OR-ing transistor body braking in phase redundancy scheme

By configuring low-side transistors and OR-ing transistors in the phase-redundant power supply, combined with inductors and a phase-redundant controller, the problem of slow current drop during no-load transients in the phase-redundant power supply is solved, achieving rapid current drop and stable output voltage, thus improving the transient performance and reliability of the system.

CN111756242BActive Publication Date: 2024-11-15ALPHA & OMEGA SEMICONDUCTOR (CAYMAN) LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202010217036.0
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Priority Date
2019-03-28
Filing Date
2020-03-24
Publication Date
2024-11-15
Estimated Expiration
2040-03-24

AI Technical Summary

Technical Problem

In modern server systems, phase redundancy power supply schemes suffer from inefficiency when dealing with transient performance and power fault isolation, especially when the power supply cannot quickly reduce current under no-load transient conditions, leading to output voltage overshoot.

Method used

By configuring low-side transistors, OR-ing transistors, and inductors in a phase-redundant power supply, and utilizing a phase-redundant controller to disconnect these transistors during no-load transients, combined with the energy release of the body diode, a rapid decrease in induced current is achieved, with the induced current decrease time being less than 2.4 microseconds.

Benefits of technology

It effectively suppresses output voltage overshoot, improves the response speed and reliability of the power supply system under no-load transient conditions, and ensures that the system can still operate stably under multiple power level failure conditions.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN111756242B_ABST
    Figure CN111756242B_ABST
Patent Text Reader

Abstract

A method, system, and computer program product for improving the fall time of an inductive current in a DC-to-DC converter having an inductor conductively coupled to a first terminal of a low-side transistor and an or-ing transistor coupled to a second terminal, wherein the DC-to-DC converter is in a phase-redundant power supply. The method includes disconnecting the low-side transistor and disconnecting the or-ing transistor when a no-load transient occurs.
Need to check novelty before this filing date? Find Prior Art

Description

Field of the Invention

[0001] Various aspects of the present invention generally relate to power converters, and more particularly to DC-to-DC power converters configured in a phase redundant scheme. Background of the Invention

[0002] Modern loads (servers, GPUs, ASICs, network switches, and other high current loads) host a need for compact, reliable, and cost-effective server systems. The primary cause of server system failure is the power supply. Reliability is improved by using phase redundant power supply solutions in many advanced servers. Phase redundant power supply solutions typically have multiple redundant power stages. If a power stage fails, the phase redundant power supply can detect and isolate the failed power stage, allowing the rest of the system to operate normally without interference. Commonly used phase redundant power supplies have N+2 power stages, where N is the number of phases. For a phase redundant power supply with N+2 power stages, the power supply can still operate normally when 2 power stages fail.

[0003] In order to implement a phase redundant power supply solution, multiple power stages are configured together so that when one of them fails, it can be isolated from the power circuit, and the remaining power stages can share the load. Therefore, compared with non-redundant power supplies, the power stages in phase redundant power supplies usually have at least one additional transistor to shut down the power converter when a failure occurs.

[0004] Another major issue for servers is transient performance. Specifically, the power supply must be able to quickly reduce current when a no-load transient occurs.

[0005] It is under this premise that various embodiments of the present invention are proposed. Summary of the invention

[0006] The present invention relates to a method for improving the fall time of an inductive current in a DC-to-DC converter, wherein a low-side transistor is electrically coupled between a first end of an inductor and ground, and an OR-ing transistor is coupled between a second end of the inductor and an output end of the DC-to-DC converter, wherein the DC-to-DC converter is in a phase redundant power supply, the method comprising: when a no-load transient occurs, disconnecting the low-side transistor and disconnecting the OR-ing transistor.

[0007] Both the low-side transistor and the OR-ing transistor are turned off within 30 nanoseconds when a no-load transient is detected.

[0008] Here, the inductor releases energy through the body diode of the OR-ing transistor and the body diode of the low-side transistor.

[0009] Among them, the inductor induced current fall time is less than 2.4 microseconds.

[0010] The present invention also relates to a DC to DC converter in a phase redundant power supply, comprising: an inductor; a low-end transistor electrically coupled between a first end of the inductor and ground; an OR-ing transistor electrically coupled between a second end of the inductor and an output end of the DC to DC converter; and a phase redundant controller interactively coupled to the OR-ing transistor, wherein the phase redundant controller is configured to disconnect the OR-ing transistor when a no-load transient occurs.

[0011] The signal received by the low-side driver comes from a pulse width modulation signal controller and is used to disconnect the low-side transistor when a no-load transient occurs.

[0012] Both the low-side transistor and the OR-ing transistor are turned off within 30 nanoseconds when a no-load transient is detected.

[0013] Here, the inductor releases energy through the body diode of the OR-ing transistor and the body diode of the low-side transistor.

[0014] Among them, the inductor induced current fall time is less than 2.4 microseconds.

[0015] The invention also includes: a high-end transistor electrically coupled between the first end of the inductor and the input end of the DC-to-DC converter. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Other features and advantages of the present invention will become apparent upon reading the following detailed description and referring to the following drawings:

[0017] Figure 1 A circuit diagram of a DC to DC voltage converter in a phase redundant power supply is shown according to one embodiment of the present invention.

[0018] Figure 2 A set of graphs showing the variation of sensed current, pulse width modulation (PWM) signal, smart power stage (SPS), gate low (GL) drive signal, and gate low (LGATE) from a phase redundancy controller (PRC) over time in accordance with various aspects of the present invention. DETAILED DESCRIPTION

[0019] Although many typical details are included in the following detailed description for the purpose of explanation, it should be understood by those skilled in the art that modifications and variations made according to the following details are within the scope of the present invention. Therefore, the typical embodiments described below do not cause any general loss to the claimed invention and do not impose any limitations on the claimed invention.

[0020] system

[0021] According to various aspects of the present invention, when a power converter in a phase redundant power supply experiences a no-load transient, an output voltage overshoot can be suppressed by disconnecting an OR-ing transistor during the transient. Figure 1 A circuit diagram of a DC-to-DC converter 100 configured in a phase redundant power supply for suppressing output voltage overshoot according to various aspects of the present invention is shown. The DC-to-DC converter herein is part of a phase redundant power supply. As part of a phase redundant power supply, the DC-to-DC converter has a plurality of components such that when a fault occurs, the DC-to-DC converter is disconnected from the power circuit, and these components are a high-side disconnect transistor Q1, an OR-ing transistor Q2, and a phase redundant controller 101. The DC-to-DC converter also includes active and passive power components configured to modulate the output voltage (V out ) and monitor load current. These active power supply elements are high-side transistor Q3, low-side transistor Q4, inductor 104 and capacitor (not shown). Additional voltage monitoring and control units, such as proportional-integral-derivative controller (PID) and pulse width modulation signal controller (PWM) are not shown, but they are at least AC-coupled and conductively coupled through PWM lead 102, respectively.

[0022] The DC-DC converter in the phase redundant power supply receives the online voltage (V in ) current on the phase redundant power supply. As an example, but not by way of limitation, the voltage source 103 can provide a 12 volt line voltage. However, it should be understood by those skilled in the art that the voltage source can be any voltage and can be any type of voltage source used in the art. The current from the voltage source 103 flows through the resistor R1, generating a voltage drop. The phase redundant controller 101 monitors the voltage drop across the resistor R1 through the current sensing leads CS+ and CS-. When a fault is detected through CS+ and CS-, the phase redundant controller 101 will send a signal to disconnect the high-side shutdown transistor Q1 and the OR-ing transistor Q2 to protect the load from the fault and enable the phase redundant power supply to continue to operate with little or no disconnection time. According to various aspects of the present invention, the phase redundant controller can receive a signal from the PWM lead 102 through PWM. The PWM signal can cause the PRC to disconnect the OR-ing transistor Q2 during a no-load transient. In addition, the phase redundant controller can monitor the reverse current through the OR-ing gate using the leads Rev+ and Rev-. The controller 101 sends a configuration signal to control the gate of the OR-ing transistor Q2 through LGATE, so that the DC-to-DC converter can disconnect the current when the Q4 FET fails. When the Q1 transistor fails, the SW node will be equivalent to V inWith the combination of R2, R3, and C1, the LX node will monitor the voltage rise with a time constant of (R2||R3)*C1. When LX rises above the internal fault threshold, the comparator in the PRC indicates that the Q1 transistor is faulty and sends a fault, disconnecting Q1 and Q2.

[0023] In some embodiments, the PRC may be configured in hardware with sufficient circuitry to perform switching of OR-ing transistors in response to an overload transient. In other embodiments, the PRC may be configured as part of an integrated circuit, chip.

[0024] When the high-side cutoff transistor Q1 is in the “on” state, current flows through the high-side gate transistor and into the smart power stage 105. The high-side transistor Q3 switches to the “on” state to charge the inductor 104 according to the load requirements determined by the PWM controller. When the OR-ing gate Q2 is in the “on” state, energy is transferred through V out Before leaving the DC-DC converter, it flows through the high-side transistor Q3, the inductor 104, and the OR-ing gate Q2. Alternatively, after the OR-ing transistor, Q2 can be a capacitor (not shown). The current flowing through the inductor creates a magnetic field in the inductor that resists current changes. When the desired voltage level reaches V out When the high-side transistor Q3 turns off, the low-side transistor Q4 turns on. The magnetic field in the inductor begins to break down, causing current to flow through the circuit to V out When the voltage level is below the required voltage level, the high-side transistor Q3 is turned on and the low-side transistor Q4 can be turned off.

[0025] The load current requirement changes according to the change of the operating state. When the load enters a light load state, the current requirement will be very low. When a no-load transient occurs, the energy stored in the inductor 104 as a magnetic field must be released quickly to avoid output voltage overshoot. In the previous method, the inductor can only be discharged by disconnecting the high-end transistor Q3 and connecting the low-end transistor Q4. However, this often results in a non-optimal discharge time. According to various aspects of the present invention, the optimal discharge time can be achieved by disconnecting the high-end transistor Q3, the low-end transistor Q4, and the OR-ing transistor Q2 according to the no-load transient. Disconnecting the high-end transistor Q3, the low-end transistor Q4, and the OR-ing transistor Q2 causes the energy to be rapidly reduced from the inductor 104 through the body diodes of the low-end transistor Q4 and the OR-ing transistor Q2. This operation is also called body braking.

[0026] How it works

[0027] Figure 2Graphs showing the operation of a DC to DC power supply with OR-ing transistor body braking in accordance with various aspects of the present invention. Figure 2 The top curve shows the variation of the inductor current over time. During normal operation, the current through the inductor will fluctuate 201 to maintain the desired voltage output level. The PWM controller can detect the no-load transient condition 1 and can request the PWM to enter the following Figure 2 The three states 202 of the second curve above are shown. Figure 2 In the third curve above, the driver in the smart power stage (SPS) responds to PWM signal 2 by pulling GL low 203 low. This sets the low-side transistor to the “off” state. Figure 2 As shown in the fourth curve above, when a tri-state PWM signal is received, the PWM signal is also received by the phase redundancy controller (PRC), pulling the LGATE signal low 204, switching the OR-ing transistor to the "off" state. It should be noted that when Figure 2 The third and fourth curves above are shown to indicate that the driver pulls GL low before the PRC pulls LGATE low, but are not limited to this, and GL can be pulled low and LGATE can be pulled low at the same time, or GL can be pulled low after LGATE is pulled low. In any case, both GL and LGATE are pulled low, and the low-side transistor and the OR-ing transistor are switched to the "off" state shortly after the no-load transient is detected. With the low-side transistor and the OR-ing transistor in the "off" state, the inductive current quickly drops 205 to the desired level, a phenomenon known as inductor deceleration. According to various aspects of the present invention, the low-side transistor Q4 and the OR-ing transistor Q2 can be disconnected within 30 nanoseconds (ns) when the no-load transient is detected.

[0028] After the sensed current reaches the desired level, the PWM controller causes the PWM to leave the tri-state 4 and send a low signal. According to various aspects of the present invention, the deceleration time can be less than 2.4 microseconds (μs). The driver receives the low PWM signal at the SPS, which drives the GL signal high 5 to switch the low-side transistor to the "on" state. The PWM low signal is also received by the PRC, which sends a high signal 6 to LGATE, causing the OR-ing transistor to switch to the "on" state.

[0029] It is noted that a conductive coupling may be any surface that provides a conductive connection between components. For example, a conductive coupling may be a wire, a conductive track on a surface, a conductive lead, or other type of conductive connection that may be coupled between two components. Similarly, a conductive coupling may be any connection that transmits information between coupled components. Components may be AC ​​coupled to other components even when a third component feeds back information, as long as the information being transmitted results in substantially the same result.

[0030] According to various aspects of the present invention, components, process steps and / or data structures may use the following different types of operating systems; computing platforms; user interfaces / displays, including personal or laptop computers, electronic game consoles, PDAs and other handheld portable devices (such as mobile phones, handheld computers, portable game consoles, etc.); and / or general-purpose machines, etc. In addition, those skilled in the art should understand that less versatile devices, such as hard-wired devices, field programmable gate arrays (FOGs), application-specific integrated circuits (ASICs), etc., may also be used without departing from the scope and spirit of the inventive concepts disclosed in the present invention.

[0031] Although the present invention has been described in detail with respect to certain preferred versions, other versions are possible. Therefore, the scope of the present invention should not be determined by the above description, but rather by reference to the appended claims and all equivalents thereof. Any option (whether preferred or not) may be combined with any other option (whether preferred or not). In the following claims, unless otherwise stated, Indefinite article "a" or "a kind" All refer to the quantity of one or more items in the following content. Unless the function is clearly indicated by "meaning to be", the attached claims should not be considered as limiting the meaning and function. Any item in the claims that does not accurately indicate a specific function should be understood as the "meaning to be".

Claims

1. A method for improving the fall time of an inductive current in a DC to DC converter, characterized in that: A low-side transistor is electrically coupled between a first end of an inductor and ground, an OR-ing transistor is coupled between a second end of the inductor and an output of a DC-to-DC converter, wherein the DC-to-DC converter is in a phase-redundant power supply, the method comprising: When a load-down transient occurs, the low-side transistor is turned off, turning off the OR-ing transistor, and the load-down transient involves a reduction in current flowing through the inductor.

2. The method according to claim 1, characterized in that Both the low-side transistor and the OR-ing transistor turn off within 30 nanoseconds when the unload transient is detected.

3. The method according to claim 1, characterized in that The inductor releases energy through the body diode of the OR-ing transistor and the body diode of the low-side transistor.

4. The method according to claim 1, characterized in that The inductor current fall time is less than 2.4 microseconds.

5. The method according to claim 1, characterized in that The low-side transistor and the OR-ing transistor are both turned off for a period of time before the low-side transistor is turned on again.

6. The method according to claim 1, characterized in that Disconnecting the low-side transistor and disconnecting the OR-ing transistor when the unload transient occurs includes turning off the low-side transistor in response to the pulse width modulated signal, and turning off the OR-ing transistor in response to the tri-state pulse width modulated signal.

7. A DC to DC converter in a phase redundant power supply, characterized in that: include: an inductor; a low-side transistor electrically coupled between the first end of the inductor and ground; An OR-ing transistor is electrically coupled between the second end of the inductor and the output end of the DC-to-DC converter; a phase redundancy controller is cross-coupled to the OR-ing transistor, characterized in that the phase redundancy controller is configured to disconnect the OR-ing transistor when a load-unloading transient occurs.

8. The DC to DC converter of claim 7, wherein: The low-side driver receives a signal from the pulse width modulated signal controller to turn off the low-side transistor when a load-down transient occurs.

9. The DC to DC converter of claim 8, wherein: Also includes: The phase redundancy controller and the pulse width modulation signal are configured to turn off the low-side transistor and the OR-ing transistor simultaneously for a period of time before turning on the low-side transistor again.

10. The DC to DC converter of claim 7, wherein: Both the low-side transistor and the OR-ing transistor turn off within 30 nanoseconds when an unload transient is detected.

11. The DC to DC converter of claim 7, wherein: The inductor releases energy through the body diode of the OR-ing transistor and the body diode of the low-side transistor.

12. The DC to DC converter of claim 7, wherein: The inductor current fall time is less than 2.4 microseconds.

13. The DC to DC converter of claim 7, wherein: Also includes: A high-side transistor is electrically coupled between the first end of the inductor and the input terminal of the DC-to-DC converter.