Hello Arney,

I also go by SilverFox on CandlePowerForums. I am heavily involved in battery testing. We have had several reports of the CR123A Lithium primary cells "rapidly venting with flame." In an effort to understand why, we set out developing a test plan and conducting a series of tests. NewBie (on CPF) did the actual testing and we were able to come up with a recipe to cause this "venting with flame" to occur.

The formula we discovered goes like this:

You need an incandescent lamp that draws about 1 C (or around 1.2 - 1.5 amps).
You need one cell at 100% capacity, and the other at around 70 - 80% capacity.
You need an ambient temperature of around 80 F.
You need an enclosure of some type. (More on this later).
You need to run the light completely down, then you need to leave the circuit connected.

Rapid venting with flame occurs 5 - 20 minutes after the lamp goes out. By the way, the temperatures exceed 1000 F for the short time the cell is venting.

This works almost every time with the inexpensive CR123A cells. It never works with the premium cells such as SureFire, Streamlight, Energizer, Duracell, and Sanyo. We haven't tested all of the brands, but we are seeing a trend here.

The problem is that the stronger cell tries to charge the weaker cell. With some temperature, the reaction is able to continue and after "cooking" for a few minutes... bang.

Cells in an open environment will not cook off. You need an enclosure to maintain the heat within the cells. We started out with a flashlight body, then progressed to simply wrapping the cells in some aluminum foil to simulate the flashlight body. This worked quite well.

It would be my "guess" that the pilot, upon discovering that his light had gone out, had pressed the switch several times and may have lost track if it was turned on or not. I believe he ended up with the light in the on position and that allowed one cell to reverse charge the other.

We have discovered that the non-premium cells often are not stable. I have been unable to determine why, but think it has something to do with the quality of raw materials, and perhaps, something to do with the sealing method of the cells. I have cells that tested (on a ZTS tester from www.ztsinc.com ) at 100% initially in April 2006, that are now testing 10% as of November 2006. While I have only observed this with 4 different brands of cells (Radio Shack is not one of them) I don't believe it to be unique. It is quite possible to end up with cells of different states of charge when you purchase them. If all of the other conditions are met, you can end up with cells that "rapidly vent with flames."

This does not seem to be the case with LED lights. We have not been able to get an LED light to vent or flame under the same conditions. I believe the LED just breaks the circuit when the voltage drops below a certain point.

To protect yourself against this, I recommend using premium cells in multi cell applications. If you want to try the "cheaper" cells, it is a good idea to load test them (ZTS tester or simulate a load with a resister and measure the voltage under load and compare it to a premium cell under the same conditions), then REMOVE them from the light when the light dims or goes out. If the circuit is broken, the reaction stops.

To summarize, we need non premium cells, a load that is high enough to heat up the chemistry within the cells, ambient temperatures high enough to sustain the chemical reaction, and a connection that is not disconnected, and of course 5 - 20 minutes for the reaction to complete.

To protect against this, when the light goes dim, remove the cells from the light, and use premium cells in multi cell applications.

Here is NewBie's test thread on CPF.

Tom