Once the retention load profile is understood, the designer faces a genuine architecture choice with real lifecycle consequences: hold the volatile memory with a capacitor, a disposable primary-lithium cell, a rechargeable source, or eliminate volatile retention by moving to non-volatile memory. Each route is defensible for a different duty, and the marketing habit of declaring one universally superior obscures the engineering. This second paper develops a rechargeable nickel-metal hydride retention module for a PLC or CNC control and places it candidly beside the alternatives. It works through the voltage matching to the 3 V or 3.6 V CMOS backup rail, the microampere-scale load and self-discharge arithmetic that set achievable retention time, the controlled trickle top-up drawn from the 24 V rail whenever the controller is running, a small hold-up capacitor that guarantees an orderly save at the instant of failure, hot-swap support that removes the CNC's dreaded five-minute replacement window, and backup-health monitoring. It then compares NiMH honestly with a supercapacitor - unmatched for short ride-through but weak across months - a primary-lithium cell - long-lived but disposable and procured as a service part - and battery-less FRAM/flash designs, which remove the cell but bring finite flash endurance and design constraints. The aim is a decision by quantified duty: to show precisely where a maintained, rechargeable NiMH retention module is the most robust and lowest-lifecycle-cost way to keep a controller's state and clock.
CMOS SRAM and RTC backup inputs usually expect a 3 V or 3.6 V rail, switched in by a power-fail comparator when the main logic supply falls below it. Nominal 1.2 V NiMH cells reach that rail with a three-cell series string (3.6 V nominal), chosen so that the end-of-discharge voltage still exceeds the SRAM and RTC data-retention minimum across temperature, while the freshly charged voltage does not exceed their absolute maximum. A low-quiescent linear regulator or ideal-diode path isolates the string from the main rail with negligible drain.
Because every series cell must share charge, the module uses matched, lot-consistent cells and conservative design margins. Paralleling strings or using larger cells raises retention time for controllers with multi-megabyte retentive areas without changing the rail, and the whole module is packaged as a pluggable, keyed part so field replacement is unambiguous and reverse-polarity-proof.
Achievable retention is usable capacity divided by the sum of the SRAM and RTC standby current, the switchover circuitry quiescent current and the cell's own self-discharge, de-rated for low temperature and end of life. As with encoders, self-discharge and temperature often matter more than nameplate capacity; low-self-discharge NiMH cells hold the large majority of their charge over a year at room temperature, which is decisive for a controller that is switched off over weekends or maintenance shutdowns.
The module is sized against the manufacturer's own published anchors - capacitor retention of about two weeks or 50 to 100 hours in compact PLCs, roughly six weeks at 40 degrees C in a high-end CPU, and years with a primary battery in larger modular systems - so the designer can target a specific horizon, such as covering a long seasonal shutdown, rather than an open-ended claim. The first animated figure steps through this sizing and design flow.
The core advantage of a rechargeable retention module is that the controller's own 24 V or logic supply is present almost every working day and can keep the NiMH string full. A current-limited charge path tops the module whenever the PLC is powered, with temperature supervision and a conservative maintenance current rather than a hard constant-voltage float, so every power-off interval - however cumulative - begins from full charge. This breaks the primary-cell countdown in which a lithium battery irreversibly depletes from the day of manufacture regardless of use.
The same path supports a fuel-gauge or voltage-monitor signal to the CPU, turning backup health into a diagnostic the program can alarm on, much as modern encoders expose a low-backup flag. Maintenance becomes condition-based and scheduled, instead of a fixed calendar replacement performed whether the cell needs it or not.
The rechargeable string is optimised for the long, microampere retention tail; the millisecond-scale orderly-save phase is best served by a small hold-up capacitor placed on the logic rail, sized from the supervisor's warning time and the energy needed to flush the scan and write critical registers to the retained domain. Splitting the two jobs - capacitor for instant ride-through, NiMH for long retention - lets each element do what it is good at and avoids forcing a battery to be a perfect pulse source.
The supervisor detects rail collapse, the hold-up capacitor sustains the CPU long enough to complete its save sequence, and the NiMH module then keeps the saved SRAM and RTC alive for the outage. This two-element architecture is more deterministic than relying on a single battery's internal resistance at the exact moment of failure.
Traditional CNC primary-lithium packs must be exchanged within a few minutes while the control stays powered, because there is no internal reservoir; miss the window and parameters and zero positions are lost. A well-designed rechargeable module carries enough local reserve - and the hold-up capacitor covers the logic side - to allow unhurried, even powered-off exchange, and because the module is recharged in normal use it does not arrive with an unknown remaining service life.
For machine-tool users this removes a recurring source of downtime and skill dependency; for the controller OEM it removes a specific disposable-lithium service part from the spare-parts catalogue and replaces it with a standard, rechargeable, non-lithium component.
The comparison must be candid. Supercapacitors offer unlimited cycles and instant ride-through but their leakage and low energy density make them poor for multi-month retention; they are the right choice for short outages and the perfect partner for the hold-up role. Primary lithium offers the longest unattended life with zero charging infrastructure and still suits sealed, rarely powered devices. Battery-less FRAM/MRAM or flash designs remove replacement entirely but constrain which data can be retentive - flash endurance around 100,000 rewrites rules out rewriting large areas every scan - and add cost and design complexity. The second animated figure scores the four options on the axes that matter.
Rechargeable NiMH is most defensible where a controller is powered most days, must retain large SRAM areas and a clock over long but intermittent shutdowns, and the customer values a rechargeable, monitored, non-lithium service part. The module is then validated by power-interruption tests across temperature, accelerated retention and cycle runs, and hold-up capture on an oscilloscope - the evidence the final paper folds into the IEC 61131-2 and battery-standard compliance campaign.
Weijiang Power designs and manufactures sealed nickel-metal hydride cells and matched industrial packs for remote, off-grid and safety-related equipment, and supports OEM partners with IEC 61951-2 performance files, IEC 62133-2 safety evidence, pulse-load characterisation, wide-temperature testing and charger/pack co-validation. Tell us your duty cycle, peak current, temperature envelope, autonomy target and the standards your product must meet, and our engineers will specify a cell-and-pack combination that protects runtime, reliability and service life. Review the range on the products page.
