Thank you for the reasoning since i dont own any cells of that format..........
I received this flashlight today. it seems very good quality. In fact i think that it is my best quality budget flashlight. Most certainly better that Ultrafire C3 and stainless steel version. It is gives about same brightness with alkaline that Ultrafire C3 with 14500. If used with 14500 its even brighter. This will be my next EDC flashlight hands down!
Highly recommended! +++++
Good, I'm glad you like it. It certainly is a high quality instrument, and ridiculously bright even on 1.2V NiMH. I wish that shipping wasn't so expensive from here to anywhere else, because I would sell mine for a song.
It is a nice light. It would be an even better light if it had no strobe and a low that is about a quarter of what it is. But the sheer brightness is very impressive. On a 14500 my one makes around 362 lumens at switch-on, 329 lumens at 30 seconds and 315 after two minutes.
Why? Don't you like it ?
Runtime. I prefer lights that have a very dim low mode, and the low mode on the R5-A3 appears brighter than my other lights on high. It only gives me about two hours of runtime on low. But it really is a nice light, just not very useful for me in particular.
Do you need that low mode for some specific purpose? I prefer my lights just hte opposite. I like when they are brighter and brighter. If i need a a low light then i can use some old non-cree led lights.
What song would you like me to sing?
And the shipping costs is?
;-) I've never even look into shipping costs from South America, but I know they're really high. Probably upwards of 50% of the value of a new R5-A3 from China. So it doesn't seem justifiable.
I see no reason why the driver makers could not design a high efficiency driver - probably this would be best with two or three NiMH cells to run an XP-G R5 for about 5-6 hours on high. Driving the LED at 350mA will still give 139-148 lumens. 3 NiMH cells should be around the forward voltage of the LED so the driver could be designed reasonably cheaply and still be of 85%+ efficiency. Heat would not be an issue at that current and it is still producing way more light than a 6D Maglite. The LED is probably less efficient below 350mA. Having checked, it isn't efficiency seems to be about the same up to 1A.
So we need 350mA at say 3V nominal (typical forward voltage of an XP-G at 350mA) which gives us 150 lumens (to simplify the sums). This should give us over 100 lumens after all the losses from reflectors etc. If the driver is 85% efficient then we use 100/85 x 350mA to get the current to the LED - 410mA so we should get an easy five hours from three Eneloops. Halve the current so we use 175mA and we get half the light (The output curve seems linear) - so 50 lumens for 10 hours or 25 lumens for 20 hours.
If we are to use a single cell layout which is more convenient for most people then in an ideal world we should divide the time by 3, so 25 lumens for 6 and a bit hours. Actually the step-up driver will likely be less efficient so we may only get 6 hours from a single cell. For me, something that gives me the option of 1, 2 or 6 hours from a single AA would be great and the 1 hour runtime would still be a lot of light.
But it seems like the motorcycle market - there are a lot of little, reasonably efficient and unglamourous devices and a lot of high-spec high-cost lunatic performance devices of no efficiency at all. And not a lot in the middle unless you spend a lot. Me, I'm in the crazy performance area myself and don't care too much about runtime but I'd have thought the market for a mid-price (by China, not Surefire standards) high efficiency device would be huge. Or maybe that market can't afford expensive LEDs - it wouldn't surprise me if the LED in this light was above half of the total cost of manufacture. Which might be less than a quarter of what we pay for them.
A single eneloop contains roughly 2.4Wh at 0.5A discharge rate (and 2.3Wh @ 1A), at a lower discharge rate the usable energy content goes up.
If we want a light with 150, 50 and 10 lm OTF modes we need roughly 30% more at the emitter.
For a XP-G R5 to make ~65lm (50 OTF) it needs ~0.43W, factor in a regulator efficacy of 85% and it will need ~0.51W
Same goes for ~200lm, 1.55W (150 OTF, 1.82W total) and ~16lm, 0.11W (10 OTF, 0,13W total)
A two cell light using a XP-G R5 and the fictional current regulated three mode (10-50-150lm) regulator would run for approximatively 37 hours in low, 9.4 in med and 2.5 in high mode using two regular eneloop cells. There are some step up regulators that can maintain efficacies of over 90%, using one of those would of course extend the run times further. If the regulator isn't a current regulated device and instead uses PWM for lower modes, then the run times in low and medium mode will be shorter.
I'd have thought power out of the cell is the same with both PWM and current regulation (The XP-G R5 output curve appears to be linear from 100-1000mA) , but then what I don't know about electronics fills several large libraries.
LEDs are less efficient when driven of off a PWM regulator for lower modes then what they would be if driven of off a current regulated regulator because LEDs are more efficient when driven of off a lower current. PWM is basically a on/off switching of the max current, so the efficacy of the LED will be that of the max output mode. Current regulated regulators utilizes this behavior to either bring you more light at the same power consumption or give you the same brightness at a lesser power consumption.
Compare this:
A PWM driver dims an XP-G R5 from 1A current to 100mA RMS (really 1A @ 10% duty cycle), the brightness is then 35lm, or 10% of what it would be at 1A.
A current regulated driver running the same XP-G R5 at 100mA constant current (100mA @ 100% duty cycle) would produce 44lm of off the same RMS current.
So 20% more light for the current regulated circuit @ the same power consumption.
If you factor in power differences due to the Vf being higher at higher current densities it gets even worse:
An XP-G R5 produces 42lm @ 0.43W RMS when PWM dimmed from 4.3W @ 10% duty cycle.
If it were consuming 0.43W at a constant current (0.43W @ 100% duty cycle) it would produce 65lm instead.
0.28W @ 100% duty cycle is enough to generate 44lm when driven of off a current regulator.
~55% more light at the same power consumption when factoring in increased Vf so we could either get a light that is 50% brighter and runs the same amount of time or have a light that is equally bright and runs for approximately 50% longer.
The efficacy curve is almost linear but it would have to be linear and not sloped (no droop at higher currents) for the output curve to be linear.
I get it now. Thanks for the explanation.
and if you heard me sing you would never ever go near a postoffice 
What does the pill look like?
Does the pill come out?
The pill looks like an aluminum pill and comes out of the front so you just have to remove the bezel, lens and reflector first.
I just got mine today from DX. It's not as bright as I thought it would be. I ran it on fresh Duraloops and the Hugsby P31 is brighter.
There's a donut hole looking hotspot on mine. The hotspot looks like the color of the emitter. Surrounding the yellow is a white ring. Anyone else have this sort of beam?
The P31 is hard to beat. That’s why its my EDC. Almost all my XP-G lights have at least some sort of a faint "donut hole" but I'm more sensitive to it than most. They are only evident to me when "whitewall hunting". Take it outside and see if the donut disappears. I never notice it outside on any of my XP-G lights, but the donuts are there when shinning on the white walls indoors. The R5-A3 is no slouch on 14500's. Do you have one to pop in there and try? It will AMAZE you! Seriously! If you still find the donut hole annoying outdoors (AA or 14500), let me know and I'll tell you how to fix it. But in doing that, you might find that there are compromises in the hot spot - beam, pattern, width and quality.
Do you have other XP-G lights to compare?
I have a 501B XPG R5 5mode put inside a Solarforce body. That one has a tight spot. I do white-walling all day long and there is no such donut hole or any color artifacts. The beam looks exactly like my Hugsby P31 except a lot brighter.
My 14500s are in the mail (still waiting for them). Is it ok to use a 10440 with spacers?