BU-807: How to Restore Nickel-based Batteries

During the nickel-cadmium years in the 1970s and 1980s, most battery ills were blamed on “memory.” Memory is derived from “cyclic memory,” meaning that a nickel-cadmium battery could remember how much energy was drawn on previous discharges and would not deliver more than was demanded before. On a discharge beyond regular duty, the voltage would abruptly drop as if to rebel against pending overtime. Improvements in battery technology have virtually eliminated the phenomenon of cycling memory.

Figure 1 illustrates the stages of crystalline formation that occur on a nickel-cadmium cell if overcharged and not maintained with periodic deep discharges. The first enlargement shows the cadmium plate in a normal crystal structure; the middle image demonstrates full-blown crystalline formation; and the third reveals some form of restoration.

Crystalline formation on nickel-cadmium cell

New nickel-cadmium cell.The anode (negative electrode) is in fresh condition. Hexagonal cadmium-hydroxide crystals are about 1 micron in cross section, exposing large surface area to the electrolyte for maximum performance.

Cell with crystalline formation.Crystals have
grown to 50 to 100 microns in cross section, concealing large portions of the active material from the electrolyte. Jagged edges and sharp corners can pierce the separator, leading to increased self-discharge or electrical short.

Restored cell.After a pulsed charge, the crystals are reduced to 3–5 microns, an almost 100% restoration. Exercise or recondition is needed if the pulse charge alone is not effective.

Figure 1: Crystalline formation on nickel-cadmium cell. Crystalline formation occurs over a few months if battery is overcharged and not maintained with periodic deep discharges.

Courtesy of the US Army Electronics Command in Fort Monmouth, NJ

The modern nickel-cadmium battery is no longer affected by cyclic memory but suffers from crystalline formation.The active cadmium material is applied on the negative electrode plate, and with incorrect use a crystalline formation occurs that reduces the surface area of the active material. This lowers battery performance. In advanced stages, the sharp edges of the forming crystals can penetrate the separator, causing high self-discharge that can lead to an electrical short. The term “memory” on the modern NiCd refers to crystalline formation rather than the cycling memory of old.

When nickel-metal-hydride was introduced in the early 1990s, this chemistry was promoted as being memory-free but this claim is only partially true. NiMH is also subject to memory but to a lesser degree than NiCd. While NiMH has only the nickel plate to worry about, NiCd also includes the memory-prone cadmium negative electrode. This is a non-scientific explanation of why nickel-cadmium is more susceptible to memory than nickel-metal-hydride.

Crystalline formation occurs if a nickel-based battery is left in the charger for days or repeatedly recharged without a periodic full discharge. Since most applications fall into this user pattern, NiCd requires a periodic discharge to one volt per cell to prolong service life. A discharge/charge cycle as part of maintenance, known as exercise, should be done every one to three months.Avoid over-exercising as this wears down the battery unnecessarily.

If regular exercise is omitted for six months and longer, the crystals ingrain themselves and a full restoration with a discharge to one volt per cell may no longer be sufficient. However, a restoration is often still possible by applying a secondary discharge called “recondition.” Recondition is a slow discharge that drains the battery to a voltage cut-off point of about 0.4V/cell and lower. Tests done by the US Army indicate that a NiCd cell needs to be discharged to at least 0.6V to effectively break up the more resistant crystalline formations. During this corrective discharge, the current must be kept low to minimize cell reversal and, as discussed earlier, NiCd can tolerate a small amount of cell reversal. Figure 2 illustrates the battery voltage during a discharge to 1V/cell, followed by the secondary discharge to 0.4V/cell.

Exercise and recondition features of a Cadex battery analyzer


Figure 2: Exercise and recondition features of a Cadex battery analyzer

Recondition restores NiCd batteries with hard-to-remove memory. Recondition is a slow, deep dis-charge to 0.4V/cell.

Courtesy of Cadex

Recondition is most effective with healthy batteries and the remedy is also known to improve new packs. Similar to a medical treatment, however, the service should only be applied when so needed because over-use will stress the battery. Automated battery analyzers (Cadex) only apply the recondition cycle if the user-set target capacity cannot be reached.

Recondition is only effective on working batteries. Best results in recovery are possible when applying a full discharge every 1–3 months. If exercise has been withheld for 6–12 months, the capacity may not recover fully, and if it does the pack might suffer from high self-discharge caused by a marred separator. Older batteries do not restore well and many get worse with recondition. When this happens, the battery is a ripe candidate for retirement.

Results of Battery Maintenance

After the Balkan War in the 1990s, the Dutch Army began servicing its arsenal of nickel-cadmium batteries that had been used for the two-way radios. The technicians in charge wanted to know the remaining capacity and how many batteries could be restored to full service using battery analyzers (Cadex). The army knew that allowing the batteries to sit in the chargers with only two to three hours of use per day during the war was not ideal, and the tests showed that the capacity on some packs had dropped to a low 30 percent. With the recondition function, however, nine out of 10 batteries could be restored to 80 percent and higher. The army uses 80 percent as a threshold for usability. At time of service, the nickel-cadmiumbatteries were two to three years old.

To analyze the effectiveness of battery maintenance further, the US Navy carried out a study to find out how user pattern affects the life of nickel-cadmium batteries. For this, the research team responsible for the program established three battery groups. One group received charge only (no maintenance); another was periodically exercised (discharge to 1V/cell); and a third group received recondition. The 2,600 batteries studied were used for Motorola two-way radios deployed on three US aircraft carriers. Table 3 summarizes the test results, including the cost factor.

Maintenance method

Annual % of batteries requiring replacement

Annual battery cost

Charge-and-use only









Table 3: Replacement rates of nickel-cadmium batteries
Exercise and recondition prolong battery life by three- and nine-fold respectively.

GTE Government Systems, the organization that conducted the test, learned that with charge-and-use the annual percentage of battery failure was 45 percent; with exercise the failure rate was reduced to 15 percent; and with recondition only 5 percent failed. The GTE report concludes that a battery analyzer featuring exercise and recondition costing US$2,500 would return the investment in less than one month on battery savings alone.


On December 6, 2011 at 11:43am
COburner wrote:

I’ve tried the deep reconditioning method advised here on some individual NiCd and some multicell tool battery packs. I’m unsure if you could deep discharge multicell packs effectively without taking the pack apart and discharging each cell individually—so far that’s what I’ve done because used cells vary so much in capacity. If you try to discharge the whole pack (even very slowly) it seems some will hold charge a long time while others go into reversal. I’m not sure if that harms the weaker cells or not. If you need to discharge each cell individually it becomes an enormous hassle to keep a pack conditioned. Any advice on this?

I’m also quite unclear from this article if the target for deep conditioning discharge of 0.4 v. is a no-load voltage or what? I’ve assumed it’s a no-load voltage, but it takes a very long time to get there at low disch rate. The battery that’s 0.4 v under 50 ma. load will repeatedly bounce back to 1.1 v. under no load. Can you clarify this?

On July 28, 2012 at 12:57am
Graham Jones wrote:

I have the same experience as COburner, and would greatly appreciate your advice. Many thanks

On July 29, 2012 at 8:50am
robert wrote:

What current should be used for deep discharge? example 0.1 C or lower?

On August 4, 2012 at 4:56am
Arek wrote:


Works with batteries ni-cd for 18 years. Much of that time I gave discharge ensure preservation. Max. discharge current for a typical AA cell is 15 mA, the High Current SC - 50mA. The current is independent of the capacity of the cell, since less capacious cells are resistant to reverse polarity.
Sorry for the english, I come from Polish.

On August 4, 2012 at 2:49pm
COburner wrote:

Agree with Mr. Arak. Slow discharge seems best for reconditioning purposes. For typical used AA NiCd of 1000 maH capacity (C), C/50 = 20 ma. For typical sub-C of 2000 maH capacity, C/50 =40 ma. One might argue that anything slower than C/20 is OK, but I’m not sure of that.

I certainly wish the scientist/engineer who wrote the article, or others at his company, would respond here. Since posting the question 7 months ago, I’ve continued to experiment. First, I think I was wrong in assuming you should drain each cell to 0.8 v. measured with no load. With no load, you’re just measuring surface charge, not true charge state. So I think that the target discharge state must be measured with a light load, 50 ma. for a sub-C cell.

It’s a tiresome endeavor to try to externally discharge a tool battery pack (e.g. 12 v, with 10 NiCl cells in it) to the above recommended 0.8 v. per cell. After a period of use, cells seem to have very different charge capacities. Not knowing whether prolonged charge reversal would hurt the weaker cells, I’ve been reluctant to keep even a slow discharge (C/40, where C = charge capacity of cell) on a pack for the 30+ hours it might take to fully drain the stronger cells in the pack. Here’s my ‘current’ protocol for pack conditioning:

  • use up the pack’s charge with the tool or device, to the point it’s depleted.

  •  open pack so you can measure each cell’s remaining charge under light load (C/40 or less; for typical 1.2 v. sub-C cell this would be 50 ma., approx. 22 ohm load); some will be at 0, many will have >1.1 v.

  • for any that still give you >0.8 v., wire each cell to its own 22 ohm 1 w. resistor, keep load on until voltage is < 0.8 v. under this load.

  • do a test recharge on pack as usual; let sit a few days and test each cell again as above; those that are below 1.24 v. probably are self-discharging and should be replaced.

  • replacing individual cells in pack is tricky to do, but possible. I don’t have the spot welder you should have, but I’ve had luck using a 250 w. soldering gun & copper strips to attach new cells. Careful prep of surface, pre-tinning attachment spots, applying minimum necessary heat has worked so far. To keep all cells in a pack as age-matched as possible, I’ll try to combine the good cells from 2 old faulty packs to make one good one.

  • keep pack in discharged state until needed; don’t keep packs fully charged and unused for months at a time.

If there are errors here, I’m happy to have them corrected by someone with more knowledge.

On August 5, 2012 at 1:45am
Arek wrote:

Hello again.

I gave the current at which there is no polarity reversal for destroying cells that are in series.
Very supportive of Mr.. Isidor to do a deep discharge from the beginning of a new battery life is very, very prolonged his life.
Unfortunately, cells that have internal short circuit can not be saved, although their performance slightly increases, it will never get rid of no fault, and the battery will discharge itself.
My oldest AA Ni-cd are now 14 years old and about 540 mAh to 600 mAh her ​​when they were new. no longer have such a low internal resistance and faster to unload but still hold 50% charge after about 40 days.
Modern Ni-Cd cells are not as good and will not live as long as manufacturers design them specifically worse. But always appropriate treatment can be extended several times of their lives in good condition.
Of course, while charging the battery must be cool, if it heats up the charge current is too high.
Developed by the Soviet army for the cell current is 120mA SC for the maximum duration of life without regard to capacity as more capacious goals less accept over-charge.
For cell-C 160mA and 350mA-D cell. However, the current welfare and civilian customers forced to design a fast charger because posiadli money and charging time was the most important.
In 1970 one AA cell in the socialist countries cost about 1/10 the average salary. The recipients were only military users. Produced very little, the price was very high and therefore had to be a lifetime maximum.

On August 5, 2012 at 1:52am
Arek wrote:

posiadli - have.

For AA-cell 45-50mA, for F-cell 700 mA (military battery for radio).

On March 17, 2013 at 5:45pm
mike wrote:

“a battery analyzer featuring exercise and recondition costing US$2,500 would return the investment in less than one month on battery savings alone.”

I’m thinking that’s stretching things. How many cells (or packs - batteries) a day can this device recondition? The graph shows a 3 hour period. Unless it can handle multiple packs at a time, that’s 8 per day - 240 per month, way less than the 2600 batteries in the study.

On May 13, 2013 at 11:30am
Eric J wrote:

Mike, I bet it sure would pay off in a month for the U.S. Navy (that’s the statement they were making. Not for your average consumer of course). They probably have guys mistreating and tossing batteries left and right when they don’t hold a perfect charge anymore. Can’t really blame ‘em for tossing ‘em if it’s something that might save their lives.

On July 5, 2013 at 2:20pm

After spending about 2 months + non stop online learning about nicd and other battery types,and brutally learning the lesson of the battery voltage that refuses to die.ill give you this example of how i would discharge an 18v cordless tool battery pack.

i would charge the battery,use it in the tool till there is a noticeable drop in power output,take it out of the tool,connect a 1000 omh resistor and 6 leds at 3.6v each blue or white and let the battery run until it totally died,then unhook the leds and short the battery with the 1000 ohm resistor and put it in storage.

On July 6, 2013 at 7:18am
COburner wrote:

P.O. Titan: are you measuring discharged voltage under a small load (C/20-50)? No load voltage for most types of cells tells you almost nothing of the true state of charge.

As written above, you’re safe to discharge sub-C cell packs at 50 mA, which would be about 400 ohm load. I don’t see the point of leaving the 1000 ohm load on the fully discharged pack either.

I’m now trying out a smart charger that can put cells thru conditioning cycles, test capacity using charge/DC cycle, charge at various currents, and measure cell resistance. Generic Chinese maker and cheap, but I’m not going to recommend until I’ve tested further.

On July 8, 2013 at 8:05pm
PO TITAN wrote:

all i do all day is work on stuff like batteries.im going to tell you what i know and what i think.i know that when a nicd is fully discharged the voltage drops to 0 under load,any load pretty much,and if you short out the battery like nasa does when they store them the voltage is 0 and it stays that way.after you remove the short the voltage quickly climbs back to close to normal voltage 1.2+- even after days of being let shorted .according to everything iv seen the best plan is to drain the battery fast and hard then drain it as slow as possible then short it out-when the danger of over heating ect is no longer present.i believe from my research that a nicd is fundamentally unstable and that this is the reason for long life (electrode regeneration especially on the cadmium side) but that when the battery is simply left lying around discharged or charged fluctuations in the open circuit voltage lead to crystal growth/distortions.if im correct eventually all nicd will short out? it would naturally stabilize the environment inside the cell.

On August 16, 2013 at 2:40pm
PO TITAN wrote:

as a follow up to my previous post’s,i took a brand new 9.6v nicd battery pack to a cordless drill and hooked 3 leds to it.it took 9 hours to charge and with a 1000ohm resistor it literally was still discharging like 2+ weeks later.i stopped it around 7v and if you left the pack sitting the voltage would climb to only about 8v but after charging again the voltage climbed back up to 10v,the initial level of the brand new pack.iv recently read papers online that say that leaving a nicd to self discharge causes crystal growth,it also recommended not leaving the pack open circuit.My strategy now is to just put enough leds together to match the voltage of the pack and let it discharge,resistors aren’t really necessary because the voltage drop on the pack controls current flow.it may take multiple sets of leds to fully discharge the pack down to 0.Example 18v pack 6 leds voltage drops to 12,connect 4 leds ect.Im now considering high voltage pulse standby float charging.

On September 19, 2013 at 9:30am
martin borjas wrote:

I had a dead battery this a.m. and put my charger on. about 5 ninutes latter I unhookeed the negative and noticed the volts on the charger climb. what is up with this?

On September 21, 2013 at 8:43pm
po titan wrote:

it’s either something weird about your charger or it sounds like what i described above the voltage of a totally dead battery is like 1.12-1.16.

a good tip to add is that if you leave a battery shorted for along time it will leak,i had 2 AA nicds leak badly.I use a breadboard and i remove 6 5mm white leds from an 18v battery 1 at a time until it is down as low as possible.the voltage will stay low but once i charge again it goes back up to normal like 20v or so.job finished in my book.

On October 29, 2013 at 2:47am
Arek wrote:

Welcome COburner .

Surely 0.4 V is measured under load.
This is certainly a very small current .

Good exercise loading - unloading can take several weeks.
It is often not cost-effective , I performed only for research purposes .

If the cell is susceptible to repair, it ‘s been two cycles of very deep discharge will significantly improve its condition .
Unfortunately, the old cells , very worn out will not be a good fix .
The electrolyte is an old and often evaporate as a result of overloading .

Repair should link misused or old that have long lay unused .

Of course, even for cell 50mA high power SC will reverse polarity sensitive . But I will not damage the electrodes and the electrolyte.
Mass anti -pole assume no problem 50mA current and turns it into heat.

I have a Panasonic SC cell 1.9 Ah ( high power ) , they can take 150mA and no damage is , it can take up to several hours. Cells already have seven years and I do it once every two weeks.
This is a very good link. Currently have a 1700mAh at 0.2 C discharge.

Poor quality cells have no mass anti-pole, even 50mA can be reversed polarity and spoil the electrolyte and the electrode. Often, these cells swell with reversed polarity even a very small current.

A friend recently got a new cell weak brand. Cells without a name. Were sold in Poland for a professional company for emergency lighting.
The cells are 1.5 Ah high temperature
After 10 cycles to 1.1 Ah.
Load current of 0.5 A at - DV (- 1 mV).
Unloaded 0.5 A to 1V per cell.

Here, no battery memory effect only corrupts the very very poor quality.
It can not be repaired. I recently sailed electrolyte from the cells, leaving white coating of potassium hydroxide.

I recommend to use good cells because they have a real capacity and you can try to fix it when we use it properly.
Poor quality cells are highly perishable and are expensive to maintain because they still need to be replaced and monitored.

The biggest impact on life expectancy than - CD cells is the temperature , especially when loading . You have to use the DV systems responsive as possible .
The only draw back is that old or unused cell long may be too early detected as charged. After a few cycles of load cells normally.
Cells can not heat up above 30st.C during charging. If that is the charge current is too high or DV system is not sensitive .
Factory battery charger cordless tools are often very poorly loaded . very hot batteries for very low detection - DV .
When you load the correct charger can even 2-3 times to extend the life of the battery and its charging speed as it saves time, which previously served only to reload the cell and its heating .
The heated cell also loses some of the accumulated capacity and shorter working device powered by a battery .

Perfect charger detects (- DV) of ZERO.
Charging takes the shortest.
Loading cares very much about the life of the battery.
The battery is the most energy.

Even with 1C charge the battery remains cool.
High current also minimizes the formation of memory effect.
A lot of cycles can be performed. More than 0.1 C charge and no - DV (charger overnight).

This charger Zero DV very difficult to buy, I myself slowly so I make the ATmega8 (learn Bascom). My professional charger has a sensitivity of-3 mV. It’s a little too little for the best possible charge.

Ni-MH batteries, just not hot. Cycles are almost useless, a mere work them enough. Discharge cycle necessary when non-stop several dozen times discharge is only 30% and then charging. Very graceful batteries, just recharge the chilly temperature 30st.C max.
The ideal fast charging is 0.3 C and ZERO DV for high power cells.

On November 14, 2013 at 4:16am
Imre wrote:

Hi Arek,
I’m opening a battery recondition business in Ireland.I would like to contact you.my email is pathfinder0628@gmail.com
Regards Imre

On November 14, 2013 at 4:27am
Imre wrote:

Hi Arek.Im opening a battery reconditioning business in Ireland,would like to contact you.
Regards Imre

On March 22, 2014 at 9:40pm
jjgarett wrote:

Id like to add that the army pays 10 times the normal cost for items. The savings is easily seen. thanks

On July 9, 2014 at 6:27pm
Chuck wrote:

So from what I am reading a owner of nicad & nimh tools are set up to have battery failures if 1) they do not use their tools on a regular basis & 2) if you do not take the battery pack apart and check the cells individually and have slow deep drainage of cells at low currents. Is this true? I think my batteries do not last because of low usage and I charge them on a one hour charger only. Do these statements sound true and what can be done as a consumer to keep our batteries longer? Thanks

On November 8, 2014 at 6:37pm
Edward wrote:

the ni-cd battery are going to disappered…...............the ni-mh battery are replaced by lithium battery. what a pity, i have been working with ni battery for so many years

On March 3, 2015 at 8:59pm
Rick wrote:

On Nov. 8th, 2014 Edward wrote “the ni-cd battery are going to disappered…...............the ni-mh battery are replaced by lithium battery. what a pity, i have been working with ni battery for so many years”.  This is true, I’m sure but there are thousands of perfectly good Ni powered tools out there and replacement batteries, if priced right, will keep these tools in use for years to come.

On March 3, 2015 at 10:36pm
Edward wrote:

Hi Rick thank you for your encouragement,  please tell me the contact information if there are the NI battery need requirement zzrm316@163.com