Terminating fast charge is not the end of charging. A NiMH cell stopped at the -delta-V peak is close to full but rarely perfectly full; it then begins to self-discharge the moment current is removed. Chargers therefore add a reduced-current top-off to complete the charge and a maintenance regime to hold it there - and it is in this seemingly gentle tail that many batteries are quietly ruined, because even a small continuous current into a full cell is continuous overcharge. This paper distinguishes top-off from trickle from maintenance, quantifies the currents and durations the cell can tolerate, and explains why pulsed maintenance outperforms a constant trickle for long service life.
Fast termination at -delta-V or the inflection point deliberately stops slightly early to avoid the hottest part of the recombination region, and at 1C or C/2 the cell may still be a few percent short of full capacity. The bq2002-family ICs offer an optional top-off precisely for chemistries that 'tend to terminate before reaching full capacity': after fast charge ends, charging continues at a reduced rate for a selected period to add the final percent without sustained high-current overcharge.
Top-off current and duration are bounded so total overcharge stays small: it is a finite, timer-terminated phase, not an open-ended taper, and its purpose is to equalise charge among cells and fill the last few percent rather than to hold the cell at full indefinitely.
A constant trickle current into an already-full cell is entirely consumed by the oxygen recombination loop and turned to heat; sustained, it keeps the negative permanently oxidising, accelerates electrolyte dry-out and corrosion, and degrades the hydride alloy's catalytic activity so internal pressure rises over life. Panasonic's charging manual is explicit that extensive trickle charging causes deterioration and is the least preferred method; if used it should be limited to roughly 0.033C to 0.05C for no more than about 20 hours.
Those numbers encode the recombination capacity of a small cell: currents near C/20 sit at the edge of what can be recombined without cumulative damage, and even then only for a bounded interval, after which charging must stop rather than continue as a 'forever full' strategy.
Pulse maintenance replaces a continuous current with short charge pulses separated by long rests, sized so the average current merely offsets the cell's self-discharge rather than forcing continuous recombination. The bq2002 class of IC offers selectable pulse-trickle rates for exactly this reason; a charger can hold a cell near full while giving the electrodes and gas balance recovery time between pulses, dramatically reducing time-averaged overcharge and heating.
Designing pulse maintenance means estimating the self-discharge current at the holding temperature - higher warm, lower cool - and setting duty cycle so net ampere-hours balance it, with the cell spending most of its time at open circuit. This is far kinder than a constant current of the same peak value because recombination is intermittent rather than sustained.
Holding a cell at 100 percent for immediate use is a legitimate goal for an always-ready device, but it is the most stressful long-term state; for cells that will be stored, best practice is to charge, use or store them at a partial state (commonly cited around 30 to 40 percent) and to allow cooling after charge or discharge before the next event. A charger design should therefore distinguish 'maintain in the charger for imminent use' from 'prepare for storage', which needs no maintenance current at all.
Low-self-discharge cells weaken the case for aggressive maintenance: a modern LSD NiMH cell loses only around a percent per month (Paper az02 in the consumer series), so the average current needed to hold charge is tiny, and over-charging to compensate for self-discharge that is already minimal is self-defeating.
Some analyzers periodically discharge and recharge cells to counter voltage depression (the so-called memory effect) and to re-condition aged cells; guidance suggests an occasional break-in or refresh every several cycles rather than every cycle, because every unnecessary deep discharge is itself wear. Maintenance logic should not conflate self-discharge compensation - small, frequent pulses - with reconditioning, which is a deliberate, infrequent full discharge/charge.
The first figure contrasts the current waveforms of top-off, continuous trickle and pulsed maintenance; the second stacks the average overcharge burden of each strategy, making visible why a constant trickle accumulates damage that pulsed maintenance largely avoids.
A complete specification states the top-off current and its timer limit, the maintenance mode (pulsed preferred), pulse amplitude and duty as a function of temperature, the prohibition on continuous current beyond the C/20-class / 20-hour bound, and a separate storage recommendation. Validation holds cells under each regime for accelerated periods and compares capacity retention, internal resistance and internal-pressure behaviour to prove the maintenance strategy is compensating self-discharge rather than driving recombination.
Weijiang characterises self-discharge and recombination-limited maintenance current for each grade so partners can size pulse duty from data. With the full fast-charge-to-maintenance sequence specified, the final termination paper shows how these criteria are integrated into a single charge-management state machine and its silicon implementation.
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.
