No stress, its way lighter duty than what the same battery would get in a laptop

Proper charging circuit (and not "higher V through big resistor" as used with NiCd) is essential for Li-xx batteries, and that prevents overcharging

Overheating can only happen in tight lanterns in which there is a working lamp at normal time. A lantern with some more air space around the gear does not get very hot, definitely less than some laptops out there. If you place the battery away from the ballast or lamp cathodes itll only be marginally warmer than room temperature. In an emergency-only lantern there is no problem, it is at room temperature at all times anyway

Controlling the charging of the LiIon batteries belongs to the easier tasks, it just means reduce or switch off, once the voltage reach certain level. The only trouble is the required accuracy. However once you overcharge it, it really tends to end up in flames.

Their aging goes mainly with cell voltages exceeding 3.9V and (as nearly every chemical reaction) is highly accelerated by the heat.
Of course, using just the ~3.8V as the float voltage will improve the lifetime to about the same as  SLA's, but then the usable energy density would become comparable to the NiCd or NiMH (when the NiMH is supposed to handle the constant overcharging, the attainable capacity is about the same as the NiCd) or more expensive than SLA.

Plus the fact you have seen a piece that have lasted unusually long does not mean the life is sufficient (well, maybe if you want to "compete" by quality with the most cheepeese makers, then it could be assumed as sufficient). It would have to be 99% of the produced batteries lasting that long, when you want to state certain period as a life rating to count on.


The sealed NiCd/NiMh are the most difficult cell to control without overcharging: Their voltage is very strongly temperature dependent, what when combined with the heat generated during the overcharge (with NiMh even with normal charging) causes the normally rising voltage as the charge level increases is reduced by the cell self heating, so e.g. simple floating voltage regulation as a means of automatic charge termination (working with lead acid, LiIon, as well as wet NiCd's) does not work (unless there is accurate cell temperature sensing coupled to the voltage regulation). This behavior was the primary reason to design the cells so, they will be able to handle long term overcharging (recover the water from the gassing overcharging reaction) and in the electrical part use just that simple charging resistor, without any charge termination at all and completely relying on the internal water recovery mechanism.

If tyhe battery voltage goes up as it charges AND up as it heats, then a simple on/off charging control with high level, low level and hysteresis would suffice.... Somehing along the lines of that 555-timer charger i designed a couple years back

How long would a battery last on that vs on float ?

If tyhe battery voltage goes up as it charges AND up as it heats, then a simple on/off charging control with high level, low level and hysteresis would suffice.... Somehing along the lines of that 555-timer charger i designed a couple years back

How long would a battery last on that vs on float ?

It goes up as it charges, but DOWN as it heats. That is then responsible for the famous "negative dV/dt" at the end of the charging: When the cell gets fully charged, the voltage related to the charging stops rising, but the heat generated by the overcharging causes the temperature to rise, so the overall voltage starts to decrease.
But that could be usable only for fast charge termination (and even that is not that much reliable), but not for the permanent maintenance - there it tends to make the setup rather unstable: Once the cell starts to overcharge a bit, it heats up, decreasing it's voltage, so drawing more current and so on.

With the maintenance charging you need ideally to keep the cell voltage just below the gassing level, but stil above the electrochemical voltage, so you keep the cell really charged. That voltage window is really small (low 10's of mV), so in order to really work, the charger voltage (regardless if it is formed by a linear regulator or a comparator switching ON and OFF the charging current - the type you present) has to track that window.
The lead acid are quite temperature stable and the charging (nor overcharging) does not generate any significant heat, so just a constant voltage is sufficient to keep them charged.
But with the NiXX you have to track the internal cell temperature, what means tightly coupled temperature sensor and very limited charging current, so the temperature difference between the core and surface is small enough. Quite complicated arrangement, mainly the battery pack itself. The constant overcharging is way easier and when supported by the cell design (so that means not exceeding the rated maintenance charge current; that means for standard cells about 50mA for AA size or ~100mA for a subC size cell; regardless if it is NiCd or NiMH - really given just by the cell size), way more reliable and leading to the longest life.

The LiXX are different mainly, because the overcharging there is extremely damaging - therefore these can never be run that far, that's why the extreme requirements for the charger termination voltage accuracy (better than 20mV on 4.2V to keep the capacity and life within the rated limits).

Then, as the battery charges initially dV/dt is positive, as it starts to heat dV/dt is negative. Between them dV/dt is zero. How about a charger that cuts out on dV/dt = 0 ?

The "dV/dt=0" point is, when the overcharging already takes the full current and the battery is already pressurized. In fact the overcharging itself does not directly heat up the battery, it "just" decomposes water to the separate hydrogen and oxygen. But it is the water recovery reaction, where these are combined back into the water, what generates the heat. And this sequence means, the heat generation always lags the actual overcharging (first the gasses get generated, then they have to pass to the extra electrode area, pressurize there a bit to gain some momentum and only then recombine back to the water and generate the heat.)
And in fact passing the "dV/dt=0" is the point, recommended to terminate the NiMH charging. The reason is, even the main charging reaction of the NiMH generates some heat (together with the Joule losses), so they get already preheated, so have less thermal margin, so the need to stop that earlier. And that generated heat means, once you want to use really fast charging, the temperature during normal charging will reach the maximum allowed, so there would be no room for the detection via the voltage. The only method used for that rapid charging is to detect the increasing pressure. Either via a strain sensor within the cell plastic sleeve, or using a pressure actuated disconnect contact inside of the cell (the "15 minute charging" cells).

With NiCd's the charging reaction actually absorbs heat, so they stay cool even beside rather high Joule losses with really very fast charge rate; therefore the tolerance of the NiCd towards that high charging currents (the only thing deteriorating the cells at high currents is the heat; no heat means no degradation). The same means, the dV/dt is way more pronounced, so it could be sensed even by just an analog circuit when high enough charging currents are involved (NiMH needs digital signal processing, as the currents can not be that high, so the slopes are way slower, so become too slow for time constants feasible only in an analog circuit). And because they get colder during the fast charging, just sensing the temperature to exceed e.g. 35degC threshold becomes the simplest fast charging control for them, when the temperature could be sensed (most battery power tools use this method, either by part of the cell being naked to be touched by the charger sensor, or a thermal switch being integrated inside of the pack)

And by the way pretty ingenious way was used in the Toyota Prius: The NiMH cells are made as wet designs, pack of few of them sealed within one pack, with the water regeneration place (serving that group of cells) being separated from the main cell's body (and attached to external cooling system). That means the heat from the overcharging is generated outside of the cells, not threatening any sensitive parts, could easily operate at very high temperatures, while the cells itself remain rather cold, while dissipating a lot of overcharge energy. That arrangements makes the regeneration dynamic breaking pretty simple: No need for extra overcharge control, dissipate all the energy required by the driver, always fully charging all the cells before redirecting the energy into a power dump (in fact not needing any separate power dump in the car at all, the cells can pretty well serve that purpose)

In a working emergency lantern the battery may be located next to a working lamp or its ballast (the mains powered one). Then it means that the battery will reach the temperature/pressure of dv/dt reverse in earlier stage

In an emergency-only lantern on outer wall under rain or snow (water proof lantern), temperatures will be low, and the battery may keep charging....

What happens in those cases ?

What happens in the event of change ? - Say the weather proof lantern is a small plastic "melt light" with high power PL or 2D, it was switched on and hot inside, and reached full charge under that temperature. Then it is switched off and starts cooling down to ambient temperature (say near 0C)

What you say about battery life seems to be true our housephone has 2 triple A Lion battery they have only lasted about 2 years because the phone spends most of its life on its charging bed. They do seem to get rather warm even when they've been on charge for days

Li-Ion cell voltage is three times that of other cell types, it mean that the user swapping it with any other type of battery will either have a phone that does not work (supplied with 2.4V instead of 7.2V), or blow up other stuff that he put the Li-Ion battery into

I really doubt any sensible manufacturer will make a home appliance that use Li-Ion batteries in a size interchangeable with other batteries. Most likely the phone uses NiMH or NiCd

Some batteries take constant charjing better than others. This issue of batteries malfunctioning due to constant charge could be fixed by a circuit that detects battery level. Some of those lanterns had those charge detection circuit damaged due to EOL batteries.

I have yet to see a lantern with an on/off charge controller.... All the ones i have are just the resistor that limit the current

And those fail just as well :

The resistor takes the voltage difference between the supply and the battery. In a charging after a power outage, in the beginnning the battery voltage is low, and high voltage fall on the resistor, then as the battery voltage goes up the resistor voltage drop comes down

In lanterns in which the resistor is marginal in power rating, it is designed so the resistor is on the edge when the battery is empty, and assuming that it does not take very long to charge, the resistor is not overloaded like that for long. But when a NiCd battery EOLs it goes shorted, then all the available voltage fall on the resistor, it heats up and burns out or damages the PCB

The new led lanterns have the battery level cutout so that it won't overcharje. I think that the portable battery powered lantern I found at a thrift store had a charge cutout for the NiCd batteries that you could install. You could use regular batteries, just don't charge them.

@Ash: Well, if the resistor is not properly rated, it indeed fails sooner. The thing is, normally the low voltage state should not happen in the first place (because of the undervoltage cutout) and even when it happens, normally it takes few minutes for a healthy battery to recover to the 0.9..1V/cell level and that the resistor should stand well, even when overloaded. The problem is indeed, when the battery fails short circuit, then the overloading is not time limited anymore and then it really may fail. The thing is, many of these fixtures are not designed for anything being really replaced over the complete fixture life. Quality NiCd's with proper undrvoltage protection tend to last there 10 years (with one discharge cycle per month in average) and that is the expected fixture life as well (plastic degradation,...)...
You may ask why not replace just the batteries? Well, if the new battery pack costs the same or even more than exactly the same fixture, plus it is rather hard to get, it does not make much sense...

@Solanaceae:
There is significant difference if the fixture is a certified emergency illumination device, or portable lantern (even when it has "emergency" feature). The first are always fixed mount lights
Depends on the battery chemistry. With the NiXX there is no need for any cutout at all, but those are not used in portable lanterns (the Cadmium is anyway banned in consumer products).
In older models were used SLA batteries, but mainly with the newer consumer products you meet nearly in 100% the LiIon, as with the portable lantern the expected use regularly on batteries.
And both SLA and LiIon require the voltage charge control (cut out or regulation).
And the LED voltage matches the LiIon's, those are really getting popular in those portable lanterns and there the voltage control is really a must...

The situation with the emergency illumination is a bit different, there the light operates nearly 100% time on the charger, whose only function is to keep the battery ready to deliver the rated illumination time at any moment. For that the 3x NiCd's are the best performance to go with LED's. The LED itself prevents the overdischarge, so no need for any special circuit, for the charging you suffice with just that resistor (well, I more like an incandescent, as at first it dissipates the power way easier, and seconds it provides very simple diagnostic about the battery state - dark => no power, dimmer glow => charged, barely red => battery EOL, bright => battery empty or when long time battery EOL; and you even do not need any ladder to check it).

In an earlier post someone put that the red LED indicators in emergency lights are in parralel with the charge resistor is this still done with newer EM lights? The reason I ask is there are some on a public building I see and the green LEDs seem very bright even though thebattsrys have been on constant charge for maybe 6 months or even longer we are lucky here powercuts are very few and farbetween

What I've seen on the "just a resistor" chargers, the LED was indeed just parallel to the charging resistor (with it's own series small resistor, of course). The intention is to allow at least some indication of the battery state (similar as the incandescent would provide).

However when the charger uses some more advanced charger circuit, the LED could be supplied separately, turned ON or OFF by the diagnostic output of the charger electronic.