Microchannel Plate (MCP) detector is a vacuum photoelectric detection device with ultra-high time resolution, spatial resolution and single-photon detection capability. Featuring microsecond-level response speed, electronic gain over 10⁶, two-dimensional imaging capability, compact size and low power consumption, it has been widely applied in cutting-edge fields such as deep-space ultraviolet detection, high-energy physics particle detection, biological fluorescence microscopic imaging, LiDAR, ultrafast time-resolved spectroscopy, and semiconductor wafer defect inspection. The high-voltage power supply is the core supporting component of the MCP detector, providing independently adjustable high-precision high-voltage bias voltages for the input electrode, multiplication electrode and output electrode of the MCP. Its output ripple level, inter-channel crosstalk suppression capability, voltage stability and electromagnetic shielding performance directly determine the electronic gain stability, dark count rate, imaging contrast, time resolution and detection sensitivity of the MCP detector. MCP detectors impose extremely stringent performance requirements on high-voltage power supplies. On the one hand, the electronic gain of MCP is exponentially correlated with the bias voltage; the gain sensitivity of a single-stage MCP to bias voltage variation can reach more than 1% gain fluctuation per 1V voltage change. This requires the long-term output voltage stability of the power supply to be better than ±0.05% and the peak-to-peak output voltage ripple to be less than 10 mV. Otherwise, it will cause severe fluctuation in MCP gain, sharp rise of dark count rate, noise spots and reduced contrast in imaging, and even fail to realize single-photon-level weak signal detection. On the other hand, an MCP detector usually requires 3~6 independent high-voltage output channels to supply different bias voltages to the photocathode, MCP multiplication electrode, and fluorescent screen/anode respectively, with voltage differences between channels up to several kilovolts. The inter-channel crosstalk is required to be less than 0.05%; otherwise, mutual interference will occur between bias voltages of different electrodes, leading to unstable MCP gain and imaging distortion. In addition, MCP detectors generally operate in a high-vacuum environment and are often used together with precision equipment such as ultrafast lasers and weak signal acquisition systems, imposing extremely high requirements on electromagnetic radiation and conducted interference of power supplies. Traditional high-voltage power supplies suffer from core pain points including high output ripple, severe inter-channel crosstalk, large electromagnetic interference and insufficient long-term stability, making them unable to adapt to the high-sensitivity detection requirements of MCP detectors. Relevant designs must strictly comply with standards such as GB/T 4793.1 Safety requirements for electrical equipment for measurement, control and laboratory use — Part 1: General requirements and GB/T 15479 Technical requirements and test methods for insulation resistance and insulation strength of industrial automatic instruments, while meeting the core demands of low noise and high stability in the field of vacuum photoelectric detection. Targeting the core application requirements and technical challenges of high-voltage power supplies for MCP detectors, this methodology establishes a full-process general technical framework covering low-ripple topology design, multi-channel low-crosstalk optimization, full-dimensional electromagnetic shielding design, high-stability closed-loop control, and vacuum environment adaptability design. It can meet the high-voltage bias demands of various single-stage, dual-stage and Generation III MCP detectors, providing standardized design criteria for the localization and performance improvement of domestic vacuum photoelectric detection devices. Aiming at the core design challenges of low-ripple output, low inter-channel crosstalk and strong electromagnetic shielding in MCP detector scenarios, this methodology adopts the main architecture of "single-channel modular independent inversion + distributed digital control + full-link multi-stage filtering", combined with hierarchical electromagnetic shielding and equipotential gradient design. It thoroughly breaks the technical bottlenecks of traditional multi-output high-voltage power supplies such as high ripple, severe inter-channel crosstalk and large electromagnetic interference, realizing multi-channel independently adjustable ultra-low-ripple high-voltage output and excellent EMC performance, which fully adapts to the full-working-condition requirements of MCP detectors. The design follows five core criteria. First, the topology adopts single-channel modular independent design. Each high-voltage output of the MCP detector corresponds to a completely independent power module, and each module is equipped with independent inversion, voltage boosting, rectification, filtering, closed-loop control and protection functions. Complete electrical isolation, physical isolation and electromagnetic shielding isolation are realized between modules, thoroughly eliminating inter-channel electrical coupling and crosstalk caused by shared inversion units and transformers in traditional single-transformer multi-output topologies. The single-channel topology adopts a quasi-resonant flyback inversion topology, which enables zero-voltage switching of power switches, greatly reducing switching losses and switching noise and suppressing ripple generation at the source. Meanwhile, this topology features simple structure, fewer components and high voltage rise ratio, easily achieving a single-channel output voltage range of 0~3 kV, which fully meets the bias voltage requirements of each electrode of MCP detectors. The inversion unit of each module adopts an independent low-temperature-drift reference source and closed-loop control circuit, enabling independent start-stop, independent voltage regulation and independent protection of each single channel. Failure of one module will not affect the normal operation of other channels, greatly improving the redundancy and reliability of the entire power supply system. In addition, the modular design allows flexible configuration of module quantity according to the channel number of MCP detectors, adapting to different types and configurations of MCP detection systems. Second, the full-link low-ripple optimization design criterion realizes full-band ripple suppression from three dimensions: noise source suppression, multi-stage filtering and linear voltage regulation, ensuring ultra-low ripple of output voltage. ① Switching noise source suppression: Quasi-resonant soft switching design ensures that power switches always operate in zero-voltage switching state under full load, completely eliminating voltage spikes and high-frequency noise caused by hard switching. Meanwhile, the switching frequency is selected outside the signal acquisition bandwidth of the MCP detector, and spread spectrum modulation technology is adopted to disperse switching frequency energy into a wider frequency band, reducing peak noise intensity and preventing switching noise from coupling into the weak signal acquisition link of MCP. In addition, SiC MOSFETs with low switching noise are selected as power switching devices, and SiC Schottky diodes with no reverse recovery loss are adopted as rectifiers to reduce noise and spikes during switching at the device level. ② Multi-stage cascaded filtering architecture: A five-stage filtering scheme of "inversion output filtering + high-voltage rectification filtering + linear voltage regulation input filtering + multi-stage π-type filtering at output terminal" is adopted. An LC low-pass filter is designed at the output of the inverter to suppress fundamental and harmonic components of the switching frequency; high-voltage thin-film capacitors with large capacity and low ESR are configured at the high-voltage rectification terminal to filter low-frequency ripple after rectification; a π-type filter is installed at the input of the linear voltage regulation unit to further suppress residual ripple of the previous stage; a three-stage cascaded RC low-pass filter network is arranged at the final high-voltage output terminal: the first stage uses large-capacity high-voltage polypropylene film capacitors to filter low-frequency ripple, the second stage adopts small-capacity NP0 ceramic capacitors to eliminate intermediate-frequency noise, and the third stage uses feedthrough capacitors to suppress ultra-high-frequency noise above 10 MHz. With this five-stage filtering architecture, the peak-to-peak output ripple can be controlled within 5 mV, far better than the industry general standard of 10 mV. ③ Post-stage linear voltage regulation optimization: A low-dropout high-voltage linear regulator is added after the high-voltage rectification and filtering unit. The linear regulator has an extremely high power supply ripple rejection ratio (PSRR ≥ 80 dB at 100 kHz), which can further suppress the residual ripple and noise of the previous switching power supply by more than an order of magnitude, while realizing precise output voltage regulation and further improving output voltage stability and low-ripple performance. Third, the multi-channel low-crosstalk optimization design criterion eliminates inter-channel crosstalk thoroughly from four aspects: architectural isolation, conducted crosstalk suppression, radiated crosstalk shielding and grounding optimization. ① Architectural isolation: Through single-channel independent modular design, the power conversion, power supply and control loops of each channel are completely independent with no shared power devices, transformers or reference sources, fundamentally eliminating crosstalk caused by electrical coupling. It ensures that load mutation, start-stop and voltage regulation of any channel will not affect the output voltage of other channels. ② Conducted crosstalk suppression: Each channel module is equipped with independent two-stage EMI filter circuits and isolation diodes at the input terminal to avoid conducted crosstalk through input power lines; independent high-voltage isolation diodes and RC filter circuits are arranged at the output terminal of each channel to prevent coupling crosstalk through output loops; optocouplers or digital isolators are used for electrical isolation of control and communication signals between channels to avoid crosstalk through control lines. ③ Radiated crosstalk shielding: Each channel module is installed in an independent metal shielding cavity made of integrally milled aluminum alloy. Cavities are completely separated by thickened metal partitions, and each cavity is grounded independently to achieve electromagnetic shielding isolation between channels with shielding effectiveness over 100 dB, preventing switching radiation noise of one module from coupling into adjacent modules and causing inter-channel crosstalk. ④ Grounding system optimization: A combined "single-point grounding + star grounding" architecture is adopted. The power ground, signal ground and shielding ground of each module are connected at a single point inside the module, and then connected to the main system grounding point through independent grounding wires to form a star grounding structure, completely eliminating grounding loops between channels and crosstalk caused by ground potential differences. Finally, the inter-channel crosstalk is controlled below 0.03%, fully meeting the high-stability requirements of MCP detectors. Fourth, the full-dimensional electromagnetic shielding optimization design criterion adopts a four-level hierarchical shielding architecture of "whole-machine shielding - module-level shielding - device-level shielding - cable shielding" to thoroughly suppress electromagnetic radiation from the power supply and external electromagnetic interference. ① Whole-machine double-layer shielding design: The whole machine adopts a double-layer sealed shielding shell with permalloy on the inner layer for low-frequency magnetic field shielding and aluminum alloy on the outer layer for high-frequency electric field shielding. The shell adopts an all-welded structure to reduce splicing gaps. All external interfaces are equipped with shielded connectors with 360° lap joint between shielding layers and the shell to ensure shielding continuity. The whole-machine shielding effectiveness exceeds 120 dB, which can effectively suppress internal electromagnetic radiation of the power supply from interfering with external precision equipment and prevent external electromagnetic interference from entering the power supply. ② Module-level independent shielding: As mentioned above, each channel module is installed in an independent aluminum alloy shielding cavity. The power inversion unit and linear voltage regulation unit are further separated by metal partitions inside the cavity to prevent switching noise of power circuits from coupling into low-noise linear voltage regulation and sampling circuits. ③ Device-level precision shielding: Noise-sensitive analog circuits such as voltage reference sources, sampling circuits and error amplifiers are secondary shielded with independent small permalloy shielding boxes to completely isolate external electromagnetic radiation interference; strong radiation noise sources such as power transformers and power switches are partially shielded with independent shielding covers to prevent diffusion of radiation noise. ④ Input and output cable shielding: All input power lines, output high-voltage lines and control signal lines adopt shielded cables. High-voltage output lines use double-layer shielded coaxial cables with single-end grounding of the shielding layer; control lines adopt twisted pair shielded cables with 360° grounding of the shielding layer to avoid electromagnetic radiation leakage and external interference coupling during cable transmission. Fifth, the high-stability closed-loop control and vacuum environment adaptability design criterion adopts an FPGA-based fully digital distributed control architecture. Each channel module is built with independent 16-bit high-precision DAC and 24-bit high-precision ADC to realize high-precision regulation and sampling of output voltage. Digital PID closed-loop control algorithm ensures that the output voltage control accuracy is better than ±0.05%, load regulation rate better than ±0.1%, and line regulation rate better than ±0.05%. Meanwhile, a full-temperature-range temperature compensation algorithm is designed to collect the internal temperature of each module in real time, dynamically adjust the output reference and closed-loop parameters, and compensate for device parameter drift caused by temperature changes. It ensures that within the full operating temperature range of 0℃~50℃, the temperature coefficient of output voltage is ≤10 ppm/℃ and the voltage drift after 8 hours of continuous operation is ≤0.05%, meeting the long-term continuous operation requirements of MCP detectors. For the high-vacuum operating environment of MCP detectors, all insulating materials at high-voltage output terminals adopt high-vacuum compatible polyimide and alumina ceramic materials with vacuum outgassing rate ≤1×10⁻⁶ Pa·m³/(s·m²) and no low-molecular volatile substances, avoiding contamination of the vacuum cavity of MCP detectors. In addition, high-voltage connection parts adopt smooth transition design to eliminate sharp corners and burrs, preventing corona discharge and micro-discharge in high-vacuum environments. Reliability and safety protection design serves as the core support of this methodology. Targeting the long-term operation demands of MCP detectors and the protection requirements of expensive devices, this methodology forms a full-process general criterion covering full-dimensional protection, reliability design and safety compliance. In terms of full-dimensional protection function design, each channel module is equipped with independent dual protection of hardware and software, including input over/under voltage protection, output overvoltage protection, output overcurrent/short-circuit protection, over-temperature protection, vacuum interlock protection and high-voltage spark discharge protection. The response time of all protection functions is less than 1 μs. When faults such as high-voltage spark discharge and short circuit are detected, the high-voltage output of the corresponding channel can be cut off instantly without affecting the normal operation of other channels, preventing expensive MCP devices and photocathodes from being damaged by high-voltage impact during faults. Meanwhile, an interlock interface with the detector vacuum system is designed; high-voltage output is allowed only when the vacuum degree reaches the set value, avoiding high-voltage discharge damage to the detector in low-vacuum environments. In terms of reliability design, all core components are designed in accordance with Class I industrial derating criteria with voltage stress ≤70% of rated value, current stress ≤60% of rated value and temperature stress ≤80% of rated value, greatly reducing the operating stress of components, delaying aging and improving long-term operating reliability. All high-voltage capacitors adopt high-stability and long-life polypropylene film capacitors to avoid the service life limitation of electrolytic capacitors. The overall mean time between failures (MTBF) is ≥50,000 hours, meeting the long-term continuous operation needs of laboratories. In terms of safety compliance design, it strictly follows safety standards for laboratory electrical equipment. Double insulation is adopted between input and output with isolation withstand voltage exceeding twice the maximum output voltage, meeting reinforced insulation requirements. All high-voltage components are installed in fully sealed shielding shells with door opening power-off and high-voltage discharge interlock. When the shell is opened, the high-voltage input can be cut off instantly and all high-voltage capacitors can be discharged quickly to ensure absolute safety of operators. In addition, high-voltage indicator lights and voltage monitoring functions are configured to display the output voltage status of each channel in real time and prevent misoperation. Targeting the core application requirements and technical challenges of high-voltage power supplies for MCP detectors, this methodology forms a full-process general technical framework from low-ripple topology design, multi-channel low-crosstalk optimization and full-dimensional electromagnetic shielding design to high-stability control. It thoroughly solves the core pain points of traditional high-voltage power supplies such as high output ripple, severe inter-channel crosstalk, large electromagnetic interference and insufficient long-term stability. The single-channel modular design eliminates inter-channel crosstalk at the source; the five-stage filtering combined with post-stage linear voltage regulation realizes ultra-low output ripple within 5 mV; the four-level hierarchical shielding architecture achieves excellent EMC performance. It fully adapts to the high-sensitivity detection requirements of various MCP detectors and can be widely applied in deep-space exploration, high-energy physics, biological imaging, LiDAR and other fields, providing core technical support for the localization and performance improvement of domestic vacuum photoelectric detection devices.