I'll respond to some of Reppans concerns....
I'd like to compare your tailcap and output measurements, with the ones I did.
For this size/weight/indestructible niche, its build quality should be compared to a Malkoff 1 or 2 AA. Gene uses dual springs, and the potting is evident. Yes, MDCs are 50% more expensive, but fit and finish is immaculate and I would bet 3:1 odds on Gene's light in a torture test. Fitment of my bezel and emitter do not improve by viewing angle or rotation, nor is it on par with my Quarks. And yes, I'm in the US and can swap my Malkoffs or Quarks in a week or two - my AT took nearly 6 weeks to arrive from China.
So here's some lightbox/tailcap measurements I got on 1 Eneloop ~1.31V - as you know I'm a low lumen guy and will focus on sub-/single digit output (lumens/ma):
- Prime Pro A1 WW: 0.02/12.4; 0.39/20; 5.8/150
- Neutron V2 NW: 0.05/19.4; 5.3/73
- SC52 CW (starting @ "0.34"): 0.07/6.8; 1.3/17.5; 7.5/81
- QPA-X CW: 0.3/10.5; 3.6/48
The current measurements match what I've (we've) seen in my various runtime tests, the NV2 and SC52 roughly doubled the PPA1 in my 5-7 lm side-by-side test. BTW, here's what I get between the PPA1 and QPAX at Max output; batts not fully charged (again lumens/amps):
PPA1 - 225/3.78 (NiMh); 385/1.68 (14500)
QPAX - 140/1.96 ""; 315/1.12 ""
We have very different results here... seems that of my samples, the Quark is more efficient on a lumens-amp basis, but yes the PPA1 is brighter (and much on NiMh).
Now this is interesting, taking the tailcap measurement on a 14500 yields (lms/ma):
PPA1 - 0.7/3.8; 5.8/29
QPAX - 0.3/4; 3.6/16.6
So the efficiency of my PPA1 sample on a 14500 now looks very good, and perhaps closer to what you are seeing given the higher 2AA voltage? Don't know if my boost driver is faulty or not, but my sample is inefficient on 1xNiMh. Given this is a wide voltage head, perhaps AT is using the same head for 1AA and 2AA (as Quarks) - can you try testing 1xNiMh on yours?