Equipment power supply system and server

Abstract

The invention discloses an equipment power supply system and a server, relates to the field of servers, and aims to solve the problem that the existing power supply design cannot meet the power supply requirement of a graphic processing unit in an electrical data peak value processing state. The system comprises a power panel, and the power panel comprises a power supply assembly and a plurality of power supply branches connected with the power supply assembly; the first capacitor module is connected to the first power supply branch; the second capacitor module and the switching module correspond to the second power supply branch; and the management module and the third capacitor module are arranged on the equipment expansion card, and the management module is used for controlling the third capacitor module to supply power to the computing equipment meeting the voltage early warning condition on the equipment expansion card. According to the invention, the power supply requirement of each computing device on the device expansion board during operation in the electrical data peak value processing state can be met, and the reliability of power supply is ensured, so that the stable operation of the server is ensured.

Description

Technical Field

[0001] The present invention relates to the field of servers, and in particular to a device power supply system and a server. Background Art

[0002] With the rapid development of cloud computing, big data and artificial intelligence technologies, GPU (Graphics Processing Unit) plays an increasingly important role in modern electronic devices. However, the existing GPU power supply solutions face the dual challenges of space and efficiency. In the server, the power board is directly connected to the GPU riser through a cable to power the GPU. In order to meet the power supply requirements of the GPU during EDPP (Electrical Date Peak Processing), capacitors are designed on each GPU power supply circuit. However, due to the diversity of server configurations, the design space left for GPU power supply on the power board is limited, resulting in the inability to place more capacitors to meet the power supply requirements of high-performance GPUs. The space on the GPU riser board is also limited. As the power consumption of the GPU continues to increase, the existing capacitor design cannot meet the power supply requirements of the GPU in the EDPP state, which may cause the GPU voltage to continue to drop below the minimum operating voltage, thereby causing the GPU to power off and affecting the normal operation of the server.

[0003] Therefore, how to provide a solution to the above technical problems is a problem that those skilled in the art need to solve at present. Summary of the invention

[0004] The purpose of the present invention is to provide a device power supply system and a server, which can meet the power supply requirements of various computing devices on the device expansion board when operating in the electrical data peak processing state, ensure the reliability of power supply, and thus ensure the stable operation of the server.

[0005] In order to solve the above technical problems, the present invention provides a device power supply system, including a power board, the power board including a power supply component and a plurality of power supply branches connected to the power supply component, the plurality of power supply branches including a first power supply branch connected to a device expansion card, and a second power supply branch connected to a server component, the device power supply system also includes:

[0006] A first capacitor module connected to the first power supply branch;

[0007] a second capacitor module and a switching module corresponding to the second power supply branch, wherein the switching module is used to connect the second capacitor module to the second power supply branch when the server component is in place, and to connect the second capacitor module to the first power supply branch when the server component is out of place;

[0008] The management module is used to control the third capacitor module on the device expansion card to supply power to the computing device on the device expansion card that meets the voltage warning condition.

[0009] Optionally, the switching module includes:

[0010] An on-site detection circuit is configured to receive a first on-site detection signal of the server component and a second on-site detection signal of the computing device, output a first trigger signal when the first on-site detection signal is an on-site signal, and output a second trigger signal when the first on-site detection signal is an out-of-site signal and the second on-site detection signal is an on-site signal;

[0011] a first switch module, configured to control the second capacitor module to be disconnected from the first power supply branch upon receiving the first trigger signal, and control the second capacitor module to be connected to the first power supply branch upon receiving the second trigger signal;

[0012] The second switch module is used to control the second capacitor module and the second power supply branch to be connected after receiving the first trigger signal, and control the second capacitor module and the second power supply branch to be disconnected after receiving the second trigger signal.

[0013] Optionally, the presence detection circuit includes a first NOT gate, an AND gate and a triode, wherein:

[0014] The input end of the first NOT gate is connected to the first presence detection signal of the server component, the output end of the first NOT gate is connected to the first input end of the AND gate, the second input end of the AND gate is connected to the second presence detection signal of the computing device, the output end of the AND gate is connected to the base of the transistor, the emitter of the transistor is connected to the first power supply, and the collector of the transistor outputs the first trigger signal or the second trigger signal.

[0015] Optionally, the first switch module includes a first switch tube, the second switch module includes a second switch tube, a control end of the first switch tube is connected to the first trigger signal or the second trigger signal, and the control end of the second switch tube performs different switching actions of the first switch tube and the second switch tube under the same trigger signal;

[0016] or,

[0017] The first switch module includes a first switch tube, and the second switch module includes a second NOT gate and a second switch tube. The control end of the first switch tube is connected to the first trigger signal or the second trigger signal, the input end of the second NOT gate is connected to the first trigger signal or the second trigger signal, and the output end of the second NOT gate is connected to the control end of the second switch tube. The first switch tube and the second switch tube perform the same switching action under the same trigger signal.

[0018] Optionally, the device expansion card includes a plurality of computing devices, and a plurality of sub-power supply branches for supplying power to the plurality of computing devices in a one-to-one correspondence;

[0019] The management module is specifically used to obtain the actual voltage value corresponding to each of the computing devices. When there is a target computing device among all the computing devices whose actual voltage value drops to a preset voltage warning value, it is determined that the target computing device meets the voltage warning condition, and the third capacitor module is connected to the sub-power supply branch corresponding to the target computing device.

[0020] Optionally, the management module includes:

[0021] A switch provided between the third capacitor module and each of the sub-power supply branches, configured to be turned on after receiving a first switch signal and turned off after receiving a second switch signal;

[0022] The manager is configured to obtain an actual voltage value corresponding to each of the computing devices, determine whether there is a target computing device among all the computing devices whose actual voltage value drops to a preset voltage warning value, and if so, generate a switching signal corresponding to the target computing device;

[0023] A controller, used to output the first switching signal to a first switching switch and output the second switching signal to a second switching switch according to the switching signal; the first switching switch is a switching switch between a sub-power supply branch corresponding to a target computing device and the third capacitor module, and the second switching switch is the switching switch other than the first switching switch.

[0024] Optionally, the management module further includes an electronic fuse provided on each of the sub-power supply branches, for collecting an actual voltage value of a computing device corresponding to the sub-power supply branch where the fuse is located;

[0025] The manager is specifically used to obtain the actual voltage value collected by each of the electronic fuses, determine whether there is a target computing device among all the computing devices whose actual voltage value drops to a preset voltage warning value, and if so, generate a switching signal corresponding to the target computing device.

[0026] Optionally, the device power supply system further includes:

[0027] A temperature monitoring component is used to monitor the first temperature information of the power board and the second temperature information on the device expansion board, and in response to the first temperature information and / or the second temperature information meeting the over-temperature condition, determine the target power supply branch and perform an over-temperature protection operation on the target power supply branch.

[0028] Optionally, the device power supply system further includes:

[0029] The indicating component corresponding to each of the second capacitor modules is used to indicate the power supply branch to which the second capacitor module is connected.

[0030] In order to solve the above technical problems, the present invention further provides a server, comprising at least one device expansion board, a plurality of server components and the device power supply system described in any one of the above.

[0031] The present invention provides a device power supply system. When any server component is not in place, the corresponding second capacitor module is connected to the first power supply branch, and the second capacitor module and the first capacitor module are used together to supply power to the device expansion card. At the same time, a management module and a third capacitor module are set on the device expansion board. When the management module monitors a computing device with a voltage warning condition on the device expansion board, the third capacitor module is controlled to supply power to the computing device, thereby meeting the power supply requirements of each computing device on the device expansion board when operating in the EDPP (electrical data peak processing) state, ensuring the reliability of power supply, and thus ensuring the stable operation of the server.

[0032] The present invention also provides a server having the same beneficial effects as the above-mentioned equipment power supply system. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] In order to more clearly illustrate the embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0034] Figure 1 A schematic diagram of the structure of a device power supply system provided by the present invention;

[0035] Figure 2 A schematic diagram of the structure of a switching module provided by the present invention;

[0036] Figure 3 This is a schematic diagram of the structure of a management module provided by the present invention. DETAILED DESCRIPTION

[0037] The core of the present invention is to provide a device power supply system and server, which can meet the power supply requirements of various computing devices on the device expansion board when operating in the electrical data peak processing state, ensure the reliability of power supply, and thus ensure the stable operation of the server.

[0038] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0039] First, please refer to Figure 1 The present invention provides a device power supply system, including a power board 1, the power board 1 includes a power supply component 11 and a plurality of power supply branches connected to the power supply component 11, the plurality of power supply branches include a first power supply branch connected to a device expansion card, and a second power supply branch connected to a server component, and the device power supply system also includes:

[0040] A first capacitor module 12 connected to a first power supply branch;

[0041] A second capacitor module 14 and a switch module 13 corresponding to the second power supply branch, the switch module 13 is used to connect the second capacitor module 14 to the second power supply branch when the server component is in place, and connect the second capacitor module 14 to the first power supply branch when the server component is out of place;

[0042] The management module 2 is used to control the third capacitor module 3 on the device expansion card to supply power to the computing device on the device expansion card that meets the voltage warning condition.

[0043] In this embodiment, the power board 1 includes a power supply component 11 and a plurality of power supply branches. The first end of the power supply branch is connected to the output end of the power supply component 11, and the second end of the power supply branch is connected to the server load. The server load includes but is not limited to device expansion boards, server components, etc. The server components include but are not limited to network cards, hard disks, etc. In this embodiment, the power supply branch whose second end is connected to the server component is determined as the second power supply branch, and the power supply branch whose second end is connected to the device expansion card is determined as the first power supply branch. Among them, the device expansion board includes but is not limited to a GPU riser card. It can be understood that the computing device installed in the slot of the GPU riser card is a GPU device. It can be understood that there can be multiple first power supply branches and second power supply branches. Figure 1 A first power supply branch and two second power supply branches are shown in FIG.

[0044] In order to ensure that server components such as network cards, hard disks, and device expansion cards in the server can be powered in time when they are running at high speed or high performance, each power supply branch is also designed with a capacitor module. In this embodiment, a first capacitor module 12 that is fixedly connected is provided on the first power supply branch, and a second capacitor module 14 that can be selectively connected is provided on the second power supply branch.

[0045] It can be understood that the server has different configuration requirements, so some hard disk configurations or network card configurations may not coexist at the same time. For example, in some configurations of the server, 3 network cards need to be configured, and in some configurations, 2 network cards need to be configured, but the power supply design still needs to be reserved, that is, 3 second power supply branches corresponding to the network cards and the second capacitor module 14 are reserved. On this basis, the present embodiment designs a switching module 13 for each second power supply branch, and the switching module 13 monitors whether the second end of the second power supply branch is configured with a server component. If the server component is configured, that is, when the server component is in place, the second capacitor module 14 is connected to the second power supply branch to ensure that the second capacitor module 14 can charge the server component in time when the server is running at a high rate or high performance. If the server component is not configured, that is, when the server is out of place, the second capacitor module 14 is connected to the first power supply branch to ensure that when the computing device is running in the EDPP state, when the power consumption increases instantly, the first capacitor module 12 and at least one second capacitor module 14 jointly meet the power supply demand of the computing device, which reduces the possibility of power failure of the computing device to a certain extent, and improves the utilization rate of the capacitor module.

[0046] In addition, in order to meet the power supply requirements of the computing device during EDPP, a capacitor will be designed on the sub-power supply branch corresponding to each computing device on the device expansion card to power the computing device. Due to the limited structural space in the server chassis and the small size of the device expansion card, the number of capacitors that can be arranged on it is limited, and the average number of capacitors per computing device is even smaller. Setting a capacitor on each sub-power supply branch alone cannot meet the power supply requirements of the computing device during EDPP. Considering that the EDPP of the computing device is a cyclic process, the computing device rises from low power consumption to peak power consumption and then to low power consumption, and the peak power consumption state lasts for milliseconds, and each computing device on the device expansion card runs separately and at different frequencies. Therefore, in order to solve the problem of insufficient power supply of the computing device during peak power consumption, this embodiment forms a third capacitor module 3 with all the compensation capacitors on the device expansion card, and monitors each computing device through the management module 2. When any computing device meets the voltage warning condition, the third capacitor module 3 is switched to the sub-power supply branch of the computing device to ensure the power supply requirements of the computing device and avoid power failure of the computing device during EDPP operation, thereby ensuring the stable operation of the server.

[0047] In this embodiment, each capacitor module is a combination of multiple capacitors connected in series and / or in parallel.

[0048] It can be seen that in this embodiment, when any server component is not in place, its corresponding second capacitor module 14 is connected to the first power supply branch, and together with the first capacitor module 12, it supplies power to the device expansion card. At the same time, a management module 2 and a third capacitor module 3 are set on the device expansion board. When the management module 2 monitors a computing device with a voltage warning condition on the device expansion board, the third capacitor module 3 is controlled to supply power to the computing device, thereby meeting the power supply requirements of each computing device on the device expansion board when operating in the EDPP (electrical data peak processing) state, ensuring the reliability of power supply, and thus ensuring the stable operation of the server.

[0049] Based on the above embodiments:

[0050] In an exemplary embodiment, the switching module 13 includes:

[0051] An in-place detection circuit, for receiving a first in-place detection signal of a server component and a second in-place detection signal of a computing device, outputting a first trigger signal when the in-place detection signal is an in-place signal, and outputting a second trigger signal when the in-place detection signal is an out-of-place signal and the second in-place detection signal is an in-place signal;

[0052] A first switch module, used for controlling the second capacitor module 14 to be disconnected from the first power supply branch upon receiving a first trigger signal, and controlling the second capacitor module 14 to be connected to the first power supply branch upon receiving a second trigger signal;

[0053] The second switch module is used for controlling the second capacitor module 14 to be connected to the second power supply branch upon receiving the first trigger signal, and controlling the second capacitor module 14 to be disconnected from the second power supply branch upon receiving the second trigger signal.

[0054] In this embodiment, the switching module 13 and the second capacitor module 14 corresponding to a second power supply branch are used as an example for explanation, and the other second power supply branches are similar. The switching module 13 includes an in-place detection circuit, a first switch module and a second switch module, the first switch module is arranged between the second capacitor module 14 and the first power supply branch, and the second switch module is arranged between the second power supply branch and the second capacitor module 14. The in-place detection circuit is respectively connected to the in-place detection pin of the server component and the in-place detection pin of the computing device to obtain a first in-place detection signal and a second in-place detection signal, wherein the first in-place detection signal is a signal from the in-place detection pin of the server component, which is used to indicate whether the server component is in place, and the second in-place detection signal is a signal from the in-place detection pin of the computing device, which is used to indicate whether the computing device is in place. When the server component is in place, the in-place detection circuit outputs a first trigger signal, and when the server component is out of place and the computing device is in place, the in-place detection circuit outputs a second trigger signal.

[0055] The first switch module is configured to control the connection state between the second capacitor module 14 and the first power supply branch according to the received trigger signal. When the first trigger signal is received, the connection between the second capacitor module 14 and the first power supply branch is disconnected, and when the second trigger signal is received, the connection between the second capacitor module 14 and the first power supply branch is connected, so that the second capacitor module 14 supplies power to the computing device.

[0056] The second switch module is configured to control the connection state between the second capacitor module 14 and the second power supply branch according to the received trigger signal. When the first trigger signal is received, the connection between the second capacitor module 14 and the second power supply branch is connected to ensure that the server component can obtain power through the second power supply branch when it is in place. When the second trigger signal is received, the connection between the second capacitor module 14 and the second power supply branch is disconnected, which usually occurs when the server component is out of place and power needs to be distributed to the computing device.

[0057] In an exemplary embodiment, the presence detection circuit includes a first NOT gate, an AND gate and a transistor, wherein:

[0058] The input end of the first NOT gate is connected to the first presence detection signal of the server component, the output end of the first NOT gate is connected to the first input end of the AND gate, the second input end of the AND gate is connected to the second presence detection signal of the computing device, the output end of the AND gate is connected to the base of the transistor, the emitter of the transistor is connected to the first power supply, and the collector of the transistor outputs the first trigger signal or the second trigger signal.

[0059] Take the example that the in-position signal is high and the out-position signal is low. When the server component is in position, the input of the first NOT gate is high and the output is low. Since the second input of the AND gate (the second in-position detection signal of the computing device) may be high or low, but since the first input of the AND gate is low, the output of the AND gate is low, the transistor is not turned on, and the collector of the transistor outputs a low level, that is, the first trigger signal. When the server component is out of position, the input of the first NOT gate is low and the output is high. If the computing device is in position at this time, the second in-position detection signal is high, and all the inputs of the AND gate are high, so the output of the AND gate is high, the transistor is turned on, and the collector outputs a high level, that is, the second trigger signal. It can be understood that the digital logic circuit constructed by the first NOT gate, the AND gate and the transistor has a fast response time, can detect the change of the device in-position state in time, and quickly output the trigger signal, and the circuit structure is simple and easy to integrate.

[0060] In an exemplary embodiment, the first switch module includes a first switch tube, the second switch module includes a second switch tube, the control end of the first switch tube is connected to the first trigger signal or the second trigger signal, and the control end of the second switch tube, the first switch tube and the second switch tube perform different switching actions under the same trigger signal.

[0061] In this embodiment, the first switch tube can be selected as an NMOS tube, and the second switch tube can be selected as a PMOS tube. Accordingly, the first switch tube can be selected as a PMOS tube, and the second switch tube can be selected as an NMOS tube. Under this structure, a trigger signal of the in-position detection circuit can trigger the first switch tube and the second switch tube to be complementary turned on. The design of complementary conduction can ensure that the switching of the power supply is more accurate, reduce the power supply fluctuation caused by the asynchronous switching action, and improve the stability and reliability of the system. In an exemplary embodiment, the first switch module includes a first switch tube, and the second switch module includes a second NOT gate and a second switch tube. The control end of the first switch tube is connected to the first trigger signal or the second trigger signal, and the input end of the second NOT gate is connected to the first trigger signal or the second trigger signal. The output end of the second NOT gate is connected to the control end of the second switch tube, and the first switch tube and the second switch tube perform the same switching action under the same trigger signal.

[0062] In this embodiment, the first switch tube and the second switch tube can both select NMOS tubes or both select PMOS tubes. Using the same type of switch tubes (all NMOS or PMOS) can simplify the circuit design because their characteristics are similar, which is convenient for design and debugging. In the production process, using the same type of switch tubes can improve assembly efficiency. On this basis, a second NOT gate is also set on the second switch module to achieve complementary conduction of the two switch tubes. The two switch tubes can be controlled by a single trigger signal to achieve a complex power management strategy.

[0063] Combination Figure 2 To explain, Figure 2It includes an AND gate U1, a first NOT gate U2, a transistor Q1, a first switch tube Q2, and a second switch tube Q3. The first switch tube selects an NMOS tube, and the second switch tube selects a PMOS tube. The emitter of the transistor Q1 is connected to the power supply 3V3. When the hard disk is in place or the GPU is not in place, the AND gate U1 inputs a low level and outputs a low level. At this time, the transistor Q1 is in a closed state, the gates of the second switch tube Q3 and the first switch tube Q2 are low levels, the second switch tube Q3 is turned on, the first switch tube Q2 is turned off, and the second capacitor module 14 is connected to the hard disk power supply branch P12V_SSD, that is, the second power supply branch; when the hard disk is not in place and the computing device is in place, the AND gate U1 inputs a high level and outputs a high level. At this time, the transistor Q1 is in the on state, the gate voltage of the second switch tube Q3 and the first switch tube Q2 is 3.3V, the second switch tube Q3 is turned off, the first switch tube Q2 is turned on, and the second capacitor module 14 is connected to the GPU power supply branch P12V_GPU, that is, the first power supply branch.

[0064] In an exemplary embodiment, the device expansion card includes a plurality of computing devices, and a plurality of sub-power supply branches for supplying power to the plurality of computing devices in a one-to-one correspondence;

[0065] The management module 2 is specifically used to obtain the actual voltage value corresponding to each computing device. When there is a target computing device among all computing devices whose actual voltage value drops to the preset voltage warning value, it is determined that the target computing device meets the voltage warning condition, and the third capacitor module 3 is connected to the sub-power supply branch corresponding to the target computing device.

[0066] In this embodiment, the first end of the sub-power supply branch is connected to the second end of the first power supply branch, and the second end of the sub-power supply branch is connected to the computing device. The management module 2 first presets a preset voltage warning value, and compares the actual voltage value of each computing device with the preset voltage warning value. If the actual voltage value drops to the preset voltage warning value, the computing device is determined to be a target computing device, and the third capacitor module 3 is connected to the sub-power supply branch corresponding to the target computing device to meet the power supply demand of the target computing device, avoid power failure of the target computing device, and ensure stable operation of the server.

[0067] In an exemplary embodiment, the management module 2 includes:

[0068] A switch provided between the third capacitor module 3 and each sub-power supply branch, configured to be turned on after receiving a first switch signal and turned off after receiving a second switch signal;

[0069] The manager is used to obtain the actual voltage value corresponding to each computing device, determine whether there is a target computing device among all computing devices whose actual voltage value drops to a preset voltage warning value, and if so, generate a switching signal corresponding to the target computing device;

[0070] The controller is used to output a first switching signal to the first switching switch and output a second switching signal to the second switching switch according to the switching signal; the first switching switch is a switching switch between the sub-power supply branch corresponding to the target computing device and the third capacitor module 3, and the second switching switch is a switching switch other than the first switching switch.

[0071] In this embodiment, the manager can be a BMC chip on the server motherboard, and the controller can be a CPLD chip on the server motherboard, which is interconnected with the device expansion card through a cable to realize signal transmission. The switching switches can all select NMOS tubes, and each switching switch is respectively set between the third capacitor module 3 and each sub-power supply branch. After the manager determines the target computing device, it generates a switching signal corresponding to the target computing device. After receiving the switching signal, the controller outputs a first switching signal to the switching switch corresponding to the target computing device to control the switching switch corresponding to the target computing device to turn on, thereby connecting the third capacitor module 3 to the sub-power supply branch corresponding to the target computing device, and outputs a second switching signal to the switching switches corresponding to other computing devices to control the switching switch corresponding to the target computing device to disconnect, thereby disconnecting the third capacitor module 3 and the sub-power supply branches corresponding to other computing devices.

[0072] In an exemplary embodiment, the management module 2 further includes an electronic fuse provided on each sub-power supply branch, for collecting the actual voltage value of the computing device corresponding to the sub-power supply branch where the management module 2 is located;

[0073] The manager is specifically used to obtain the actual voltage value collected by each electronic fuse, determine whether there is a target computing device among all computing devices whose actual voltage value drops to a preset voltage warning value, and if so, generate a switching signal corresponding to the target computing device.

[0074] In this embodiment, the electronic fuse can select an electronic fuse chip with PMbus function, which can monitor voltage, current and other information in real time. The electronic fuse can also protect the computing device. It can be understood that the access point of the third capacitor module 3 on the sub-power supply circuit should be set at the rear end of the electronic fuse to respond more quickly and power the computing device.

[0075] Take the device expansion board as GPU riser, which includes four GPUs as an example, refer to Figure 3 As shown, the management module 2 is mainly constructed by a manager U3, a controller U4, a first switch Q4, a second switch Q5, a third switch Q6, and a fourth switch Q7. Figure 3In the text, the sub - power supply branch of the first computing device is represented by P12V_GPU1, the sub - power supply branch of the second computing device is represented by P12V_GPU2, the sub - power supply branch of the third computing device is represented by P12V_GPU3, and the sub - power supply branch of the fourth computing device is represented by P12V_GPU4.

[0076] On the GPU riser, an eFuse (electronic fuse) with PMbus function is used. The eFuse can obtain the real - time current and voltage information of each GPU and transmit it to the manager U3 through the I2C signal. The preset voltage warning value of the manager U3 is 1.02×V min value (V min value is the minimum operating voltage of the GPU). When it is monitored that the actual voltage value of a certain GPU drops to the voltage warning value, the manager U3 transmits a signal to the controller U4. The controller U4 controls the gate voltages of the first switching switch Q4, the second switching switch Q5, the third switching switch Q6, and the fourth switching switch Q7 to switch the capacitor bank to the power supply circuit of this GPU.

[0077] Combined Figure 2 and Figure 3 are described as follows:

[0078] After the server system is powered on, when it is detected that the hard disk / network card is not in place and the GPU is in place, the first NOT gate U2 outputs a high level, the AND gate U1 outputs a high level, and the triode Q1 conducts. The gate voltages of the first switching transistor Q3 and the second switching transistor Q2 are 3.3V. At this time, for the first switching transistor Q3, Vgs < Vth, so the first switching transistor Q3 is turned off; for the second switching transistor Q2, Vgs > Vth, so the second switching transistor Q2 conducts, and the second capacitor module 14 is switched to the first power supply branch corresponding to the GPU, and the second capacitor module 14 starts to store electricity. The GPU starts to run, and the power supply powers the GPU. At the same time, the first capacitor module 12 and the second capacitor module 14 on the power supply board 1 also power the GPU to ensure that the voltage of the GPU is maintained within the operating voltage range. Each eFuse on the GPU riser transmits information such as the voltage and current of the GPU to the manager U3 through the I2C signal. When it is monitored that the voltage V GPU 1≤1.02×V minWhen the voltage of GPU 01 is no longer reduced, the BMC determines to switch the third capacitor module 3 on the riser card to the sub-power supply branch of GPU 01, and transmits the information to the controller U4. The controller U4 outputs a high-level signal to control the first switching switch Q4 to be turned on, and outputs a low-level signal to control the second switching switch Q5, the third switching switch Q6, and the fourth switching switch Q7 to be turned off. At this time, the third capacitor module 3 charges GPU 01 to ensure that the voltage of GPU 01 does not drop, so that the GPU can run stably during EDPP. Repeat the above steps. When the actual voltage value of a GPU is monitored to reach the voltage warning value, the controller U4 controls the switching switch to cut the third capacitor module 3 into the corresponding sub-power supply branch to power the GPU.

[0079] In an exemplary embodiment, the device power supply system further includes:

[0080] The temperature monitoring component is used to monitor the first temperature information of the power board 1 and the second temperature information on the device expansion board. In response to the first temperature information and / or the second temperature information meeting the over-temperature condition, the target power supply branch is determined and an over-temperature protection operation is performed on the target power supply branch.

[0081] Among them, monitoring the first temperature information of the power board 1 involves measuring the temperature of the power supply component 11 and each power supply branch. Monitoring the second temperature information on the device expansion board involves measuring the temperature of the GPU or other computing devices. The temperature monitoring component evaluates the monitored temperature information to determine whether a predetermined over-temperature condition is met. The over-temperature condition can be a fixed temperature threshold or a threshold that is dynamically adjusted according to system load and environmental conditions.

[0082] When the temperature monitoring component detects overtemperature, it is necessary to determine which power supply branch is the cause of the overtemperature, or which power supply branch needs to reduce the load the most to reduce the temperature, and perform overtemperature protection operations on it, including but not limited to reducing the supply voltage or current, disconnecting the power supply branch, temporarily stopping power supply to specific devices, adjusting fan speed, increasing heat dissipation, issuing alarms, notifying users or system administrators, etc. By adding temperature monitoring components, the device power supply system can manage power distribution more intelligently, prevent equipment damage due to overheating, and extend the service life of the system. In addition, this design also helps to improve system reliability and user sense of security.

[0083] In an exemplary embodiment, the device power supply system further includes:

[0084] The indicating component corresponding to each second capacitor module 14 is used to indicate the power supply branch to which the second capacitor module 14 is connected.

[0085] In this embodiment, considering that the second capacitor module 14 can be connected to the first power supply branch or the second power supply branch, the power supply branch connected to the second capacitor module 14 is indicated by the indicating component, so that the power supply status of each second capacitor module 14 can be intuitively understood. When power-related problems occur, the indicating component can help technicians quickly locate the problem, such as determining whether a capacitor module is not correctly connected to the power supply branch. By indicating the power supply status of the capacitor module in real time, equipment damage caused by incorrect power configuration can be prevented.

[0086] The indicating component may be an LED lamp, a display screen or other indicating devices, and a suitable type may be selected according to actual needs.

[0087] In a second aspect, the present invention further provides a server, comprising at least one device expansion board, a plurality of server components, and the device power supply system described in any one of the above embodiments.

[0088] For an introduction to a server provided by the present invention, please refer to the above-mentioned embodiment, and the present invention will not be described in detail here.

[0089] A server provided by the present invention has the same beneficial effects as the above-mentioned equipment power supply system.

[0090] It should also be noted that, in this specification, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "comprises", "comprising" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the statement "comprising a ..." does not exclude the presence of other identical elements in the process, method, article or device including the element.

[0091] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A device power supply system, characterized in that: The device power supply system includes a power board, the power board includes a power supply component and a plurality of power supply branches connected to the power supply component, the plurality of power supply branches include a first power supply branch connected to a device expansion card, and a second power supply branch connected to a server component, and the device power supply system also includes: A first capacitor module connected to the first power supply branch; a second capacitor module and a switching module corresponding to the second power supply branch, wherein the switching module is used to connect the second capacitor module to the second power supply branch when the server component is in place, and to connect the second capacitor module to the first power supply branch when the server component is out of place; The management module is used to control the third capacitor module on the device expansion card to supply power to the computing device on the device expansion card that meets the voltage warning condition.

2. The device power supply system according to claim 1, characterized in that: The switching module comprises: An on-site detection circuit is configured to receive a first on-site detection signal of the server component and a second on-site detection signal of the computing device, output a first trigger signal when the first on-site detection signal is an on-site signal, and output a second trigger signal when the first on-site detection signal is an out-of-site signal and the second on-site detection signal is an on-site signal; a first switch module, configured to control the second capacitor module to be disconnected from the first power supply branch upon receiving the first trigger signal, and control the second capacitor module to be connected to the first power supply branch upon receiving the second trigger signal; The second switch module is used to control the second capacitor module and the second power supply branch to be connected after receiving the first trigger signal, and control the second capacitor module and the second power supply branch to be disconnected after receiving the second trigger signal.

3. The device power supply system according to claim 2, characterized in that: The in-position detection circuit comprises a first NOT gate, an AND gate and a triode, wherein: The input end of the first NOT gate is connected to the first presence detection signal of the server component, the output end of the first NOT gate is connected to the first input end of the AND gate, the second input end of the AND gate is connected to the second presence detection signal of the computing device, the output end of the AND gate is connected to the base of the transistor, the emitter of the transistor is connected to the first power supply, and the collector of the transistor outputs the first trigger signal or the second trigger signal.

4. The device power supply system according to claim 3, characterized in that: The first switch module includes a first switch tube, the second switch module includes a second switch tube, the control end of the first switch tube is connected to the first trigger signal or the second trigger signal, and the control end of the second switch tube performs different switching actions under the same trigger signal; or, The first switch module includes a first switch tube, and the second switch module includes a second NOT gate and a second switch tube. The control end of the first switch tube is connected to the first trigger signal or the second trigger signal, the input end of the second NOT gate is connected to the first trigger signal or the second trigger signal, and the output end of the second NOT gate is connected to the control end of the second switch tube. The first switch tube and the second switch tube perform the same switching action under the same trigger signal.

5. The device power supply system according to claim 1, characterized in that: The device expansion card includes a plurality of computing devices and a plurality of sub-power supply branches for supplying power to the plurality of computing devices in a one-to-one correspondence; The management module is specifically used to obtain the actual voltage value corresponding to each of the computing devices. When there is a target computing device among all the computing devices whose actual voltage value drops to a preset voltage warning value, it is determined that the target computing device meets the voltage warning condition, and the third capacitor module is connected to the sub-power supply branch corresponding to the target computing device.

6. The device power supply system according to claim 5, characterized in that: The management module includes: A switch provided between the third capacitor module and each of the sub-power supply branches, configured to be turned on after receiving a first switch signal and turned off after receiving a second switch signal; The manager is configured to obtain an actual voltage value corresponding to each of the computing devices, determine whether there is a target computing device among all the computing devices whose actual voltage value drops to a preset voltage warning value, and if so, generate a switching signal corresponding to the target computing device; A controller, used to output the first switching signal to a first switching switch and output the second switching signal to a second switching switch according to the switching signal; the first switching switch is a switching switch between a sub-power supply branch corresponding to a target computing device and the third capacitor module, and the second switching switch is the switching switch other than the first switching switch.

7. The device power supply system according to claim 6, characterized in that: The management module also includes an electronic fuse provided on each of the sub-power supply branches, for collecting the actual voltage value of the computing device corresponding to the sub-power supply branch where the fuse is located; The manager is specifically used to obtain the actual voltage value collected by each of the electronic fuses, determine whether there is a target computing device among all the computing devices whose actual voltage value drops to a preset voltage warning value, and if so, generate a switching signal corresponding to the target computing device.

8. The device power supply system according to claim 1, characterized in that: The equipment power supply system also includes: A temperature monitoring component is used to monitor the first temperature information of the power board and the second temperature information on the device expansion board, and in response to the first temperature information and / or the second temperature information meeting the over-temperature condition, determine the target power supply branch and perform an over-temperature protection operation on the target power supply branch.

9. The device power supply system according to any one of claims 1 to 8, characterized in that: The equipment power supply system also includes: The indicating component corresponding to each of the second capacitor modules is used to indicate the power supply branch to which the second capacitor module is connected.

10. A server, characterized in that: It comprises at least one device expansion board, a plurality of server components and the device power supply system according to any one of claims 1 to 9.

Images