IN Brief:
- Apacer’s CorePower architecture detects an unexpected supply failure and switches to stored capacitor energy.
- Cached DRAM data can be transferred into NAND before the SSD loses operating power.
- The hardware targets edge-AI and industrial systems that cannot depend on an orderly software shutdown.
Apacer has introduced CorePower hardware power-loss protection for industrial solid-state drives, using a dedicated detection circuit and onboard tantalum capacitors to preserve data during an unexpected supply interruption.
When the input voltage falls outside the drive’s operating range, the controller detects the failure and switches to stored capacitor energy. That reserve allows information held in volatile DRAM cache to be transferred into NAND flash before the SSD loses sufficient power to complete its internal operations.
The technology is available in the UK through Astute Group and is intended for edge-AI platforms, smart-manufacturing equipment, transport systems, healthcare devices, and installations exposed to mains failure, loose connections, battery faults, or abrupt equipment shutdowns.
An SSD write involves more than placing the latest user data into flash memory. The controller may also be updating address-mapping tables, metadata, error-correction information, wear-levelling records, and flash-translation-layer structures, any of which can be left incomplete when power disappears.
DRAM caching improves performance by allowing the controller to organise writes before committing them to NAND, although volatile memory loses its contents when the supply fails. Hardware power-loss protection gives the controller a defined interval in which to complete essential transactions and establish a recoverable internal state.
The available protection period depends on capacitor energy, controller behaviour, the volume of data in flight, supply fall time, operating temperature, and component condition. Design margin has to accommodate worst-case operation near the end of the intended service life rather than only a new drive tested at room temperature.
Edge systems are retaining larger volumes of operational information locally as machine vision, predictive maintenance, autonomous equipment, medical imaging, and transport monitoring move processing closer to the source. Models, event logs, images, measurements, and configuration data may be difficult or impossible to reconstruct after corruption.
That expansion is taking place while AI infrastructure reshapes the wider memory market. A recent warning over DRAM and high-bandwidth-memory capacity showed how investment and wafer allocation are being redirected towards the largest compute workloads, adding another variable to industrial memory availability and qualification.
Industrial storage has a markedly different operating profile from a conventional office computer. Drives may write continuously, run at elevated temperatures, experience vibration, or operate for years without planned downtime inside sealed, fanless, or remotely deployed equipment.
Power-loss protection consequently sits alongside endurance, data retention, error correction, bad-block management, thermal throttling, secure erase, firmware support, and lifecycle availability. Preserving one interrupted write offers limited value if sustained workloads exceed the drive’s endurance or long-term temperature exposure shortens data retention.
The host power system remains part of the protection chain. Main-supply hold-up capacitance, supervisory circuits, battery backup, shutdown signals, and operating-system behaviour can provide time to stop applications and close file systems, while drive-level protection covers transactions that have already reached the storage controller.
Tantalum capacitors provide compact energy storage within the drive, although capacitance and equivalent series resistance vary with temperature, age, and electrical stress. The shutdown sequence must retain enough margin to complete reliably after prolonged operation in the intended environment.
Validation should interrupt power repeatedly at different points within realistic workloads rather than rely on a single removal test. File-system integrity, address-table recovery, data consistency, restart behaviour, and progressive corruption after repeated events all need examination, using supply decay profiles that represent the host equipment.
Sealed and remote systems gain particular value from protection within the storage device because the SSD retains responsibility for completing its own critical internal work. The architecture reduces dependence on every software and power-management layer above it responding correctly during a sudden fault.
Storage comparisons often concentrate on capacity, throughput, and interface speed, yet the final milliseconds of available power can determine whether an industrial system restarts normally or requires recovery. As edge platforms take on more autonomous functions and retain more operational evidence, controlled failure behaviour is becoming part of the storage specification.
