Detection circuit for locked state of heat sink in server and server

Abstract

The application provides a detection circuit for a locked state of a heat sink in a server and a server, and relates to the technical fields of server heat dissipation detection and electronic circuit detection. The detection circuit comprises an asymmetric Wheatstone bridge unit, a signal amplification unit and a baseboard management controller. The asymmetric Wheatstone bridge unit comprises a plurality of force-sensitive resistors and a zero-setting resistor. The plurality of force-sensitive resistors are respectively arranged at each locking point of the heat sink, and are used for converting pressure changes of each locking point into resistance value changes. The zero-setting resistor is used for setting the output voltage of the bridge to zero when the heat sink is normally locked. The signal amplification unit is used for amplifying the differential voltage signal output by the asymmetric Wheatstone bridge unit and biasing the differential voltage signal into a single-ended positive voltage signal. The baseboard management controller is used for judging whether the heat sink is locked according to the voltage amplitude of the single-ended positive voltage signal. The application effectively reduces the circuit complexity and hardware cost by using a single bridge to replace multiple independent sensor acquisitions.

Description

Technical Field

[0001] This application relates to the fields of server heat dissipation detection and electronic circuit detection technology, and in particular to a detection circuit for the locking state of the heat sink in a server and a server. Background Technology

[0002] As the core hardware of data centers, servers' stable operation directly depends on an efficient cooling system. In high-density server racks, critical components such as the Central Processing Unit (CPU) work together with heat sinks, heat pipes, and fans to quickly dissipate heat. The tightness of the heat sink is crucial for ensuring efficient heat transfer: if the heat sink is not evenly attached to the CPU surface or the tightness is insufficient, the thermal interface material (TIM) will fail, reducing heat dissipation efficiency and potentially leading to CPU overheating, performance throttling, or even hardware damage. How to quickly and accurately determine whether the heat sink is properly secured without disassembling the server has become an important issue in the field of server hardware health monitoring.

[0003] In related technologies, a common approach is to use digital pressure sensors to detect the locking status of the heatsink. Specifically, an independent digital pressure sensor is installed at each stud of the CPU heatsink. Each sensor communicates with the Baseboard Management Controller (BMC) via a two-wire synchronous serial bus, such as the Inter-Integrated Circuit Bus (I²C) or the System Management Bus (SMBus). The BMC reads the pressure data by polling the addresses of each sensor, and determines whether each locking point is properly locked. However, this approach suffers from complex communication and high hardware costs. Summary of the Invention

[0004] This application provides a detection circuit and server for detecting the locking status of a heat sink in a server, in order to improve the problems of complex communication and high hardware cost of related technologies.

[0005] In a first aspect, this application provides a detection circuit for the locking state of a heatsink in a server, comprising: an asymmetric Wheatstone bridge unit, a signal amplification unit, and a baseboard management controller; wherein:

[0006] The asymmetric Wheatstone bridge unit includes multiple force-sensitive resistors and a zero-adjustment resistor. The multiple force-sensitive resistors are respectively set at each locking point of the heat sink to convert the pressure change at each locking point into the resistance value change. The zero-adjustment resistor is connected in series in one arm of the asymmetric Wheatstone bridge unit to adjust the output voltage of the asymmetric Wheatstone bridge unit to zero when the heat sink is in the normal locking state.

[0007] The differential input terminal of the signal amplification unit is connected to the differential output terminal of the asymmetric Wheatstone bridge unit, which is used to amplify the differential voltage signal output by the asymmetric Wheatstone bridge unit and bias it into a single-ended positive voltage signal.

[0008] The analog-to-digital converter pin of the baseboard management controller is connected to the output of the signal amplification unit, and is used to determine whether the heat sink is locked based on the voltage amplitude of the single-ended positive voltage signal.

[0009] In one possible embodiment, the plurality of force-sensitive resistors are designated as a first force-sensitive resistor, a second force-sensitive resistor, a third force-sensitive resistor, and a fourth force-sensitive resistor. The first terminal of the first force-sensitive resistor is connected to the power supply voltage, and its second terminal is connected in series with the third force-sensitive resistor and a zero-adjustment resistor before being grounded. The first terminal of the second force-sensitive resistor is connected to the power supply voltage, and its second terminal is connected in series with the fourth force-sensitive resistor before being grounded. The common terminal of the first and third force-sensitive resistors serves as the positive output terminal of the asymmetric Wheatstone bridge unit, and the common terminal of the second and fourth force-sensitive resistors serves as the negative output terminal of the asymmetric Wheatstone bridge unit.

[0010] In one possible embodiment, the signal amplification unit includes an instrumentation amplifier and a reference source; the differential input terminal of the instrumentation amplifier is connected to the positive output terminal and the negative output terminal of the asymmetric Wheatstone bridge unit, respectively; the reference source is connected to the reference voltage pin of the instrumentation amplifier to provide a reference bias voltage to the instrumentation amplifier, so that the amplified differential signal is biased to a positive voltage, thereby realizing single power supply for the instrumentation amplifier.

[0011] In one possible embodiment, the instrumentation amplifier is provided with a gain resistor, and the voltage amplification factor of the instrumentation amplifier is adjusted by adjusting the resistance value of the gain resistor.

[0012] In one possible embodiment, a window comparator unit is further included; the input terminal of the window comparator unit is connected to the output terminal of the signal amplification unit, the window comparator unit is used to compare the voltage amplitude output by the signal amplification unit with a preset voltage threshold, and the output terminal of the window comparator unit is used to output a level signal characterizing the comparison result.

[0013] In one possible embodiment, the window comparator unit includes a first comparator, a second comparator, a first voltage divider resistor, a second voltage divider resistor, and a third voltage divider resistor; the three voltage divider resistors are connected in series between the power supply voltage and ground; the common terminal of the first voltage divider resistor and the second voltage divider resistor is connected to the non-inverting input terminal of the first comparator, serving as the upper threshold input terminal; the common terminal of the second voltage divider resistor and the third voltage divider resistor is connected to the inverting input terminal of the second comparator, serving as the lower threshold input terminal; the output terminal of the signal amplification unit is connected to the inverting input terminal of the first comparator and the non-inverting input terminal of the second comparator, respectively; the output terminal of the first comparator and the output terminal of the second comparator are connected together to serve as the output terminal of the window comparator unit.

[0014] In one possible embodiment, a power enable control unit is further included; the power enable control unit is connected to the output terminal of the baseboard management controller and the output terminal of the window comparator unit, respectively, and is used to control the on / off of the load power supply in the server according to the monitoring and diagnostic results of the baseboard management controller and the level signal representing the comparison result output by the window comparator unit.

[0015] In one possible embodiment, the plurality of force-sensitive resistors are ring-shaped force-sensitive resistors, which are respectively disposed between each locking screw of the heat sink and the heat sink body.

[0016] Secondly, this application provides a detection device for the locking state of a heat sink in a server, including an interface circuit for connecting to a server motherboard, and a detection circuit for the locking state of a heat sink in a server provided above, disposed on the interface circuit.

[0017] Thirdly, this application provides a server, including a server chassis, a server motherboard, and a detection circuit or device for detecting the locking state of a heat sink in a server, as described above.

[0018] This application provides a detection circuit and server for detecting the locking state of a heatsink in a server. The detection circuit includes: an asymmetric Wheatstone bridge unit, a signal amplification unit, and a baseboard management controller. The asymmetric Wheatstone bridge unit includes multiple force-sensitive resistors and a zero-adjustment resistor. The force-sensitive resistors are respectively disposed at each locking point of the heatsink to convert pressure changes at each locking point into resistance changes. The zero-adjustment resistor is connected in series in one arm of the asymmetric Wheatstone bridge unit to adjust the output voltage of the asymmetric Wheatstone bridge unit to zero when the heatsink is in a normal locking state. The differential input terminal of the signal amplification unit is connected to the differential output terminal of the asymmetric Wheatstone bridge unit to amplify and bias the differential voltage signal output by the asymmetric Wheatstone bridge unit into a single-ended positive voltage signal. The analog-to-digital converter pin of the baseboard management controller is connected to the output terminal of the signal amplification unit to determine whether the heatsink is locked based on the voltage amplitude of the single-ended positive voltage signal.

[0019] This application utilizes an asymmetric Wheatstone bridge unit, including multiple force-sensitive resistors and zero-adjustment resistors, to convert the pressure changes at each locking point of the heat sink into differential voltages output by the bridge circuit. When the heat sink is in a normal locking state, the output voltage is adjusted to zero. This allows for a comprehensive characterization of the force state at multiple locking points using a single bridge circuit, eliminating the need for independent sensor and communication bus configurations for each locking point. This reduces the communication complexity and high hardware costs associated with multi-channel sensor acquisition and communication. The differential voltage signal is amplified and biased into a single-ended positive voltage signal by a signal amplification unit. The analog-to-digital converter pin of the baseboard management controller determines whether the heat sink is locked based on the voltage amplitude of this single-ended positive voltage signal. Compared to relying on a digital bus to transmit pressure data, differential signal transmission improves the stability and anti-interference capability of signal acquisition, thereby enhancing the reliability of identifying abnormal locking states of the heat sink and improving the overall monitoring efficiency. Attached Figure Description

[0020] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0021] Figure 1 A schematic diagram of a detection circuit for the locking state of a heat sink in a server, provided as an exemplary embodiment of this application;

[0022] Figure 2 A schematic diagram of an asymmetric Wheatstone bridge unit provided for an exemplary embodiment of this application;

[0023] Figure 3 A schematic diagram of a signal amplification unit provided as an exemplary embodiment of this application;

[0024] Figure 4Another schematic diagram of a detection circuit for the locking state of a heat sink in a server, provided as an exemplary embodiment of this application;

[0025] Figure 5 A schematic diagram of the structure of a window comparator unit provided for an exemplary embodiment of this application;

[0026] Figure 6 This is another schematic diagram of a detection circuit for the locked state of a heat sink in a server, provided as an exemplary embodiment of this application.

[0027] In the diagram, 10—detection circuit for the lock-up status of the heat sink in the server; 11—asymmetric Wheatstone bridge unit; 12—signal amplification unit; 13—board management controller; 14—window comparator unit; 15—power enable control unit.

[0028] The accompanying drawings illustrate specific embodiments of this application, which will be described in more detail below. These drawings and descriptions are not intended to limit the scope of the concept in any way, but rather to illustrate the concept of this application to those skilled in the art through reference to particular embodiments. Detailed Implementation

[0029] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application.

[0030] The terms “first,” “second,” etc., used in some aspects of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that embodiments of the application described herein can be implemented, for example, in orders other than those illustrated or described herein. Furthermore, the terms “comprising” and “having,” and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, products, or apparatus.

[0031] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties. Furthermore, the collection, use and processing of the relevant data must comply with relevant laws, regulations and standards, and corresponding operation entry points are provided for users to choose to authorize or refuse.

[0032] In traditional solutions, during server testing or on-site maintenance, if a malfunction (such as failure to power on) is suspected due to a loose heatsink, maintenance personnel must disassemble the server chassis, remove the heatsink for visual inspection, or reinstall it. This process is time-consuming, labor-intensive, and carries the risk of human error and hardware damage. In related technologies, solutions using digital pressure sensors to detect heatsink locking status involve the control unit collecting and judging the status information of each detection point to identify whether the heatsink is properly locked. These solutions typically use multiple sensors to output detection data, which is then uploaded to the control unit for analysis via a communication link. While this achieves locking status monitoring, it often relies on multiple digital sensors and a bus communication architecture. This not only requires handling sensor addressing, polling, and protocol interaction but also consumes significant controller interface and software resources, making the communication process complex. Furthermore, configuring detection devices and associated circuits at multiple locking points increases hardware costs and introduces more potential failure points. Additionally, in the strong electromagnetic interference environment inside the server, the digital communication link may experience signal anomalies, affecting the stability of the detection results.

[0033] To address the aforementioned issues, this application provides a detection scheme for the locking status of heat sinks in servers. By employing an asymmetric Wheatstone bridge unit comprising multiple force-sensitive resistors and zero-adjustment resistors, the pressure changes at each locking point of the heat sink are converted into differential voltage signals. These signals are then amplified and biased into single-ended positive voltage signals by a signal amplification unit. Finally, the analog-to-digital converter pin of the baseboard management controller (BDC) acquires and judges the signal. This simplifies the circuit architecture for detecting the heat sink's locking status, eliminating the need for independent sensor and communication bus configurations for each locking point. This reduces the communication complexity and high hardware costs associated with multi-channel sensor acquisition and communication. Furthermore, the differential voltage signals are amplified and biased into single-ended positive voltage signals by the signal amplification unit. The BDC pin of the BDC determines whether the heat sink is locked based on the voltage amplitude of this single-ended positive voltage signal. Compared to relying on digital bus transmission of pressure data, differential signal transmission improves the stability and anti-interference capability of signal acquisition, thereby enhancing the reliability of identifying abnormal heat sink locking status and improving overall system monitoring efficiency.

[0034] The technical solution of this application and how the technical solution of this application solves the above-mentioned technical problems are described in detail below with specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments. The embodiments of this application will now be described with reference to the accompanying drawings.

[0035] Figure 1 A schematic diagram of a detection circuit for the locked state of a heat sink in a server, provided as an exemplary embodiment of this application. Figure 1 As shown, the detection circuit 10 for the locking state of the heat sink in the server includes an asymmetric Wheatstone bridge unit 11, a signal amplification unit 12, and a baseboard management controller 13.

[0036] The asymmetric Wheatstone bridge unit 11 includes multiple force-sensitive resistors and a zero-adjustment resistor. The multiple force-sensitive resistors are respectively set at each locking point of the heat sink to convert the pressure change at each locking point into a resistance value change. The zero-adjustment resistor is connected in series in one arm of the asymmetric Wheatstone bridge unit 11 to adjust the output voltage of the asymmetric Wheatstone bridge unit 11 to zero when the heat sink is in the normal locking state.

[0037] The differential input terminal of the signal amplification unit 12 is connected to the differential output terminal of the asymmetric Wheatstone bridge unit 11, and is used to amplify the differential voltage signal output by the asymmetric Wheatstone bridge unit 11 and bias it into a single-ended positive voltage signal.

[0038] The analog-to-digital converter pin of the substrate management controller 13 is connected to the output terminal of the signal amplification unit 12, and is used to determine whether the heat sink is locked based on the voltage amplitude of the single-ended positive voltage signal.

[0039] The asymmetric Wheatstone bridge unit 11 is a differential sensing circuit composed of multiple force-sensitive resistors and a zero-adjustment resistor. It converts pressure changes at each locking point of the heat sink into voltage change signals that can be amplified and sampled by subsequent stages. This unit performs the primary signal conversion function in the overall circuit. Multiple force-sensitive resistors are installed at each locking point of the heat sink, allowing the clamping force at each locking point to act on the stress-sensitive resistors. When the locking force changes, the resistance value of each force-sensitive resistor shifts with the change in force state, forming a differential voltage output in the bridge structure. The zero-adjustment resistor is connected in series in one of the bridge arms to compensate for the static imbalance of the bridge circuit under normal locking conditions, bringing the output reference closer to zero, thus facilitating the subsequent identification of abnormal offsets.

[0040] The signal amplification unit 12 is an analog front-end circuit that amplifies and performs level biasing processing on the differential output signal of the bridge. It is used to convert the differential voltage caused by the change in the locking state into a single-ended positive voltage signal that can be recognized by the board management controller 13. The differential input terminal of this unit is electrically connected to the differential output terminal of the asymmetric Wheatstone bridge unit 11, and the output terminal is connected to the analog-to-digital converter pin of the board management controller 13 to form an intermediate signal processing link from pressure sensing to voltage amplification. When the bridge output signal is a reference signal near zero, the output of the signal amplification unit 12 remains near the corresponding bias level; when the locking state shifts, the output voltage changes accordingly.

[0041] The board management controller 13 refers to a board-level management and control unit with analog-to-digital conversion sampling capability and the ability to determine the status of the sampling results. It receives the single-ended positive voltage signal output by the signal amplification unit 12 and identifies whether the heatsink is in a locked state based on the voltage amplitude. The board management controller 13 is connected to the output of the signal amplification unit 12 via its analog-to-digital converter pin. The sampled voltage value is processed by internal comparison logic or firmware algorithm to determine whether the heatsink is locked. The board management controller 13 can integrate corresponding sampling and judgment functions.

[0042] Based on the above analysis, it can be seen that the detection circuit 10 for the locking state of a heat sink in a server provided in this application, during operation, has multiple force-sensitive resistors whose resistance values ​​change with the clamping force at each locking point of the heat sink. The asymmetrical Wheatstone bridge unit 11 converts this resistance change into a differential voltage output, and the zero-adjustment resistor compensates for the static deviation of the bridge circuit under normal locking conditions, so that the bridge outputs a signal close to zero in the reference state. When the force on a certain locking point of the heat sink decreases or increases, or when multiple locking points change simultaneously, the balance between the two arms of the bridge is disrupted, and the output differential voltage changes accordingly according to the direction of pressure offset. This change signal is amplified by the signal amplification unit and a bias level is applied to form a single-ended signal located in the positive voltage range, which is then sampled by the analog-to-digital converter of the baseboard management controller 13. Since the front end uses a bridge differential detection method, the change in the locking state can be reflected as a voltage offset; since the subsequent stage amplifies and biases the differential signal, the sampled signal can be adapted to the analog-to-digital converter interface of the board management controller; therefore, the board management controller 13 can determine whether the heat sink is locked based on the sampled voltage amplitude.

[0043] It should be noted that the aforementioned multiple force-sensitive resistors are, for example, four force-sensitive resistors; correspondingly, the four force-sensitive resistors are respectively installed at the four locking points of the heat sink.

[0044] It should be understood that the four locking points are merely an example. Those skilled in the art can also apply the technical solution of this application to two, six, or other numbers of locking points based on the actual heat sink fixing structure, simply by connecting the corresponding number of force-sensitive resistors according to the asymmetric bridge topology.

[0045] The detection circuit for the locking state of a heatsink in a server provided in this application embodiment uses an asymmetric Wheatstone bridge unit, including multiple force-sensitive resistors and zero-adjustment resistors, to convert the pressure changes at each locking point of the heatsink into differential voltages output by the bridge circuit. When the heatsink is in a normal locking state, the output voltage is adjusted to zero. This allows for a comprehensive characterization of the force state at multiple locking points using a single bridge circuit, eliminating the need for independent sensor and communication bus configurations for each locking point. This reduces the communication complexity and high hardware costs associated with multi-channel sensor acquisition and communication. The differential voltage signal is amplified and biased into a single-ended positive voltage signal by a signal amplification unit. The analog-to-digital converter pin of the baseboard management controller determines whether the heatsink is locked based on the voltage amplitude of this single-ended positive voltage signal. Compared to relying on a digital bus to transmit pressure data, differential signal transmission improves the stability and anti-interference capability of signal acquisition, thereby enhancing the reliability of identifying abnormal locking states of the heatsink and improving the overall monitoring efficiency.

[0046] In some embodiments, the multiple force-sensitive resistors are designated as a first force-sensitive resistor, a second force-sensitive resistor, a third force-sensitive resistor, and a fourth force-sensitive resistor. The first terminal of the first force-sensitive resistor is connected to the power supply voltage, and its second terminal is connected in series with the third force-sensitive resistor and a zero-adjustment resistor before being grounded. The first terminal of the second force-sensitive resistor is connected to the power supply voltage, and its second terminal is connected in series with the fourth force-sensitive resistor before being grounded. The common terminal of the first and third force-sensitive resistors serves as the positive output terminal of the asymmetric Wheatstone bridge unit, and the common terminal of the second and fourth force-sensitive resistors serves as the negative output terminal of the asymmetric Wheatstone bridge unit.

[0047] The first, second, third, and fourth force-sensitive resistors can be implemented using piezoresistive thin-film elements, carbon-based composite piezoresistive elements, or semiconductor piezoresistive sensing sheets, respectively. Piezoresistive thin-film elements can be formed by laminating a flexible substrate with a pressure-sensitive conductive layer. Carbon-based composite piezoresistive elements can be made of carbon ink, conductive adhesive, or polymer composite materials. Semiconductor piezoresistive sensing sheets can use silicon-based piezoresistive materials or similar miniature pressure-sensitive chips. Their shape can be sheet-like, ring-like, or washer-like to fit the installation space around the heat sink locking screw and minimize interference with the original mounting height and locking stroke. The thickness of a single force-sensitive resistor is typically less than the installation gap between the locking screw head and the heat sink, usually controlled within the range of 0.1mm to 2mm. The planar dimensions can be designed to be 3mm to 20mm square or a similar diameter depending on the force-bearing surface of the locking point, ensuring that pressure is effectively transmitted to the pressure-sensing area. The four force-sensitive resistors can be matched in resistance according to the bridge balance condition. Under normal locking conditions, they, together with the zero-adjustment resistor, meet the requirement that the output voltage is approximately zero. The resistance value of the zero-adjustment resistor can be finely adjusted according to the component dispersion and installation tolerance to compensate for the initial pressure difference at each locking point. It should be understood that the above example is only for demonstration and not a limitation. In practical applications, strain gauges, piezoresistive thin-film sensors, or miniature pressure-sensitive elements can be used as equivalent replacements, and electrical coupling with the bridge circuit can be achieved through welding, wire connection, flexible circuit connection, or conductive adhesive connection.

[0048] For example, Figure 2 A schematic diagram of an asymmetric Wheatstone bridge unit provided for an exemplary embodiment of this application. (See diagram below.) Figure 2 As shown, R1, R2, R3, and R4 are the first, second, third, and fourth force-sensitive resistors, respectively, and R5 is the zero-adjustment resistor. The power supply voltage is applied to the first terminals of the first force-sensitive resistor R1 and the second force-sensitive resistor R2. The four force-sensitive resistors R1, R2, R3, and R4, along with the zero-adjustment resistor R5, form two parallel bridge arms according to the connection relationship of an asymmetrical Wheatstone bridge unit. The second terminal of the first force-sensitive resistor R1 is connected in series with the third force-sensitive resistor R3 and the zero-adjustment resistor R5 before being grounded. The second terminal of the second force-sensitive resistor R2 is connected in series with the fourth force-sensitive resistor R4 before being grounded. The common terminal of the first and third force-sensitive resistors R1 and R3 serves as the positive output terminal V0+ of the asymmetrical Wheatstone bridge unit, and the common terminal of the second and fourth force-sensitive resistors R2 and R4 serves as the negative output terminal V0- of the asymmetrical Wheatstone bridge unit.

[0049] With the above connection method, when the heat sink is in the normal locked state, the pressure distribution on the four locking points is relatively stable, and the resistance value of each force-sensitive resistor remains within the preset balance range. By adjusting the resistance value of the zero-adjustment resistor R5, the bridge output voltage is adjusted to near zero, thereby reducing circuit power consumption. When a locking point becomes loose, the pressure decreases, or the force is uneven, the resistance value of the stress-sensitive resistor shifts. Since the zero-adjustment resistor R5 does not change its resistance value with the pressure change of the locking point, even if the four locking points loosen synchronously in the same proportion, a differential voltage corresponding to the degree of loosening will be generated between the two output terminals of the bridge, thereby eliminating the detection blind zone of the traditional symmetrical bridge in the case of equal-proportional loosening. When a single locking point undergoes an abnormal change, it can also be reflected as a measurable voltage deviation at the output terminal. When multiple locking points change simultaneously, the bridge output corresponds to the comprehensive differential result, thereby improving the ability to identify the overall abnormal state.

[0050] Correspondingly, still refer to Figure 2 The output voltage of an asymmetric Wheatstone bridge unit satisfies the following formula:

[0051]

[0052]

[0053]

[0054] in, This is the differential output voltage of the asymmetric Wheatstone bridge unit.

[0055] Furthermore, the differential voltage output by the asymmetric Wheatstone bridge unit is amplified by the subsequent signal amplification unit and biased into a single-ended positive voltage signal. This signal is then acquired by the analog-to-digital converter pin of the board management controller. The board management controller performs threshold judgment based on the voltage amplitude of this single-ended positive voltage signal to identify whether the heat sink is in a locked state.

[0056] In this embodiment, by adopting the above-described connection method, the circuit structure is simplified, eliminating the need for independent configuration of digital pressure sensors and corresponding communication buses for each locking point. This avoids the system complexity caused by multi-channel sensor addressing and polling communication. Simultaneously, the differential signal transmission method reduces interference from the bus link in the server's strong electromagnetic environment, thereby improving the stability and reliability of locking status detection. Furthermore, by employing an asymmetric bridge design, even slight loosening or simultaneous proportional loosening at any one of the multiple locking points will disrupt the pre-calibrated bridge balance conditions and generate a significant output voltage change. This effectively eliminates the detection blind zone of traditional symmetrical bridges under proportional loosening conditions, further enhancing detection sensitivity.

[0057] Optionally, the multiple force-sensitive resistors are ring-shaped force-sensitive resistors, which are respectively disposed between each locking screw of the heat sink and the heat sink body.

[0058] Correspondingly, when the system starts up, the four ring-shaped force-sensitive resistors are located between the four locking screws of the heat sink and the heat sink body, respectively. During the assembly and tightening of the locking screws, they undergo compressive deformation under the axial preload. Because the ring structure allows for a more uniform distribution of force along the circumference, the pressure changes at each locking point can be mapped to corresponding resistance changes relatively consistently. As the heat sink completes normal locking, the resistance values ​​of the four pressure-sensitive elements are within the range corresponding to the predetermined clamping state, allowing the subsequent bridge circuit to form a stable differential voltage baseline. When any locking screw becomes loose, unbalanced, or insufficiently clamped, the pressure on the corresponding ring-shaped force-sensitive resistor decreases, and its resistance changes accordingly. This change, along with the state differences at other locking points, is reflected in the bridge output. Because the pressure-sensing element adopts a ring structure and is directly arranged between the locking screw and the heat sink body, its force path is short and its response is clear. This can reduce the bias error caused by local contact of the locking screw head, and at the same time reduce the detection fluctuation caused by changes in installation posture. Furthermore, the four locking points correspond to four independent ring force-sensitive resistors, so that the pressing state of the four corners or four locking positions of the heat sink can be sensed synchronously and in a distributed manner. This is conducive to improving the consistency and reliability of locking state detection, and reducing the probability of misjudgment caused by uneven single-point contact or local vibration.

[0059] Based on the above analysis, it can be seen that this ring arrangement can directly sample the tightening pressure of the locking screws without increasing complex mechanical linkages, and provide a stable input for subsequent bridge detection, signal amplification and board management controller determination, thereby improving the adaptability and anti-interference capability of server heat sink locking status detection.

[0060] In some embodiments, the signal amplification unit includes an instrumentation amplifier and a reference source; the differential input of the instrumentation amplifier is connected to the positive and negative outputs of the asymmetric Wheatstone bridge unit, respectively; the reference source is connected to the reference voltage pin of the instrumentation amplifier to provide a reference bias voltage to the instrumentation amplifier, so that the amplified differential signal is biased to a positive voltage, thereby realizing single-supply power supply for the instrumentation amplifier.

[0061] Optionally, the instrumentation amplifier is equipped with a gain resistor, and the voltage amplification factor of the instrumentation amplifier can be adjusted by adjusting the resistance value of the gain resistor.

[0062] For example, Figure 3 This is a schematic diagram of a signal amplification unit provided as an exemplary embodiment of this application. (See diagram below.) Figure 3As shown, the non-inverting input (pin 3) of the instrumentation amplifier is connected to the positive output terminal VO+ of the asymmetric Wheatstone bridge unit, and the inverting input (pin 2) of the instrumentation amplifier is connected to the negative output terminal VO- of the asymmetric Wheatstone bridge unit, thereby converting the differential voltage signal output by the bridge into a signal. Connect the instrumentation amplifier; a gain resistor Rg is connected between pins 1 and 8 of the instrumentation amplifier to set the voltage amplification factor; the output of the reference source is connected to the reference voltage pin (pin 5) of the instrumentation amplifier to provide a reference bias voltage to the instrumentation amplifier. This biases the amplified differential signal to a positive voltage, enabling single-supply operation of the instrumentation amplifier; pin 4 of the instrumentation amplifier is connected to reference ground GND, and pin 7 is connected to the power supply voltage. Pin 6 serves as the output terminal of the signal amplification unit, outputting an amplified and biased single-ended positive voltage signal V_amp. This output terminal is connected to the analog-to-digital converter pin of the board management controller.

[0063] The instrumentation amplifier can be implemented using a dedicated integrated chip with a three-op-amp structure, an instrumentation amplifier chip with an internal resistor network, or a differential amplifier circuit composed of discrete operational amplifiers. For example, the AD620 instrumentation amplifier chip has differential input terminals at pins 2 and 3, a gain resistor connected between pins 1 and 8, and a reference voltage introduced through pin 5. The reference source can be implemented using a precision reference voltage chip, a voltage divider reference network, or a programmable reference source. For example, the REF3025 reference source chip provides a reference bias voltage of, for example, 2.5V to pin 5 of the instrumentation amplifier at its output. The relevant components can be packaged in various forms, including Small Outline Transistor (SOT-23), Small Outline Integrated Circuit (SOIC), Mini Small Outline Package (MSOP), or Quad Flat No-lead Package (QFN), and are mounted on a printed circuit board (PCB) and connected to the bridge and substrate management controller via copper traces. The substrate material can be FR-4, ceramic substrate, or low-noise high-frequency board material. In terms of layout, the instrumentation amplifier and reference source are typically positioned near the bridge output and the PCB analog-to-digital converter (ADC) interface to shorten the analog signal path and reduce noise introduction. Their dimensions are generally in the range of several millimeters to over ten millimeters. The reference voltage should be set according to the PCB ADC input range, usually near the midpoint of the range or ensuring that the output signal does not exceed the single-supply limit under maximum differential voltage conditions. Sufficient zero-point margin should be provided in the normal locking state to distinguish between normal locking, loosening, and unlocking conditions. In addition, the instrumentation amplifier and reference source can be arranged using methods such as shielded grounding and differential routing, depending on the board layout requirements, to further improve anti-interference capabilities.

[0064] Based on the above connection relationship, the instrumentation amplifier amplifies the differential voltage signal output by the bridge, specifically satisfying the following formula:

[0065]

[0066] in, , The reference bias voltage provided to the reference source; the differential voltage output of the bridge. After being amplified by the instrumentation amplifier, the reference bias voltage provided by the reference source is superimposed. It is biased as a single-ended positive voltage signal. This ensures that the output signal remains positive and falls within the effective sampling range of the board management controller's analog-to-digital converter, even when the instrumentation amplifier is powered by only a single power supply (such as +5V).

[0067] The gain resistor Rg is a resistor element placed between the gain setting terminals of the instrumentation amplifier to determine the differential signal amplification factor. It can be externally connected to the gain control pin of the instrumentation amplifier or integrated into the signal amplification unit using a built-in programmable resistor structure, thus forming an adjustable gain amplification structure. The function of this gain resistor is to appropriately amplify the weak differential voltage output from the asymmetric Wheatstone bridge unit in the front stage, based on the sensitivity required for heat sink locking state detection. This ensures sufficient amplitude variation within the input range of the analog-to-digital converter of the subsequent board management controller, facilitating stable identification of locking state changes corresponding to pressure variations by the controller. In terms of location and relationship, the gain resistor is typically connected between the gain setting terminal of the instrumentation amplifier or the corresponding feedback network, forming a complete signal conditioning loop together with the differential input terminal, reference voltage pin, and output terminal. Its installation position is preferably close to the instrumentation amplifier chip to reduce the impact of parasitic inductance and noise coupling on gain stability. In terms of form and material, the gain resistor can be implemented using one or more of the following methods: surface mount resistor, precision metal film resistor, adjustable potentiometer, or digital resistor. Surface mount resistors facilitate mass assembly, precision metal film resistors have low temperature drift and high accuracy, adjustable potentiometers facilitate factory calibration, and digital resistors facilitate programmed adjustment by a controller. The material can be metal film, alloy film, thick film ceramic, or programmable resistive material to meet different accuracy, power consumption, and environmental adaptability requirements. The package form can be common surface mount specifications such as 0402, 0603, and 0805, or a package form with a fine-tuning knob. Regarding size and proportion, the resistance value of the gain resistor can be configured within several ranges according to the target amplification factor. Its accuracy should generally be higher than that of ordinary voltage divider resistors to reduce zero-point drift and gain error, and its dimensions should match the board layout density used. In one possible embodiment, if the instrumentation amplifier needs to amplify millivolt-level differential signals to the volt range, the gain resistor can be selected from a resistance range of several thousand ohms to tens of thousands of ohms, and the final gain is determined by the ratio with the internal set resistor. It should be understood that the above examples are for illustrative purposes only and are not intended to be limiting.

[0068] In this embodiment of the application, by adopting the above-described circuit structure, the high common-mode rejection amplification of the bridge differential signal by the instrumentation amplifier can effectively reduce the impact of internal power supply ripple, switching noise and electromagnetic interference on the detection results. Meanwhile, the bias voltage provided by the reference source ensures that the amplified signal falls stably within the effective sampling range of the single-supply analog-to-digital converter of the board management controller. Thus, without increasing the complexity of multi-channel digital acquisition and communication links, stable, continuous and reliable voltage detection of the heat sink locking state can be achieved.

[0069] Based on the above embodiments, in some embodiments, the detection circuit for the locking state of the heat sink in the server further includes a window comparator unit; the input terminal of the window comparator unit is connected to the output terminal of the signal amplification unit, the window comparator unit is used to compare the voltage amplitude output by the signal amplification unit with a preset voltage threshold, and the output terminal of the window comparator unit is used to output a level signal characterizing the comparison result.

[0070] For example, Figure 4 Another schematic diagram of a detection circuit for the locked state of a heat sink in a server, provided as an exemplary embodiment of this application. Figure 4 As shown, the detection circuit 10 for the locking state of the heat sink in the server also includes a window comparator unit 14. The input terminal of the window comparator unit 14 is connected to the output terminal of the signal amplification unit 12 and is used to receive the single-ended positive voltage signal output by the signal amplification unit 12.

[0071] In terms of specific structure, the window comparator unit 14 is used to compare the voltage amplitude output by the signal amplification unit 12 with a preset voltage threshold, and output a corresponding level signal according to the comparison result. Specifically, the window comparator unit 14 can be implemented in several ways: it can be composed of an upper limit comparator and a lower limit comparator by independent comparator chips, which, together with a reference voltage network, form a window detection circuit. The upper limit comparator is used to determine whether the input voltage is higher than the upper limit threshold, and the lower limit comparator is used to determine whether the input voltage is lower than the lower limit threshold. The outputs of the two comparators are logically combined to form the window discrimination result; it can also be implemented using an integrated window comparator chip, which directly uses the internal reference source, comparator, and output stage of the chip to complete the detection. The carrier of the window comparator unit 14 can be a printed circuit board on the server motherboard, a small functional board located near the baseboard management controller 13, or integrated with the signal amplification unit 12 into the same module to reduce the wiring length and improve the anti-interference capability. The device materials used may include complementary metal-oxide-semiconductor (CMOS) process devices, bipolar analog devices, or low-power comparator devices. The comparator output stage can be selected as open-drain output, push-pull output, or open-collector output according to the interface form of the subsequent stage to adapt to different level standards.

[0072] Regarding size and proportion, the preset voltage thresholds of the window comparator unit 14 include an upper threshold and a lower threshold, which should be set to match the normal operating voltage range of the signal amplification unit 12. The upper threshold corresponds to the voltage boundary when the heat sink clamping force is too large or abnormal, while the lower threshold corresponds to the voltage boundary when the heat sink clamping force is insufficient or loose. When the voltage amplitude output by the signal amplification unit 12 is between the upper and lower thresholds, it indicates that the pressure at each clamping point is within the preset normal range; when the voltage amplitude is higher than the upper threshold or lower than the lower threshold, it indicates a clamping abnormality. For implementations using discrete components, the dimensions of the comparator unit and its peripheral voltage divider resistors, feedback resistors, and decoupling capacitors can be arranged according to the motherboard space. Surface mount devices are typically used, with resistor and capacitor values ​​used to form the upper and lower thresholds and an appropriate response speed. For integrated implementations, the window comparator unit 14 can be packaged in a small integrated circuit to shorten the signal path and reduce parasitic interference. The ratio of the comparator input threshold setting should ensure that the normal state voltage is in the middle of the threshold range, thereby leaving a margin for device deviation, temperature drift and noise fluctuation.

[0073] During operation, when the system starts up or runs, the heat sink is pressed against the heat-dissipating component (e.g., CPU) in the server through four locking points. The force-sensitive resistors at each locking point change their resistance values ​​under pressure, and the asymmetric Wheatstone bridge unit 11 converts this pressure change into a differential voltage signal. The signal is amplified and biased into a single-ended positive voltage signal by the signal amplification unit 12. Subsequently, one path is sent to the analog-to-digital converter pin of the substrate management controller 13 for software judgment, and the other path is sent to the input terminal of the window comparator unit 14. Correspondingly, the window comparator unit 14 compares the voltage amplitude output by the signal amplification unit 12 with the upper threshold and the lower threshold respectively: when the voltage amplitude is within the window range between the upper threshold and the lower threshold, the output terminal of the window comparator unit 14 outputs a first level signal (e.g., high level), which indicates that the heat sink is in a normal locking state; when the voltage amplitude is higher than the upper threshold or lower than the lower threshold, that is, deviates from the window range, the output terminal of the window comparator unit 14 outputs a second level signal (e.g., low level), which indicates that the heat sink is in an abnormal locking state. Thus, the output terminal of the window comparator unit 14 outputs a level signal representing the comparison result. The subsequent control circuit, such as a complex programmable logic device (CPLD), can quickly identify whether the heat sink is in a normal locking state based on this level signal, and can further realize power enable control, fault alarm or maintenance prompts, etc.

[0074] Based on the above working process, it can be seen that the window comparator unit directly converts the amplified analog voltage signal into a binary level signal through pure hardware comparison, and can complete the hardware-level discrimination of the locking state without the participation of the board management controller. Compared with the existing scheme that relies on the digital bus to poll each locking point sensor, this application does not need to report the state of all locking points separately and perform complex software polling judgment. Instead, it directly performs hardware threshold discrimination on the amplified analog signal through the window comparator unit, thereby improving the real-time performance and anti-interference capability of locking state recognition, and reducing the occupation of sampling resources and software processing cycle of the board management controller. This helps to reduce the complexity of the overall detection circuit and improve heat dissipation reliability.

[0075] In some embodiments, the window comparator unit includes a first comparator, a second comparator, a first voltage divider resistor, a second voltage divider resistor, and a third voltage divider resistor; the three voltage divider resistors are connected in series between the power supply voltage and ground; the common terminal of the first voltage divider resistor and the second voltage divider resistor is connected to the non-inverting input terminal of the first comparator, serving as the upper limit threshold input terminal; the common terminal of the second voltage divider resistor and the third voltage divider resistor is connected to the inverting input terminal of the second comparator, serving as the lower limit threshold input terminal; the output terminal of the signal amplification unit is connected to the inverting input terminal of the first comparator and the non-inverting input terminal of the second comparator, respectively; the output terminal of the first comparator and the output terminal of the second comparator are connected together to serve as the output terminal of the window comparator unit.

[0076] In this embodiment, the window comparator unit is implemented, for example, using an LM393 dual-channel differential comparator chip. For example, Figure 5 A schematic diagram of a window comparator unit provided for an exemplary embodiment of this application. For example... Figure 5 As shown, the window comparator unit includes a first comparator U3.1, a second comparator U3.2, a first voltage divider resistor R6, a second voltage divider resistor R7, and a third voltage divider resistor R8. The first voltage divider resistor R6, the second voltage divider resistor R7, and the third voltage divider resistor R8 are connected in series with the power supply voltage. A voltage divider branch is formed between the voltage divider resistor R6 and the ground (GND); the common terminal of the first voltage divider resistor R6 and the second voltage divider resistor R7 serves as the upper threshold terminal, outputting the upper threshold voltage. The common terminal of the second voltage divider resistor R7 and the third voltage divider resistor R8 serves as the lower threshold voltage terminal, outputting the lower threshold voltage. The output of the signal amplification unit The inverting input of the first comparator U3.1 and the non-inverting input of the second comparator U3.2 are connected respectively; the non-inverting input of the first comparator U3.1 is connected to the upper threshold terminal to receive the upper threshold value. The inverting input of the second comparator U3.2 is connected to the lower threshold input to receive the lower threshold value. The output of the first comparator U3.1 is connected to the output of the second comparator U3.2 and then serves as the output of the window comparator unit.

[0077] In this embodiment, the upper limit threshold and lower threshold They respectively satisfy the following relations:

[0078]

[0079]

[0080] By selecting different resistance values, the upper and lower threshold values ​​of the window comparator can be flexibly set to adapt to the voltage range corresponding to different heat sink locking pressures.

[0081] Correspondingly, when the system starts up, the first voltage divider resistor R6, the second voltage divider resistor R7, and the third voltage divider resistor R8 are first divided by the power supply voltage. A stable reference potential distribution is established between the voltage divider resistor R6 and the ground (GND). The common terminal of the first voltage divider resistor R6 and the second voltage divider resistor R7 forms the upper threshold. The common terminal of the second voltage divider resistor R7 and the third voltage divider resistor R8 forms the lower threshold. The voltage output by the signal amplification unit The signal is simultaneously input to the inverting input of the first comparator U3.1 and the non-inverting input of the second comparator U3.2 for comparison. When the heat sink is in a normal locked state, the output voltage after amplification and biasing of the preceding signal is... Within the preset window, i.e. At this time, the outputs of both the first comparator U3.1 and the second comparator U3.2 are in a high-impedance or high-level state, and the output of the window comparator unit is high, indicating that the locking state is normal; when the pressure at a certain locking point of the heat sink is abnormal, causing the output voltage to exceed the upper limit threshold ( or below the lower threshold ( When the lockout state is abnormal, at least one of the first comparator U3.1 or the second comparator U3.2 will change its output state, causing the output of the window comparator unit to output a low level, indicating an abnormal locking state, triggering an alarm or preventing power-on. By adopting the above circuit structure, the output of the signal amplification unit can be windowed without complex bus communication and multiple independent detection links, and the lockout state of the heat sink can be directly converted into a level signal output, thereby improving the real-time performance, stability and anti-interference capability of the detection circuit, while reducing the number of system interfaces and control complexity.

[0082] Based on the above embodiments, in some embodiments, the detection circuit for the heat sink locking state in the server further includes a power enable control unit; the power enable control unit is connected to the output terminal of the baseboard management controller and the output terminal of the window comparator unit respectively, and is used to control the on / off of the load power supply in the server according to the monitoring and diagnostic results of the baseboard management controller and the level signal representing the comparison result output by the window comparator unit.

[0083] For example, Figure 6 This is another schematic diagram of a detection circuit for the locked state of a heat sink in a server, provided as an exemplary embodiment of this application. Figure 6 As shown, the detection circuit 10 for the locking state of the heat sink in the server also includes a power enable control unit 15, which is connected to the output of the baseboard management controller 13 and the output of the window comparator unit 14.

[0084] In one possible embodiment, the power enable control unit 15 is a logic control component used to receive the monitoring and diagnostic results from the baseboard management controller 13 and the comparison result level signal output by the window comparator unit 14, and control the enable state of the server load power supply accordingly. Its purpose is to establish a linkage between the heatsink locking state detection result and the power supply operation of the server load power supply, thereby promptly prohibiting the load power supply from starting or maintaining its off state when the heatsink is not locked, the clamping force is abnormal, or the detection link determines an anomaly, thus achieving hardware-level protection for the server's critical loads. The power enable control unit 15 is physically connected to the output terminals of the baseboard management controller 13 and the window comparator unit 14, respectively. Its control output terminal is connected to the enable terminal, gating terminal, or power switch control terminal of the load power supply in the server, so that the monitoring and diagnostic information from the baseboard management controller 13 and the level signal representing the comparison result output by the window comparator unit 14 are logically synthesized to form the final control command for turning the load power supply on and off. In one exemplary embodiment, the power enable control unit 15 can be composed of at least one of logic gate circuits, programmable logic devices, field-programmable gate arrays, metal-oxide-semiconductor switch control circuits, or relay drive modules, or a combination of the above devices. The logic gate circuits can be used to perform AND, OR, and NOT logic operations on the enable signal; the programmable logic devices can combine and determine different state signals according to a preset logic table; the metal-oxide-semiconductor switch control circuits can be used to directly drive the high-side or low-side switches on the load power path; and the relay drive module is suitable for achieving electrical isolation between the relay and high-voltage or high-current paths. When an isolation-required structure is used, the logic control section and the power path should preferably be connected through an optocoupler isolator, a drive buffer, or an isolated power management chip to reduce the coupling effect of internal electromagnetic interference on the control signal and meet the withstand voltage and immunity requirements of the on / off control loop. Based on the above analysis, the power enable control unit 15 can be arranged in a local logic control area on the motherboard, an independent small board module, or a power management submodule. Its shape can be an integrated circuit chip, a surface-mount control board, or a modular plug-in unit. The device materials used typically include silicon-based CMOS logic devices, copper foil circuit boards, epoxy resin encapsulation, power MOSFETs, and metal contacts or electromagnetic relay contacts, etc. In some embodiments, the size of the unit is matched with the power level of the server power control path. The layout area of ​​the logic control part is usually smaller than the main power circuit area of ​​the power supply, while the control signal line width and device spacing meet the corresponding electrical safety specifications. If necessary, sufficient creepage distance and electrical clearance can be reserved between the control end and the power end to ensure long-term operational reliability.Specifically, the circuit board size of this unit can be configured as a small module according to the space arrangement of the motherboard, with its length and width on the order of several centimeters. If a relay or electronic switch is used, the rated current of its contacts or switching transistors is adapted to the peak current of the server load power supply to avoid damage to the device due to overcurrent. During operation, when the baseboard management controller 13 determines that the current state is normal based on the sampling result of the heat sink locking state, it outputs an enable signal. At the same time, the window comparator unit 14 outputs an effective level indicating that the amplified signal is within the preset window range. The power enable control unit 15 then turns on the load power supply, allowing the server load to enter the normal power supply state. When the heat sink is not locked, the clamping force is insufficient, there is local pressure loss, or the comparison result exceeds the preset threshold range, the window comparator unit 14 outputs the corresponding abnormal level, or the baseboard management controller 13 outputs an inhibit signal based on the diagnostic result. The power enable control unit 15 then cuts off, keeps the load power supply off, or inhibits the enable path of the load power supply, thereby preventing the server from continuing to operate when the heat dissipation conditions are not met. Because this control process directly links the heat sink locking detection with the power on / off control, it can promptly stop the load power supply when the heat sink malfunctions, reducing the risk of device overheating caused by improper heat sink installation, and reducing the probability of delay and misjudgment caused by software polling or multi-level communication links, thereby improving the server's power-on safety, abnormal response speed, and overall heat dissipation reliability.

[0085] In summary, this application has at least the following advantages:

[0086] First, by using an asymmetric Wheatstone bridge unit comprising multiple force-sensitive resistors and zero-adjustment resistors, the pressure changes at each locking point of the heatsink are converted into differential voltages output by the bridge circuit. When the heatsink is in a normal locking state, the output voltage is adjusted to zero using the zero-adjustment resistors. This allows for a comprehensive characterization of the force state at multiple locking points using a single bridge circuit, eliminating the need for independent sensors and communication buses for each locking point. This avoids the circuit complexity and hardware costs associated with multi-channel sensor acquisition and polling communication. Furthermore, differential signal transmission, compared to methods relying on digital bus transmission, possesses common-mode noise suppression capabilities, improving signal acquisition stability and anti-interference capabilities, thereby enhancing the reliability of identifying abnormal locking states of the heatsink.

[0087] Second, the differential voltage signal is amplified and biased into a single-ended positive voltage signal by a signal amplification unit. This voltage amplitude is directly acquired by the analog-to-digital converter pin of the baseboard management controller for judgment. This eliminates the need for complex bus communication protocols, device address allocation, and software polling overhead, significantly reducing the input / output (I / O) resource consumption and software development complexity of the baseboard management controller. Furthermore, the core detection element consists only of a force-sensitive resistor and an operational amplifier, eliminating the need for independent pressure sensor chips and communication circuits for each locking point. This results in lower material costs, and the force-sensitive resistor offers superior lifespan and reliability compared to digital sensor chips, making it more suitable for the harsh environment of continuous operation within servers.

[0088] Third, by employing an asymmetric Wheatstone bridge design, even a slight loosening of any one of the four locking points, or a simultaneous proportional loosening of all four locking points, will disrupt the bridge's balance and produce a significant change in output voltage. Compared to the inherent defect of traditional symmetrical bridges, which cannot detect proportional loosening, this design effectively eliminates the detection blind zone and further improves detection sensitivity. Simultaneously, the amplified voltage signal is also sent to a window comparator unit. This unit compares the voltage amplitude with a preset voltage threshold and outputs a level signal representing the comparison result. This level signal can be directly used to control the server motherboard's power-on. When the bridge output signal exceeds the preset normal range, hardware logic directly prevents the motherboard from powering on. Compared to solutions that rely on BMC software for judgment before shutdown, this design can complete the response within nanoseconds, fundamentally eliminating the risk of overheating and burnout caused by improper heatsink installation, thus providing high-priority safety protection for the server.

[0089] This application also provides a detection device for the locking state of a heat sink in a server, including an interface circuit for connecting to a server motherboard, and a detection circuit for the locking state of a heat sink in a server provided above, disposed on the interface circuit.

[0090] This detection device is electrically connected to the server motherboard via an interface circuit, allowing the detection circuit, located on the interface circuit, to be directly integrated into the server's overall hardware system. The detection circuit converts, amplifies, and outputs electrical signals corresponding to the force changes at each locking point of the heatsink, thus converting the heatsink's locking state into detection information recognizable by the server motherboard. Because the detection circuit is integrated into the interface circuit, the signal transmission path is shorter, and the connection is simplified, making device deployment easier and improving hardware compatibility. This reduces the design burden caused by multiple independent detection devices and complex communication links, improving anti-interference capabilities and detection stability. Therefore, reliable detection of the server heatsink's locking state can be achieved at a lower cost.

[0091] This application also provides a server, including a server chassis, a server motherboard, and a detection circuit or device for detecting the locking state of a heat sink in a server, as described above.

[0092] The server chassis forms the overall installation space, while the server motherboard houses the processor, baseboard management controller, and related interface circuits. Detection circuitry or devices are mounted on the server motherboard and positioned corresponding to the heatsink locking area. By converting the force changes at each locking point of the heatsink into collectable electrical signals, which are then sampled and judged by the baseboard management controller, real-time detection of the heatsink locking status is achieved during server operation. The simplified detection architecture reduces controller interface requirements and communication processing needs, thereby decreasing hardware costs and debugging complexity. It also improves detection stability in the electromagnetic interference environment inside the server, thus enhancing the reliability of heatsink installation anomaly identification and ensuring processor cooling performance and overall system operational safety.

[0093] Other embodiments of this application will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of this application that follow the general principles of this application and include common knowledge or customary techniques in the art not disclosed herein. The specification and examples are to be considered exemplary only, and the true scope and spirit of this application are indicated by the following claims.

[0094] It should be understood that this application is not limited to the precise structure described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope.

Claims

1. A detection circuit for the locking state of a heat sink in a server, characterized in that, Includes: an asymmetric Wheatstone bridge unit, a signal amplification unit, and a board management controller; wherein: The asymmetric Wheatstone bridge unit includes multiple force-sensitive resistors and a zero-adjustment resistor. The multiple force-sensitive resistors are respectively disposed at each locking point of the heat sink to convert the pressure change at each locking point into a resistance change. The zero-adjustment resistor is connected in series in one arm of the asymmetric Wheatstone bridge unit to adjust the output voltage of the asymmetric Wheatstone bridge unit to zero when the heat sink is in the normal locking state. The differential input terminal of the signal amplification unit is connected to the differential output terminal of the asymmetric Wheatstone bridge unit, and is used to amplify and bias the differential voltage signal output by the asymmetric Wheatstone bridge unit into a single-ended positive voltage signal. The analog-to-digital converter pin of the substrate management controller is connected to the output terminal of the signal amplification unit, and is used to determine whether the heat sink is locked based on the voltage amplitude of the single-ended positive voltage signal.

2. The detection circuit for the locking state of a heat sink in a server according to claim 1, characterized in that, The plurality of force-sensitive resistors are designated as a first force-sensitive resistor, a second force-sensitive resistor, a third force-sensitive resistor, and a fourth force-sensitive resistor; wherein: The first end of the first force-sensitive resistor is connected to the power supply voltage, and the second end of the first force-sensitive resistor is connected in series with the third force-sensitive resistor and the zero-adjustment resistor and then grounded. The first end of the second force-sensitive resistor is connected to the power supply voltage, and the second end of the second force-sensitive resistor is connected in series with the fourth force-sensitive resistor and then grounded. The common terminal of the first force-sensitive resistor and the third force-sensitive resistor serves as the positive output terminal of the asymmetric Wheatstone bridge unit, and the common terminal of the second force-sensitive resistor and the fourth force-sensitive resistor serves as the negative output terminal of the asymmetric Wheatstone bridge unit.

3. The detection circuit for the locking state of a heat sink in a server according to claim 1, characterized in that, The signal amplification unit includes an instrumentation amplifier and a reference source; The differential input terminals of the instrumentation amplifier are respectively connected to the positive output terminal and the negative output terminal of the asymmetric Wheatstone bridge unit. The reference source is connected to the reference voltage pin of the instrumentation amplifier to provide a reference bias voltage to the instrumentation amplifier, so that the amplified differential signal is biased to a positive voltage, thereby realizing the single power supply of the instrumentation amplifier.

4. The detection circuit for the locking state of a heat sink in a server according to claim 3, characterized in that, The instrumentation amplifier is equipped with a gain resistor, and the voltage amplification factor of the instrumentation amplifier can be adjusted by adjusting the resistance value of the gain resistor.

5. The detection circuit for the locked state of a heat sink in a server according to any one of claims 1 to 4, characterized in that, It also includes a window comparator unit; the input terminal of the window comparator unit is connected to the output terminal of the signal amplification unit, the window comparator unit is used to compare the voltage amplitude output by the signal amplification unit with a preset voltage threshold, and the output terminal of the window comparator unit is used to output a level signal characterizing the comparison result.

6. The detection circuit for the locking state of a heat sink in a server according to claim 5, characterized in that, The window comparator unit includes a first comparator, a second comparator, a first voltage divider resistor, a second voltage divider resistor, and a third voltage divider resistor; Three voltage divider resistors are connected in series between the power supply voltage and ground. The common terminal of the first and second voltage divider resistors is connected to the non-inverting input of the first comparator as the upper threshold input. The common terminal of the second and third voltage divider resistors is connected to the inverting input of the second comparator as the lower threshold input. The output terminal of the signal amplification unit is connected to the inverting input terminal of the first comparator and the non-inverting input terminal of the second comparator, respectively. The output of the first comparator is connected to the output of the second comparator and then serves as the output of the window comparator unit.

7. The detection circuit for the locking state of a heat sink in a server according to claim 5, characterized in that, It also includes a power enable control unit; The power enable control unit is connected to the output terminal of the baseboard management controller and the output terminal of the window comparator unit, respectively, and is used to control the on / off of the load power supply in the server according to the monitoring and diagnostic results of the baseboard management controller and the level signal representing the comparison result output by the window comparator unit.

8. The detection circuit for the locked state of a heat sink in a server according to any one of claims 1 to 4, characterized in that, The force-sensitive resistors are ring-shaped force-sensitive resistors, which are respectively disposed between the multiple locking screws of the heat sink and the heat sink body.

9. A device for detecting the locking state of a heat sink in a server, characterized in that, It includes an interface circuit for connecting to a server motherboard, and a detection circuit for the locked state of a heat sink in a server, as described in any one of claims 1 to 8, disposed on the interface circuit.

10. A server, characterized in that, It includes a server chassis, a server motherboard, and a detection circuit for the locking state of a heat sink in a server as described in any one of claims 1 to 8, or a detection device for the locking state of a heat sink in a server as described in claim 9.

Images