The 'memory effect' is among the most cited and least understood phenomena in nickel batteries, and modern NiMH suffers a real but narrower version than the notorious NiCd problem. Repeated shallow cycling between similar states of charge can depress a cell's operating voltage - and, if entrenched, its apparent capacity - even though no active material has been lost. This paper explains the crystalline and thermodynamic origins of voltage depression, distinguishes genuine memory from ordinary ageing and undercharge, shows why a controlled full discharge and recharge restores the voltage profile, and sets out how a smart charger should use occasional reconditioning without turning the cure itself into a source of wear.
When a NiMH cell is repeatedly cycled over a narrow, partial state-of-charge band, portions of the nickel active material that never see a full charge or discharge adopt a more ordered, lower-energy configuration whose reaction potential is slightly reduced; on subsequent discharge the cell delivers the same energy through a lower voltage plateau, and equipment reading a voltage-based fuel gauge reports 'empty' early - an apparent capacity loss that is really a voltage-profile change.
Crucially the material is not destroyed: a full, slow charge to completion followed by a controlled deep discharge re-homogenises the electrode and restores the higher plateau, which is what distinguishes reversible memory from the irreversible corrosion and drying of Paper 21.
The effect is most associated with applications that live in a narrow band - hybrid vehicles operating in HRPSoC windows, devices always returned to the charger before empty, backup packs floated at a fixed state. Because the unused electrode regions never undergo the full phase transition, they 'remember' the restricted cycle; the longer and more monotonously a cell occupies one band, the more the depressed plateau stabilises.
This links memory directly to charge practice: a profile that never quite reaches full charge (an overly early termination) compounds the problem by leaving the upper electrode region un-refreshed every cycle, so undercharge and shallow cycling together entrench voltage depression faster than either alone.
Reconditioning applies a controlled, low-current discharge to a defined cutoff voltage - per cell, well below the normal operating cutoff but above the deep-discharge reversal danger zone - followed by a complete, careful charge. Sweeping the full stoichiometric range forces the ordered regions back through their phase transitions, restoring the normal voltage plateau; analyzer products automate this and report recovered capacity, distinguishing cells that recondition well from those whose loss is genuine ageing.
Current must be modest and per-cell voltage monitored in series strings, because an aggressive or unmonitored deep discharge can reverse the weakest cell and cause the very damage the procedure aims to avoid; the cure is a controlled measurement event, not a short circuit.
Reconditioning is not free: every deep discharge is a modest wear event, and routine full discharges accelerate mechanical and corrosion ageing - the opposite of the shallow cycling that maximises NiMH cycle life. Best practice is therefore occasional rather than automatic: analyzer guidance suggests a break-in/refresh every several cycles (on the order of ten) or when a depressed plateau is actually observed, rather than every charge. A charger that force-discharges batteries on every insertion over-treats a condition most cells do not have.
Low-self-discharge and modern alloy formulations exhibit milder memory than older NiCd and early NiMH, reinforcing a conservative trigger: monitor for the symptom and recondition on evidence, not on a fixed aggressive schedule.
Three problems look alike on a fuel gauge. Memory is reversible by one or a few full cycles; undercharge is a charger fault cured by correct termination, appearing as low delivered charge with a normal plateau once properly filled; and irreversible ageing (raised resistance, lost active material, dried electrolyte) is not restored by reconditioning and shows rising resistance and heat. Logging capacity before and after a controlled refresh separates them: full recovery indicates memory, partial recovery mixed causes, no recovery end of life.
The first figure shows the depressed and restored discharge-voltage plateaus; the second sequences a safe reconditioning cycle and its decision gate, the logic a smart analyzer implements.
Good charger firmware offers reconditioning as a user-selected or evidence-triggered mode: it verifies cell health and temperature, discharges at low current with per-cell cutoff, recharges with the standard multi-criterion profile, and compares recovered capacity against the prior reading to update health. It avoids automatic deep cycling, never deep-discharges a hot or faulty pack, and records the event so refresh frequency itself is tracked. Weijiang provides the recommended reconditioning cutoff, currents and expected recovery behaviour for each grade so partners implement refresh safely.
With reversible memory distinguished from irreversible wear, the next paper examines the surface-level and corrosion changes that accumulate irreversibly at the hydride electrode and how charge protocol accelerates or slows them.
Weijiang Power designs and manufactures nickel-metal hydride cells, matched packs and charging-ready configurations for consumer, industrial, medical and mobility customers, and supports partners with charge-protocol guidance, IEC 61951-2 performance files, IEC 62133-1 safety evidence and charger co-validation. Share your cell format, charge rate, thermal envelope and cycle target and our engineers will specify a cell-and-charge combination that protects both runtime and service life. Review the range on the products page.
