Can you run CR123s in parallel?

I need to know the best way of dealing with excessive current drawn through CR123s.

I am experimenting with 20W and 50W 12V halogen bulbs and a stack of CR123s. I am overdriving the bulbs to 15V with one extra CR123 cell. The bulbs seem to handle it without complaint. The cells cannot handle the current, though.

Tonight I ran 5 CR123s in series through a 50W bulb. The cell temperature climbed steadily to 100°C+ over the course of 10 mins, before one popped, spewing its grey molten guts in a fine spray. Up until that point, the head cast a beautiful uniform intense yellow-white light with no hot-spot. In terms of lighting and longevity, it's all I need, but I don't like my cells exploding at random.

I then ran a duplicate cell pack through the 20W bulb, and they seemed to handle that fine. The temp peaked at 59°C and the lux dipped after about half an hour. That'll do for me in terms of longevity, but the light output is a bit weak.

What I learned was that I need the 50W bulb, but with the (per cell) current flow of the 20W bulb.

For my next experiment, I plan to run two 15V chains of CR123s and wire them in parallel. Now, if the cells aren't uniformly charged, the circuit could make one stack charge the other with unhappy results.

As a safety, I could put in two protection diodes. Do I have to? Is it safe in any case?

Advice appreciated,

Dan

If you can wire the switch so that the two stacks are not put into parallel until your load is applied, you should have no problem - assuming the cells are well balanced to begin with. Even in that scenario, a weak cell somewhere int the stack of 5 could get seriously reverse charged, with ? results.

I wonder how SF handles this in "The Beast"?

The reverse protection diodes are a really good idea. I did this on my side by side 123 setup - similar to where you are heading.

Series strings of five CR123As don't sound like such a good idea on their own, much less in parallel. With this many CR123As together, one weak cell will result in reverse charging and a 'vent with flame' incident.

Have you considered an AGM lead acid battery or a protected li-ion pack for this application? Both options would be safer than an armload of CR123As--and cheaper to operate to boot.

Just some observations. The SF M6 casette has two seriesed-strings of 3 cells in parallel. This is full-time parallel whether on, off, or locked-out. I see no evidence of any diodes with my meter.

My beast has four seriesed-strings of 5 cells in parallel. If I'm reading the meter right, (tail cap off, Ohms between the negative ends of the strings) It is also "full time," with no diodes evident. /ubbthreads/images/graemlins/icon3.gif

Larry

Hi tvodrd, I'm still around but just quieter now.



Higher voltage cells give less room for error. If the cell are the same capacity (you should test them), you don't overtax them and be nice to them they will work fine.

One weak cell in the bunch is not noticeable and will be "pulled up" in voltage by the stronger cells but that one cell will always be weaker than the rest of the cells with just a loss of run time. When you drain the cells to the limit the weak cell will die first. As you continue to drain the cells the weak cell will reverse charge and really be ruined. In your case it hissed and fizzed slime.

You are taking your chances with multiple higher voltage cells. And the more cells you use and how far you drain the cells gives higher risks.

I am just a newb, and I am not argueing... I just don't understand.

How then, can the Beast, run four strings of five cells all in parallel... and not blow them up?

Thank you

SS

[ QUOTE ]
ZDP189 said:
I need to know the best way of dealing with excessive current drawn through CR123s.

I am experimenting with 20W and 50W 12V halogen bulbs and a stack of CR123s. I am overdriving the bulbs to 15V with one extra CR123 cell. The bulbs seem to handle it without complaint. The cells cannot handle the current, though.

[/ QUOTE ]

I think your problem was too simply too much current.

Energizer and GE-Sanyo rate their CR123A's at a maximum of 1500maA continuous discharge. Duracell's site is a flash nightmare, but I imagine they're not too far from 1500mA continuous discharge.

50W / 15V == 3333mA - no wonder you lost a cell in one chain... and you're overdriving that bulb at 15V, so I imagine your current was signifigantly more that.

If you assume a constant resistance in the filament (usually an error), I calculate that the filament is 2.88 ohms ... 15V / 2.88 ohms == ~5200mA. Using the same (dumb) assumption, I get 7.22 ohms and ~2000mA through your 20W rig.

3500mA is an acceptable pulse for Energizer and Sanyo, but I'd keep pulses under 15s.

[ QUOTE ]
For my next experiment, I plan to run two 15V chains of CR123s and wire them in parallel. Now, if the cells aren't uniformly charged, the circuit could make one stack charge the other with unhappy results.

[/ QUOTE ]You're going to be past the margins on that with the 50W bulb. I'm sure that most cells will be able to supply a bit more than 1500mA continuous, but what sort of (catastrophic) failure rate is acceptable to you?

/ubbthreads/images/graemlins/thumbsup.gif Idle! (Off to post the latest USL beamshots. /ubbthreads/images/graemlins/jpshakehead.gif )

Larry

123's will not comfortably handle more than 1.5 amps of draw and 2 amps is really pushing it!

3.3 is out of the question. Parallel strings will come close to 3.3.

Surefire does not seem to worry about the problem in their cartridges, but if I recall correctly, they also specify that all cells should be replaced at the same time.

Bill

why would you use 5 123s.. that sounds really expensive.. get some higher voltage 18500s or 18650s.. leeme see.. 5 x 3v = 15V or 4 x 3.7 = 14.8V.. so maybe 4up LiON.. with a charging /discharging protection circuit? can easily handle the power.. probably even just one string.. pretty sure an 18650 can handle 5W

-awr