123 Primary Lithium cell info/testing/links

[Chronos]

OMG That was scary! 711 degrees!?! That's hot enough to ignite anything in close vicinity to the cells; thinking the rubber tailcaps here. This could ignite any rubber or plastic parts in the light, causing a fire.

I'm probably just restating the obvious here. But still, yikes.


Something that has me worrrying... the ADT wireless alarm setup in my home uses CR123 cells for many (if not all) of the wireless fire alarms (and I think other monitors too). I haven't seen a test (that I remember) demonstrating a single-cell vent- is this possible? I have to believe these devices are low-draw, but I have already "burned" (sorry-bad pun) through a couple of 123 cells on my fire alarms. What a terrible irony that would be...

 

[SilverFox]

In an effort to expand the scale of Newbie's testing, I will be sending him some mismatched cells.

Newbie has shown that if you partially drain a cell and couple it in series with a new cell, venting can occur. We have been using the ZTS tester to match cells, but the question keeps coming up "Will ZTS mismatched new cells vent in the same way as when you physically mismatch cells?"

I will be sending Newbie cells that tested at 60%, 40%, 20%, 10%, and one cell that came in at 0%. Along with these cells will be those that tested at 100% to pair up with the mismatched cells.

All of these cells are brand new and have never been used. They are Titanium cells that came from the stock that Wayne pulled off the shelves when he discovered that he had a bad batch of cells back in April. I did some testing on the cells demonstrating what parameters the ZTS tester was looking at, and decided that they may be able to assist us in understanding what is going on with the venting of cells.

Wayne took some cells from this batch back to the factory, along with a ZTS tester and a SureFire M6, to demonstrate the problem to them. As I recall, the result was that while the factory does a test similar to the ZTS test, they were using different resistor values. I believe these values were adjusted to subject the cells to a higher load. I have heard that Wayne has been continuing to do random checks of the new cells he has been receiving, including ZTS testing, and has been pleased with the results.

Since this batch of cells has some known issues, it is my hope that we can get some meaningful information from them.

Newbie asked me if the testing of these cells could be delayed until he got his data logging device functional. I left the decision up to him. We are all interested in the ability of the ZTS tester to add to the safety of using multi-celled lights, but we also have to keep in mind that this is a destructive test with a limited amount of samples. Also, keep in mind that no testing like this has been done, so we don't know how the results will play out.

It is my hope that this will not be a total waste of Newbie's time and efforts...

Tom

 

[nakahoshi]

what a violent reaction, that is HOT, I cant believe that little cell is possible to do something so powerful. Imagine if 2 or 3 cells did that in a light? I have 2 cells sitting here in front of me, and im thinking they are going to disapear as soon as possible. Doesnt flashlight reviews recomend using Titanium cells??? Doesnt seam like a good idea any more... Good work newbie! Im glad that diddnt happen inside of my HD45!!! I used to keep my stock of cells next to my bed near my lights...

-bobby

 

[NewBie]

what a violent reaction, that is HOT, I cant believe that little cell is possible to do something so powerful. Imagine if 2 or 3 cells did that in a light? I have 2 cells sitting here in front of me, and im thinking they are going to disapear as soon as possible. Doesnt flashlight reviews recomend using Titanium cells??? Doesnt seam like a good idea any more... Good work newbie! Im glad that diddnt happen inside of my HD45!!! I used to keep my stock of cells next to my bed near my lights...

-bobby

Keep in mind, the flames and sparks are even hotter. The white color sparks you see in some of the videos are lithium on fire, which is about 3000F.

The container has a limited amount of oxygen in it, so it serves to extinguish most the events.

 

[Archangel]

Would an event in a waterproof torch be much less then since there's correspondingly less oxygen?

 

[NewBie]

Would an event in a waterproof torch be much less then since there's correspondingly less oxygen?

Doubtful. Even a Surefire 9P, the big heavy duty robust light that it is, had it's side ripped open by one of these events.

It takes alot to contain one of these 123 cells...

 

[Lost_In_Beijing]

NewBie,

Just curious if you are keeping track of the lot numbers on the batteries. I'm wondering if maybe the difference in chemistry between batches may be part of the problem (and if so, why are they so sensitive).

Also, just saw that Dell is recalling 4.1 million laptops with lithium-ion batteries made by Sony. They have been overheating and catching fire. It covers laptops sold from April 2004 through July 2006.

Weldon

 

[jsr]

Newbie - after purposely draining some cells to create a mismatch, are you using a ZTS tester to check the remaining capacity? If so, wouldn't that be a very similar, if not identical, scenario to testing unused cells that show a mismatch on the ZTS? Of course, testing both cases is best, but just to try to get more information on the current test results, this would be good to know. Thanks for all the work!

 

[SilverFox]

Hello JSR,

Battery Station, Titanium, and several other CR123 cells have a tendency to initially drop in voltage, then pick back up when the chemistry becomes activated. Testing those cells reveals that there is some difference in capacity, but not the 30-40% that is indicated by initial ZTS testing.

Once you use the cells, the ZTS results are more accurate.

When Newbie tests these mismatched new cells, we will have another data point on this problem.

Tom

 

[HarryN]

Nice work Jar - That is going to be one charred bucket. :laughing:

The secondary venting is interesting. I have seen that on some other reactions before (not cells) but it took more like 30 minutes for it to happen. Interestingly, sometimes, the secondary reaction is worse than the first.

 

[NewBie]

Hello JSR,

Battery Station, Titanium, and several other CR123 cells have a tendency to initially drop in voltage, then pick back up when the chemistry becomes activated. Testing those cells reveals that there is some difference in capacity, but not the 30-40% that is indicated by initial ZTS testing.

Once you use the cells, the ZTS results are more accurate.

When Newbie tests these mismatched new cells, we will have another data point on this problem.

Tom

Quickbeam did some testing on this:
https://www.candlepowerforums.com/threads/123801&highlight=zts

 

[SilverFox]

Hello Newbie,

I saw that, and went to Doug's review site to see what he had done. He did one sample of mixing a 100% with a 20% and shut things down when the light went out. I believe you have demonstrated that you need to keep the circuit connected for a while after the lamp goes out to get the venting to occur.

It is interesting to note that run times with the mismatched cells was similar to using cells that tested at 100%, however the performance was not the same.

Tom

 

[mdocod]

not sure if it has been mentioned already... But I think it is important to note... and emphasize, that before that tailcap nearly went through an oak cabinet door- it went through several layers of plastic (plastic bins in the cabinet)... So had the inicial force, been focused only on the door, It probably WOULD HAVE gone right through it... The power behind that tailcap could be estimated in the ballpark of small caliber weapon fire.

 

[NewBie]

Update:

I got the ZTS tested Titanium cells from Tom- Thanks!
I have another batch of SF and Titanium cells from McGizmo- Thanks!
I have another batch of BS cells from CY- Thanks!
I have another batch of BS cells, newer date code, from Kevin at BS.
I ordered and received in Ultralast cells and Samsung cells.
I'm waiting for my order of Tenergy and ?Powerizer? cells from Battery Junction.

I finished getting the uC programmed for the datalogger, and it is reporting all the ADC channels nicely.

I started building the signal conditioning circuits.

I'm short one ice point temperature reference for the K-Type thermocouple conditioning circuit. The part number I'm waiting for is LTC1025 from Linear Tech.

Almost in business!

 

[SilverFox]

Hello NewBie,

I hope you are on good terms with your garbage collector... A sudden increase in dead lithium cells may raise some questions.

This should be a very informative run of testing. Hopefully it will provide us with some answers and not add to the questions we already have.

Thanks for taking the time and effort to do this.

Tom

 

[NewBie]

Update:

I have three channels of the signal conditioning circuitry built up, and have connected and calibrated two of the voltage channels. The third channel is for the current measurement signal conditioner circuit.

I still don't have my ice point reference parts.

I may go ahead and run one set of cells to see what data I can gather for now, once I get the current channel going and calibrated.


I decided to switch to coding in C, as it is *much* easier, than dealing with indirect registers, memory management and various other items, in assembler. It also makes code changes pretty simple. A sample of one of my earlier screen captures:

 

[NewBie]

Okay, well, I'd gotten the signal conditioning circuits all wired up for both battery voltages, and for the current. I'm still waiting for the parts to build the temperature monitoring signal conditioning circuits. Anyhow, I put it all together and did a run, logging the data.

For starters, I decided to start with a particularly discharged cell, 45% discharged, 55% remaining, and figured I'd work my way up until I hit the discharges that would cause venting. As it turned out, even this combination in the Amondotech Titanium caused a venting incident. I will post the video tomorrow.

This incident was one of those where the cell vented before the bulb went completely out.

As I had long suspected, the venting cell was being reverse charged in this scenario.

Here is the plot of the data:

It will be very interesting, to see how these plots vary with the venting or not, the severity of the vent, as well as temps. I'm hoping to get my ice point chips soon.

So far, it has been a good exercise, I got my feet wet with both Assembler and C coding.
-----------------

Venting video that goes with this dataset is found here, it is also a very unique and interesting video:
http://www.molalla.net/~leeper/amondo10.wmv

 

[McGizmo]

Real interesting and informative system you have developed!! :thumbsup:

Seeing the relatively straight line of the current with the voltage of the one cell dropping is quite illuminating. It will be ineresting to see if you have a similar plot with a constant current driver instead of a hotwire.

Facinating! Thanks!!!

 

[Ben H]

Awesome info Newbie! Thanks for sharing all your effort, time, and money.

 

[SilverFox]

Hello NewBie,

Looking good...

It is interesting that the voltage of the 45% cell did not hold up well at all.

Tom