A 32650 to 26650 sleeve can be made from a 3/4” PVC coupling by removing the center stop.
Tom- pull R13 and bridge it totally. Also did you already have a stack at R12 or just a single resistor? Mine pictured above with a stack of 2 R150’s is totally stock.
Shit, hooked my test LED up backwards on accident and now the driver seems to be [near] dead, it lights very very dim on mode 3 and 4, nothing on 1 and 2. I don’t know what to mess with first or if I even want to bother or just put one of my own in it. I see it uses LEDA QXxxxx Buck controller and I have plenty of those and parts for complete circuits but I don’t have parts for a complete SEPIC so IDK
I did find you do NOT have to remove that one screw to remove the driver, all it does is locate the pill in the body.
Edit: solved it-
All the legs are broken off this FET and the connections were made with blobs of solder. WTF lumintop?! I’m looking for a similar (better) FET to replace it with but in the mean time it’s repaired with another solder blob.
Yes I did, it was lowish at 2.2..2.4A -> reported to Rey back then. The manufacturer boosted the amps to about 2.6 after that (that might be the reason for stacking at R12).
JM got one, which was actually working better than my sample. Confusing indeed.
Still waiting for my GB lights to arrive.
Ok now that I’ve got the driver repaired I’ve been doing some testing and resistor mod’s dont seem to change anything, I stacked a R150 on top of the stock stack and that didnt do anything at all to the current (measured I out to the LED) so then I removed those three (my one added to the factory two) and replaced it with a stack of two R060’s, it was still the same 2.5A so I added yet another R060 and it was STILL the same 2.5A. I then removed all 3 and put a single R150 (just to see) and that gave me 2.04A, it seems to me like 2.5A is all this driver can manage to do regardless of total sense resistance.
I also tested with both the R13 in place and removed with a solder jumper instead, again no difference.
I hope that a better SOIC8 FET can provide a I boost
Edited to add-
Switch LED swapped for a blue 0603 on this one, the other one is getting either pink or white.
I'm afraid if you want to get some power, would have to sandwich in a good driver. I think the only thing lost would be single D cell capability? Would the 4 AA carrier work with a ATTiny13A driver?
Mine showed a R100 on the top of the 2 stack - no idea what was below it.
So bridging R13 makes no difference?
Review comments:
- this light is exceptionally light and small. Makes you wonder why a SupFire L5 or K2 is sooo big? It's lighter with a 32650 in it than a K2 with a 26650 by far.
- the pill top is only 1.2 mm thick - quite thin, and a poor heat path to the body, but the body is quite thin anyway.
- this is not a light that would make a good hot rod for both it's small mass and poor heat path. Of course you can, but it will have serious thermal sag and will heat up quick.
- It's all about pros/cons of light weight, small size, but the separate pill design that doesn't even directly thread into the body is not ideal.
- The big TrustFire 32650 is a tight fit - I'm afraid with repeated inserts I may damage the threads of the endcap, since it's struggle to get the threads to catch. The 26700 cell definitely doesn't fit. Ideally I'd love to rip off the protection circuit off a TF 32650. Think there's a post somewhere of someone who did it (Edit: led4power here: https://budgetlightforum.com/t/-/23275, but no wrapper info)
Can't find unprotected TF 32650's, though WallBuys used to sell them - may have to look around...
About the 13A based driver question- it would work with the 3AA carrier (it’s 3s) in or if you convert the 4AA carrier to 4S but in 2s2p config the voltage will be to low.
EDIT- I did this to my 4AA carrier, very simple. See below.
They should of dumped the QXxxxx (LEDA) buck controller and made the SEPIC out of a different controller… Preferably from a boost controller.
18650 26650 32650(D) all are fine.
but if you try the battery holder, you should change another spring,it's also in the packing box.
Like I’ve mentioned I really didnt like how there was not only no moonlight, but that there was a dead spot in the UI runnig 4 NiMH in the Al carrier. I decided to do this with the main goal of making it so all 4 modes light up, it would of been pretty hard to explain what was going on and how to get around it to my mother, who I actually got one of my lights for. Moonlight is a good amount higher, no longer a “moon” mode at all really, but at least there isnt a dead spot in the UI now that I converted to 4s.
step 1: Mark the PCB and the plate, we will only be working on one of the two PCB’s, the one marked “Battery tray -A” is the one that need’s the work. MAKE SURE YOU MARK THE POSITION OF THE PCB relative to the carrier, the positing on the end plate doesnt matter but the PCB its self must be returned to the same position!
step 2: Disassemble the carrier, again no need to remove the bottom.
after you remove the 4 screw’s and the plate there is a 5th screw that is the + connection
step 3: Ignore the red marks, these 3 black line’s are where we’re going to cut
cut
step 4: Ensure you’ve cut threw, there should be no continuity between these points
step 5: scrape corner’s of two sections to connect
step 6: connect your cut free pad’s, I used a piece of used solder wick. Recheck for continuity at the same two positions as above, if needed run blade over cut’s again.
Step 7: Reassemble the thing, make 100% sure you have the plate in the correct orientation!
4S NiMH with + and - at both end’s!
Moonlight mode using the 4AA carrier running NiMH’s, all four modes work!
Guys, I'm really sorry. Life has been absolutely crazy for me the last 3 months or so and are finally getting sorted out. My sample only did 1.7A with a Samsung 25R cel and since it was going to be changed again for the final version I didn't see any sense in reporting my findings as they had no bearing on the final product and would only get people upset. I did report my findings to Rey and =the= as for the modes working properly as well as all the cells working properly. The sample I had also went all around the world and by the time I finally got it =the= reported on his and Rey said they werer making changes to the driver anyway.
I just got my official copy I purchased yesterday and just now tested the tail amps. I only get 2.1-2.2A at the tail with a 25R. The tint is quite lovely and I've yet to put it on the IS but I will and will update once I do.
My final mod and #'s for now:
- reflowed the stock LED XM-L2 NW (who knows if T6 or U2 but looks like 3C) onto a 20mm Noctigon
- had to file out 2 more side holes on the Noctigon to accommodate the ground wire (the grnd wire is needed because the driver itself doesn't appear to be grounded)
- sanded down pill top all the way to 2500 GRIT, used GC Extreme for the MCPCB
- added piggyback resistors that didn't seem to do much, if anything (added two R100's to R12, 1 R220 to R13)
- replaced 26 AWG LED wires with 24 AWG, replaced 26 AWG grnd wire with 22 AWG
- added spring bypass wire to + driver spring (24 AWG)
- added spring bypass wire to tail spring (22 AWG)
- used NO-OX-ID where applicable, and additional thermal grease on the outside of the pill
On a TF32650 @4.20v, tailcap measured amps and modes:
Hi: 2.85A, lumens: 1003 @start, 925 @30 secs (stock light measured lumens: 833 @start, 768 @30 secs)
Med: 0.84A, lumens: 388
Lo: 0.23A, lumens: 117
Moon: 0.02A, lumens: 4
What's interesting is I believe the performance mimics the LD-1 driver, so seems to have very efficient moon, lo, med modes, maybe not using PWM's then. I'm pretty happy with the performance now. It probably does a little better on a high performance cell, like the EFEST 3500 mAh 26650, or the EFEST 35A 18650, or similar.
Thanks JM for the update, and thanks Dan - looks like a real interesting carrier mod.
Note: all my lumens and kcd measurements are "my own" taken on an older LX1330B meter - kcd taken indoors and lumens in a PVC constructed lightbox, accuracy can be freely debated
I thought I read somewhere the electronic side switch was suppose to illuminate. If so, that is the only thing that doesn’t work on mine. Other than that, I really like this light and glad I made the purchase.
Your item has been processed through our sort facility in ISC NEW YORK NY (USPS) at 4:09 pm on January 8, 2015. :~
Look closely, it lights up green very faintly when pressed. Notice the pic above of mine swapped to blue, the 0603 blue LED I used is much brighter than the factory green one is. I also press it with a thin stick for the picture so you can really see it.
Maybe it’s a low voltage indicator, as-is only lighting up when pressed isn’t really useful in any way.
Not sure why it went to Grand Rapids. Usually goes to Allen Park outside of Detroit. I live about a hour from the D. 
Do we get a replacement driver since everyone has issues?
Thank you Cereal Killer, you are exactly right! Than all works perfectly as it should.
Has anyone compared tailcap current with led current?
My light arrived today.
The 4 aas will not fit in the body in the supplied carrier. The driver also seems to be under performing as stated above.
How are you going to fix this Rey? I am not happy.
Same here. Mine also has some defects in the anno near the logo and the reflector is not seated properly (hopefully just an assembly error that I can fix).
Pretty disappointing at $39.