Quantity I would buy: 2 (maybe 3 if it’s $10 or less)
Emitter: My priorities would be warmth, efficiency, CRI, in that order, then max output last.
Reflector/Optics: Whatever gives a reasonable beam pattern and a happy medium beam width.
Charging: Two ideas people have already mentioned - an external charger if it’d help with making the light more compact, or a USB-C input if doing internal - could both be interesting ideas. But built in micro-usb seems more realistic.
Battery: Looks like 10180 is the only way to get something replaceable and significantly smaller than AAA. Just wish someone could make 10180s with capacity per volume more in line with 14500s & 16340s (would be ~120mAh instead of 80). Needs to be replaceable but it’s ok if doing so takes some work.
Modes & Brightness: L-M-H with something like 1, 8, 40 lumens, and turbo at 100-120lm requiring more twists rather than being in sequence with the 3 normal modes. This takes the power-law relationship of perceived brightness changes (which Noir mentioned) into account. It makes sense for possible tasks- one mode for close up reading, one for e.g. ‘around camp’/’around home’, one for walking around in less controlled environments (where you need more sense of your surroundings), and one for ‘ok what was that’. This would also give the light another clear selling point over most 10180 lights, which have just two modes: “too bright to be a moonlight mode but still too dim to walk around” and “using all the power in the 10180 in no time flat.” I think a lot of the people who are questioning the usefulness of low lumen settings and suggesting 300 lumen modes have forgotten what basic lights did for them in the days before all their lights ran off 18650s or 21700s
Dimensions:As small as reasonably possible but no smaller
For a light like this, which really will go on keychains, getting the tail cap and keychain attachment right matters more than usual.
I prefer gray or silver to other colors.
As Noir says, a titanium version would be optimal but design decisions about the surface of the aluminum one will need to take into account how it’ll hold up to jangling against other metal objects continually.
Tilion You’re not late to the party at all. It was supposed to get started after Chinese New Year, but of course delays are inevitable due to COVID-19.
Thanks for your sensible and relevant suggestions.
Quite. I still have a childhood incandescent flashlight which takes 2 X AA, has terrible run time, and seems to put out less light than the 3 lumen mode on the Sofirn C01S. It’s incredible to call it a flashlight today, and yet it still had its use.
Taking a look at things, it really does seem like most 10180s are a little worse than one would expect for their size, and the slightly larger battery sizes aren’t quite as uncommon as I thought.
Some people claim to be selling 100mAh 10180s, which would be a big improvement over the ~80 mAh typical- if it’s accurate. But if those aren’t really 100mAh (or they’re impractical/expensive to source), it could be worth looking at 10220s (~130mAh) or possibly even 10280s (~200mAh). The DQGSpy-X uses a 10220, and looks like a worthy competitor (except it’s hard to find and 3x the cost).
Looks like 12300s (~350mAh) are a real step up - not much over twice the volume of a 10180, and over four times the energy! More energy than typical 10440s (AAA size) despite slightly lower volume and much shorter length. Similarly both lower volume and more energy than the 14250 (half AA) used by some small lights. It’s also rated for much higher discharge current than a 10180, higher even than normal 10440s- you could get 350+lm, though a keychain light would have no good way to dissipate that much heat. The 12300 might be near the upper limit of what I’d like on a keychain, and gets a little uncomfortably close to competing with the handy availability of aaa sized batteries and lights, but it could be neat.
I don’t know how the cost on these would compare if you were ordering in quantity.
I guess I’ll say It’s worth saying a bit about why I think runtime is worth taking this seriously when other people seemed focused on max output. If I really anticipate using a flashlight, I’ll bring something bigger. If I only need light for a moment, cell phone lights these days are ~50lm, floody, and fairly workable, but chewing through phone battery and holding the phone awkwardly get old when it’s more than momentary. So though it may be fun to use a keychain light in a variety of circumstances, the bother of carrying it is largely justified by emergencies - ‘I didn’t think I’d be out in the dark’ or ‘two is one.’ And in those circumstances one doesn’t want to be stuck worrying about whether the light will run out after 30 minutes.
Nice research. I agree that runtime is a higher and more useful priority than peak (unregulated…) brightness.
Canister type batteries, especially in small sizes, have the disadvantage of the weight and space constraints of the steel canister. It seems the pack style offers more advantages with more energy capacity and density for a given space and weight.
Having said that, overall physical size is still one of the highest priorities. I would rather have a 30 minute light that is unnoticed on my keychain than a 1hr light that I keep perceiving in my pocket. One “hack” for longer runtime is to choose brightness modes that are useful but not too high.
The Olight i1R (2) is a good benchmark to try to beat. Going just a bit larger opens up interesting options such as Rovyvon and Nitecore TIKI, but I’m not sure I’m willing to accept the size penalty, and also the price penalty.
I wonder if a pack style battery could be made to fit in an i1R type form factor with USB charging. It seems challenging.
While pack style batteries are more efficient per volume, it’s much less help than one might think. Dimensions, not just volume, determine the effective profile of the flashlight in your pocket, and a thin but wide rectangular object will be more obtrusive than its volume suggests - sometimes almost the same effective profile as the circumscribing cylinder would have. See Nitecore’s TINI which has low volume but not much advantage re profile on a keychain. And while ‘buy it for life’ may be less of a thing if you’re aiming for a extra low price point, it still seems like having the possibility (even with tools and some care) of replacing the battery, rather than viewing the light as disposable, is a boon.
Looking one more time, I think common 10180 lights may be pushing their batteries harder than manufacturers’ recommended max continuous discharge currents, which would be part of why output drops rapidly. I think the 10220 sounds like a good compromise. It only increases the length of the flashlight by 4mm, which is 1/8 the size of the smallest lights or 1/11th the length of the i1R. But it gives anywhere from 1.25x to 1.6x the capacity and, according to one manufacturer, twice the rated discharge current. Again I think the Spy-X is a good comparison.
Note that the TIKI’s docs say its pack battery is 130mAh, which is the same as the nominal capacity of many 10220s, but it has a 2mm wider diameter (15) and is 21mm longer (62!) than the DQG Spy-X.
Came across some old data; for perspective, here’s what incandescents the non-turbo modes I suggested for a 10220 light might compare to in runtime and brightness:
1 lm: like a Maglite Solitaire with ~15x the runtime and half the weight&size
8 lm: better than a 2AA Mini-Mag, 1/9 the weight
40 lm: as bright a 3D Maglite, and a bit better time to 50% (~75m v 1h). Of course it then drops swiftly off in the next 15m while the Mag tails off gradually for another 9h, but it’s 1/70th as heavy
And of course the beam quality would be vastly better.
I agree 100% with the idea of the 12300 cell. I think that’s the perfect capacity, maybe with a Turbo of ~300 (stepdown to 150ish?), but then well spaced modes, like a Moonlight, 15, 60/75, then Turbo. I know some people don’t like turbo, but I think as long as it’s not a crazy stepdown or way out of line with the other modes I would like the option to be there. I have the Nitecore TIKI (which I love) and that’s how it works, stepdown is after 1 minute 20 seconds, down to 120, and I think with this we could do a bit better, probably a full 2 minutes if we have a bit more size for heat dissipation due to the 12300 size. Just my opinion, of course.
It would need an e switch (I dislike twisties- my only gripe with the Mag Solitaire), USB C, and a reasonable battery capacity, maybe 350-400 maybe? I like the idea of TIR/optics as well. For the emitter, why not do something like a Samsung LH351D high cri?
Yes I’ve been edc’ing a 1aaa light in my watch pocket for years. It’s ok to have 2 or 3 brightness levels but no blinky modes please. Don’t need 100s of lumens. A 350 lumen 10180 powered keychain light (SC01) is almost scary and crazy. 500 lumens is a traditional car headlamp. But, a 1aaa light IME is too big for a literal keychain light. Even 10180 is too big. I have a custom 1xLR44 light there now, but before that always had something like a Photon.
The rectangular lights that look like car alarm remotes are interesting as pocket lights in their own right, but I wouldn’t want one on my keychain either.
Keychain lights are less important than they used to be, now that everyone has a flashlight in their phone.
Added: I’ve always had trouble with e-switch lights turning on by accident, to the point that lots of them have lockout modes and the ones that don’t, should. Twisties IME have not had that problem. So I do like twisties.
Ok, thinking about this some more, forget Anduril (I’m guessing this light will use way too small a processor if there is one at all) but keep the 10180 battery, keep the light as small as possible, twisty is good, and I’d say eliminate the built-in USB charger to keep the size down (see Lumintop EDC Pimi for how this would look). There should be either 1 mode of about 10 lumens, or 2 modes of say 3 and 30 lumens. No blinking, no 300 lumens, make the light really nice rather than worrying about the cost within reason ($20+ is fine). I see this as a slightly upscale replacement for the classic Photon Microlight, rather than a photon grenade as TK put it.
Why don’t I buy a Pimi? I’m tempted though 1) copper seems like a stupid material for a flashlight, because of dings and tarnishing; Ti is much nicer. 2) it is too bright / kills battery too fast. I’d want it backed off to about 10 lumens as mentioned above. I wonder if I can do that by changing a resistor or something.
I wouldn’t want the 12300 cell because of size and because it’s hard to find. At that point you’re making a pocket light. I use 1aaa for that (fits in my watch pocket ok) but ideally I’d like something thinner. The Nitecore Tube is just 7mm thick and was quite tempting, but I’m told it is not a very good light, and the levels are poorly chosen, so meh.
Lockout procedure takes too long and PWM is visible on low, else this is a good light. I’ve had countless of them open to swap the LED. It was somehow rumored that it does not have LVP, but that’s not true. I like the UI (except for lockout) for it’s ramping and momentary features. At the same time it is easy to operate. Modes are moon and max (i.e., low ), but as I said, there’s “stepless” ramping. Press and hold from off and stop when it’s bright enough.
Gave away some of them and never heard anything negative.
The small 10180 are terrible cells. Take away lots of space for, what, 80 mAh? My latest baby, the TF Minix, is about the size of the Tube, somewhat thicker, and has 300 mAh. I only would kick mode memory out, maybe use a warmer CCT LED (BLF Edition? ), but that are minor complaints. Anyway, this little beast made me rethink my AAA choice for keychain lights.