Quark AA2 mod

In a previous post (here) I was asking how I might revive a dead Quark AA2. There were some helpful suggestions - thanks. Ultimately there was no simple fix as the driver was kaput and I didn't really want to disassemble a stacked driver to re-attach the inductor or troubleshoot some other component. Into the yet another rainy day box.
So, I'm looking at replacing "things." This is my first go at this so I'll jump in with suggestions Cereal_Killer made in my previous post - "...note 2AA boost driver selection is non-existent in the size you need. You’d have to consider other battery setups, perhaps one 14500 plus a dummy cell.... That or go for a low drive current MT-G2 and run a pair of 14500 with a great driver." If, after having reviewed many modding and informational threads, my thinking is taking me down a dead end path, please say so.
First, given the advances in emitters and drivers since I bought the flashlight ten years ago, is the Quark AA2 even a suitable host in terms of heat sinking and the tail switch (given the options I'm looking at below)? I'm using Mountain Electronics to reference parts. Other suggestions welcome.
Option 1: single 14500 and dummy
- QLITE REV.A 7135*8 3.04A LED DRIVER
- 17mm; Nichia 219C D240 on 16mm MTN DTP CU MCPCB - 90+ CRI 4000K;
- Sanyo UR14500P 840mAh Battery or Efest Purple 14500 IMR 650mAh.
With this choice of driver and emitter, are there pros/cons/cautions about using a high drain (the Efest) battery?
Option 2: two 14500s
- QLITE 6V+ Zener Mod - For 6V MT-G2/XHP;
- or, FET Driver - 17mm Zener 6V mod with LVP;
- Cree XHP50.2 J4 6500K on 16mm DTP Copper MCPCB; (Mtn Elec. warning for this emitter - "due to extremely low forward voltage, it may be damaged by use with an FET driver and high-drain 18650/26650 batteries." Would this still be an issue with lower capacity, but high drain, 14500 batteries?)
- The reason I'm choosing a XHP and not C_K's MT-G2 is that the hole in the base of the reflector is 4.5mm which, with only a tiny enlargement, would fit the 5mm of the XHP50.2. The MT-G2 is 9mm and enlarging the hole to that size would destroy the offset/raised base of the reflector.
Thanks for having a look at the post and any suggestions or comments.

Wellp, it is an AA-sized light, after all. I wouldn’t go pushing it very hard, ’cause even if the LED doesn’t cook itself to death, it’ll be like holding the wrong end of a curling-iron.

A boost driver and pair of NiMH cells (eg, eneloops) could power a nice high-CRI chip quite well, and the quality of light (vs quantity) would probably be very worthwhile. If you like the CT/tint/CRI, you’ll appreciate the light more than if you cooked an XHP50 to its maximum output (but could only use the light for 30sec before dropping it in pain).


The one mode I found most useful on the Quark was moonlight. I have found only one 17mm AAx2 boost driver on Fasttech (suggestions welcome). There's no moonlight and it starts on HI (my least preferred startup mode). I'll order/install one until something better comes along, or have it on hand as a backup.
A look through the single AA/14500 forum shows any number of lights using hosts considerably smaller than the Quark AA2. The only suitable driver (with moonlight) that I have found is the Qlite 17mm at Mountain Electronics (not in stock as of this date - other suggestions welcome). Even though I would be limited to just 14500 lithiums (plus dummy cell) with this driver, wouldn't this be a workable solution? Driving a Nichia 219C D240 90+CRI? (to quote Texas_Ace "...work great in the small AA/AAA and 10440/14500 lights...). Battery - advantages/concerns between regular and high drain?

If you are ok with using 1 14500 with a dummy then that is an easy solution. You could use a FET+1 driver. These also have good moonlight modes. I would recommend bistro firmware.

Regarding the 14500 cell, if you use a FET driver you will get higher brightness on turbo with the high drain cell. The light will get hot quickly. With the higher capacity sanyo cell it will not be as bright but you’ll get more runtime.