sensible way to drive an XM-L w/ 2x CR123

the reason for my question:

There are a couple of folks who I’d love to give modified SK-98s as presents. The crux is I don’t trust them with rechargeable Li-Ions, as they have worked with alkalines for their past 40 years or so and don’t know anything about battery-charging safety precautions at all. Donating them recs and a reliably safe charger as well would surely overrun the budget, and I still would’t trust them with the usual safety precautions we all care about here, like never leaving the cells unobserved whilst being charged.

So, I thought about the ubiquitous (and not too terribly expensive when you know where to buy them) CR-123A primary lithium cells.

1st observation: they do fit (d’uh, who’d have thought! A given no-brainer…)

2nd observation: the SK-98 actually does work with 2x3V primaries - at least for a second. I did not leave it on longer, as I was not in the mood of deliberately frying a driver (and the SK-98’s one is quite decent despite it’s 1.0A reading, except for the silly strobe)

Is there a driver which would work from, let’s say, 3-6volts, and deliver around 2500-2800 mA in high mode? I’d take the fabled NANJG-105C, but it is rated to like 5V input max. Apparently, some guys here have powered 105C-equipped lights with 2xCR123s, and found out they would hold up to the voltage, but run terribly inefficient (or something like that, my memory tends to be fuzzy at times).

If not, could anyone here make (build, fabricate) such a driver? Runtime is not an issue. The people I have in mind do use their lights for short “bursts”, mostly. Maybe five minutes switched on constantly, at the most. A low mode would be nice, but it surely is not mandatory.

Simon

G4 MCU group buy light runs 2*123 - not a zoomie but if you’re willing to mod, I’m thinking an XPG R5 would be a pretty nice fixed reflector light and the starting price is right while the code works.

thanks a lot for your input, Horst and GottaZoom.

Still, I am looking for a way to run a single XM-L around 2800-ish mA on high (and probably one or more less aggressive lower modes), driven by 2x 3V CR-123-A. Heatsinking is a problem I can take care of quite well, to a certain extent.

Generally, matching a given battery voltage to a driver’s input voltage efficiently is still really, really difficult for me.

A simple master-slave or flip-flop circuit was about the best I could (barely!) master in physics, and that has been 25 years ago.

You might ask E1320 or others here if you could add 7135s to the MCU driver to raise the power draw like Erik has done with the 2100s in his group buy thread. Key thing is this driver handles the higher volts.

Bear in mind that power/watts to the led could increase proportionately with 2 cells in series depending on driver design, so the 123s probably won’t need 2.8A like a single 18650.

I didn’t suggest the XML because the XML c8 uses a very slightly longer head, but if you can offset with a better heat-sinking pill that might be a fine solution. It might be a bit more floody with this reflector, but even a regular C8 can usually be found for not a lot more.

Servus!

As far as I know, CR123 batteries are rated at a maximum continuous discharge rate of 1.5A. You might want to look for a buck driver with a lower current to stay on the safe side...

Yay! more valuable information.

Thank you all!

The reasons why I chose the Sipik:

1st: form factor: I actually do like its shape
2nd: aspherics - I never thought anyone could make a convex lens which would support the XM-L emitter, ever. Really. I stand corrected!
3rd: so easy to modify! - unscrew the retainer from the tailcap, then punch out the existing switch and second plastic retainer (do the Chinese like redundancy, all of a sudden?), and put in whichever switch will fit.
4th: The pill is much easier to modify - the bore is not tapered, for instance.
5th: It already comes with an XM-L! Not as much oomph! as a SST90, but so much more efficient. I love it as much as I love the XR-E- R2… which I like a lot, by the way.

downside: They designed the belt clip badly, though. One hole has been drilled (or rather punched) far too large, weakening the whole clip. Well, you can’t always get what you want - and as I said, that’s just a tiny obstacle with a “kit torch” like the Sk-98

I would NEVER EVER give a 2xCR123 primary cell light to ANYONE or own one! The problem is what happens if you have a good cell and a weak/bad/flakey cell in the same light. CR123’s can contain metallic lithium. IF you have a good cell and a bad cell in the same light, under load the bad cell gets reverse voltage on it and them BLAMMO! I know of a guy that just about blew his dick off when he put one in his pocket (still on) for a second to move a box. The thing got very angry and that was that. YMMV…

Surely, all kinds of electric batteries pose certain potential risks to the user. Just the night before one of our night watchmen had one 9000 mAh D-cell NiMH go bad; it suddenly heated up in his lightbulb-equipped baton (i.e. Mag 4D) for no apparent reason, then popped with a loud bang just seconds after being removed from the battery tube; spattering moist, crumbly wet stuff everywhere - like wet charcoal ashes from a furnace.

Anyways, I think I’ll rig up a test for runtime, reliability, excess heat buildup and amperage for a 105C powered by two 3V lithium primaries. Then again, I am suffering from procrastination from time to time, so that test eventually might never be performed.

I also read a bit more on the forum and now I have learned about the Xtar MP1 and its delightfully low price.
My own WP2 did cost three times as much!

Conclusion after thinking it over (thanks to your input): the most sensible way to power a NANJG-105c-driven XM-L is by using a 3.7V 18650 li-Ion rechargeable cell, period.

(off to buy a couple MP1s)

Simon