PSM Integration for Enhanced Module Solutions in 48 V Data Centers

×

Figure 1

   

Question:

Is a power management system (PSM) critical in a 48 V data center?

Answer:

PSM is indispensable for 48 V data center power systems. Without it, the infrastructure cannot reliably detect faults, coordinate power delivery, manage logic operation, maintain redundancy, or protect critical loads. The system becomes unpredictable, unsafe, and operationally unmanageable at scale.

Introduction

Telecom and datacom power systems increasingly rely on high efficiency 48 V intermediate buses using hot-swappable DC-to-DC converters in standard quarter brick (QB) footprints. Emerging industry standards now mandate integrated PMBus^® interfaces and robust protection circuitry, reflecting a clear shift toward digitally managed power solutions for rack servers and network equipment.

The LTC2971 is ideally suited to meet these requirements. It sequences and trims two regulated rails from –60 V to +60 V, implements programmable fault responses, and logs critical events to on-board EEPROM. Its native PMBus interface enables seamless integration with standard system monitoring architectures, while its digital power system management (DPSM) functions—sequencing, margin testing, fault detection, and telemetry—are now considered essential in complex power platforms. Internal EEPROM and an independent watchdog further support autonomous operation with minimal host intervention, improving overall system reliability.

QB solutions typically provide an analog adjust input (VADJ) for voltage trimming. The power system manager (PSM) interfaces directly with RUN/ENABLE, VOUT_SNS/FB, and temperature-sense pins, while its ISENSE inputs accurately measure output current using inductor DCR or an external shunt. Its precision VDAC output enables fine output-voltage adjustment through resistor dividers. By integrating the power system manager (PSM) on the power board, designers gain centralized, high accuracy digital telemetry of all rail voltages, currents, and board temperatures via PMBus offering far greater precision and flexibility than legacy analog monitoring solutions, which often struggle to meet strict data center performance specifications.

Quarter Brick Solution

The 2 kW quarter brick solution¹ reference design employs a high reliability four-phase architecture utilizing Analog Devices’ latest DC-to-DC intermediate bus converter (IBC) and coupled inductors for exceptional power delivery. Integrated telemetry provides continuous voltage monitoring, rapid fault detection, and real-time configuration via I²C/PMBus. Designed on a standardized footprint, the solution enables seamless integration into a wide range of customer platforms, enhancing system flexibility, efficiency, and ease of adoption. See Figure 1.

Key Findings on the Quarter Brick Solution and the Need for Integration

The integration of the LTC2971 with quarter brick converters has several key benefits.

Precision Telemetry

The PSM features high resolution ADCs that accurately monitor voltage and temperature. On the QB board, it demonstrated output voltage and temperature reporting within approximately ±0.5 to 1.0% accuracy, enabling precise power budgeting. Reliable measurements of voltage, current, and temperature give engineers deep insight into solution behavior under varying conditions, allowing performance to be pushed safely to its limits. By programming EEPROM parameter limits and fine-tuning warning thresholds based on the telemetry, engineers can ensure the solution consistently delivers maximum performance without triggering unnecessary faults. See Figure 2.

Digital Control and Sequencing

A critical requirement in data centers is precise power sequencing to ensure that the QB solution supplies and powers up processors safely in a controlled and timely manner. The PSM fully automates multirail startup, shutdown, and voltage margining, supporting both time-based and tracking sequencing across rails. It can also cascade multiple QB solutions for scalable designs. The PSM DRV_EN/RUN and VOUT_EN parameters, including power-on delays, rise times, and sequencing policies, are fully programmable via PMBus commands, guaranteeing that interconnected rails always power up and down in a coordinated manner. This controlled sequencing minimizes voltage overshoot and undershoot during every power cycle, enhancing the reliability of data center racks.

Fault Management

A key advantage of the PSM is its robust, integrated fault management and event logging. Voltage or temperature violations can be configured to trigger latching shutdown, automatic retry, or a controlled ramp-down. Importantly, the device continuously buffers status and telemetry in RAM and, upon detecting a fault, automatically commits and saves this data to nonvolatile EEPROM. This black-box functionality preserves a complete snapshot of voltages, currents, and temperatures at the moment of failure, enabling precise root-cause analysis. As illustrated in the LTC2971’s data sheet’s conceptual diagram, the circular buffer automatically transfers to EEPROM on fault events. Unlike basic PMBus fault registers or external watchdogs, the PSM records the full operating state rather than just flagging errors. These advanced features strengthen QB solution reliability, support detailed lifecycle assessment, and extend operational longevity through continuous monitoring and logging. See Figure 3.

PMBus Integration

The data center specification requires a robust communication interface with PMBus protocol support. As a PMBus-compliant device, the PSM adheres to industry standards and integrates seamlessly with existing system-management platforms. Its status monitoring and control functions are fully accessible through standard PMBus commands, and it supports LTpowerPlay^® for a streamlined configuration. This allows design teams to define register settings, such as voltage rails, limits, and sequencing, offline and program them directly into the EEPROM, significantly reducing development time. Additionally, the device’s bidirectional FAULTB pin enables flexible fault management by allowing multiple devices to share or isolate fault lines as needed. Overall, the digital control capabilities, including servo trimming and sequencing, provide a far more efficient path to advanced power management implementation compared to purely analog solutions.

Implementation Outcomes

On the QB solution board performance,² the PSM demonstrated reliable monitoring of the IBC’s operation, input voltage, and performance metrics. PMBus telemetry remained stable and aligned with calibrated bench digital multimeter (DMM) measurements. Fault injection tests confirmed that the device correctly latched off faulty channels and accurately recorded events in the MFR_FAULT_LOG registers. System-level validation showed that the integrated digital power management introduced no additional noise or instability, while the controlled soft start algorithm effectively prevented gate-driver shoot-through, resulting in clean output waveforms. The I²C interface is programmable from 100 kHz to 400 kHz to accommodate varying communication requirements, and voltage and temperature readings closely matched manual data within ±0.1% and ±0.5%, respectively. See Figure 4.

Collectively, these results show that the LTC2971 enhances quarter brick power solutions by adding high fidelity monitoring, fine-grained control, and detailed diagnostic data. Its PMBus compatibility ensures it can be adopted in modern telecom systems with minimal interoperability issues. See Figure 5.

Conclusion

Integrating the LTC2971 PSM into the QB solution reference design delivers substantial benefits for DC power systems. Its on-chip 16-bit telemetry inputs provide high precision monitoring that meets stringent DC accuracy requirements. Built-in sequencing and closed-loop trim enables precise multirail power-up and power-down control, while the PMBus interface ensures seamless integration with modern rack management infrastructure. Autonomous fault logging offers advanced diagnostics that are difficult to achieve with analog circuits alone. These capabilities do introduce additional design considerations, including careful configuration, calibration, and reliable bus operation. However, the resulting improvements in reliability, visibility, and control typically outweigh these challenges. As telecom and data center applications increasingly demand automated, network-managed power subsystems, devices like the PSM are poised to become standard building blocks, enabling robust digital control and precise telemetry for next-generation power shelves.

It should be noted, however, that for solutions requiring more than nine power solutions in parallel, the PSM may be less advantageous due to PMBus addressing limitations. In such cases, other ADI PSM ICs with advanced addressing capabilities are better suited.

References

¹Karl Audison Cabas and Christian Cruz. “Enabling Future Innovations: Intermediate Bus Converter—Part 1: Benefits.” Analog Devices, Inc., July 2025.

²Karl Audison Cabas and Christian Cruz. “Enabling Future Innovations: Intermediate Bus Converter—Part 2: Performance.” Analog Devices, Inc., July 2025.

关于作者