Self-checking method of multimode redundancy dual-state machine supporting logic built-in self-testing

Through the multi-mode redundant dual-state machine self-test method, the state machine takes turns to perform LBIST and voting modules to determine the flow state, which solves the problem that traditional dual-state machines cannot support LBIST, improves the reliability and detection capabilities of the state machine, and ensures that Safe transition in case of failure.

CN119002344APending Publication Date: 2024-11-22SUZHOU HUAXIN SEMICON TECH CO LTD
View PDF 0 Cites 1 Cited by

Patent Information

Application Number
CN202411076332.8
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-07
Publication Date
2024-11-22

AI Technical Summary

Technical Problem

Traditional digital circuit dual-state machines cannot effectively support built-in logic self-testing in application scenarios with high reliability requirements, and may lead to serious consequences when logic circuit failures, and cannot perform LBIST at any time.

Method used

The multi-mode redundant dual-state machine self-test method is adopted, and the LBIST is performed in turn through state machine A and state machine B, and the voting module is used to determine the flow state to ensure that the reliability of the state machine can be effectively detected and guaranteed at any time.

Benefits of technology

The state machine circuit is LBIST at any time, which improves the accuracy and reliability of state machine flow, ensures safe conversion in the event of a failure, and avoids serious consequences.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119002344A_ABST
    Figure CN119002344A_ABST
Patent Text Reader

Abstract

A self-checking method of a multimode redundancy dual-state machine supporting logic built-in self-testing comprises a state machine A, a state machine B and a voting module, and the reliability of state codes and key control signals of each set of state machine is reinforced through multimode redundancy. When the two sets of state machines have normal working permissions, the circulation is completely independent, and output is carried out through the voting module; when one set of state machine and the voting logic related to the state machine in the voting module are subjected to LBIST, the self-check overturning of the related combinational logic and the sequential logic does not influence the independent circulation of the other set of state machine with the normal working authority and determines the output of the whole state machine. When one set of state machine does not pass the normal work permission due to the fact that the LBIST does not pass, and only the other set of state machine has the normal work permission, the state machine with the normal work permission directly determines the output of the whole state machine, and LBIST can be carried out on the state machine only when the state machine is in a relatively safe state.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of integrated circuit safety, and in particular to a self-test method for a multi-mode redundant dual-state machine supporting logic built-in self-test. Background Art

[0002] In application scenarios with higher reliability requirements, such as driverless taxis, the digital state machine in the chip may be in normal working mode for a very long time and control the key signals of the system. The extension of time also greatly increases the probability of errors in the state machine. On the other hand, when the long-term operation of the logic circuit is approaching its own life span, the probability of error flipping will gradually increase. In order to avoid unpredictable serious consequences caused by the failure of the logic circuit of the chip, the chip needs to support LBIST (Logic Built-In Self-Test) at any time, that is, to allow the digital logic circuit (or part of it) on the chip to detect whether this part of the logic circuit has a fault through self-testing at the appropriate time when the chip is working.

[0003] Although the traditional single-state machine of digital circuit can reduce the probability of state machine flow errors caused by single particle radiation or accidental interference through dual-mode redundancy or triple-mode redundancy of state coding, the state machine itself must always remain valid and cannot support the flipping required by LBIST. Therefore, serious consequences may occur when the logic in the state machine circuit itself fails.

[0004] There are generally two ways to implement the traditional dual-state machine of digital circuits. One is called standby redundancy, that is, one of the dual-state machines is in normal operation and the other is in standby state. When the entire digital circuit is in normal working mode, they will not be enabled at the same time. Although this method can save power consumption by using one of the state machines as a standby, in application scenarios with higher reliability requirements, the enabling time of the standby state machine itself will be an unacceptable problem. Another dual-state machine circuit is called running redundancy. They work at the same time, but one of the state machines is responsible for leading the state flow of the entire digital circuit (called the leading state machine), and the other is also in operation, but is only responsible for monitoring whether the work of the aforementioned leading state machine has a fault (called the monitoring state machine). Once the leading state machine fault is detected, the monitoring state machine can change its role in a very short time and take over. However, in this design, a single state machine is not reinforced with multi-mode redundancy, and the other state machine is only used for monitoring, rather than two sets of state machines acting in the same role. In addition, both of the above two dual-state machines cannot well support LBIST at any time. Summary of the invention

[0005] In view of the deficiencies of the prior art, the present invention provides a self-test method for a multi-mode redundant dual-state machine supporting logic built-in self-test, which overcomes the deficiencies of the prior art and can effectively ensure the reliability of the normal operation of the entire state machine, while supporting LBIST of the state machine circuit at any time.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions:

[0007] A self-test method for a multi-mode redundant dual-state machine supporting logic built-in self-test, comprising a state machine A, a state machine B and a voting module, wherein both the state machine A and the state machine B flow according to the current state output by the voting module to output the state of the next clock and a key control signal;

[0008] The self-test method includes the following steps:

[0009] After the power-on reset stage, the state machine A and the state machine B are automatically triggered to perform LBIST in turn. At the same time, the circuits related to the state machine A and the state machine B in the voting module also perform LBIST in turn.

[0010] Step S1: After the power-on reset stage, the state machine A and the circuits related to the state machine A in the voting module are automatically triggered to perform LBIST;

[0011] Step S2: Determine whether the LBIST of state machine A passes, if yes, proceed to step S4; if no, proceed to step S3;

[0012] Step S3: Determine whether the number of LBIST failures of state machine A reaches a threshold; if so, proceed to step S4; if not, wait for a preset time and return to step S1;

[0013] Step S4: automatically triggering the state machine B and the circuits related to the state machine B in the voting module to perform LBIST;

[0014] Step S5: Determine whether the LBIST of state machine B passes, if yes, proceed to step S7; if no, proceed to step S6;

[0015] Step S6: Determine whether the number of LBIST failures of state machine B reaches a threshold; if yes, proceed to step S7; if no, wait for a preset time and return to step S4;

[0016] Step S7: Determine whether the LBIST of state machine A and state machine B are both passed. If so, grant normal working rights to both state machine A and state machine B, and start the normal flow of the state machine; if not, proceed to step S8;

[0017] Step S8: Determine whether one of state machine A and state machine B passes LBIST. If yes, the state machine that passes LBIST is granted normal working authority, and the state machine that passes LBIST determines the flow of the entire chip state machine. The state machine that fails LBIST is set as abnormal working authority, and the processor is notified through an interrupt, and the processor is allowed to read the LBIST result and perform subsequent operations; if no, both state machine A and state machine B enter fail-safe mode.

[0018] Preferably, it also includes a counter and a threshold register, wherein the counter is used to count the number of incomplete matches that occur when state machine A and state machine B are both in normal working authority and flow independently; the threshold register has a built-in threshold, and when the number of times counted by the counter reaches the threshold set by the threshold register, state machine A and state machine B and circuits related to state machine A and state machine B in the voting module are started in turn to perform LBIST.

[0019] Preferably, the step S7, when the LBIST of the state machine A and the state machine B are both passed, the state machine A and the state machine B are both in normal working authority and the flow of the state machine is started, further comprises the following steps:

[0020] Step S71: Record the number of times the flow of state machine A and state machine B do not match;

[0021] Step S72: Determine whether the number of mismatches between state machine A and state machine B reaches a threshold. If yes, proceed to step S74; if no, proceed to step S73;

[0022] Step S73: Determine whether the state machine as a whole enters or leaves a specific state to trigger the LBIST; if so, proceed to step S74; if not, return to step S71;

[0023] Step S74: state machine A and circuits in the voting module related to state machine A perform LBIST;

[0024] Step S75: Determine whether the LBIST of state machine A passes, if yes, proceed to step S77; if no, proceed to step S76;

[0025] Step S76: Determine whether the number of LBIST failures of state machine A reaches a threshold; if so, set state machine A to abnormal working authority, and let state machine B alone determine the flow of the entire chip state machine, and use an interrupt to notify the processor and allow it to query the LBIST result; if not, return to step S74;

[0026] Step S77: restore the normal working authority of state machine A, and make state machine B and the circuits related to state machine B in the voting module perform LBIST;

[0027] Step S78: Determine whether the LBIST of state machine B passes, if yes, proceed to step S710; if no, proceed to step S79;

[0028] Step S79: Determine whether the number of LBIST failures of state machine B reaches a threshold; if so, set state machine B to abnormal working authority, and let state machine A alone determine the flow of the entire chip state machine, and use an interrupt to notify the processor and allow it to query the LBIST result; if not, return to step S77;

[0029] Step S710: Restore the normal working authority of state machine B, clear the number of mismatched flows between state machine A and state machine B, notify the processor through an interrupt, and allow the processor to read the LBIST result and perform subsequent operations; then return to step S71.

[0030] Preferably, the step S8, when it is determined that only one of the state machines A and B passes the LBIST and has normal working authority, further comprises the following steps:

[0031] Step S81: granting the state machine that passes the LBIST normal working authority, and making the state machine that passes the LBIST determine the flow of the entire chip state machine, and setting the state machine that fails the LBIST as an abnormal working authority;

[0032] Step S82: determine whether the state of the state machine of the normal working authority is capable of performing LBIST; if so, proceed to step S83; if not, return to step S81;

[0033] Step S83: determine whether the processor triggers LBIST; if so, proceed to step S85; if not, return to step S84;

[0034] Step S84: determine whether the state machine triggers LBIST due to entering or leaving a specific state; if so, proceed to step S85; if not, return to step S81;

[0035] Step S85: performing LBIST on the state machine in normal working authority and the circuits related thereto in the voting module;

[0036] Step S86: Determine whether the LBIST of the state machine has passed, if so, return to step S81; if not, proceed to step S87;

[0037] Step S87: Determine whether the number of LBIST failures of the state machine reaches a threshold; if so, set the state machine to abnormal working authority, and switch both state machine A and state machine B of the entire chip into fail-safe mode; if not, return to step S85.

[0038] Preferably, the state codes and key control signals of the state machine A and the state machine B are respectively reinforced with multi-mode redundancy for reliability.

[0039] The invention provides a self-test method for a multi-mode redundant dual-state machine supporting a logic built-in self-test. It has the following beneficial effects: by setting two sets of state machines, and when both sets of state machines are in a normal working mode, both sets of state machines have the same authority, each set of state machines uses the output of a voting module as the current state to circulate completely independently, and the reinforcement effect of the dual multi-mode redundancy on the state machine will further improve the correctness of the state machine circulation. And in the LBIST mode, one set of state machines continues to maintain normal operation and directly determines the state output, while the voting logic in the other set of state machines and its voting module enters the LBIST mode, so when the other set of state machines performs LBIST, the flipping of the relevant combination logic and the sequential logic will not affect the entire circuit. Then the reliability of the normal operation of the entire state machine is guaranteed, and at the same time, the state machine circuit is supported to perform LBIST at any time. In addition, when the chip life is near the end, even if one set of state machines cannot work normally due to a fault, it will give the entire system enough time to enter a relatively safe state without serious problems. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to more clearly illustrate the present invention or the technical solutions in the prior art, the drawings required for describing the prior art are briefly introduced below.

[0041] Figure 1 Operation flow chart of LBIST after power-on reset in the present invention;

[0042] Figure 2 Operation flow chart of LBIST when both state machine A and state machine B are in normal working authority in the present invention;

[0043] Figure 3 The operation flow of the LBIST in the present invention is when only one state machine has normal working authority. DETAILED DESCRIPTION

[0044] In order to make the purpose, technical solutions and advantages of the present invention more clear, the technical solutions of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of the present invention.

[0045] Embodiment 1, as Figure 1-3 As shown, a self-test method of a multi-mode redundant dual-state machine supporting logic built-in self-test includes a state machine A, a state machine B and a voting module. Both the state machine A and the state machine B flow according to the current state output by the voting module to output the state of the next clock and the key control signal;

[0046] The self-test method includes the following steps:

[0047] After the power-on reset stage, the state machine A and the state machine B are automatically triggered to perform LBIST in turn. At the same time, the circuits related to the state machine A and the state machine B in the voting module also perform LBIST in turn.

[0048] Step S1: After the power-on reset stage, the state machine A and the circuits related to the state machine A in the voting module are automatically triggered to perform LBIST;

[0049] Step S2: Determine whether the LBIST of state machine A passes, if yes, proceed to step S4; if no, proceed to step S3;

[0050] Step S3: Determine whether the number of LBIST failures of state machine A reaches a threshold; if so, proceed to step S4; if not, wait for a preset time and return to step S1;

[0051] Step S4: automatically triggering the state machine B and the circuits related to the state machine B in the voting module to perform LBIST;

[0052] Step S5: Determine whether the LBIST of state machine B passes, if yes, proceed to step S7; if no, proceed to step S6;

[0053] Step S6: Determine whether the number of LBIST failures of state machine B reaches a threshold; if yes, proceed to step S7; if no, wait for a preset time and return to step S4;

[0054] Step S7: Determine whether the LBIST of state machine A and state machine B are both passed. If so, grant normal working rights to both state machine A and state machine B, and start the normal flow of the state machine; if not, proceed to step S8;

[0055] Step S8: Determine whether one of state machine A and state machine B passes LBIST. If yes, the state machine that passes LBIST is granted normal working authority, and the state machine that passes LBIST determines the flow of the entire chip state machine. The state machine that fails LBIST is set as abnormal working authority, and the processor is notified through an interrupt, and the processor is allowed to read the LBIST result and perform subsequent operations; if no, both state machine A and state machine B enter fail-safe mode.

[0056] In this embodiment, the state encoding of state machine A and state machine B both use traditional dual-mode redundancy or triple-mode redundancy for reliability reinforcement, and key control signals related to the state machine flow can also use traditional dual-mode redundancy or triple-mode redundancy.

[0057] The working mode of the voting module is:

[0058] a) When state machine A and state machine B are both in normal working order: the voting module votes on the final output based on the output states of state machine A and state machine B.

[0059] When the state codes in state machine A and state machine B use dual-mode redundancy, that is, each bit in the non-redundant state code has a corresponding redundant bit, the two state machines have a total of four bits eligible to participate in the vote. The voting module can vote in one of the following two ways:

[0060] The first type: All four of them participate in the vote, and the minority obeys the majority. However, when the two-to-two situation occurs, it means that two of the four have failed, and the state machine should enter the fail-safe mode. The fail-safe mode is different from the normal working mode, and the key control signal is fixed to a value that can make the system safe.

[0061] The second method is to select three of the above four for voting and adopt the traditional voting method of minority obeys majority. The three can be selected in a fixed way, that is, two of the state machine A and a fixed one of the state machine B, or a fixed one of the state machine A and two of the state machine B; or a dynamic method can be adopted, that is, three are selected from the four in turn or randomly for voting each time.

[0062] When the state codes in state machine A and state machine B use triple-module redundancy, that is, each bit in the non-redundant state code has two corresponding redundant bits, the two state machines have a total of six bits eligible to participate in the vote. The voting module can use one of the following four methods:

[0063] The first one: All the above six people participate in the voting, and the minority obeys the majority. However, when a three-to-three situation occurs, it means that two of the six people have failed. At this time, the state machine should enter the fail-safe mode.

[0064] The second method is to select five of the above six digits for voting and adopt the traditional voting method of minority obeys majority. The selection of the five digits can be fixed, such as three digits of state machine A and two fixed digits of state machine B, or two fixed digits of state machine A and three digits of state machine B; or a dynamic method can be adopted, that is, five digits are selected from the six digits in turn or randomly for voting each time.

[0065] The third method: Take four of the above six bits for voting, and adopt the voting method of minority obeys majority. However, when a two-to-two situation occurs, it means that two of the four bits have failed, and the state machine should enter the fail-safe mode. The selection of the four bits can be fixed, that is, a certain N bits of state machine A and a certain 4-N bits of state machine B, N = 1, 2, 3; or a dynamic method can be adopted, that is, four bits are selected from the six bits in turn or randomly for voting each time.

[0066] The fourth method: three of the six above are selected for voting, and the traditional minority obeys the majority voting method is adopted. The three bits can be selected in a fixed way, that is, a certain N bits of state machine A and a certain 4-N bits of state machine B, N = 1, 2, 3; or a dynamic way can be adopted, that is, three bits are selected from the six bits in turn or randomly for voting each time.

[0067] b) The above method can also be used to vote on key control signals related to the state machine flow.

[0068] c) When state machine A is in normal working authority and state machine B performs LBIST, the state output of state machine A in normal working mode and key control signal are directly output as current output state and current key signal output by multiplexer in the voting module. At this moment, the multiplexer except being relevant to state machine A in the voting module also performs LBIST with state machine B.

[0069] d) When state machine B is in normal working authority, when state machine A performs LBIST, the state output of state machine B in normal working mode and key control signal are directly output as current output state and current key signal output by multiplexer in the voting module. At this moment, other multiplexers except the multiplexer relevant to state machine B in the voting module also perform LBIST with state machine A.

[0070] Embodiment 2. In this embodiment, a counter and a threshold register are implemented in the chip, and the counter is used to count the number of incomplete matches that occur when state machine A and state machine B are both in normal working authority and flow independently; the threshold register has a built-in threshold, and when the number of times counted by the counter reaches the threshold set by the threshold register, state machine A and state machine B and circuits related to state machine A and state machine B in the voting module are started in turn to perform LBIST.

[0071] like Figure 2 As shown, when the LBIST of state machine A and state machine B are both passed, state machine A and state machine B are both in normal working authority and start the flow of state machine, the following steps are also included:

[0072] Step S71: Implement a counter in the chip to count the number of incomplete matches that occur when both state machine A and state machine B are independently circulating in normal working authority;

[0073] Step S72: Determine whether the number of mismatches between state machine A and state machine B reaches a threshold. If yes, proceed to step S74; if no, proceed to step S73;

[0074] Step S73: Determine whether the state machine as a whole enters or leaves a specific state to trigger the LBIST; if so, proceed to step S74; if not, return to step S71;

[0075] Step S74: state machine A and circuits in the voting module related to state machine A perform LBIST;

[0076] Step S75: Determine whether the LBIST of state machine A passes, if yes, proceed to step S77; if no, proceed to step S76;

[0077] Step S76: Determine whether the number of LBIST failures of state machine A reaches a threshold; if so, set state machine A to abnormal working authority, and let state machine B alone determine the flow of the entire chip state machine, and use an interrupt to notify the processor and allow it to query the LBIST result; if not, return to step S4;

[0078] Step S77: restore the normal working authority of state machine A, and make state machine B and the circuits related to state machine B in the voting module perform LBIST;

[0079] Step S78: Determine whether the LBIST of state machine B passes, if yes, proceed to step S710; if no, proceed to step S79;

[0080] Step S79: Determine whether the number of LBIST failures of state machine B reaches a threshold; if so, set state machine B to abnormal working authority, and let state machine A alone determine the flow of the entire chip state machine, and use an interrupt to notify the processor and allow it to query the LBIST result; if not, return to step S77;

[0081] Step S710: Restore the normal working authority of state machine B, clear the number of mismatched flows between state machine A and state machine B, notify the processor through an interrupt, and allow the processor to read the LBIST result and perform subsequent operations; then return to step S71.

[0082] Embodiment three, as Figure 3 As shown, when it is determined that only one of the state machines A and B passes the LBIST and has normal working authority, the following steps are also included:

[0083] Step S81: granting the state machine that passes the LBIST normal working authority, and making the state machine that passes the LBIST determine the flow of the entire chip state machine, and setting the state machine that fails the LBIST as an abnormal working authority;

[0084] Step S82: determine whether the state of the state machine of the normal working authority is capable of performing LBIST; if so, proceed to step S83; if not, return to step S81;

[0085] Step S83: determine whether the processor triggers LBIST; if so, proceed to step S85; if not, return to step S84;

[0086] Step S84: determine whether the state machine triggers LBIST due to entering or leaving a specific state; if so, proceed to step S85; if not, return to step S81;

[0087] Step S85: performing LBIST on the state machine in normal working authority and the circuits related thereto in the voting module;

[0088] Step S86: Determine whether the LBIST of the state machine has passed, if so, return to step S81; if not, proceed to step S87;

[0089] Step S87: Determine whether the number of LBIST failures of the state machine reaches a threshold; if so, set the state machine to abnormal working authority, and switch both state machine A and state machine B of the entire chip into fail-safe mode; if not, return to step S85.

[0090] Through the above-mentioned multi-mode redundant dual state machine self-test method, when state machine A and state machine B are both in normal working authority, state machine A and state machine B have the same authority, and each state machine uses the output of the voting module as the current state for completely independent flow. The dual multi-mode redundancy will further improve the correctness of the state machine flow by reinforcing the state machine.

[0091] In addition, when in the state transfer stage, when one of the state machines is in LBIST mode, the multi-mode redundancy of a single state machine can also provide protection during this period. And in LBIST mode, one of the state machines continues to work normally and directly determines the state output, while the voting logic in the other state machine and its voting module enters LBIST mode, so when the other state machine performs LBIST, the flip of the related combinational logic and sequential logic will not affect the entire circuit. Then the reliability of the normal operation of the entire state machine is guaranteed, and at the same time, the state machine circuit is supported to perform LBIST at any time. In addition, when the chip life is nearing the end, even if one state machine cannot work normally due to a fault, it will give the entire system enough time to enter a relatively safe state without serious problems.

[0092] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A self-test method for a multi-mode redundant dual-state machine supporting logic built-in self-test, Features: It includes state machine A, state machine B and a voting module. State machine A and state machine B are transferred according to the current state output by the voting module to output the state of the next clock and key control signals. The self-test method includes the following steps: After the power-on reset stage, the state machine A and the state machine B are automatically triggered to perform LBIST in turn. At the same time, the circuits related to the state machine A and the state machine B in the voting module also perform LBIST in turn. Step S1: After the power-on reset stage, the state machine A and the circuits related to the state machine A in the voting module are automatically triggered to perform LBIST; Step S2: Determine whether the LBIST of state machine A passes, if yes, proceed to step S4; if no, proceed to step S3; Step S3: Determine whether the number of LBIST failures of state machine A reaches a threshold; if so, proceed to step S4; if not, wait for a preset time and return to step S1; Step S4: automatically triggering the state machine B and the circuits related to the state machine B in the voting module to perform LBIST; Step S5: Determine whether the LBIST of state machine B passes, if yes, proceed to step S7; if no, proceed to step S6; Step S6: Determine whether the number of LBIST failures of state machine B reaches a threshold; if yes, proceed to step S7; if no, wait for a preset time and return to step S4; Step S7: Determine whether the LBIST of state machine A and state machine B are both passed. If so, grant normal working rights to both state machine A and state machine B, and start the normal flow of the state machine; if not, proceed to step S8; Step S8: Determine whether one of state machine A and state machine B passes LBIST. If yes, the state machine that passes LBIST is granted normal working authority, and the state machine that passes LBIST determines the flow of the entire chip state machine. The state machine that fails LBIST is set as abnormal working authority, and the processor is notified through an interrupt, and the processor is allowed to read the LBIST result and perform subsequent operations; if no, both state machine A and state machine B enter fail-safe mode.

2. A self-test method for a multi-mode redundant dual-state machine supporting logic built-in self-test according to claim 1, Features: The invention also includes a counter and a threshold register, wherein the counter is used to count the number of incomplete matches when the state machine A and the state machine B are both in normal working authority and flow independently; the threshold register has a built-in threshold, and when the number of times counted by the counter reaches the threshold set by the threshold register, the state machine A and the state machine B and the circuits related to the state machine A and the state machine B in the voting module are started in turn to perform LBIST.

3. A self-test method for a multi-mode redundant dual-state machine supporting logic built-in self-test according to claim 2, Features: The step S7, when both the LBIST of the state machine A and the state machine B are passed, the state machine A and the state machine B are both in normal working authority and start the flow of the state machine, further comprises the following steps: Step S71: Record the number of times the flow of state machine A and state machine B do not match; Step S72: Determine whether the number of mismatches between state machine A and state machine B reaches a threshold. If yes, proceed to step S74; if no, proceed to step S73; Step S73: Determine whether the state machine as a whole enters or leaves a specific state to trigger the LBIST; if so, proceed to step S74; if not, return to step S71; Step S74: state machine A and circuits in the voting module related to state machine A perform LBIST; Step S75: Determine whether the LBIST of state machine A passes, if yes, proceed to step S77; if no, proceed to step S76; Step S76: Determine whether the number of LBIST failures of state machine A reaches a threshold; if so, set state machine A to abnormal working authority, and let state machine B alone determine the flow of the entire chip state machine, and use an interrupt to notify the processor and allow it to query the LBIST result; if not, return to step S74; Step S77: restore the normal working authority of state machine A, and make state machine B and the circuits related to state machine B in the voting module perform LBIST; Step S78: Determine whether the LBIST of state machine B passes, if yes, proceed to step S710; if no, proceed to step S79; Step S79: Determine whether the number of LBIST failures of state machine B reaches a threshold; if so, set state machine B to abnormal working authority, and let state machine A alone determine the flow of the entire chip state machine, and use an interrupt to notify the processor and allow it to query the LBIST result; if not, return to step S77; Step S710: Restore the normal working authority of state machine B, clear the number of mismatched flows between state machine A and state machine B, notify the processor through an interrupt, and allow the processor to read the LBIST result and perform subsequent operations; then return to step S71.

4. The self-test method of a multi-mode redundant dual-state machine supporting logic built-in self-test according to claim 1, Features: The step S8, when it is determined that only one of the state machines A and B passes the LBIST and has normal working authority, also includes the following steps: Step S81: granting the state machine that passes the LBIST normal working authority, and making the state machine that passes the LBIST determine the flow of the entire chip state machine, and setting the state machine that fails the LBIST as an abnormal working authority; Step S82: determine whether the state of the state machine of the normal working authority is capable of performing LBIST; if so, proceed to step S83; if not, return to step S81; Step S83: determine whether the processor triggers LBIST; if so, proceed to step S85; if not, return to step S84; Step S84: determine whether the state machine triggers LBIST due to entering or leaving a specific state; if so, proceed to step S85; if not, return to step S81; Step S85: performing LBIST on the state machine in normal working authority and the circuits related thereto in the voting module; Step S86: Determine whether the LBIST of the state machine has passed, if so, return to step S81; if not, proceed to step S87; Step S87: Determine whether the number of LBIST failures of the state machine reaches a threshold; if so, set the state machine to abnormal working authority, and switch both state machine A and state machine B of the entire chip into fail-safe mode; if not, return to step S85.

5. The self-test method of a multi-mode redundant dual-state machine supporting logic built-in self-test according to claim 1, Features: The state codes and key control signals of the state machine A and the state machine B are respectively reinforced with multi-mode redundancy for reliability.

Citation Information

Cited By

  • State restorer of logic controller

    CN120406291A