers 34 (FIG. 2) of the integrated circuits 20 allows testing of all of the integrated circuits 20 at the same time but avoids bus contention by storing test results in the fail data register 34. A second advantage of the embodiment of FIGS. 2 and 3 is that the amount of data that needs to be read from each of the integrated circuits 20 is reduced. Reading read data from, for example, every memory location in a memory array involves more data then merely reading the addresses of failed memory cells. A third advantage of the embodiment of FIGS. 2 and 3 is that testing time is reduced because the comparisons between expect data and read data are performed simultaneously in each of the integrated circuits 20 being tested, rather than being performed by the automated tester 12 of FIG. 1. A fourth advantage is that each additional integrated circuit 20 being tested only requires a single additional dedicated control line 52_nthat is separate from the common bus 52, rather than separate buses 14 each dedicated to one integrated circuit 10.

[0030] When testing of the integrated circuits 20 is complete, the automated tester 50 may read the contents of each of the fail data registers 34, as is described in more detail below.

[0031] FIG. 4 is a flow chart describing a process 60 for testing integrated circuits 20 in accordance with one embodiment of the present invention. The process 60 tests functionality of the function circuit 28 of FIG. 2, and may use the automated tester 50 of FIG. 3 to do so, although it will be recognized that other types of controllers might be used. In a step 62, the automated tester 50 sends control signals to the integrated circuits 20 that are to be tested to set the integrated circuits 20 to the test mode. In a step 64, the automated tester 50 sends signals to the integrated circuits 20 that clear the fail data registers 34. In a step 66, the automated tester 50 writes the expect data into all of the expect data registers 30 using the common bus 52. In a step 68, the automated tester 50 writes the background data to the function circuits 28, again using the common bus 52.

[0032] In a step 70, the on-chip test circuit 22 obtains read data from the function circuit 28 and corresponding expect data from the expect data register 30. In a query task 72, the on-chip test circuit 22 compares the read data to the corresponding expect data to determine if a failure of the function circuit 28 has occurred, i.e., the read data do not match the corresponding expect data. When the on-chip test circuit 22 determines that a failure has occurred, the on-chip test circuit 22 initiates a step 74. In the step 74, the on-chip test circuit 22 stores data describing the failure in the fail data register 34. Control passes to a query task 76 when either the query task 72 determines that no failure of the function circuit 28 has occurred or after the failure data have been recorded in the step 74. The query task 76 determines if the testing has been completed.

[0033] When the query task 76 determines that the testing has not been completed, a step 78 increments the expect data register 30 and the function circuit 28 to provide new expect and read data, respectively. The on-chip test circuit 22 then returns to the step 70 and continues testing the function circuit 28. When the query task 76 determines that the testing has been completed, a step 79 returns the integrated circuit 20 to a normal mode of operation and the process 60 ends.

[0034] FIG. 5 is a flow chart describing a process 80 for reading stored failure data from the fail data registers 34 (FIG. 2) of integrated circuits 20 that have been tested in accordance with an embodiment of the present invention. In one embodiment, the automated tester 50 (FIG. 3) may initiate the process 80 after the process 60 (FIG. 4) ends. In a step 82, one of the integrated circuits 20 that has new failure data to be downloaded is selected. In one embodiment, a signal is coupled to the control line TME of the selected integrated circuit 20 from the automated tester 50 by the control line 52_n(FIG. 3) that is dedicated to the selected integrated circuit 20.

[0035] In a step 84, the on-chip testing circuit 22 in the selected integrated circuit 20 downloads data describing the fail status of the function circuit 28 from the fail data register 34 to the automated tester 50 through the common bus 52. A query task 86 then determines if all of the failure data have been downloaded. When the query task 86 determines that not all of the failure data have been downloaded, control passes back to step 82 to select another one of the integrated circuits 20, allowing all of the integrated circuits 20 to be selected in turn. When the query task 86 determines that all of the failure data have been downloaded, the process 80 ends.

[0036] In another embodiment, the integrated circuit 20 may initiate the process 80 each time a failure occurs, or, alternatively, each time the fail data register 34 has accumulated data relevant to a predetermined number of failures. In these embodiments, the integrated circuit 20 executes the step 82 by sending an interrupt to the automated tester 50 through the control line 52_n(FIG. 3) that is dedicated to the selected integrated circuit 20. In this embodiment, control passes to step 84 when the query task 86 determines that not all of the failure data have been downloaded.

[0037] FIG. 6 is a flow chart describing an integrated circuit speed testing process 90 in accordance with an embodiment of the present invention. The speed testing process 90 tests a group of integrated circuits 20 at M many different clock frequencies to determine a maximum clock frequency for reliable operation of each of the integrated circuits 20. In a step 92, the automated tester 50 of FIG. 3 sets an index variable m to 1. In a step 94, the automated tester 50 sets an m^THclock frequency f_mfor a clock signal that is coupled to the input line CLK_INof the clock buffer 38 in the integrated circuit 20 of FIG. 2. A step 96 invokes the process 60 of FIG. 4. An optional step 98 invokes the process 80 of FIG. 5.

[0038] A query task 100 determines if m=M; when m≠M, a step 102 increments m and control then passes back to the step 94. When m=M, an optional step 104 may invoke the process 80 of FIG. 5 to download any stored failure data from the fail data registers 34 of the integrated circuits 20 that are being tested. Either the step 98 or the step 104 may be used to download failure data, however, the step 104 will download failure data for all M many tests, which may be separately stored, e.g., in each of the sub-registers FDR₁, FDR₂etc. of FIG. 2, while the step 98 downloads a group of failure data for each of the m clock frequencies f_mat the conclusion of the testing at each of the clock frequencies f_m. In either case, the step 106 evaluates failure data for each of the integrated circuits 20 to determine a maximum clock frequency for each of the integrated circuits 20 to be able to operate reliably. A step 107 then returns the integrated circuits 20 to the normal mode of operation, and the process 90 then ends.

[0039] Speed testing of integrated circuits 20 (e.g., the process 90 of FIG. 6) differs from functional testing (e.g., the process 60 of FIG. 4) because speed testing is typically carried out with integrated circuits 20 that have previously been functionally tested and that are therefore known to be functional. In the case of integrated circuits 20 that are read-write memories such as DRAMs, the integrated circuits 20 have previously been tested and defective memory cells have previously been replaced as is conventional. Accordingly, speed testing may not need addresses for memory cells that fail, and may need only the total number of failures, such as memory cell failures, in order to determine a maximum error-free clock frequency for the integrated circuit 20. When only the total number of failures is needed, the fail data register 34 may include a counter (when step 98 is used to download failure data) or a group of M many counters (when step 104 is used to download failure data). This may permit simplification of the on-chip testing circuit 22 for some applications.

[0040] FIG. 7 is a simplified block diagram of a portion of a computer system 120 including the integrated circuit 20 of FIG. 2 in accordance with an embodiment of the present invention. The computer system 120 includes a central processing unit 122 for performing various computing functions, such as executing specific software to perform specific calculations or tasks. The central processing unit 122 is coupled via a bus 124 to a memory 126, a user input interface 128, such as a keyboard or a mouse, and a display 130. The memory 126 may or may not include a memory management module (not illustrated) and does include ROM for storing instructions providing an operating system and read-write memory for temporary storage of data. The processor 122 operates on data from the memory 126 in response to input data from the user input interface 128 and displays results on the display 130. The processor 122 also stores data in the read-write portion of the memory 126.

[0041] The integrated circuit 20 is particularly useful when it is a memory integrated circuit in the read-write memory portion of the memory 126, because it may then allow the memory 126 to be tested more rapidly (e.g., using the process 60 of FIG. 4) while booting. Following testing of the memory integrated circuits, the processor 122 may extract the failure data from the memory 126 (e.g., using the process 80 of FIG. 5) in order to form a memory map describing the addresses of the defective memory cells. The memory map allows the processor 122 to avoid writing data to or reading data from the memory cells that were identified as being defective.

[0042] Examples of systems where the computer system 120 finds application include personal/portable computers, camcorders, televisions, automobile electronic systems, microwave ovens and other home and industrial appliances.

[0043] It is to be understood that even though various embodiments and advantages of the present invention have been set forth in the foregoing description, the above disclosure is illustrative only, and changes may be made in detail, and yet remain within the broad principles of the invention. Therefore, the present invention is to be limited only by the appended claims.