# Voltage drop vs output

Hi, i always wondered at what point does modern mid/high end flashlights with decent circuit start to lose a considerable amounts of output +10% and so on.

I’m perfectly aware that you cant notice this buy your eyes, but i just need some reference.

So i will put a small list of lights as a examples and would like to here some experience of users who measured the lumens drop after the cells are lets say around 3.6/3.7v?

TK75, Niwalker MM15, TN32, TM36, TM26, BTU Shocker, RC40 and so on.

I only know at what point i cant use the turbo on my TK75, its at about 3.6v….but i never measured the output difference, before it actually drop to high, will do and report back later on, please post your findings thanks.

My rule of thumb is under 3.6V under load an XM-L starts losing brightness, you can check HKJs battery tests and find the chart and since most of my lights use 2.8 or 3.04A 105C drivers he has a nice 3A on the graph.

Yes i think the same, but i would like to know the numbers behind ~ 3.6 vs 4.2v……

Does it drop 20? 15? we both know that expensive flashlights often have advanced circuit/driver and can offer crazy amount of efficiency, not 100% sure but i think my TK75 measured same lumens with freshly charged cells ~ 4.15v vs 3.6v IIRC.

most budget lights use linear drivers, the LED will only run at full current if it has enough voltage, so if the voltage falls below what the LED needs then it draws less current making it less bright. The brightness falls linearly, you can use the cree characterization tool here to determine what the output will be at a given input voltage which is the voltage under sag on HKJs graph at your chosen current minus about 0.15V for the driver circuitry until the battery runs out of power or the low voltage warning kicks in (on the 105C it typically kicks in at 2.7-3V)

High end lights can have boost drivers so they maintain constant brightness as the battery is depleted, this type of circuitry is more expensive and usually proprietary

For practical use, it is better for it to get gradually dimmer than to leave you suddenly in total darkness.

Check the flashlight tests on CPF as many of them have run time graphs which show the light output versus time for the different modes on the lights tested. Whether the individual test does depends on the author but a couple of them routinely include that information in their light tests. An excellent indicator for how well regulated a given light is.

this is a single-cell light, falls out of regulation in like 10min on a panasonic

most multi-cell in series do better, but not all

that’s why 900+ lumen 1-cell flashlights are a joke they simply can’t sustain it for any uselful amount of time. exceptions include the buck/boost flashlights like zebralight, some armyteks (with FLAT regulation), and thrunite Lynx

That tests are ANSI (with drops to high after few minutes) so sadly they are not rl tests, i’m taling about using the light on turbo on low voltage from the very start…

Also P12 is far from high end flashlight, i was talking about expensive lights which are known to have good regulation. Thanks anyway for the answers guys…

Yes, i’m interested in rl usage of this lights, not entry level.

For whatever specific light you want to know more about look it up specifically on google and try to find a chart of output vs runtime.

Thanks will look around……

The only way to get true results is to measure the cell’s voltage during operation under load, removing batteries and testing them in air before/after the test is very inaccurate and wont give you the data you need to truly know much of anything about the regulation.

I dont need 100% accuracy, i just need to find out, at what point i should recharge my cells, 3.8 or 3.6V generally

And should i charge them to 4.2V or 4.1V, because if the output is same at 3.6V, no need of 4.2V, charging to 4.1V will give the cell MUCH longer shell life.

I’ve now had two SK68 clones that were run for ten-fifteen minutes at full brightness, til they got too hot (lighting up a workspace) — and now both of them, when turned on bright, will run a minute or so then drop to half brightness and flicker at that level. Tried a couple of freshly charged Li-ion cell tested in both (these are the kind that won’t light up at all with 1.5v). Is this a damaged driver, most likely?

That's especially true under high load.

If someone is relying on a flashlight with no low voltage warning as their only source of light with only one set of batteries then it’s kind of their fault… That being said, most lights from the premium brands (Eg. not ultrafire, ultraOK, jacobs) have low voltage warnings.

The only good way I see that answered is if someone tested your flashlight in an integrating sphere while powered by a dc power supply. You can probably figure out how the battery will perform by looking at existing discharge curves, but it'd be a good idea to redo those too.

Even then, it can be awfully hard to separate voltage sag from thermal sag, so you might need the emitter replaced by a dc load tester like HKJ uses.

Based on the graphs, both TK61/TK75 have insane regulation, like fully flat line until you can’t run in on turbo anymore, but i will test it out to have an general idea.

recharge when you like, like someone said high end lights shoudl have low voltage warning, 3.6V no load means dead. If they are fully regulated then charge to whatever you want, the brightness will be the same, just runtime will be less.

Yes, that is correct. I was just to worried that i lose output if i go under 3.7V….