Choosing how an absolute encoder remembers revolutions while unpowered is one of the most consequential, and most often defaulted, decisions in motion-system design. The catalogue offers three architectures - a battery-buffered electronic counter, an energy-harvesting Wiegand stage and a mechanical gear multi-turn - and within the battery-buffered route the designer must still choose the source itself: a disposable primary lithium cell, a rechargeable chemistry or a capacitor. This second paper develops a rechargeable nickel-metal hydride backup for the multi-turn counter and places it candidly beside each alternative. It begins with the architecture decision, showing where a battery-buffered encoder is the right answer despite its service requirement; it then works through the microampere-scale retention budget, the trickle or top-up charging that keeps a rechargeable cell ready whenever the axis is powered, the self-discharge arithmetic that decides achievable off-power life, switchover and low-battery circuitry, and wide-temperature behaviour. It is deliberately honest: for an encoder that is powered only once every few years, a primary lithium cell's low self-discharge is hard to beat, whereas for an axis that is energised on most working days - the vast majority of servomotor, robot and machine-tool applications - a properly managed rechargeable NiMH backup eliminates planned primary-cell replacement, avoids disposable-lithium disposal and turns multi-turn memory into a rechargeable, monitored and serviceable subsystem.
The architecture choice precedes chemistry. A battery-buffered electronic multi-turn is compact, freely programmable in resolution and cheap, but its memory is only as durable as its source. A Wiegand energy-harvesting multi-turn needs no external cell because rotation induces the pulse that powers the count, but it requires movement to register and stores state in a capacitor that ages over roughly a decade; a mechanical gear multi-turn encodes revolutions physically and needs no electrical memory but adds gear trains with their own wear, torque and backlash.
For mainstream factory automation - servomotors, positioning stages, warehouse and machine-tool axes that are powered most days - the battery-buffered design remains attractive because it is resolution-flexible, fast and inexpensive, and its one weakness, source life, is manageable with a well-designed rechargeable backup. The rest of this paper addresses that design.
Sizing a backup begins not with ampere-hours but with the retention current specified by the encoder ASIC, typically in the microampere region, plus the self-discharge of the chosen cell. Off-power life is the usable capacity divided by the sum of counter current and self-discharge current, de-rated for low temperature and end of life. Because self-discharge often exceeds the counter's own draw, a chemistry's charge-retention characteristic matters more than its nominal capacity - a large cell that leaks its own charge quickly is no better than a small, stable one.
Sealed NiMH cells have steadily improved in charge retention; low-self-discharge variants retain the large majority of their charge over a year at room temperature, which is ample when the encoder is trickle-topped whenever the machine is on. The first animated figure steps through the backup design flow from retention budget to validated service interval, making the self-discharge term explicit rather than hidden.
The decisive advantage of a rechargeable backup in a regularly powered axis is that the active 5 V or 24 V supply is available to maintain it. A charge-management path tops the NiMH cell or small pack with a controlled, current-limited maintenance charge while the encoder is running, so the backup begins every power-off interval full regardless of cumulative off-time. This is fundamentally different from a primary lithium cell, which is a finite, irreplaceable-on-board reservoir counting down from the day of manufacture.
The charge path must be conservative - NiMH is maintained with a low current and temperature supervision rather than a hard constant-voltage float - and it must never back-feed or load the sensitive retention rail. Integrating this with the encoder chip's built-in supply switching yields a clean architecture in which the main supply powers the active interface and quietly refreshes the backup, while the backup powers only the counter when the main rail is absent.
A reliable design pairs the rechargeable source with glitch-free power-path switching - preferably the encoder ASIC's integrated switchover - and exposes the source state over the diagnostic channel. The low-battery flag available on modern chips becomes a predictive-maintenance datum: if the backup is repeatedly found low despite normal machine running, it signals an ageing cell or a charge-path fault long before any position is lost.
A small redundant capacity margin and a defined service interval turn the classic 'dead coin cell after N years' failure into a scheduled, observable event. For an OEM this is also a supply-chain benefit: the backup becomes a standard, replaceable, rechargeable component rather than a specific primary-lithium coin that must be procured, stocked and disposed of under evolving battery and waste rules.
Primary lithium is the incumbent for a reason: extremely low self-discharge and high energy density support years of unattended retention, which is why it suits devices powered rarely. Its costs are a fixed end-of-life, disposable-cell replacement, lithium transport and waste considerations, and a voltage that offers little graceful warning. Supercapacitors deliver virtually unlimited cycle life and instant switchover but their energy per volume is far too low for months of retention and their leakage is significant, so they suit ride-through of seconds, not years.
Wiegand harvesting removes the cell but adds the first-turn-after-power-off limitation and capacitor ageing. The second animated figure scores the options on the axes that matter; rechargeable NiMH is not the winner for a once-a-year-powered remote valve, but it is the most defensible choice for a daily-powered industrial axis because it is refreshed in normal use, avoids disposable lithium, operates over a wide temperature span with simple charging and ships under non-lithium transport rules. The selection should follow the actual powered-duty profile rather than catalogue habit.
Before release the design is validated by cycling the main supply while rotating the shaft in the off-power window to prove no revolution is lost, measuring retention life against the self-discharge-and-current model at high and low temperature, accelerating charge-discharge cycling to confirm the cell ages predictably under maintenance charging, and injecting low-battery and overspeed conditions to verify the diagnostic flags. Switchover is captured on an oscilloscope to confirm the counter rail never dips below its minimum during the millisecond handover.
The result is a characterised backup with a documented off-power retention figure, a maintenance-charge specification and a demonstrated diagnostic path - ready for the EMC, vibration and battery-standard qualification campaign that the final paper describes, and a credible, serviceable alternative to the disposable coin cell in daily-powered motion equipment.
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.
