End of discharge voltage for "refreshing" of NiMH cells

[Thomas Anderson]

I have been reading some books on NiMH batteries. Some authors state that for optimum performance and longevity NiMH batteries need to be "exercised" or "refreshed" or "reconditioned", that is subject to a complete discharge, followed by a complete charge.

For instance, from "Ni-MH Technical Handbook (Panasonic, 2017)":
"As with Ni-Cd batteries, repeated charge and discharge of these batteries under high discharge cut-off voltage conditions (more than 1.1V per cell) causes a drop in the discharge voltage (which is sometimes accompanied by a simultaneous drop in capacity). The discharge characteristics can be restored by charge and discharge to a discharge end voltage of down to 1.0V per cell."

This is form the "Handbook of Batteries (Linden, 2002)":
"The voltage drop occurs because only a portion of the active materials is discharged and recharged during shallow or partial discharging. The active materials that have not been cycled change in physical characteristics and increase in resistance. The active materials are restored to their original state by the subsequent full discharge-charge cycling."

One author states that such an "exercise" should be done for NiMH batteries at least once every three months.

There is some disagreement as to what "end of discharge voltage" should the cells be discharged to. Some authors suggest 1.0V per cell, others 0.9V per cell. Most authors agree that overdischarge is detrimental to NiMH batteries, but the rationale that they provide is only applicable for cells connected in series: if you discharge several cells connected in series, then because of the capacity mismatch one or more of the cells might get their polarity reversed, if the cut-off voltage is too low. To prevent that from happening people suggest setting the cut-off voltage to 1.0V per cell, or even higher when more cells are connected in series.

Polarity reversal is no doubt detrimental, but it could not happen if cells are discharged individually in a "constant resistance" mode (that is being connected to a load of a "constant" resistance). It is established that for NiCd batteries, discharging to a voltage below 0.9V per cell frequently works better for restoring the battery's performance.

In addition to this I have not found info on how this end of discharge voltage should be measured. By logic it should be the open circuit voltage, but if you look at the discharge plots, it seems as if they measure it with the load connected to the cell.

If you connect a load to a cell, the voltage will keep decreasing, but as soon as you disconnect the load the voltage will start increasing back, not to the original value, but considerably.

Say, I had been discharging a NiMH cell (an AA battery) in a "constant" resistance mode for about 30 hours and the voltage has dropped to 4.8 mV with the load still connected. After I disconnected the load, the cell's voltage started to increase rapidly and in an hour the cell's voltage has risen to 1.1V. Does this mean that the cell was not fully discharged?

So I have these questions:
1) What is the best end of discharge voltage for "refreshing" of a single NiMH cell in a "constant resistance" mode?
2) What is safe end of discharge voltage for discharging of a single NiMH cell in a "constant resistance" mode?
3) Is this an open circuit voltage, or a voltage with the load connected to the cell?
4) If this is an open circuit voltage then how one can practically measure it while discharging a single NiMH cell?

 

[crutschow]

For discharging one cell by itself:

1) 0.9V
2)0V
3)Load connected
4)NA

But unless you have a application where you have experienced a drop-off in capacity, I don't think this "refreshing" is worth the effort.

 

[Thomas Anderson]

3)Load connected

But if you measure the end of discharge voltage with the load still connected then how can you ensure a complete discharge of the cell? If you set the cut off voltage to 0.9V, the actual depth of discharge will depend on the discharge current: the higher the current, the smaller the depth of discharge. In one example that I gave the voltage with a load dropped to almost zero (4.8mV), and once the load was disconnected the cell's voltage rose to 1.1V. Thus, the cell was not completely discharged.

I don't think this "refreshing" is worth the effort

Since people research this topic, it kind of hints that it is worth the effort. Maybe not for everyone. Some people just use primary batteries and throw them away.

 

[crutschow]

once the load was disconnected the cell's voltage rose to 1.1V. Thus, the cell was not completely discharged.

The amount of charge that's left in the battery, once it reaches 0.9V under load (for reasonably small discharge currents) it likely too small to be of concern.
But if you want to complicate the discharge protocol by disconnecting the load before the voltage measurement, you are welcome to do that.

Since people research this topic, it kind of hints that it is worth the effort. Maybe not for everyone

I just have doubts that the memory effect is large enough to be worth the "refresh" effort for most typical applications of the battery.

 

[Thomas Anderson]

I just have doubts that the memory effect is large enough to be worth the "refresh" effort for most typical applications of the battery.

I use NiMH batteries with my photo camera, a lot. When the batteries get older, the camera won't start up with them (voltage depression), until I "refresh" them. But eventually I have to buy a new batch. Recently, after buying yet another batch, I started to wonder that maybe I was doing it wrong and decided to go scientific on them.

But in the books I've read I never saw a statement that 0V is a safe end of discharge voltage. All they say is that NiMH cells are sensitive to overdischarge, and a typical EODV is 1.0V or 0.9V.

Do you by any chance have a reference where this 0V EODV was tested and shown to be safe?

 

[ChrisP58]

When the voltage drops rapidly, it's done, regardless of what the voltage is.

https://www.powerstream.com/AA-tests.htm

 

[MrAl]

I have been reading some books on NiMH batteries. Some authors state that for optimum performance and longevity NiMH batteries need to be "exercised" or "refreshed" or "reconditioned", that is subject to a complete discharge, followed by a complete charge.

For instance, from "Ni-MH Technical Handbook (Panasonic, 2017)":
"As with Ni-Cd batteries, repeated charge and discharge of these batteries under high discharge cut-off voltage conditions (more than 1.1V per cell) causes a drop in the discharge voltage (which is sometimes accompanied by a simultaneous drop in capacity). The discharge characteristics can be restored by charge and discharge to a discharge end voltage of down to 1.0V per cell."

This is form the "Handbook of Batteries (Linden, 2002)":
"The voltage drop occurs because only a portion of the active materials is discharged and recharged during shallow or partial discharging. The active materials that have not been cycled change in physical characteristics and increase in resistance. The active materials are restored to their original state by the subsequent full discharge-charge cycling."

One author states that such an "exercise" should be done for NiMH batteries at least once every three months.

There is some disagreement as to what "end of discharge voltage" should the cells be discharged to. Some authors suggest 1.0V per cell, others 0.9V per cell. Most authors agree that overdischarge is detrimental to NiMH batteries, but the rationale that they provide is only applicable for cells connected in series: if you discharge several cells connected in series, then because of the capacity mismatch one or more of the cells might get their polarity reversed, if the cut-off voltage is too low. To prevent that from happening people suggest setting the cut-off voltage to 1.0V per cell, or even higher when more cells are connected in series.

Polarity reversal is no doubt detrimental, but it could not happen if cells are discharged individually in a "constant resistance" mode (that is being connected to a load of a "constant" resistance). It is established that for NiCd batteries, discharging to a voltage below 0.9V per cell frequently works better for restoring the battery's performance.

In addition to this I have not found info on how this end of discharge voltage should be measured. By logic it should be the open circuit voltage, but if you look at the discharge plots, it seems as if they measure it with the load connected to the cell.

If you connect a load to a cell, the voltage will keep decreasing, but as soon as you disconnect the load the voltage will start increasing back, not to the original value, but considerably.

Say, I had been discharging a NiMH cell (an AA battery) in a "constant" resistance mode for about 30 hours and the voltage has dropped to 4.8 mV with the load still connected. After I disconnected the load, the cell's voltage started to increase rapidly and in an hour the cell's voltage has risen to 1.1V. Does this mean that the cell was not fully discharged?

So I have these questions:
1) What is the best end of discharge voltage for "refreshing" of a single NiMH cell in a "constant resistance" mode?
2) What is safe end of discharge voltage for discharging of a single NiMH cell in a "constant resistance" mode?
3) Is this an open circuit voltage, or a voltage with the load connected to the cell?
4) If this is an open circuit voltage then how one can practically measure it while discharging a single NiMH cell?

Hi,

The answer is almost in your own post, but maybe not too obvious.

The standard test current is C/20, so that's probably the standard for choosing either 0.9v or 1.0v for the end of discharge voltage. Let's say it's 1.0v for now.
As you noted though, at a higher current discharge, once you reach 1.0v the internal resistance is dropping some of that voltage so at a lower C/20 current the voltage might have been actually 1.1v (just for example). If you cut out at 1.0v, that means you have not discharged the cell all the way yet.
This would however be the place where the application specifications start to kick in over the spec's of the cell itself.

That is, can the application device still work at 1.0v or not. If it can, then you might go down to 0.9v before cutout.
Probably the only way to check this is when the voltage gets down to 1.0v with higher current load, switch to a load of C/20 and measure the voltage. If the voltage is 1.1v, then you can go down (probably) to 0.9v, as long as the application still runs normally of course.

What this boils down to then is a measurement. As the voltage gets down to the original target voltage (we are saying 1.0v for that for now) then switch to a load of C/20 and measure the voltage, and if it is higher (and it will be) then as a first approximation take the difference and subtract that from 1.0v and use that as the new higher current discharge ending voltage. Again, if that comes out to 1.1v and the original target was 1.0v, then the difference is 0.1v, then subtract that from the original target voltage of 1.0v and that gives you 0.9v as the new target.
If you like, you can test again at 0.95v and recalculate the difference and the new target voltage, which may go up a little. For example, if at 0.95v at full current, then at C/20 current it comes out to 0.97v, then the new difference voltage would be 0.97-0.95=0.02v, and 0.95-0.02v=0.93v which would be the new target voltage for discharge at full current.
That's probably the only way to be sure, because although the original target voltage will always be the same (like 1.0v) the final actual cutout voltage will be lower when there is a higher current draw. This idea seems to be reflected in the discharge curves also.

 

[camerart]

Hi T,
I used to work for a battery charger company, and a few things were explained to me. From long ago memory!

They have a memory.

Many chargers can do recycling, so it kind of exercises a cell, and kin of pump it up. You can see that it should hold more milli Watt hours each loop, or it wasting it's time.

I thnk you can run them right down to 1% as long as the charger can 'see' them.

When they are full any more charging will need more ci=urrecnt and they will get warm so once over the peak, then stop charging.

Good luck.
C

 

[rjenkinsgb]

I just connect a ten ohm resistor across each single cell and leave them for a week or so.

The memory effect is not as bad as with NiCd, but still exists - when they are never "deep discharged, the plates crystallise to some extent and when discharged to the point the fresher amorphous metal is depleted, the internal resistance increases - which makes them appear flat when used with anything but low current loads.

The 1.1V per cell limit is the lowest discharge in a pack, where cells could be imbalanced and the weakest one be reverse polarised by a deeper overall pack discharge.

That cannot happen to a single cell, so leave until totally dead then fully charge and leave on low trickle (less than C/20) for a good few hours.

 

[Tony Stewart]

Just as in Li-Ion cells, you will get more lifetime mAh capacity using far less capacity than 90~100% between charges to prevent the acceleration of aging. All batteries have thousands of Farads of capacitance, C not to be confused with C/# charge rates, which are related to the milliohms of effective series resistance ESR and heat rise. Heat then under/over voltage are the main accelerants to aging.

Aging is not a constant. Overstress/abuse is bad.

It can be simulated that the imbalance of C in % kFarads and/or % of mohms ESR will accelerate going from 0.1% to 1% to 10% mismatch to dead for series cells with an exponential rate.

The memory effect is different.

It is due to many parallel networks of RC equivalent circuits (main and fringe) Some weak several second time constants have very low energy capacity. (multiple double-layer-effect capacitance properties of electrolytic and high density ceramics have this memory effect. Like shorting a big e-cap in a fast arc then measuring new voltage later with a zap from old TV caps on your finger. Ni-Cads were different and had a bigger memory effect and depended on monthly draining to recondition the secondary charge chemistry but Lithium also has memory but a smaller percentage but don't try that. Instead use only from 80% to 20% (give or take) to get 10x the number for cycles from full discharge. You can float also Li Ion near 66% voltage with a standby regulator and cycle full range every month or few.

The weakest cell in series gets under/overcharged 1st. So the more conservative you are on SoC range or % DoD Range , the bigger the product of cycle number * capacity used to 10x to 20x rated cycle number.

Batt. U site has more details.

To measure C and ESR, there are tools or try a simple way knowing the multiple RC products can be averaged to an equivalent or minimum with changes in burst current levels.

Load cell with C/10 current and measure ΔV/ΔI=ESR Ω
Load cell with constant current and C[F]=I Δt/ΔV in [amp-seconds per volt = Farads]

then measure C rate of change dropping and ESR rate rising .
Again don't confuse capacitance from charge ratio, C related to Amps in A-h rating. Normally cells are rated in a 20 hour time scale for fast reuse types. Others usually have longer time constant ratings for smaller loads and with longer usage cycles.

 

[Tony Stewart]

"constant" resistance mode for about 30 hours and the voltage has dropped to 4.8 mV with the load still connected. After I disconnected the load, the cell's voltage started to increase rapidly and in an hour the cell's voltage has risen to 1.1V"

I would expect that memory storage energy is 1% of your mAh rating. You can verify ESR by choosing an R to cut the voltage quickly by 50%. It will probably be 100 to 1k times bigger ESR from fully 0V discharged to fully charged.