Does this light exist?...
-Floody beam without a diffuser, faint hot spot from an OP reflector would be nice
-90 Lumen ball park with NiMH
-Under-driving the XML to maximize Lumens per watt efficiency
????
Why not just use and XP-G or something like that? I'm interested to hear too, since I don't see, for the life of me, why the XM-L is used so much as an underdriven LED. I think it's just because it's "the newest hot emitter" on the scene. 1AAA, you're talking about a very low ma, less than 500, probably no more than 250 for any longevity. Might as well use a lower output emitter, maybe even an XP-C. Less cost.
Interested to hear what the reasoning is for the XM-L from others.
Doesn't the XML emit more lumens than the XPG at lower drive levels? (and conversely use less current to achieve the same lumen output).
I have also found the XPG to color shift towards green when under-driven. I only own two XPG lights though.
I am looking for a very floody wall of light kind of beam (like a Nichia 5mm). I was thinking the large die XML would be best for that, without losing Lumens with diffusers and a smaller LED.
There are (kinda) this: http://www.candlepowerforums.com/vb/showthread.php?323564-Peak-Vesuvius-AAA-sized-capable-of-producing-600-OTF-lumens! and this: http://www.candlepowerforums.com/vb/showthread.php?329823-500-Lumen-AAA-Flashlight-(photos)-Work-In-Progress-Part-2
The first is out of production, I think, and the second is not in production yet.
I think the first is 10440 only and the second is geared more towards 10440 than NiMH.
Don't know if there exists any other such light.
I wanted to pull apart a Maratac AAA and replace the emitter with a XM-L , unfortunately the pills a glued tight , and its too small a host to force open .
I got into the Vesuvius nonsense a little bit and what I see exactly what I expected from the weird advertisement from Curt on CPF. Now I don't mean to be rude but how the guy presents his product reminds me of some late night commercials for some random cr** on TV.
Anyway, the light is not for sale on their site (anymore or whatever), they don't seem to have any proper business correspondence... in other words: sounds like junk business to me.
the reasoning is efficiency.
it is also, imo, has the best beam profile, ie the floodiness hides some artifacts in cheap reflectors
i recently took the mini c8 looking AAA light from meritline and modded it. tried xpe and xml with stock driver. xml drew more current, was significantly brighter and had much better beam profile. Didn't have an xpg on hand but do now so maybe i'll try it.
this graph was created with match's numbers from his emitter testing thread
@dthrckt Thanks for posting that. It pretty much nails what I am getting at. My ~3 year old ITP-EOS has served me well, but its time for an upgrade. I was thinking about what were some things I'd like to change about it. Floodier beam and more lumens per watt drawn I think would be the main benefits from the XML. Its not about huge Lumen numbers, a light this small doesn't have the mass to sink heat away anyways. So its more about low current efficiency and color tint too. From what I have seen the XML does not color tint shift as much as the XPG when under-driven. Its almost like an XML on a small stick kind of design.
Based on that graph you get about a 20% increase in output driven at 850ma (proper for single AAA?) compared to an XP-G. Who wouldn't want that? 
Doesn't the the driver efficiency trump the variance in emitter efficiency by a huge margin? I remember 50% runtime/brigthness difference in drivers especially on a single NiMH. (Zebralight reviews come to mind... here and here for in depth explanation)
Something I hadn't thought about. But yeah, the light I am thinking about would be optimal with an eneloop AAA, designed for efficiency around a 1.2-1.5V source. If necessary, 10440 compatibility could be sacrificed in favor of NiMH efficiency. In that scenario, wouldn't the superior XML Lumen/watt efficiency still be a worthy investment (over the other LEDs)?
Yeah of course... but look for the good driver first and then at the emitter efficiency... wanting a XM-L for the different beam profile over an XP-G makes more sense imho. But seeing/owning 8-10 1AAA lights primarily for use on NiMHs... I can tell you that emitters aren't the end all aspect to consider. I would choose the DGQ tiny II with a TIR (XP-G R4 NW) or my Brinyte PD03A ("obsolete" XR-E with a short reflector) any day over my Maratac/ITP/Blackcat/Tank 703 when it comes to " general beam usefulness". Unluckily I don't have a store around the corner, where I can try the lights first. I think people concentrate too much on single aspects these days rather than looking at light engines as a system consisting of driver/emitter/optics. The problem is that recommendations are so subjective, so we tend to cling to hard evidence in form of tech-data and reviews, because these can be verified.... but more than once I was totally mislead and my expectations were not met. After buying 50+ lights and keeping about 30 of them... I still fail to translate the reviews/tech-data I read into a ruff idea what to expect in that little envelope in the mail.
The LED would see 850mA from a 10440 (and that's more than 2C-3C typically, so probably less than 850mA), but for a NiMH it would be much less, probably around 350mA if the driver is efficient, because of the necessary voltage conversion.
So probably not much gain from using XM-L over XP-G in this area.
There's not a lot in the way of drivers for 1AAA anyhow.
I have not used it.
If someone with the knowledge and ability to test it wants to see what the effeciency is, I would send it to them.
That's a offer, if anyone didn't notice. 
that's true, about 8%, although I think you'll often draw (a little) more current from the same driver/battery w/ an xm-l
what vectrex said is right. I was thinking along the lines of modding a light already owned...
awesome thanks!!
Think about an already efficient & small EDC 1 x AAA upgraded -> XM-L WITHOUT :
-trying to make it a thrower
-trying to get most lumens with no battery cap, no sinking etc
-PWM, strobes, too many modes (1 or 2)
...and that would be about it.
Imagining DQG Tiny AAA with XM-L
Some 30-60 lumens with nimh, not heating up badly and considerable runtime!! 8)