Configurable scan path structure

Not Applicable

Not Applicable

Applicant claims priority from European Patent Application No. 02290839.6 filed Apr. 4, 2002.

1. Technical Field

This invention relates in general to electronic circuits and, more particularly, to an electronic circuit with a configurable scan path.

2. Description of the Related Art

As circuit designs become denser and more complicated, the need for testing increases. Scan path testing, where test data is input to various circuit modules and the resultant output is compared to expected results, is one of the most powerful testing schemes. Unfortunately, scan path testing is one of the larger costs involved in manufacturing an electronic device.

To reduce scan path testing costs, low cost testers have been developed to reduce tester costs. Typically, however, the low cost testers are limited in function vis-à-vis more expensive testers, particularly in the number of scan chains that can be simultaneously tested at the most efficient mode of test operation for the tester.

Accordingly, the circuit designer must make a decision regarding whether to design a circuit with fewer, long scan chains for optimization of testing on a low cost tester, or a circuit with more, shorter scan chains for optimization of testing on a higher cost tester. This scenario places the designer in a difficult position, since it is most efficient to test many circuit designs on the low cost tester during the beginning stages (on die) and on the more expensive tester at later stages (packaged chip).

Therefore, a need has arisen for a circuit design that can be optimized for multiple testers.

In the present invention, an integrated circuit includes logic circuitry having a plurality of defined logic cores. A plurality of input scan core ports and a plurality of a plurality of output scan ports are coupled to the logic cores using routing circuitry for selectively routing test signals on the input scan ports through specified logic cores, such that the number of input scan ports used to pass test signals can programmably configured to optimize testing using various testers.

The present invention has significant advantages over the prior art. Most importantly, the invention allows optimization of scan test time to any tester being used. Thus, if different testers are used during different stages, the scan configuration can be dynamically configured for each tester. This reduces test time and hence reduces costs due to testing.

Further, if it is necessary, or desirable, to change testers over a term of production of a device, the new tester can be accommodated at its optimum configuration without redesign of the device.

For a more complete understanding of the present invention, and the advantages thereof, reference is now made to the following descriptions taken in conjunction with the accompanying drawings, in which:

FIGS. 1 a and 1 b illustrate prior art scan test structures;

FIG. 2 illustrates a block diagram of an integrated circuit with a configurable scan test architecture for accommodating various tester scan access bus widths and more complicated scan paths;

FIG. 3 illustrates a block diagram of a scan core used in the logic of FIG. 2 ;

FIG. 4 illustrates a block diagram of the routing circuitry in relation to the scan cores;

FIG. 5 a through FIG. 5 d illustrates various scan paths that may be implemented using the architecture of FIG. 2 .

The present invention is best understood in relation to FIGS. 1–5 a–d of the drawings, like numerals being used for like elements of the various drawings.

FIGS. 1 a and 1 b illustrate the inflexibility of scan path design in a circuit 10 when using a limited logic tester as in the prior art. In the embodiment illustrated in FIGS. 1 a and 1 b , it is assumed that the logic tester can test a maximum of eight scan chains in serial mode. In this embodiment, a single scan input port SI of eight scan chains [ 7 : 0 ] is serially passed through four logic modules 12 (referenced individually as logic modules 12 a–d ), resulting in a single scan output port SO[ 7 : 0 ].

In the embodiment of FIG. 1 a , in order to test all modules 12 , the data must pass serially through the modules. Accordingly, the test time is very long.

As shown in FIG. 1 b , the prior art also allows testing a single module at a time. In FIG. 1 b , only module 12 c is tested. While this allows a faster test, for a complete test, a single module test would be required on each module, and the total test time would be even greater than that of FIG. 1 a.

FIG. 2 illustrates a block diagram of a circuit architecture that allows flexible routing of scan chains, such that a single circuit design can use both low-end and high-end testers efficiently.

In FIG. 2 , a circuit 20 has inputs for N input scan ports 22 and N output scan ports 24 . For the illustrated embodiment, each input scan port 22 and output scan port 24 has eight scan chains; however, for a given design, each scan port could accommodate any desired number of scan chains. The scan ports 22 and output scan ports 24 are coupled to a scan configuration module (SCM) 26 , which includes routing circuitry 28 and control circuitry 30 . Control circuitry 28 operates responsive to control word scanconfig, which is received through TAP (test access port) circuitry 32 . TAP circuitry 32 includes external ports TDI (test data in), TDO (test data out), TMS (test mode select), TCK (test clock), and TRST (test reset). The TAP circuitry 32 can be used to instruct the control circuitry 30 on a desired setting for the routing circuitry 28 .

The SCM is coupled to the logic 34 . The logic 34 includes the various scan cores 36 (individually referenced as 36 0 through 36 N−1 ) that can be tested.

In operation, the routing circuitry 28 , shown in greater detail in connection with FIG. 4 , serves to selectively route multiple scan ports through sequences of scan cores 36 . The SCM 26 can reconfigure the structure of the scan chain paths so that the highest degree of parallelism can be achieved, based on the capabilities of the tester.

FIG. 3 illustrates a block diagram of a scan core 36 . Each scan core 36 includes a logic module 40 . The scan core 36 has three modes of operation. In the first mode (scan mode), the logic module 40 receives data over the scan chains at its scan port SI. In the second mode (bypass mode), the logic module 40 is bypassed by the scan chains. In the third mode (boundary scan mode), data is received at a separate boundary scan input port, BSI.

In scan mode, data received via the scan chains on the scan input, SI, propagates through the logic module 40 and is output at the scan output SO. In bypass mode, the data on the scan chains is output without propagation through the logic module 40 . In boundary scan mode, data received via the boundary scan input, BSI, propagates through the logic module 40 and is output from the boundary scan output, BSO.

FIG. 4 illustrates a preferred embodiment for routing scan chains through the scan cores 36 using the routing circuitry 28 . A first set of multiplexers 42 (individually referenced as multiplexers 42 1 through 42 N−1 ) selectively pass the signal from the input scan ports SI( 0 ) through SI(N−1) to the scan cores 36 . The multiplexers 42 are arranged such that each input scan port SI( 1 ) through SI(N−1) is coupled to the respective multiplexer 42 1 through 42 N−1 . The output of each multiplexer 42 1 through 42 N−1 is coupled to the other input of the next sequential multiplexer 42 ; hence, the output of multiplexer 42 1 is coupled to the input of multiplexer 42 2 along with input scan port SI( 2 ), the output of multiplexer 42 2 is coupled to the input of multiplexer 42 3 along with input scan port SI( 3 ), the output of multiplexer 42 3 is coupled to the input of multiplexer 42 4 along with input scan port SI( 4 ), and so on. The output of each multiplexer 42 1 through 42 N−2 is also coupled to the input of a respective multiplexer 44 1 through 44 N−2 . Input scan port SI( 0 ) is coupled to an input of multiplexer 44 0 . The other input to each multiplexer 44 0 through 44 N−2 is coupled to the output of the next higher scan core 36 . Hence the output of scan core 36 1 is coupled to an input of multiplexer 44 0 , the output of scan core 36 2 is coupled to an input of multiplexer 44 1 , and so on. The output of each multiplexer 44 0 through 44 N−2 is coupled to the input of scan cores 36 0 through 36 N−2 , respectively. The output of multiplexer 42 N−1 is coupled to the input of scan core 36 N−1 .

In operation, the signal on an input scan port SI(x) can propagate through the multiplexers 42 to any scan core 36 y , where y≧x. The signal on the input scan port SI(x) can then pass through a chain of scan cores 36 y through 36 z via multiplexers 44 , where y≧x≧z, bypassing selected scan cores 36 . In this case, z denotes the last scan core 36 that does not receive signals from another scan port. For example, if SI( 3 ) is coupled to scan core 36 3 and SI( 4 ) propagates through multiplexers 42 to scan core 36 7 , then the test signals on SI( 4 ) could pass through scan cores 36 7 through 36 4 .

As described above, the control circuitry 30 controls the routing circuitry 28 , responsive to tester commands issued via the TAP circuitry 32 . Table 1 provides a preferred embodiment for signaling to set up the routing circuitry 28 and scan cores 36 . In this table, the N denotes the number of scan ports and n denotes the number of scan chains per scan port.

The SCM 26 is designed to be completely generic. There is no limitation on the number of “Scan Cores” that can be hooked-up. A “N — CORES” (equal to N) generic constant must be defined into the package section of RTL code to indicate the number of cores. Also a “N — CHAINS” (equal to n) constant indicating the Scan Core number of chains must be defined. The SCM 26 is controlled by the scanconfig control signal. As shown in Table 2, scanconfig is divided into three fields, each field being N — CORES bits wide:

The Cores — te and Cores — bndryte fields set the mode of operation for each scan core 36 —i.e., bypass, internal scan chain or boundary scan chain. Each scan core 36 has an associated bit in each of the Cores — te and Cores — bndryte fields. For a given scan core 36 , if the respective bits of the Cores — te and Cores — bndryte fields are both “0”, then bypass mode is used for that scan core. If the respective bit in Cores — te is a “1” and the respective bit in Cores — bndryte is a “0”, an internal scan chain is set. If the respective bit in Cores — te is a “0” and the respective bit in Cores — bndryte is a “1”, a boundary scan chain is set.

The remaining field of scanconfig is the Scan — pads — en field. By setting a bit in the Scan — pads — en field respective Scan Port is activated. These are also used to code different configurations.

Table 4 provides examples of different configurations that can be enabled using the invention. This is a non-exhaustive list of possible scan configurations possible using the definitions of Cores — te, Cores — boundryte, and Scan — pads — en set forth above.

FIGS. 5 a through 5 c illustrate examples of different scan test structures that may be programmably implemented on a device using the invention. During on die testing, a scan test similar to that shown in FIG. 1 a could be used in conjunction with low-cost (limited scan port) tester. With an N — CORES of four, such a test could be implemented using Cores — te[ 3 : 0 ]=1111, Cores — bndryte[ 3 : 0 ]=0000 and Cores — pads — en[ 3 : 0 ]=0001.

For high speed testing of the scan cores, which may be performed, for example, on the packaged devices with a high speed tester, a parallel test of each scan core could be used, as is shown in FIG. 5 a . In this case, each scan core 36 is tested separately. With an N — CORES of four, such a test could be implemented using Cores — te[ 3 : 0 ]=1111, Cores — bndryte[ 3 : 0 ]=0000 and Cores — pads — en[ 3 : 0 ]=1111.

A higher speed (relative to the test of FIG. 1 a ), yet less parallel, test is shown in FIG. 5 b . In this case, scan cores 36 0 and 36 1 are tested in parallel with scan cores 36 2 and 36 3 . Such a test could be implemented using Cores — te[ 3 : 0 ]=1111, Cores — bndryte[ 3 : 0 ]=0000 and Cores — pads — en[ 3 : 0 ]=0101.

In FIG. 5 c , a group of three scan cores 36 0 , 36 1 , and 36 2 , are tested in parallel with a single scan core 36 . Such a test could be implemented using Cores — te[ 3 : 0 ]=1111, Cores — bndryte[ 3 : 0 ]=0000 and Cores — pads — en[ 3 : 0 ]=1001.

In FIG. 5 d , scan cores 36 0 and 36 2 (with scan core 36 2 bypassed) are tested in parallel with scan core 36 3 . This test could be implemented using Cores — te[ 3 : 0 ]=1011, Cores — bndryte[ 3 : 0 ]=0000 and Cores — pads — en[ 3 : 0 ]=1001.

It should be noted that an almost unlimited variety of scan test structures could be implemented using the present invention; FIGS. 5 a–d illustrate only of few of these structures.

The present invention has significant advantages over the prior art. Most importantly, the invention allows optimization of scan test time to any tester being used. Thus, if different testers are used during different stages, the scan configuration can be dynamically configured for each tester. This reduces test time and hence reduces costs due to testing.

Further, if it is necessary, or desirable, to change testers over a term of production of a device, the new tester can be accommodated at its optimum configuration without redesign of the device.

Although the Detailed Description of the invention has been directed to certain exemplary embodiments, various modifications of these embodiments, as well as alternative embodiments, will be suggested to those skilled in the art. The invention encompasses any modifications or alternative embodiments that fall within the scope of the claims.