Li-Ions and over-heating

[Braka]

I just thought I'd mention this and see what folks think.

Yesterday I was doing a runtime test on a flashlight which gets rather absurdly hot, and I remember noting, "After 9 minutes I could hardly pick it up". Whereas I don't have anything else this bad, it's certainly not uncommon for (especially cheaper cr123 lights) to get got enough to be uncomfortable to hold after a while. At this point anyway, I switched it off to cool down, and took the batteries out. They were warm - almost hot. I'd guess 60+ celsius, and I suddenly found myself thinking, "do I really want to be sitting here with a couple of batteries in my hand which are known being unstable if they get too hot?"

Today I was calling a battery store here in town, on another matter, and the guy seemed to have made it his mission to scare the hell out of my about Li-Ions. He said that they had one, which he described as 'about half the size of an AA' explode in their recharging room. He said this room was specially designed to wityhstand such accidents, but the explosion basically blew the room apart. At the time he was talking about this, I was holding the phone in one hand, and two 18650's in the other. I suddenly became aware of the fact that the room I was in was filled with torches which, the way this guy was talking, were potential hand grenades.

If I'd not been so consumed with this thought, I might have thought to ask him how come his store SELLS the things, but anyway.

Apart from the warnings of dropping, kicking, or looking at Li-Ions the wrong way, a more reasonable question seems to be - are Li-Ions really unsafe in a torch that gets really damn hot? I mean, if you're not supposed to leave them in the car on a hot day, I'm pretty sure you wouldn't be meant to walk around holding a couple of them in your hand inside an uliminium tube which you've had to switch off because it's starting to burn your fingers.

I'm not talking about unmonitored, overdischarged, unprotected cells catching fire in the recharger. That's a pretty widely accepted risk factor. I'm talking about putting the things inside devices which generate a tremendous amount of heat, and then conduct it to the interior of the casing.

I'd be more than happy to be calmed down after this guy's panic-mongering (which I expect to get more of tomorrow), but also interested in others' opinions and/or experiences.

Braka


EDIT: BTW after talking to the store manager, it turns out the reason the Li-Ion exploded on the charger is because the guy who I'd been speaking to before put it on the wrong charger in thje first place. Exactly what that constitutes - whether it was just the wrong Li-Ion charger, or whether he tried to put it on a different family of charger altogether, I don't know. In any case the manager seemed considerably less worried about cells exploding than his employee is.

 

[LightBright]

There are several threads in the "Batteries Included" area discussing this very subject, so browse around and do some reading. I have been using bare (no protection circuits) Lithium Ion cells for a few years now and I have not had any problems whatsoever.

I handle them carefully and I do not over charge them, drain them too far (less than 3.0V), let them run at high temps (over 120F) or pull more current out of them than specified (each cell can have different ratings). I also do not let anybody else around the house handle them unless they know what they're getting into.

 

[SilverFox]

Hello Braka,

Welcome to CPF.

60 C is considered the maximum temperature for "safe" Li-Ion cell use.

Runtime tests are hard because the light is generally just sitting there. In actual use, your hand may draw a lot of the heat away. If I find a light getting hot during a runtime test, I will run some water over it to cool it off, or stand it up in a glass of water. Some people position a fan to blow over it during the test.

I believe you need another 20 C before you get to the ignition point of the chemistry, but above 60 C you do some damage to it resulting in reduced capacity.

Tom

 

[Braka]

Thanks for the responses and welcome. I think part of the problem is it comes down to guess work a lot of the time. It's not so bad if you're just working around the house, but if you're out walking at night, how do you know when an unprotected cell has dropped below a specific voltage unless you're carrying a multimeter (which you wouldn't be able to read anyway)? 'When the light dims' is a rule of thumb, but not a precise one. I suppose you could check the cells when you got back again before trying to charge them, but that doesn't help with the original problem. Similarly, how do you know exactly when a flashlight is so hot as to have damaged the cells? There are plenty of lights out there which generate enormous amounts of heat, and dispose of it partly by conducting it back along the body of the light. I think that 'if it's getting uncomfortable to hold' is a good rule of thumb too, but, again not a particularly precise one.

Braka

 

[LightBright]

If the flashlight doesn't have electronics to monitor the cell's temp and voltage, it's easiest to use a "protected" 18650 battery. It's a good thing that those are available now. Of course the protected cells won't give you much warning before they turn off, and also they seem to have different cutoff voltages depending on the brand you buy. That's why I'm designing my own board to address some of these issues.

 

[SilverFox]

Hello Braka,

Most lights will run for roughly an hour. It is up to you to check this out for your particular light. You can start by using it for half an hour and then check to see where you are. If you are going for a long walk, remember to bring a spare set of batteries to change when your light dims. If it is really dark, bring along a small light to see with while changing your batteries.

To get an idea of heat, bake yourself a potato. When you pull it out of the oven, toss it back and forth in your hand. This is an example of too hot.

Tom

 

[LED61]

Braka, it would really help if you can specify what light you used these li ions on, what specific li-ions, and how many were there in series. That can tell you if the heat came from the flashlight or the cells. 60C is pretty hot.

 

[Braka]

Hello Braka,

Hi

Most lights will run for roughly an hour. It is up to you to check this out for your particular light.

Wow. OTTOMH I can only think of one Li-Ion run light I have which definately runs for an hour or more. 30 minutes would be closer for a lot of them (well, so I'm told. I have a tendency to recharge cells well before they get close to their supposed curl-up-and-die point)

You can start by using it for half an hour and then check to see where you are. If you are going for a long walk, remember to bring a spare set of batteries to change when your light dims. If it is really dark, bring along a small light to see with while changing your batteries.

Yep, always take spare batts, and almost inevitably there will be a little emergency light on me somewhere. Quite often I'll actually have two or three on me, because I'm testing out some new light against an old 'benchmark' one.

To get an idea of heat, bake yourself a potato. When you pull it out of the oven, toss it back and forth in your hand. This is an example of too hot.

LOL! Well, let's see.... I'd probably set the oven to at least 180 degrees celsius (fan-forced) to bake a potato, which is about twice the boiling point of water. If you don't mind, I'm not going to be taking them out of the oven with my bare hands, to get a tactile idea of what you mean

Braka

 

[Braka]

Hi.

If the flashlight doesn't have electronics to monitor the cell's temp and voltage, it's easiest to use a "protected" 18650 battery.

Yeah, I know. Just that I only have one light which uses that size cell, and at the moment all I know is that it won't work. I'm taking it to the battery store tomorrow, who claim they can diagnose what the problem is. By my crude experiments the batteries seem charged up, and I can't find anything obviously wrong with the light either. If it turns out to be the batteries I'll fladly order some protected ones.

It's a good thing that those are available now. Of course the protected cells won't give you much warning before they turn off

Right, but I'd actually prefer that to guesswork. Like I said, there's bound to be another light on my person somewhere anyway.

Braka

 

[Braka]

Braka, it would really help if you can specify what light you used these li ions on, what specific li-ions, and how many were there in series. That can tell you if the heat came from the flashlight or the cells. 60C is pretty hot.

Sure. If you're talking about the one I was trying to a runtime on, it was just a cheap incan from 5thunit (I keep forgetting what they're called now) called the Hemse. According to them it puts out something absurd like 280 lumens from two 3.6v 123's (though it works just as well on 3.0's IMO). Clearly it's nowhere near that, and I didn't expect it to be. I'd guess it was closer to the low to mid 100's, but it's hard to tell, because it's as floody as hell, and floody lights tend to seem deceptively dim (at least to me), compared with something which stabs out a blinding pencil thin beam.

Whatever the case, there is certaihly a lot of heat/ebergy emerging from the front of that light, as if you hold it cupped in your hands for bout 20 seconds it starts to feel like you're holding a magnifying glass up to your hand in the midday syb. I've got a light in 4-figure lumens that doesn't feel like that.

The main problem with the 'Hemse' is that it's a cheap design and has no real heat dissipation, so the whole damn thing just heats up. The body gets nearly as hot as the head.


That's what I was doing the runtime test on, if that's what you meant. I wanted to see if the figures were as bad as I'd seen quoted (15-20 mins)


Braka


(as usual, please excuse typos)

EDIT: sorry, you asked which specific Li-Ions. I don't remember. They were probably either Ultrafire (either 3.6 or 3.0)'s, since I'd just received them, or Powerplus 3.0's, which is what I normally use. Probably only the major difference (pther than the Ultrafires being less reliable) is that I think they're 800mAh, whereas the PP's are 700.

 

[LED61]

OK, if two R123´s are being discharged completely in15-20 minutes that´s over the limit of safe discharge for Li ions. You have the incan producing heat as well as the batteries for too high a discharge rate. I would stop doing that. You don´t want to charge or discharge the Li ions too fast.

 

[Braka]

OK, if two R123´s are being discharged completely in15-20 minutes that´s over the limit of safe discharge for Li ions. You have the incan producing heat as well as the batteries for too high a discharge rate. I would stop doing that. You don´t want to charge or discharge the Li ions too fast.

Why?

Sorry - I don't mean to come across in the least bit confrontational, and I apologise if it seems that way. I'm obviously ignorant about this, and am genuinely interested. I realise of course that you don't want to discharge the cells below a certain point, but I wasn't aware that it was a problem to discharge them to a safe level in say 15 minutes rather than 30. In another thread today someone mentioned that their Suefire only had a 20 minute battery life (he was probably talking about throw-aways, though)

Most probably you're completely right. Would you mind just explaining the reason? I'd appreciate it.

Thanks for your help


EDIT: 0h, sorry - I forgot to include this. When you say it's beyond the safe limit, do you mean 'unsafe' as in shortening the life of the cell, or unsafe as in the thing catching fire during recharge? (I realise the two concepts are related to a degree)

Braka

 

[LED61]

the term "C" is used to define the battery´s capacity in Amp Hours. A Lithium Ion 123 will typically have about 700 mah stored capacity, or .7 amp-hours. The maximum safe discharge rate for Li ion 123´s is 2 times their capacity, or 1.4 amps. Assuming the correct math, and everything according to numbers, you could theoretically assume a discharge rate of 2C is 30 minutes. But, as additional burden and or heat is placed on the cells you could still be drawing 1.4 amps and the batteries would discharge in less time, so it is better to use the time it takes the cells to discharge as the guideline for C rate discharge. Your light´s current demand is apparently too much for the Li ions.

I recommend you do some research here on the forum using the search function to be able to use those Li ions safely.

 

[LightBright]

When you say it's beyond the safe limit, do you mean 'unsafe' as in shortening the life of the cell, or unsafe as in the thing catching fire during recharge?
Braka

YES!

 

[FlashCrazy]

I have the Hemse as well and did a runtime test on it yesteday....ran it for 20 minutes and then let it cool. The body got to 112 deg F...pretty warm, but not really hot. I did test the current being pulled from the batteries when I first put them in, the light pulled 1500 mAh. I'm using Dealextreme's 800 mAh 3.6V unprotected cells. This puts it just under a 2C discharge rate.

I have the light next to me now with freshly charged batts...I'll test it again...

 

[FlashCrazy]

Just did another test of the Hemse. Yep...it gets pretty hot...almost too hot to hold if you're touching the head, not too hot to hold when holding by the body --but pretty darn hot. Here's the results:

Runtime and temp in deg F

5 min 100 body / 104 head

10 min 108/113

18 min 117/122

20 min 120/125

Light still very strong at this point, but I stopped the test to be safe. Weird chemical/burning smell when removing batteries. Battery temp 119 deg F. I noticed the same smell when doing a runtime test of my Glimt model LED light from Dealextreme. It pulls 2.0 amps on high....the body of it was also 120 degrees after 20 minutes. I thought it was the driver board really heating up and releasing the smell...like maybe the varnish on the PC board. Now I'm wondering if it's from the batteries. Anyone ever smell something like this?

I'd say the Hemse is fine for a few minutes use at a time...but not for 15-20 minutes continuous. I REALLY love the light though!!

 

[VidPro]

if you cant hold it in your hand, then it cant be good for batteries that are designed for "room temperature" operations.
mabey you need a magma-zinc-trioxide battery instead

 

[LED61]

I have the Hemse as well and did a runtime test on it yesteday....ran it for 20 minutes and then let it cool. The body got to 112 deg F...pretty warm, but not really hot. I did test the current being pulled from the batteries when I first put them in, the light pulled 1500 mAh. I'm using Dealextreme's 800 mAh 3.6V unprotected cells. This puts it just under a 2C discharge rate.

I have the light next to me now with freshly charged batts...I'll test it again...

If you are discharging the batteries in 20 minutes your real discharge rate is 3C. Don´t use the current to determine C rate. Heavier loads increase internal resistance in the cells making it harder for them to deliver the power.

 

[FlashCrazy]

If you are discharging the batteries in 20 minutes your real discharge rate is 3C. Don´t use the current to determine C rate. Heavier loads increase internal resistance in the cells making it harder for them to deliver the power.

Oops, forgot to finish that story...I let it run for 20 minutes then let it cool...that wasn't the end of the test, the light was still strong. I let it run another 10 minutes, then 5 minutes after that. Battery voltage was 3.0 volts upon checking them immediately after taking out. I don't like to run the lithiums this low normally, because under load the voltage was much less I'm sure. With my Li-Poly 11.v pack I use with my radio-controlled airplanes, the speed controller is set to shut off the motor when the pack reaches 9.0v under load. When I check the voltage of the battery right after a flight (no load), it shows 11.1 vots which means 3.7v per cell. Granted, the motor pulls 12 amps, but these cells are 2000 mAh and rated for a 10C load. I figure that with my 123 batts that I should only run them down to a resting voltage of 3.6v or so. Any ideas?

 

[AndyTiedye]

Is the Hemse even designed for 3.7v li ions?
If not, everything could get really hot (or worse) due to overvoltage.

How does that light do on a single 17670?