Server Power Supplies
Assess PFC, isolated conversion, output rectification, standby rails, and ORing by topology and operating profile.
Server PSU Conversion and Output Stages
A server PSU may combine an input power-factor-correction stage, an isolated DC-DC converter, secondary rectification, output filtering, standby conversion, and redundant-output ORing. The exact topology and control strategy determine the electrical stress on each switch, driver, sensor, and magnetic component.
Evaluate efficiency and heat over the PSU's input and load range, including startup, standby, and transitions between redundant supplies. A component's headline voltage or current rating does not establish its suitability for the switching waveform or thermal location in the design.
For alternates, identify the stage and reference design position, isolation boundary, controller interface, switching conditions, output requirements, and protection role. Keep power conversion and source redundancy as separate checks in the schematic review.
Component Requirements
Typical functions, devices, and parameters in Server Power Supplies.
| System function | Typical devices | Key parameters | Browse |
|---|---|---|---|
| Input and PFC Conditions the PSU input and establishes the high-voltage DC link used by the isolated stage. | PFC controllers, SiC or silicon MOSFETs, bridge and boost devices | Input range, switching frequency, efficiency, surge, thermal limit | View category |
| Isolated DC-DC conversion Transfers energy across the PSU isolation boundary to its main output rail. | PWM or resonant controllers, MOSFETs, gate drivers, isolated sensors | Topology, isolation, switching stress, regulation, startup | View category |
| Secondary rectification and filtering Converts the secondary switching waveform and filters the regulated output. | Synchronous rectifiers, output controllers, capacitors, current sensors | Current, ripple, reverse recovery, thermal path | View category |
| Standby and redundant output Supplies auxiliary rails and coordinates parallel or redundant power sources. | Flyback controllers, supervisors, ideal-diode and ORing controllers | Standby load, handover, reverse current, power-good | View category |
Design & Selection Considerations
Topology and switch position
Name the PFC, primary isolated converter, synchronous rectifier, standby converter, or ORing stage and provide its switching and voltage conditions.
Load and thermal points
Compare input and load ranges, airflow, ambient temperature, output sharing, and the location of heat-producing components.
Isolation and measurement
Trace primary-to-secondary isolation and identify which voltage and current measurements feed regulation, telemetry, or protection.
Supply Chain Considerations
Power stage and controller must match
Include the topology, controller, switching frequency, drive method, and magnetic or transformer design when screening a switch or driver alternate.
PSU metrics need test conditions
Provide input voltage, load profile, airflow, temperature, and measurement method for efficiency or thermal comparisons.
Redundant output is a system circuit
Share the output diagram and source-handover requirements before comparing an ORing controller or ideal-diode device.
Supply Chain Support
BOM organized by stage
Separate input protection, PFC, isolated conversion, secondary rectification, standby, sensing, and redundancy parts.
Topology-specific replacement data
Provide controller and transformer details, switching conditions, isolation requirements, package, and thermal constraints.
Platform and production context
List PSU revision, output configuration, redundancy mode, build quantity and date, and manufacturer or compliance restrictions.
Frequently Asked Questions
The PFC stage shapes input current and establishes a DC link; the isolated DC-DC stage transfers energy across the isolation boundary to the PSU output. They have different switch stresses, controller functions, sensing, and protection needs, so identify the actual stage and topology before comparing devices.
Switching, conduction, magnetic, and control losses change with input voltage and delivered load. Those losses determine heat and cooling needs across the product's operating profile. Compare measured or specified conditions that match the target platform instead of using a single peak-efficiency point.