I think that a 2A drive limit on a 14500 is sensible, it gives you around 700 lumen output which is ample for a small light and will already need a stepdown after a few minutes because of the heat. Also, most 14500 cells will not even deliver more than 2A for any length of time (but depending on led voltage).
I wonder what a sensible driver limit would be? If they used the LH351D, the high CRI supposedly gets ~360-400lm@1050mA(bin dependent), that’d probably end up around 640-870lm@3v/2A. Driving it at 3v/1.5A would get ~485-690lm and keep load down on 14500. An eneloop can push what, 3A@1.2-1.5v? boosted to 3V~1.5A sounds doable ish.
An eneloop can push what, 3A@1.2-1.5v? boosted to 3V~1.5A sounds doable ish.
Unfortunately it’s a bit more complicated.
Let’s have a look at the images below.
At higher currents, the internal resistance becomes noticeable in that the output voltage instantly drops considerably. (If you have a Li-ion and a NiMH with the same IR, the NiMH drops a much bigger fraction of its V_out at a given current, where it counts)
Then there’s losses everywhere – springs, contacts, inductors, semiconductors etc etc – and we’re losing even more precious mVs.
The third graph is from this very chip – it’s as good as it gets – and we can see that at these voltages we’re already dipping into VERY inefficient territory @1A out – and now, if you want to keep this 1A out up, you pull 6 amps from the poor battery. —> lower voltage, even higher losses, even lower Eff, in short – everything shits the bed at once.
It’s just the way it is. ¯\_(ツ)_/¯
(Oh and it’s much much worse with Alkalines.)
That’s also the reason why the lights with the highest claimed output on AA/NiMH don’t keep that output up for long.
Thanks for that explanation kikkoman!
Efficiency really just falls screaming off a cliff in the third graph. I guess it entirely makes sense, boosting voltage is so much more difficult with a high difference between in/out.
So if i understand correctly then we’d be converting 1.5v->3.3v in ideal condtions we’d need 2.2A at 100% efficiency to get 1Aout, but factoring in driver dipping below 50% we’d need 4.4A, then toss in rando V loss for another 33% to end up in the 6A range?
Poor battery holy crap.
Also probably explains why Zebralights are so much more efficient than Armyteks by only boosting to 6v for their emitters instead of boosting all the way to 12v.
So if i understand correctly then we’d be converting 1.5v->3.3v in ideal condtions we’d need 2.2A at 100% efficiency to get 1Aout, but factoring in driver dipping below 50% we’d need 4.4A, then toss in rando V loss for another 33% to end up in the 6A range?
My back-of-the head calculation was: 1.1V (already under load! +losses) to 3V@1A = 3A in —> Eff. dipping below 50% —> 6A in. (And then further V_in drop etc.)
This particular converter shuts down at 4A I_sw, that’s where it ends.
When the efficiency curve is that steep it doesn’t make much sense to make detailed calculations anyway. It’s a dynamic system, and it just goes bad, fast. One problem exacerbates the other and vice versa. That’s the bottom line.
You could, in theory, run 2 switchers in parallel and share the load, but then the battery still remains the weak link.
—————————————————-
With all this in mind, what’s a reasonable output?
I think the sweet spot is around 200 lumen, a good balance between output and runtime. Seems to be what most AA lights have settled at. (Although some still step down from that, for some reason.)
Keep in mind, 90CRI has a bit of an impact too.
This is turning out great. Now I’m definitely interested in 4 or 5 of these especially if the price is not so far off the SP10B. Also interested in a couple of drivers if they become available for purchase.
Attiny MCU and AA Voltage wont work without a separate boost powering the MCU, which would make likely a way too high standby drain
not sure how the original driver handles this
I wonder what a sensible driver limit would be? If they used the LH351D, the high CRI supposedly gets ~360-400lm@1050mA(bin dependent), that’d probably end up around 640-870lm@3v/2A. Driving it at 3v/1.5A would get ~485-690lm and keep load down on 14500. An eneloop can push what, 3A@1.2-1.5v? boosted to 3V~1.5A sounds doable ish.
Unfortunately it’s a bit more complicated.
Let’s have a look at the images below.
At higher currents, the internal resistance becomes noticeable in that the output voltage instantly drops considerably. (If you have a Li-ion and a NiMH with the same IR, the NiMH drops a much bigger fraction of its V_out at a given current, where it counts)
Then there’s losses everywhere – springs, contacts, inductors, semiconductors etc etc – and we’re losing even more precious mVs.
The third graph is from this very chip – it’s as good as it gets – and we can see that at these voltages we’re already dipping into VERY inefficient territory @1A out – and now, if you want to keep this 1A out up, you pull 6 amps from the poor battery. —> lower voltage, even higher losses, even lower Eff, in short – everything shits the bed at once.
It’s just the way it is. ¯\_(ツ)_/¯
(Oh and it’s much much worse with Alkalines.)
That’s also the reason why the lights with the highest claimed output on AA/NiMH don’t keep that output up for long.
Thanks for that explanation kikkoman!
Efficiency really just falls screaming off a cliff in the third graph. I guess it entirely makes sense, boosting voltage is so much more difficult with a high difference between in/out.
So if i understand correctly then we’d be converting 1.5v->3.3v in ideal condtions we’d need 2.2A at 100% efficiency to get 1Aout, but factoring in driver dipping below 50% we’d need 4.4A, then toss in rando V loss for another 33% to end up in the 6A range?
Poor battery holy crap.
Also probably explains why Zebralights are so much more efficient than Armyteks by only boosting to 6v for their emitters instead of boosting all the way to 12v.
Put a sliding cover on the side button to prevent pocket light. They’re being made for cell phones:
https://www.google.com/search?client=firefox-b-1-d&q=phone+button+slide+...
Extremely interested!
My back-of-the head calculation was: 1.1V (already under load! +losses) to 3V@1A = 3A in —> Eff. dipping below 50% —> 6A in. (And then further V_in drop etc.)
This particular converter shuts down at 4A I_sw, that’s where it ends.
When the efficiency curve is that steep it doesn’t make much sense to make detailed calculations anyway. It’s a dynamic system, and it just goes bad, fast. One problem exacerbates the other and vice versa. That’s the bottom line.
You could, in theory, run 2 switchers in parallel and share the load, but then the battery still remains the weak link.
—————————————————-
With all this in mind, what’s a reasonable output?
I think the sweet spot is around 200 lumen, a good balance between output and runtime. Seems to be what most AA lights have settled at. (Although some still step down from that, for some reason.)
Keep in mind, 90CRI has a bit of an impact too.
Interested
This would be awesome!
Correction
This is what would be awesome!
I’m interested!
Interested
This is turning out great. Now I’m definitely interested in 4 or 5 of these especially if the price is not so far off the SP10B. Also interested in a couple of drivers if they become available for purchase.
Interested.
Are you talking about something like webcam cover?
or more like remote control cover?
The second design would fit more this new 2AA flashlight from Sofirn
Interested
uuuh interested
Look ma, no
handsseparate boost converter1.4V in, this is standby
Anduril on an AA light? If that happens, sign me up for several! I guess I’d better keep a closer eye on this thread!
Interested.
The best light is the one in your pocket when you need some.
Nice project :). Interested
Interested
interested
>sliding cover
either kind, whatever’s available that could be adapted without reinventing the wheel.
Thanks to Rost333 above, for finding those examples.
Here’s another
https://www.aliexpress.com/item/32898609281.html
Interested
interested
When I'm spending ₼on€¥ foolishly, I like to do it ₩i$el¥...
I have all the flashlights I need, but not as many as I want...
An AA-size flashlight with Anduril would be great.
I would definitely prefer high efficiency over high lumen output.
Regards Frank
Interested!
interested
Smile, you cannot kill them all.
Interested!
Pages