Multi-layer circuit board temperature sensing alarm control method and system

Abstract

The invention relates to the technical field of temperature monitoring and alarming of electronic equipment, in particular to a multilayer circuit board temperature sensing and alarming control method and system. The temperature sensor module comprises a plurality of micro temperature sensors embedded in a multilayer circuit board and is used for collecting temperature data of key parts in real time; utilizing a machine learning algorithm to generate an expected temperature distribution model, and dynamically adjusting an alarm threshold value; temperature deviation reasons are evaluated through an intelligent decision engine, and false alarms are reduced. Besides, the system also has the functions of adaptive learning, performance evaluation and external environmental factor feedback, ensures stable operation in different environments, and designs a redundant sensor network to enhance reliability. According to the invention, the effects of improving the temperature monitoring precision and timely finding and early warning potential faults are achieved.

Description

Technical Field

[0001] This application relates to the technical field of temperature monitoring and alarm for electronic devices, and in particular to a temperature-sensing alarm control method and system for multi-layer circuit boards. Background Art

[0002] In the design and operation of electronic devices, temperature management of circuit boards is crucial, especially in circuit boards with dense component layouts or multi-layer stacked designs. Excessive temperature not only leads to performance degradation but may also cause equipment failures, seriously affecting overall reliability and lifespan. With the increasing complexity of modern electronic devices, precise management and real-time monitoring of circuit board temperature become particularly important. Existing circuit board temperature monitoring technologies have to a certain extent met the basic temperature detection requirements, but there is still room for improvement in intelligent early warning and precise adjustment. To address the situation of abnormal circuit board temperature, the currently commonly used solution is to trigger an alarm through a fixed temperature threshold. Specifically, when the temperature of a certain area on the circuit board reaches the preset threshold, the system will automatically send an alarm signal. Common means include using a temperature sensor network to monitor the temperature of specific areas and implementing simple temperature comparison and alarm functions through hardware circuits. Another common method is to install temperature-sensitive components at key positions on the circuit board, and once the temperature exceeds the preset value, the circuit is triggered to issue an alarm. There are also some more advanced methods, such as using a single-chip microcomputer or microcontroller in cooperation with temperature sensors to collect and simply process temperature data through software, and then trigger an alarm. However, these traditional fixed-threshold methods have obvious deficiencies. First, the fixed temperature threshold cannot adapt to the actual heat generation characteristics of the circuit board under different working environments, resulting in a high false alarm rate. For example, in some cases, the circuit board may generate transient high temperature due to short-term high load, but this does not mean that there is a fault, so an alarm should not be triggered. Second, the setting of the fixed threshold is often relatively conservative, prone to alarm delay, and unable to detect potential faults in a timely manner. These problems seriously affect the reliability and effectiveness of the circuit board temperature monitoring system. Summary of the Invention

[0003] The purpose of this application is to overcome the above technical problems and provide a temperature-sensing alarm control method and system for multi-layer circuit boards A temperature-sensing alarm control method for a multi-layer circuit board, including but not limited to: a plurality of tiny temperature sensors embedded in the multi-layer circuit board, which are distributed at key parts of the circuit board; collecting real-time temperature data of the temperature sensors; using machine learning algorithms to analyze historical temperature data to generate an expected temperature distribution model of the circuit board under the current load; dynamically adjusting the alarm threshold according to the expected temperature distribution model under the current working conditions; when it is detected that the temperature in a certain area exceeds the expected temperature range and the duration exceeds a predetermined threshold, starting an early warning mechanism; using an intelligent decision-making engine to evaluate the cause of the temperature deviation, distinguishing between temporary fluctuations and potential faults, and reducing false alarms caused by misjudgment. By adopting the above technical solutions, the occurrence frequency of false alarms is significantly reduced, the effectiveness and response speed of the system early warning are improved; by using intelligent decision support and historical data analysis, real threats can be quickly identified, unnecessary maintenance actions are reduced, and operating costs are saved; the system performs better in complex environments and can maintain stable monitoring quality under harsh conditions such as high temperature and humidity, extending the service life of the equipment. Preferably, it further includes: in the system initialization stage, based on the temperature data calibrated in the factory and the knowledge base formulated by experts, setting preliminary temperature threshold parameters; during operation, continuously collecting temperature data in the actual working environment, and using reinforcement learning algorithms to gradually optimize the threshold to form a personalized alarm standard adapted to the current environment. By adopting the above technical solutions, it is possible to set preliminary temperature threshold parameters based on the temperature data calibrated in the factory and the knowledge base formulated by experts in the system initialization stage, ensuring that the system has a reasonable alarm benchmark during the initial operation. During operation, continuously collect temperature data in the actual working environment, and use reinforcement learning algorithms to gradually optimize the threshold, enabling the system to adapt to the specific changes in the current environment and form a personalized alarm standard, thereby improving the accuracy and effectiveness of the alarm and reducing the false alarm rate. Preferably, it further includes: regularly evaluating the model performance every month, by retrospectively verifying the historical data of the past month and comparing it with the newly collected data in real time, ensuring that the prediction accuracy and reaction speed are in the best state. By adopting the above technical solutions, regularly evaluating the model performance, retrospectively verifying the historical data of the past month and comparing it with the newly collected data in real time, ensuring that the prediction accuracy and reaction speed are in the best state, thereby enhancing the stability and reliability of the system during long-term operation. Preferably, it further includes: specifically considering the changes in environmental temperature and humidity, adjusting the temperature threshold calculation formula to improve the robustness of the system under different environmental conditions. By adopting the above technical solutions, specifically considering the changes in environmental temperature and humidity and adjusting the temperature threshold calculation formula can significantly enhance the robustness of the system under different environmental conditions, ensuring that stable monitoring quality and reliable alarm performance can still be maintained under extreme environments such as high temperature and humidity, thereby extending the service life of the equipment. Preferably, it further includes: designing at least two groups of redundant sensor networks, so that even if some sensors fail, the continuous operation of the system can still be guaranteed, ensuring the integrity of the data.By adopting the above technical solutions, at least two sets of redundant sensor networks are designed. Even if some sensors fail, the continuous operation of the system can still be ensured, and the integrity of data can be guaranteed. This solution enhances the stability and reliability of the system. Especially in complex and harsh working environments, it can effectively avoid the failure of the entire system caused by single-point failures, and ensure the persistence and accuracy of temperature monitoring data. A multi-layer circuit board temperature sensing and alarm control system includes: a plurality of micro temperature sensors embedded in the multi-layer circuit board and distributed at key parts of the circuit board; a central processing unit for collecting real-time temperature data of the temperature sensors and generating an expected temperature distribution model of the circuit board under the current load using machine learning algorithms; an alarm module for dynamically adjusting the alarm threshold according to the expected temperature distribution model under the current working conditions, and starting an early warning mechanism when it detects that the temperature in a certain area exceeds the expected temperature range and the duration exceeds a predetermined threshold; an intelligent decision-making module for evaluating the reasons for temperature deviation, distinguishing between temporary fluctuations and potential failures, and reducing false alarms caused by misjudgment. By adopting the above technical solutions, the alarm threshold can be dynamically adjusted according to the expected temperature distribution model under the current working conditions, and the early warning mechanism can be quickly started when the circuit board temperature is abnormal, reducing the occurrence of false alarms and improving the effectiveness and response speed of the system's early warning. At the same time, the intelligent decision-making module can evaluate the reasons for temperature deviation, distinguish between temporary fluctuations and potential failures, further reduce the false alarm rate, and ensure the stable operation of the system in complex and changing working environments. Preferably, it further includes: a data storage module for storing factory-calibrated temperature data, a knowledge base formulated by experts, and temperature data collected during operation; an adaptive learning module for gradually optimizing the temperature threshold using reinforcement learning algorithms to form a personalized alarm standard adapted to the current environment. By adopting the above technical solutions, the system can store factory-calibrated temperature data, a knowledge base formulated by experts, and temperature data collected during operation, ensuring the integrity and availability of the data. At the same time, using reinforcement learning algorithms to gradually optimize the temperature threshold makes the alarm standard more adapted to the current environment, improving the accuracy and reliability of the alarm. Preferably, it further includes: a performance evaluation module for regularly evaluating the model performance every month. By retrospectively verifying the historical data of the past month and comparing it with the newly collected data in real time, it ensures that the prediction accuracy and reaction speed are in the best state. By adopting the above technical solutions, the model performance can be regularly evaluated every month. By retrospectively verifying the historical data of the past month and comparing it with the newly collected data in real time, it ensures that the prediction accuracy and reaction speed are in the best state. This helps to maintain the long-term stability and reliability of the system, timely discover and correct possible deviations, and thus improve the accuracy and response efficiency of the alarm system. Preferably, it further includes: an external environmental factor feedback module for specifically considering changes in environmental temperature and humidity, adjusting the temperature threshold calculation formula, and improving the robustness of the system under different environmental conditions.By adopting the above technical solution, it is possible to specifically consider the changes in environmental temperature and humidity, adjust the temperature threshold calculation formula, and improve the robustness of the system under different environmental conditions. This improvement enables the system to maintain stable monitoring performance in high-temperature, humid, or other extreme environments, ensuring the accuracy and reliability of temperature alarms, thereby extending the service life of the equipment and enhancing the overall stability of the system. Preferably, it further includes at least two groups of redundant sensor networks, so that even if some sensors fail, the continuous operation of the system can still be ensured, and the integrity of the data can be guaranteed. By adopting the above technical solution, at least two groups of redundant sensor networks are added. Even if some sensors fail, the continuous operation of the system can still be ensured, and the integrity of the data can be guaranteed, thereby improving the stability and reliability of the system, avoiding monitoring interruptions caused by single sensor failures, and ensuring the continuous effectiveness of the multi-layer circuit board temperature sensing alarm control system.

[0004] In summary, the present application includes at least one of the following beneficial technical effects: 1. By dynamically adjusting the alarm threshold and based on the expected temperature distribution model under the current working conditions, the occurrence frequency of false alarms is significantly reduced, and the effectiveness and response speed of the system warning are improved; 2. Using the intelligent decision-making engine to evaluate the reasons for temperature deviation, distinguishing between temporary fluctuations and potential faults, reducing false alarms caused by misjudgment, and improving the accuracy and reliability of the system; 3. Considering the changes in environmental temperature and humidity, adjusting the temperature threshold calculation formula, enhancing the robustness of the system under different environmental conditions, and ensuring stable operation in complex and changeable working environments. Specific implementation manner

[0005] The technical solutions in the embodiments of the present invention will be described clearly and completely below. The described embodiments are only possible technical implementations of the present invention, not all possible implementations. Those skilled in the art can fully combine the embodiments of the present invention to obtain other embodiments without creative labor, and these embodiments are also within the protection scope of the present invention. The inventors of this application have found that in the design and operation of electronic devices, the temperature management of the circuit board is crucial. Especially in circuit boards with dense component layouts or multi-layer stacked designs, overheating may lead to performance degradation or even failures, thus affecting the overall reliability and lifespan of the device. Therefore, this application mainly adopts the following technical solutions: collect real-time temperature data through a network of tiny temperature sensors embedded in the multi-layer circuit board, analyze historical temperature data using machine learning algorithms to generate an expected temperature distribution model of the circuit board under the current load, dynamically adjust the alarm threshold based on the expected temperature distribution model under the current working conditions. When it is detected that the temperature in a certain area exceeds the expected temperature range and the duration exceeds a predetermined threshold, start the early warning mechanism. Use an intelligent decision-making engine to evaluate the cause of the temperature deviation, distinguish between temporary fluctuations and potential failures, and reduce false alarms caused by misjudgment. This solution significantly reduces the occurrence frequency of false alarms, improves the effectiveness and response speed of the system early warning, and at the same time can quickly identify real threats, reduces unnecessary maintenance actions, and saves operating costs. The following is a further detailed description of this application. Embodiment 1 A temperature-sensing alarm control method for a multi-layer circuit board provided by an embodiment of this application includes a plurality of tiny temperature sensors, a central processing unit, an alarm module, and an intelligent decision-making module. These components cooperate together to achieve real-time monitoring and intelligent alarm of abnormal temperatures in the multi-layer circuit board. Specifically, a plurality of tiny temperature sensors are embedded in the multi-layer circuit board and are distributed at key parts of the circuit board. These sensors can be thermistors, thermocouples, or other types of temperature-sensing elements. For example, a thermistor can be selected because it has the characteristics of high sensitivity and fast response; a thermocouple can also be selected, which is suitable for temperature measurement in high-temperature environments. These sensors are connected to the central processing unit through wires or wireless communication methods to ensure real-time transmission of temperature data. The central processing unit is responsible for collecting the real-time temperature data of these temperature sensors and using machine learning algorithms to generate an expected temperature distribution model of the circuit board under the current load. The central processing unit can be a high-performance microcontroller or a dedicated temperature monitoring chip. For example, an ARM Cortex-M series microcontroller can be selected, which has powerful data processing capabilities and low power consumption characteristics; a dedicated temperature monitoring chip such as the TMP102 of Texas Instruments can also be selected, which is specifically used for high-precision temperature measurement. The machine learning algorithm can be implemented in various ways, such as support vector machine (SVM), neural network, or random forest, etc. These algorithms can predict the normal temperature distribution of the circuit board under the current load based on historical temperature data.For example, deep learning frameworks such as TensorFlow or PyTorch can be used to generate a model that can accurately predict the temperature distribution of the circuit board by training a large amount of historical temperature data. The dynamic adjustment of the alarm threshold is achieved based on the expected temperature distribution model under the current working conditions. When the temperature in a certain area is detected to exceed the expected temperature range and the duration exceeds a predetermined threshold, the early warning mechanism is activated. The setting of the alarm threshold can be dynamically adjusted based on historical data and the current load conditions. For example, it can be set to increase or decrease a certain percentage from the upper and lower limits of the expected temperature, and the specific percentage can be flexibly adjusted according to the actual application scenario. The intelligent decision-making module is used to evaluate the reasons for the temperature deviation, distinguish between temporary fluctuations and potential faults, and reduce false alarms caused by misjudgment. The intelligent decision-making module can be a rule-based expert system or a complex machine learning model. For example, a Bayesian classifier can be used to determine whether the temperature deviation is caused by a short-term current surge or a long-term hidden danger based on historical maintenance logs and data from other sensors, so as to decide whether immediate cooling measures need to be taken or an emergency repair needs to be arranged. The implementation principle of this embodiment is as follows: real-time temperature data is collected through a network of tiny temperature sensors embedded in the multi-layer circuit board, and a machine learning algorithm is used to generate an expected temperature distribution model of the circuit board under the current load. The alarm threshold is dynamically adjusted based on the expected temperature distribution model under the current working conditions. When the temperature in a certain area is detected to exceed the expected temperature range and the duration exceeds a predetermined threshold, the early warning mechanism is activated. The intelligent decision-making module evaluates the reasons for the temperature deviation, distinguishes between temporary fluctuations and potential faults, and reduces false alarms caused by misjudgment. This method significantly reduces the occurrence frequency of false alarms, improves the effectiveness and response speed of the system early warning, and at the same time can quickly identify real threats, reduces unnecessary maintenance actions, and saves operating costs. Embodiment 2 The difference between this embodiment and the above embodiment is that a function of setting initial temperature threshold parameters based on factory-calibrated temperature data and a knowledge base formulated by experts is added during the system initialization phase. In addition, during operation, temperature data in the actual working environment is continuously collected, and a reinforcement learning algorithm is used to gradually optimize the threshold to form a personalized alarm standard adapted to the current environment. Specifically, the initial temperature threshold parameters can be set during the system initialization phase. These parameters can be based on factory-calibrated temperature data and a knowledge base formulated by experts. For example, the factory-calibrated temperature data can be obtained by testing the circuit board multiple times under standard laboratory conditions to ensure the accuracy and reliability of the data. The knowledge base formulated by experts contains rich experience data and fault cases, which can help the system better understand and predict the temperature behavior of the circuit board. The reinforcement learning algorithm is used to gradually optimize the temperature threshold during the system operation. For example, algorithms such as Q-learning or DQN can be used to gradually adjust the temperature threshold through continuous trial and error and feedback to make it more suitable for the actual working environment.This can ensure that the system maintains a high alarm accuracy and response speed under different working conditions. The implementation principle of this embodiment is as follows: In the system initialization stage, based on the temperature data calibrated in the factory and the knowledge base formulated by experts, preliminary temperature threshold parameters are set. During operation, the temperature data in the actual working environment is continuously collected, and the reinforcement learning algorithm is used to gradually optimize the threshold to form a personalized alarm standard adapted to the current environment. This method not only improves the initial setting accuracy of the system but also continuously optimizes during operation, ensuring the long-term stability and reliability of the system. Embodiment 3 The difference between this embodiment and the above embodiment is that a function of regularly evaluating the model performance monthly is added. By retrospectively verifying the historical data of the past month and comparing it with the newly collected data in real time, the prediction accuracy and response speed are ensured to be in the best state. Specifically, the performance evaluation module regularly evaluates the model performance monthly. This module can be an independent subsystem or integrated into the central processing unit. The main function of the performance evaluation module is to retrospectively verify the historical data of the past month and compare it with the newly collected data in real time to ensure that the prediction accuracy and response speed are in the best state. For example, statistical analysis methods such as mean squared error (MSE) or mean absolute error (MAE) can be used to evaluate the prediction accuracy of the model. If it is found that the model performance has declined, the model can be retrained or the parameters can be adjusted to restore its best state. Retrospective verification is carried out by analyzing the historical data of the past month. These historical data can be stored in the local data storage module or uploaded to the cloud server for centralized management. Through retrospective verification, it can be found that the model performs poorly under certain specific working conditions, and thus targeted optimization can be carried out. Real-time comparison is to compare the newly collected data with the historical data to ensure the real-time performance of the model. For example, the sliding window technique can be used to compare only the data of the most recent period each time to ensure that the model can adapt to new working condition changes in a timely manner. The implementation principle of this embodiment is as follows: Regularly evaluate the model performance monthly, retrospectively verify the historical data of the past month and compare it with the newly collected data in real time to ensure that the prediction accuracy and response speed are in the best state. This method can not only timely detect the decline in model performance but also ensure that the system always maintains good performance during long-term operation through targeted optimization. Embodiment 4 The difference between this embodiment and the above embodiment is that a function of specifically considering the changes in environmental temperature and humidity and adjusting the temperature threshold calculation formula is added. This can improve the robustness of the system under different environmental conditions. Specifically, the external environmental factor feedback module is used to specifically consider the changes in environmental temperature and humidity and adjust the temperature threshold calculation formula. This module can be an independent subsystem or integrated into the central processing unit. The main function of the external environmental factor feedback module is to obtain the environmental temperature and humidity data through sensors and incorporate them into the temperature threshold calculation formula. For example, temperature and humidity sensors such as DHT11 or DHT22 can be used to monitor the environmental temperature and humidity in real time.These data can be transmitted to the central processing unit by wired or wireless means. The temperature threshold calculation formula can be adjusted according to the changes in ambient temperature and humidity. For example, methods such as linear regression or polynomial regression can be used to establish a relationship model between the temperature threshold and ambient temperature and humidity. The specific formula can be expressed as: \[ T_{\text{threshold}} = f(T_{\text{ambient}}, H) \] where \( T_{\text{threshold}} \) is the temperature threshold, \( T_{\text{ambient}} \) is the ambient temperature, \( H \) is the ambient humidity, and \( f \) is a function that can be adjusted according to the actual application scenario. The implementation principle of this embodiment is: specifically considering the changes in ambient temperature and humidity, adjusting the temperature threshold calculation formula, and improving the robustness of the system under different environmental conditions. This method can not only adapt to different environmental conditions, but also improve the reliability and stability of the system, ensuring effective operation under various harsh conditions. Embodiment 5 The difference between this embodiment and the above embodiments is that at least two sets of redundant sensor networks are added, so that even if some sensors fail, the continuous operation of the system can still be guaranteed, ensuring the integrity of the data. Specifically, the redundant sensor network designs at least two sets of sensor networks, so that even if some sensors fail, the continuous operation of the system can still be guaranteed. These two sets of sensor networks can be distributed in different areas of the circuit board to ensure the comprehensiveness and integrity of the data. For example, the first set of sensors can be distributed in the core area of the circuit board, and the second set of sensors can be distributed in the edge area. Each set of sensors includes multiple micro temperature sensors, and these sensors are connected to the central processing unit through wires or wireless communication. The redundant design ensures the high availability and fault tolerance of the system. When a certain sensor in a set of sensors fails, the other set of sensors can continue to work, ensuring the continuity and integrity of the data. For example, redundant communication protocols such as CAN bus or RS485 can be used to ensure the reliability of data transmission. Data integrity is achieved through redundant design. When there are problems with the data of a set of sensors, the system can switch to the other set of sensors to ensure the integrity and accuracy of the data. For example, data fusion technology can be used to perform weighted averaging on the data of the two sets of sensors to improve the reliability and accuracy of the data. The implementation principle of this embodiment is: designing at least two sets of redundant sensor networks, so that even if some sensors fail, the continuous operation of the system can still be guaranteed, ensuring the integrity of the data. This method not only improves the high availability and fault tolerance of the system, but also ensures stable operation in various complex environments, extending the service life of the device. The above are all preferred embodiments of this application. It does not limit the protection scope of this application accordingly. Therefore, all equivalent changes made according to the structure, shape, and principle of this application should be covered within the protection scope of this application.

Claims

1. A multi-layer circuit board temperature alarm control method, characterized in that: Including but not limited to: multiple tiny temperature sensors embedded in the multi-layer circuit board, which are distributed in key parts of the circuit board; collecting real-time temperature data of the temperature sensors; using machine learning algorithms to analyze historical temperature data and generate an expected temperature distribution model of the circuit board under the current load; dynamically adjusting the alarm threshold based on the expected temperature distribution model under current working conditions; when it is detected that the temperature of a certain area exceeds the expected temperature range and the duration exceeds the predetermined threshold, the early warning mechanism is activated; using an intelligent decision-making engine to evaluate the cause of temperature deviation, distinguish temporary fluctuations from potential failures, and reduce false alarms due to misjudgment.

2. The multi-layer circuit board temperature alarm control method according to claim 1, characterized in that: Also includes: During the system initialization phase, preliminary temperature threshold parameters are set based on factory-calibrated temperature data and a knowledge base developed by experts. During operation, temperature data from the actual working environment is continuously collected, and the threshold is gradually optimized using a reinforcement learning algorithm to form a personalized alarm standard that adapts to the current environment.

3. The multi-layer circuit board temperature sensing alarm control method according to claim 2 is characterized in that: Also includes: Model performance is evaluated regularly every month by back-verifying historical data from the past month and comparing newly collected data in real time to ensure that prediction accuracy and response speed are at their best.

4. The multi-layer circuit board temperature sensing alarm control method according to claim 1, characterized in that: Also includes: Specifically consider the changes in ambient temperature and humidity, adjust the temperature threshold calculation formula, and improve the robustness of the system under different environmental conditions.

5. The multi-layer circuit board temperature sensing alarm control method according to claim 1, characterized in that: Also includes: Design at least two sets of redundant sensor networks to ensure continuous operation of the system and data integrity even if some sensors fail.

6. A multi-layer circuit board temperature alarm control system, characterized in that: include: A plurality of tiny temperature sensors are embedded in the multi-layer circuit board and distributed in key parts of the circuit board; a central processing unit is used to collect real-time temperature data of the temperature sensors and generate an expected temperature distribution model of the circuit board under the current load using a machine learning algorithm; an alarm module is used to dynamically adjust the alarm threshold according to the expected temperature distribution model under the current working conditions, and activate the early warning mechanism when it is detected that the temperature of a certain area exceeds the expected temperature range and the duration exceeds the predetermined threshold; Intelligent decision-making module to evaluate the causes of temperature deviations, distinguish temporary fluctuations from potential failures, and reduce false alarms due to misjudgment.

7. The multi-layer circuit board temperature sensing alarm control system according to claim 6, characterized in that: Also includes: A data storage module for storing factory-calibrated temperature data, an expert-developed knowledge base, and temperature data collected during operation; The adaptive learning module is used to gradually optimize the temperature threshold using a reinforcement learning algorithm to form a personalized alarm standard that adapts to the current environment.

8. The multi-layer circuit board temperature sensing alarm control system according to claim 7, characterized in that: It also includes: a performance evaluation module that regularly evaluates model performance every month, back-checks historical data from the past month and compares newly collected data in real time to ensure that prediction accuracy and response speed are at their best.

9. The multi-layer circuit board temperature sensing alarm control system according to claim 6, characterized in that: Also includes: The external environmental factor feedback module is used to specifically consider changes in ambient temperature and humidity, adjust the temperature threshold calculation formula, and improve the robustness of the system under different environmental conditions.

10. The multi-layer circuit board temperature sensing alarm control system according to claim 6, characterized in that: Also includes: At least two sets of redundant sensor networks can ensure the continuous operation of the system and the integrity of the data even if some sensors fail.