When to recycle LiIon cells?

[cave dave]

What are the characteristics of a cell nearing the end of its safe or useful life? It will help greatly if I don't need fancy equipment to figure this out.:thinking:

I was hoping for a sticky on this topic or a link in Threads of Interest but I couldn't find one.

I have some 2 or 3+ yr old LiIon Rcr123's. Some blue label ones, AW and Battery station.

I've noticed that my WF139 charger has never charged these over 4.17v, even when everything was new. I pull them when the light turns green and as a backup I have a pin-type timer connected to it to automatically cut power after about 3hrs from start.

I thought I might get a little better safety and performance by getting a Pila IBC which I received from Kitelites GB last week (great guy to deal with).

When I put the batteries in the Pila the max voltage on the best battery after removal is less than the WF 139 by about 0.02v. In example case about 4.15v. Some of them are lower. I think the worst was about 4.09v.

My first thought was that the Pila was bad. But I'm now thinking its the batteries after reading some silverfox comments in other threads. I thought that if a charger charges to 4.20v that meant that the batteries would come off at 4.20v, but I may have been mistaken.

I have also read that LiIon cells can loose 50% of their capacity over the course of 3 yrs whether they are used or not.

Also when using my Ra170cn flashlight most of the batteries will trip off the low battery indicator and upon removal will read about 3.24v. Battery Station Rcr123 (BS2) has a shorter runtime and upon removal will read 3.66v or so.

The runtimes I have are:
BS1 = 59 min (bought May'07)
BS2 = 49 min (bought Dec '06, end voltage 3.66v)
AW1 = 55 min (bought Dec '07)
AW2 = 63 min (bought Dec '07)

The testing equipment I have at my disposal is the Pila, a voltmeter, and the Ra flashlight. I have no way to drain cells except for flashlights.

Thanks, and hopefully the resulting discussion will achieve sticky status!

 

[Turbo Guy]

The lower open / resting volage at the end of charge,reduced run time and increased bounch back (higher) voltage when the load is removed all point to tired cells which have developed higher IR over the years.

Most likely if you charge them up and let them set while taking voltage readings every 24 hours for a week or so you will see that they like many tired NiMH cells have developed a much higher self discharge than newer cells.

Best storage for LiIon cells is at approx. 50% stste of charge and at cool temps. 32F being ideal. The higher the stste of charge and the higher the temperture the faster they bevelop higher IR.

Cells with high IR can be deceptive as they will often still show excelent capacity if discharged at very low rates but will drop like a rock under heavier loads. Much like a car battery that will power the dome light or radio but not start the engine.

 

[cave dave]

I looked up when I bought the cells and added that data. I also wrote the info on the cells.

The BS cells have spent the last year or so in the fridge at 3.8v. I haven't realy used them since I got the AW's. I let them come up to room temp before charging.

Still, at what point should I "let them go? I think I will recycle BS#2. It is currently charged so I guess I will just do a last runtime test in the Ra to drain off the energy and recycle it if someplace will take it.

 

[Turbo Guy]

The age old standard is when capacity drops to 80% of orginal.

For reasons lost in time, the manufacturers seem to be using 80% capacity as the end of life. If most Li Po flyers are as frugal as we are, I doubt that a pack might be discarded until that 60 % is reached.

However, test work does indicate that cell decline goes almost asymptotic when the 80% level is reached. That may be why manufacturers are using that figure. There is good reason: The Failure Mode Effect Analysis for a failing cell is that salts begin to form at pressure points or in areas of heat concentration as the mechanism of failure. Think about it for a moment: As a salt crystal forms, it is a very rough sphere. The surface area is Pi X R cubed. Not quite exponential but close. In other words the crystal that starts forming at 80% capacity grows as the cubed so the cell isn't long for this world. Cell unbalance and safety are affected by this decline. It is smart to keep track of flight time achieved and discard a pack when 80% of the original time is reached.

The above quote pertains directly to LiPoly however the same holds true for all rechargable cells types in general.

 

[zx7dave]

Just about all of my Surefires are on rechargeables...I date the batteries as I buy them from AW, but when I see that a cell is down to about 50% of original runtime I add additional marking to the battery so I know to never use it outside the house...I continue to use it usually till it gets down to about 35% of the original runtime then dispose of it to the proper recycle stream.

 

[cave dave]

The age old standard is when capacity drops to 80% of original.

Right, good info. And I determine that how?

I have not kept track of "flight" time and the lights I had then are not the lights I have now.

I just labeled all my batteries with purchase date. I may have to keep a reference runtime light around?

 

[SilverFox]

Hello Cave dave,

The 80% value is very easy to approximate. When the cell is at 4.00 volts, it has about 80% of its capacity left.

When you finish charging the cells, let them sit for a few hours, then measure the voltage. If it is below 4.00 volts, it's time to recycle the cell.

Tom

 

[cave dave]

Hello Cave dave,

The 80% value is very easy to approximate. When the cell is at 4.00 volts, it has about 80% of its capacity left.

When you finish charging the cells, let them sit for a few hours, then measure the voltage. If it is below 4.00 volts, it's time to recycle the cell.

Tom

I will do that for my flashlight LiIons.

But, what is the danger if it is below 80% (4.00v) and I keep using it. I don't dare measure my cellphone or camera battery, it is for sure below 80%. :green:

 

[Turbo Guy]

See quote in post #4.

It is a simple fact that all rechargeable nbatteries are less safe during charging when they start to deteriate,lose capacity,develop higher internal resistence.

Li Ion and LiPoly are inherently more dangerious than Nicad,NiM and PB.

Higher IR causes increased heating during charging and discharging. Increased heating causes increased gassing and danger of venting.

 

[SilverFox]

Hello Cave dave,

As Charles has indicated, when the internal resistance increases to the point where your voltage after charging is below 4.00 volts, damage has been done to the cell and this causes the cell to heat up during charging.

If it gets hot enough, bad things can happen.

I would be surprised if your cell phone and camera would function if the batteries got that bad. You may get a couple of pictures or a few minutes of talk, but then the device would signal low battery and shut off.

Tom

 

[cave dave]

I would be surprised if your cell phone and camera would function if the batteries got that bad. You may get a couple of pictures or a few minutes of talk, but then the device would signal low battery and shut off.

Tom

Well that's not really 80% then unless we are measuring capacity in different ways. If my camera would take 200 pictures on a fresh battery but now only takes 160 pictures that would be 80%. What am I missing?

 

[Turbo Guy]

If 4.2 volts no load is 100% and 4.0 no load is 80% that is some form of modern math. .8X7.2=3.36.

Problem is that two different things are really be represented here. In the first place most Li cells can not deliver all of their rated capacity in normal useage. A nonimal voltage of 3.7 volts is really mostly empty as in very little useable capacity remaining so you really only have the range from 3.7 to 4.2.

Another problen is that there are many different nonimal voltage for Li cells. LiPoly 3.7 discharged / 4.2 charged. Same for newer Li Ions. Old liIons 3.6 / 4.1. LifePO4 3.3 / 3.6 but useable down to 2.0 if A123 Systems cells. Same A123 System cells useable down to 0.5volts at below 32F. and can be charged to 4.2 below 32F.

As an example Ni cells are nonimal listed as 1.2 volts however they charge to approx. 1.45 and can be discharged down to 0.9V however many devices stop functioning when they reach approx. 1.0-1.1 volts and often only approx. 50 to 60% of their useable(?) capacity has been used.

 

[SilverFox]

Hello Cave dave,

The difference is that in one case we are making an estimate based on open circuit voltage, and in the other case we are looking at performance under load.

With healthy cells, 4.0 volts is around 80% of a full charge, however as the internal resistance goes up, this ratio can change. The idea is to have a rule of thumb to alert us to the fact that the cells are on their way out.

I don't think a cell that has high enough internal resistance that it only charges to 4.0 volts still has 80% of its initial capacity, but I haven't tested that case.

The rule of thumb is based on new, healthy cells. To get actual numbers you would have to do a discharge and compare actual capacities.

Tom

 

[cave dave]

Silverfox, I agree about the runtime data. That's why I included the runtimes in the first post.

The runtimes I have are:
BS1 = 59 min (bought May'07)
BS2 = 49 min (bought Dec '06, end voltage 3.66v)
AW1 = 55 min (bought Dec '07)
AW2 = 63 min (bought Dec '07)

As we can see BS2 lasted 83% of BS1 and BS1 is a couple years old so it probably isn't top capacity either. So by the 80% runtime standard this cell is already toast. If I base current capacity off labled capacity (900mah) it was toast when I bought it

However I get 49min of 120lm light out of it and that ain't too shabby. Also I charged it last night and 21 hrs later it read 4.14v

So, do I keep it or recycle it?

From now on when I get new cells I'm going to do a runtime and mark the date of purchase (wish I had manf date) on the cells.

 

[SilverFox]

Hello Cave dave,

Even though our system is imprecise, it is still a good rule of thumb. If your cell ends up above 4.0 volts after charging, it is not yet "crap."

When checking capacity drop it is best to have a baseline run on the cell when it was initially purchased. This way you can compare the actual capacity with the labeled capacity, as well as having a benchmark to compare later tests to.

Tom

 

[cave dave]

I think this information is all useful. I had no idea that my LiIon chargers wouldn't actually charge to 4.2v if the cell was starting to fade and I've been around the forums a long time and read a lot of LiIon safety threads. I've seen the 80% capacity before but I'm thinking a 4.0v after charging cell wouldn't provide anywhere near 80% runtime would it?

When you think about it LiIon flashlight cells for light just haven't been around that long. I looked up the first FS info on MP and it was back in May 2004.

Rechargable 123's Pre Sale :
These will be the first protected cell 123s in existence

I imagine most people really started using LiIon cells back in 2005 or 2006 at the earliest. So we are just hitting the 3-5 year window when these cells should start dying on us and we should know what to look for. Lots of people including me just hate to throw anything away. Just look at the number of people who try to cycle life back into their dying old Nimh cells. But it sounds like that is exactly the wrong thing to do with LiIon at the end of their life. If LiIon becomes more dangerous at the end of its life cycle and we are just now approaching end of life for the first LiIon batteries bought by CPFers we should see the rate of "incidents" climb dramatically.

I'm on the third test cycle of the BS2 and the capacity has gone up by 3 minutes though, I think I'll keep it around.

 

[VidPro]

if they dont even get TO the charger voltage when charging, and they are allowed to fully charge, then that is BAD. EX: your light on one of them leetel chargers never turns the other color, indicating it reached some voltage.

Note: which has nothing to do with not having fully charged them via the chargers own max voltage, or users manuel termination. i am referring specifically to the battery itself not taking the charge.

li-ion usually "accept" most of the charge your handing them, if it isnt then that is going to come out as waste, and it will be heat, and that will be bad. after all by the spec design, they are not ever "overcharged" to reach max capacity , but limited to a max voltage instead. and they are supposed to be slowly charged at the end, so acceptance without gassing is maintained.

a Wasted one, however it is wasted , will not charge fully (as said) but if you hit it hard enough with a charge it will go up in voltage more , but then it starts heating up, and heating up badly, and i assume would explode if charged too quickly, I never seen a even really bad one cause a problem when charged very very slowly, i have caused a lot of problems for bad really old ones, and they handle a lot of abuse, but you can tell that given enough power (charging) you could cause a thermal situation.

most of the little li-ion chargers themselves cant charge "too Fast" and the primo ones that can charge fast i ASSUME (for the bloody price of them) would realise it wasnt charging, and cool out, or have a thermal probe, timer, or something. depending on the speed of the charge.

Most chargers your li-ion battery shouldnt be getting warm, to the touch , like it will on the end of charge on a ni-mh , if it was getting HOT on charge, then that would be a bad battery for SURE. Quantitive warming could be determined by the user knowing how warm a good one would ever get on thier charger. for most of my stuff that would be, Not ever even noticably warm to the touch. if i felt junior was getting a fever i could probably tell with my hand on its forehead most of the time , me being 98.6*F and them being LESS than that, they are usually cooler then my hand.

also the self discharge is way worse, charge it up, take it off the charger and is visually goes down on the voltmeter, up to that bad even. obviously to test that on all types of chargers, you would rest it for a half hour or so, then see if it keeps tanking in voltage all by itself.

if it ever opens up, Even if it doesn't Properly anode disconnect, a cell might stink, then right after that it dries out quickly and goes to heck fast. while most bats have an anode disconnect, they all didnt, and the stench was one indication. a stinkey one is another warning flag. that stench on an open one, can be smelled from quite a distance , junior is gonna throw a fit if he dont get a diaper change that indicates that it Gassed at some time, could be overcharge, reverse charge , or just to far aged, or damage, or overheat , or the seal just broke because it wasnt perfect.

when your device just sucks in runtime, and you charge battery, and everything reads right when you charge it, but the runtime just sucks anyways, and a new one in same device has a much better runtime. nobody wants a battery with reduced runtime anyways, and chances are the other things are happening to it also.

There is no putting old bad li-ions in lesser devices, if they suck, then they should be recycled, they should not be used again, i cant think of any instances where a person would want to take a cell that was no longer usable for one purpose, and put it to another. if your super duper hotwire failed, then you probably tortured the battery, and putting it in your LED light will work, but charging it your still taking unnessisary risk. the decline from any of the issues, doesnt get better, it only goes down hill.

if when it has been parked for a long time it has no power. if your sure your device doesnt have a parasitic draw, and the device FULLY shut off, but the li-ion you had in there doesnt have any power in it, and its been less than 6mo, then suspect something, and check something. we have lights here that fully shut off, and 6MONTHs later, you wouldnt know it wasnt just charged yesterday. so self-discharge again. and if your device is discharging them, then mabey you need a lockout, or to pay more attention to that item, because discharged all the way down for a long time , they dont like that.

most of this info i found out a Long time ago, by using "used" batteries, not via nice good new ones. most of the new ones are lasting longer than it would have been expected. so there you go, no meter, there are still some indicators for you.

 

[cave dave]

Even though our system is imprecise, it is still a good rule of thumb. If your cell ends up above 4.0 volts after charging, it is not yet "crap."...

I have some additional data for some other cells I have that may disagree with the 4.0v rule. I dug some old blue label (BL) CR123's out of cold storage (3.8v in the fridge) let them warm up and did couple tests.

Runtimes on 120lm setting w/ 170lm burst on for Ra 170cn

AW2 Dec'07 recap for comparison:
Runtime = 63 min

BL1, purchased Spring 2006:
4.11v off charger, dropping to 4.10v after ~1hr:
Runtime = less than 1 min.
I then changed to 42lm setting and got 36min on that.

BL2, purchased Spring 2006:
4.15v dropping to 4.12v after 2 hr.
Runtime = 23 min.

BL3, purchased Dec '05
4.07v dropping to 4.02v
I'm going to turn burst off and just try 120lm. (The 10second, 1amp draw may be torture for these old cells)
Runtime = 14 min. Then 83min on 42lm setting. I think turning burst off made a difference. Otherwise it would drop out of max in less than 1 min.


Also to note is the behavior vs. a good cell on the Ra 170. With a good cell it will run about an hour, then after the first drop, it will do several more drop until it reaches min in less than 15 min. With a bad cell, it quickly drops out of max, but the subsequent drops last a lot longer. I gave up after an hour as I couldn't stand the double blinking every min, and didn't want the cell to go in my precious.

 

[Turbo Guy]

I've been around the forums a long time and read a lot of LiIon safety threads. I've seen the 80% capacity before but I'm thinking a 4.0v after charging cell wouldn't provide anywhere near 80% runtime would it?

The lower the discharge rate the less a cell will recover when the load is removed. Here is a graph which shows a LiPoly discharged in 20% steps . It was allowed to recoved 2 min. at the end of each discharge step and the resting voltage was recorder. This was a very high C rate battery so held voltage under load much better than most.

http://www.rcgroups.com/forums/showpost.php?p=10981333&postcount=1

One thing that many over look is that at 3.7 volts under a moderate load there is no real useable capacity left and if you discharge below this point with a light load the cells take a beating or at least LiPolys cells do. I am still learning about Li ion and esopecially their use in flashlights.

 

[mdocod]

I figured I'd share some useful tidbits here...

The WF-139, when charging a protected RCR123, will *probably* never achieve a full 4.20V charge because of the way in which it charges. The WF-139 uses what is essentially a constant current charge through the duration of the charge. The charge voltage actually rises above 4.2V by a fair bit as this is happening. The charger cycles the charging on and off every second or so, with each moment with the charging "off" it takes a resting cell voltage reading. When that reading is 4.20V, the charger terminates. The problem is that on smaller cells, with higher cell resistance, this charging routine results in a high differential between charging and resting voltage. So large in fact, that it winds up tripping the PCB on the cell before the cell ever reaches a 4.20V resting voltage. (Yes, it reaches 4.35V charging voltage when charging RCR123s).

The Pila IBC was originally developed for use with larger format cells. It uses a more true Constant Current followed by Constant Voltage charge method. When designing a charger with this type of proper charging algorithm, the reality is that the final termination will actually only be perfect for a cell of a particular capacity/resistance, and all other cells will have minor "imperfections" in the level of charge they will take on that charger. When an 18500 or 18650 is inserted into an IBC, the charge is 600mA for the majority of the charge time, then towards the end, when the cell reaches 4.20V, the charger holds the voltage there until, theoretically, the current drops to ~0.1C for the cell size/capacity it was intended to be used with. So I would guess that the final current termination point is probably set somewhere between 140-220mA. For those larger format 18500 and 18650 cells, finishing off the charge at 0.1C rate with 4.20V CV results in cells with a resting voltage very close to 4.20V after termination. When the same algorithm is used on an RCR123, the final charge current before termination is actually closer to ~0.3C, which isn't slow enough to really top off the cell in the CV stage.