I have the 2D mod with 2x26650 and a 6D (7.2V) bulb. Works great.
Can i use 3x26650 in a bigger Maglite and just drop an osram 12v bulb in there?
Depends on the bulb. If it fits the maglite bulb socket, maybe.
I forget of the top of my head but the incandescent maglite switches have limited current handling capabilities. So if the bulb draws too many amps it’ll melt and destroy the switch. Also you can potentially melt the reflector and plastic window too.
Back in the old days before we had such a nice lithium ion batteries we’d mod the switches to take more current and mod the sockets to take bi-pin projector bulbs; Metal reflectors and borosilicate glass windows; sub-c cells or multiple AA battery holders. Bored out bodies to accommodate 4xAA holders in a d cell body. Look over at CPF archives for 5megas stuff.
1 Thank
Oh, ok, so i will not do that then 
Want to be safe, just wondered if there was an easy mod with just dropping three batteries and a bulb in the old mag 
Thanks for the answer , cheers
Yes you can, i’ve done it many times, some 12v bulbs can take 5 cells and be overdriven to almost twice of power rating, the more hours it is rated for the more overdrive it could take, 3 cells however is basically rated voltage, i have not blown any 12v bulb running on 3 cells.
Actual 12V bulb, or 14V but listed “12V” bulb (eg, automotive)?
Mag reflectors are plastic, and the dinky come-with bulbs don’t overcook them too much, but try to do a Roar Of Pelican dealy on them, and I wouldn’t count on the reflector lasting too long.
Besides, overdriving a bulb flattens out the brightness, but drastically reduces its life.
Back in the old days, I made a big stank about driving my car’s H4s (4×6 4-headlight system) with overkill paralleled wire and 40A relays for lowest resistance and highest current/wattage… until I found out the above. So I’d be gaining an absolutely imperceptible increase in brightness but lowering lifetimes of the bulbs.
Kept the relays (way nicer than the usual 30A ones), but swapped the 12ga wires back to 16ga, and went back to just threading them through the come-with harness instead of the 2nd one I added on (6 chonky wires + split-loom tubing).
You’d only be hitting the bulb with 12.6V right out of the charger, which shouldn’t be too bad. Use protected cells to a) keep any cell from getting overdischarged, and b) add a teeny bit of extra resistance to prolong the bulb life with near-zero impact on brightness.
The former seems very counterintuitive to me: the brightness of a blackbody radiator, at low temperatures (say below 4000K), should increase faster-than-linear as a function of power, since the higher power induces a higher CCT that has greater spectral efficacy, i.e., more lumens per watt. Would you explain to me why overdriving flattens out the brightness?
I definitely believe the lifetime reduction. Once put a 18650 in a old 2AA incan; the bulb made a beautiful 3500-ish K for a few seconds before it blew.
If you have the budget, get the Lumencraft mag- upgrade for the sheer ooh-aahh moment 
Cheers 
Filament resistance increases with temperature. That’s why inrush current when cold is N times operating current. It’s also why a 12V bulb starts glowing at like 2V or so, is pretty bright at 6V, is near full brightness (visually) at 9V, and doesn’t seem to get noticeably brighter at full rated 12V.
Pushing it to 14V won’t get noticeably brighter because filament resistance rose to counter much more current-draw. But the higher filament temp burns off material that much faster.
That’s why I could have “dead” batteries in an old-timey flashlight that would still glow the filament orange. They’d just “regenerate” enough to handle the inrush current and start the filament glowing, before stabilising at that dull orange.
Anyone who has/had old-timey Lionel train stuff where all the lighting was based on “track voltage” is well familiar with this. Just a smidgen above “off” gets those bulbs a-glowing, even if not brightly. And those variacs were pretty linear for “lever angle”.
1 Thank
Thank you for the explanation! I see that you meant voltage as the input, not power. If the resistance increases with temperature/power, it does make sense that increasing the voltage past a certain point does not increase power substantially, which would cause the brightness to level out.
Yeh, no worries. I was harping on voltages, 12V vs 14V, etc., and figured that was understood. All that happens with increasing voltage beyond spec is that power increases, CT increases, localised heat increases, and the filament burns off faster, thus reducing bulb/filament life.
There are some circuits that actually use small hotwire bulbs as “dynamic resistors” as a sort of AGC to keep an oscillator in its sweet spot, ie, most gain without saturating.
1 Thank
This thread has tests of pretty much every widely available bulbs till they flash.
1 Thank
This is a tremendous resource, such a shame that the image hosting is down.
Search for “Mag1185”.
Post 53 in that thread has all destruction tests tables in a zip file, feel free to download.
1 Thank
If anyone gives a cares, here’s an example of using a small hotwire blub to dynamically controlling gain in an oscillator. 
It’s that “flattening out” with increasing voltage that adjusts the relative gain to keep the little beastie stable.
2 Thanks