Which would you choose and why?

I’ve ordered some odds and ends from DX, and I’n going to be building some 16340 lights for EDC rotation.
I’m at a bit of a crossroads in my planning of these projects, and I’d like to get some opinions.
Basically, I have a choice of 1000mA or 700mA drivers, and XP-G or XM-L emitters.
I’d like a good balance between runtime and output.
I’d like to use 1000mA and XM-L’s in all the lights, but I’m concerned about budget, runtime and current draw from the cells.
Is the XM-L a big step up from the XP-G at these currents?
Would I get significantly greater output from 1000mA over 700mA, or would the loss in runtime outweigh the increase in output?
Is it healthy to draw 1000mA from a 16340?
What would you do?

At 1000mA, a T6 XM-L will be equivalent to a XP-G S2 bin in terms of total output.

http://flashlight-wiki.com/Brightness_Bins

But the T6 will have a larger hotspot, not as bright as the XP-G. So unless you just want to light up a room, you will get better results with a XP-G, even if it is a R4 or R5. For the XM-L you will get 37.5% more out at 1000mA vs. 700mA. Without looking it up in the XP-G datasheets, I would think the XP-G would be about the same percentage.

If you run a XP-G at 700mA or a XM-L at 700mA, you should get exactly the same runtime. The maximum draw from a li-ion is supposed to be 2C where C is the capacity in mAh. The Trustfire flames are rated at 880mAh, so as long as you keep the draw below 1.76A, you should be okay. Honestly, 2C is pretty high, so 1000mA sounds a lot better, but you'll still be lucky to get about 45 minutes of runtime.

Thanks guys, that clarifies things a bit for me.
The drivers I’m using are the 16 mode DX ones that seem to have a few fans here.
They have (I think) 1040mA and can be upgraded to ~1400mA (I’m typing on my phone so I can’t look it all up at the moment :D)
I know that 1400mA is way too much, but I am undecided about 1040mA.
The reason is that, from what I gather, the suppliers are rather 'optimistic ’ about their batterie’s capacities.
I was given to understand that most of these cells are more like 400-500mAh, so 2C is more like 800-1000mA. That’s why I am a bit concerned about current draw.
Right now, I’m running a Trustfire TR-803 with the 16 mode driver and a XP-G R4 neutral white.
I removed one of the chips so the driver is now putting out 700mA.
I am really happy with the light, but I didn’t know whether to go for a little more ‘oomph’ with my next builds.
I was all set on the XP-G because you can get neutral tints, but then they release the XM-L in neutral to confuse things!
I think I will use XP-G’s because the benefits don’t appear to outweigh the drawback (cost and availability, remember I’m building several lights, 5 in fact)
I’m still on the fence about the power though.
I have some Trustfire ‘flames’ on the way, but I just don’t trust the manufacturer’s claimed capacity.
Has anyone done an exhaustive test to see what the capacity truly is?
I’m very wary of Li ion cells and I’d rather play it safe…

I said I am wary of lithium ion cells, not that I don’t trust them.
They are perfectly safe when handled correctly.
Running them over 2C is not handling them correctly, it is asking for trouble, which is why I was asking if 1040mA is safe.
As the above poster mentioned, according to the manufacturer’s specs, 1040mA is fibe, but unfortunately the manufacturers are known to lie about their specs.
If I didn’t remove one of the chips,

(sorry, something weird happened and my typing started to go off the screen!)
If I hadn’t removed one of the chips from the driver, then I would have lost a mode.
Why would I want to have a mode I didn’t want and have to skip through it when changing modes?
On that light, I wanted the highest mode to be ~700mA, so I removed one of the chips.
I don’t see how using a lower mode would have been preferable to modifying the driver to behave exactly how I wanted it to?

(same problem again!)
Why would I want to sacrifice lo-med-hi to have lo, hi, and (unwanted, possibly cell-damaging) ‘turbo’ modes?
I’m now thinking (hence my original post) that maybe I might be able to get away with not removing a chip on my other drivers, but still stay within the parameters of the cells I’ll be using.
I expect an ultrafire C3 could probably be made into a 3xAA with the extension tubes.
Not my cup of tea, as I think it would be a bit too long for edc in the pocket… :smiley:

Don't you have to take into account the difference in Vf? If the XP-G's Vf is, say 3.5V at 700 mA (I'm too lazy to look it up right now), and the XM-L's Vf is 3.0 V at the same current, then the power drawn by each LED is:

- XP-G: 0.7 x 3.5 = 2.45 W

- XM-L: 0.7 x 3 = 2.1 W

Therefore the XP-G requires more power from the battery, reducing run-time. The only way I can see this being false is if the driver circuitry doesn't adjust for the lower Vf well, and becomes very inefficient at lower Vf (which may be the case--I don't know much about drivers).

If it's a linear regulator, then the regulator is going to burn off all the extra voltage the battery is supplying anyway, so I don't think the differing Vf would matter (I could be wrong, but that's my understanding). Whatever the LED doesn't use the regulator will. This is why linear regulator drivers are less efficient when used with 6V than with 4V, because the extra voltage is just burned off anyway. If you had a boost or buck circuit the lower Vf might be more efficient because then you would be delivering less power.

Would be interesting to see , side by side results ...

Considering the variable vF for instance ..

Lets say 1A driver : Driver supplies vF of say 3.7v ,

For arguments sake , lets call the XP-G 3.5vF and the XM-L 3.2vF ...

In the same light / driver / battery , which would perform better ?

Lower vF would introduce some over volting with a 3.7vf driver , would this juice up the emitter some ?

I agree.
I think that the voltage supplied by the driver is governed by the driver itself not the LED being driven by it.
I had this problem when I made a red flashlight.
Red LED’s have lower Vf than blue (which is what white ones are under the phosphor) so the voltage was too high for my red emitter.
The weird part is that the emitter itself is fine. Because the driver is current controlled, it doesn’t really matter if the Vf is too high.
The odd part was that the modes stopped workibg, and I still haven’t figured out why.
In fact, I may make another thre

Ffs! My typing keeps disappearing off the screen!
Is this a problem with the forum or my iphone’s rendering of the page?
I was going to say I’ll post anoer thread about my red emitter/driver problem :smiley:

The linear regulator driver only allows a certain current through. The voltage drop across the LED is then determined by the LED given that current. In fact, that's what Vf is: at a given current, the voltage drop is Vf. If you supply a higher voltage, the LED would want to draw more current, but it can't do that because the current is limited, so it settles in at the Vf for the current that is available.

Here are some measurements I took on the AK47 linear regulator driver. Note that Vout stays at 3.28V, which must be the Vf at 1050mA. But once the battery voltage drops below some threshold (the Vf plus the linear regulator's minimum overhead) you're just seeing direct drive (at 3.63V on the battery):

At rest battery voltage 3.63 3.83 4.05 4.19
High V in (volt) 3.55 3.71 3.91 4.04
I in (mA) 850 1050 1060 1050
P in (watt) 3.02 3.89 4.14 4.24
V out 3.28 3.37 3.38 3.37
I out 900 1050 1060 1070
P out 2.95 3.54 3.58 3.61
Efficiency 98% 91% 86% 85%

Ah, I see!
Thanks for explaining that, I understand now, I had it backwards!
I expect that’s the reason I lost my modes when I swapped a red emitter into a 1xAA light, the driver must have had a fit because it wasn’t getting the voltage drop it was expecting.
What should I do about this? Could I put a resistor in series with the red emitter and get the modes to work?
I actually put in a single mode driver, but I’d prefer to have multi mode.
I soldered a white emitter to the driver once I had removed it, and the modes worked fine again, so it’s definitely the red emitter that caused the problem

Not sure why the low Vf would cause you to lose modes unless the Vf was low enough that the regulators had to burn off so much excess voltage that they were overheating (in which case I think a resistor would help, but I'm getting in over my head once I start talking about circuits). If it was a 1xAA light, were you using 1 AA or a 14500? If it was an AA, then the driver couldn't be a linear regulator because it would need to be boosting the voltage.

That's weird that once you swapped the LED back out it worked again.

Wow, I haven't heard the term "linear regulator" in about 10 yrs. (I used to work in electronics but now I don't) My memory is foggy but if I remember correctly, a linear regulator is exactly as you described: it outputs a fixed voltage, and if your input voltage increases, regulator efficiency decreases since it just "burns off" the excess, as you say. If the output of the regulator is attached directly to a diode (such as an LED) with no resistor, the output will be clamped to the diode's Vf at whatever maximum current the regulator is capable of, at which point the regulator has essentially fallen out of regulation and is probably getting stinkin' hot. Is that right? It's been a long time since these synapses have fired.

I didn't know that linear regulators were used in flashlights. I assumed that flashlights were all using switching regulators for better efficiency, but that was just my assumption based on no research. I hadn't looked into flashlight drivers until very recently when I started thinking about a mod of my own (which uses a boost switching regulator, so the lower Vf of an XM-L does make a difference).

Maybe I'm misunderstanding something, or am not aware of newer technology in regulation. (by "newer", I mean newer than 10 yrs ago!) Why would someone use an inefficient linear regulator when they could use an efficient switching regulator?

I may be using the wrong term. The chips being used are AMC7135 current regulators. I know what you're talking about. I have a linear regulator that outputs a constant 5V, but requires at least 7V of input, so it isn't efficient. It has a metal tab on it with a hole that is used to radiate heat off the chip because it can get pretty hot. That's not what these are.

As you can see in the table above, I was measuring efficiencies of 85-92% while the driver was in regulation (99% when it goes into direct drive). It is very efficient when the supplied voltage is just above the Vf of the LED. And lithium-ion batteries fall right into that sweet spot.

It would appear that the driver was overheating, it would stay on hi for 30 seconds or so and then go dim. Leaving it off for a while then switching it back on would give hi again for another 30 seconds or so.
Seems to me like it was overheating and then cooling down.
If this is the problem, could a series resistor provide a fix?
If so, how would I calculate the value of the resistor?

Just a little update, I decided I liked my red light so much that I transplanted the emitter into one of my CR123 lights.
It’s now running at 700mA on high, which is the maximum for this emitter, and it’s really bright.
In fact high mode is too bright for what I made it for, namely navigating dark rooms…
Anyway, long story short, I used the 16 mode DX driver modified to put out 700mA instead of 1050mA, and the modes work just fine!
No Idea why the other driver didn’t work, but this one does so I’m happy :slight_smile:

http://www.tmart.com/Ultrafire-MUCC88-500700-Lumens-Aluminium-LED-Flashlight_p108158.html

http://www.dealextreme.com/p/saik-sa-305-cree-q2-wc-3-mode-140lm-white-led-flashlight-3-aa-42672?r=65116799