Nice work KB!
Any throw numbers?
Iām curious to know numbers too. Great mod btw Kawiboy 
here is an example of the beam from the stock reflector, that I removed
and an example of the beam from the upgraded reflector, that you noticed has the quad effect
this reflector has a smaller hole at the bottom, both LED are 219b
He just wanna make us beg... K, i'm begging... 
Just over 1.3Mcd so far. Looks like more to be had. Got 13500 LUX at 32.4ft. Prolly need to measure further out, but itās raining now, maybe after or later tonight?
. I came up with a new or better way to shim the reflector upā¦ and keep from spinning the shims and the big spacer bar, to keep from sheering the leads offā¦ got real lucky didnāt fry the driver 
. Also just a couple thousands off center JACKS the spot up, with this tiny emitter to this big ass reflector it is a BIG DEAL. I maybe have .009ā slop in the mcpcb diameter to the centering block I.D. Got a couple Ideaās on how to tighten it up, so I donāt keep losing itā¦after bumping it around during shim changesā¦
Donāt let the pic fool yaā¦ that spot is just a Micro Dotā¦ unbelievably smallā¦
Hahaā¦shit, you can see how small the beam is narrowing down at that range 
ā¦ā¦ 
Kind of sad, Iām just doing this for something to doā¦ it will all come down and go away, once I get my partsā¦ Itās been funā¦ 
That tree looks like a big candle 
Nice work
Might be a first for KawiBoy1428ās local fire department. āI was just shinning my flashlight up in that there tree and it caught fireā.
I grabbed a Convoy M21C with the intention of finding something interesting to slap in the host. I only played with it for a few minutes before tearing into it. The SST40 is a good performer but I donāt really care for the tint.
I have been wanting to try a Luxeon V 4000K so this was a good opportunity. I figured with the relatively large die and medium size reflector it would make a reasonably throwy beam; great for my nighttime trail walks. The M21C has very thick walls on the head and it takes a long time for it to heat up. Perfect for sustaining high currents for a long time.
I used one of the nifty remote MOSFET drivers from led4power. I donāt know if it really makes much of a difference as far as thermal performance but itās something different at it looks neat so I figured Iād give it a try. I set the driver to CC mode with a max of 7A.
The M21C is designed for a 22mm driver but the LD-B4 is only 20mm. This means I had to solder the driver to the retaining ring. This isnāt really a problem, but I wanted to solder the wires onto the MCPCB first because the head is very deep and I didnāt want to be fiddling with all those tiny wires down inside there. After soldering the wires to the driver, I gave it four or five anti-clockwise rotations before screwing the driver/retaining ring into position. I was worried Iād twist off some of the tiny signal wires but it ended up working out just fine. There was plenty of space in the cavity for all the extra length of the wires.
I really like the Luxeon V. The tint is very neutral; not pink but doesnāt appear green either. Itās very uniform with minimal tint shift across the beam. Itās not high CRI but colors appear natural. The hotspot ends up being just a hair tighter than a C8+ with XM-L2.
Sunwayman V10a
mod to 219b 3000k 9080, Sapphire lens, relubed control ring with Nyogel 767a, attached some tiny trits to the body.
Makes about 90 lumens on AA Eneloop, about 180 lumens on LiIon. The minimum mode is an Ultra Low Firefly of 0.005 lumens. 
Great mod jon_slider 
Those little trits looke nice too
Nice job Rayoui!
Yes! I love the Luxeon V. Iāll take it over SST40 every time, and a lot of HCRI emitters too.
This is an Eye10 modded to triple N219b 9080 by pinkpanda
It arrived with the ball and spring removed, to eliminate the detentes on the dial. It had very light damping grease, and was totally smooth, like melted butter.
I decided to reinstall the ball and spring, after filling in the detente holes with JB Weld and repacking with heavy damping grease, Nyogel 767a
The heavy damping works well to prevent the dial from moving unintentionally.
However, I could still feel a slight feedback from the ball crossing the filled in detente holes, and the heavy grease prevented the one detente I actually wanted, at off, from working.
So, I took the light apart again, and removed the ball and spring. Now Im happier. 
Nodded a cyan s2+ today
Nichia 219c , copper 1 cent as the spacer , fatter glass , it fits very nicely and doesnāt get as hot as the ylp t95 from where I got the nichia from
Iām surprised that it doesnāt get super hot with the bits I could use
I did things today, and yesterday. I managed to fix everything apart from the unfixable ti shelf. It became so good that this is the first titanium light to become my EDC.
1) the 10mA drain when off in the stock light was a guarantee that your battery is always empty when you need your EDC (and that it possibly drains your battery to death if you forget it a bit longer). So on the new driver I installed a 680 Ohm bleeder resistor, the two resistors in the tail that are inline with the tail leds were changed to 8.4 kOhm each and the leds were replaced by 485nm leds that are more visible to the eye than blue, btw at these low currents the colour they emit seems sea-green: a bit over 500nm to me. Also bits of aluminium tape were sticked to the sides of the switch to guide a bit more light out of the tail. The current when off is now 0.36 mA which is 30 times lower than stock and should last for 3 months on a battery charge. The amount of tail light seems a bit less than stock but is well visible in the dark as locator light.
2) the led was replaced with a 3500K 90CRI LH351D (from the batch I bought from contactcr) that I sliced so that the final CCT is about 3000K. I wanted to create a new customised ledboard but decided for the stock board, it seemed to me that the gains in thermal performance would not be high (at most from very very bad to very bad). I did very carefully apply a thin and even layer of AS5 to at least do something for the heat transfer.
3) the stock driver would have been fine if it did not have an ancient 2012-type user interface (memory without direct acces to low which is terrible for my EDC use, low is way too high, bad mode spacing) so it had to be replaced. The new driver is a Banggood BLF-A6 driver, my all-time favorite driver that is in all my EDCās for years now. But since this host needs a 15mm driver I transferred the components from a BG BLF-A6 driver to a custom Oshpark board (designed years ago by wight, OSH Park ~ ), only the FET was replaced by a smaller LFPAK33-size FET (from Mtn). The edge of the driver was re-shaped to fit the non-circular cavity in the Tool AA ti.
4) the components and traces on the driver are fairly close to the edge, so the driver cavity was lined with a strip of kapton tape to prevent possible shorts with the shell. The stock driver was glued in place, the new driver is pulled in place by exact fitting short led wires, then clamped by the battery tube.
5) the driver board is a bit thicker than stock, which caused a small gap between the battery tube and head, with the disc sander I sanded about 0.5mm from the tube to close the gap.
6) (not visible in the picture because it was not done yet) after assembly the beam appeared to have a faint but annoying ring a bit outside the hotspot, it may already have been present in the stock light. The ring originates from light reflected from the side of the narrow reflector hole which has considerable thickness and is close to the led die so it catches some light. The solution was ditching the stock centerpiece, replace it with a standard ābutterflyā type 3535 centerpiece and reaming the reflector hole to exactly fit the new centerpiece. The new centerpiece appears to give perfect focus to the domeless led and the beam is also perfect, a nice evenly illuminated but distinct hotspot (the way I like best) and perfect even spill.
7) I did not like the golden clip, and even the black clip of my alu Tool AA is not great, I do not use the clip anyway so it will remain clip-less and will make a nice custom lanyard from thin paracord.
Performance
A direct drive LH351D is easily too much for the heat trap that a all-titanium head is, so the tail spring was not bypassed and I used a yellow ā3Aā Vapcell 14500 instead of the ā10Aā H10.
In 4-mode setting I measured 6.5lumen~~65lumen~~>350lumen->650lumen. I have not measured current yet but my guess is that the 350lumen setting has about the same current as the 650lumen max of the stock XP-L2/driver, so already not really sustainable in this host. The 650lumen max in the new situation (guess=3Acurrent) is dangerously hot now, although I let the light run until the 40-second stepdown (built-in feature of the BLF-A6 driver) and that did not cause a meltdown yet.
The setting that I will probably use most is the 65 lumen. The light is pretty throwy so that already gives a fairly bright hotspot. Here is the hotspot tint at 65 lumen setting:
Summary of the tint: :heart_eyes:
I always wanted a titanium EDC, I had an Eagletac D25A and D25C once but I thought they looked not nice and the user interface was too complicated and mediocre. The Tool AA ti is pretty, very compact and has my ideal form factor, and now with a new nice led, beam, and my favourite driver Iām really happy and will check it out as EDC the coming period.
The nice hotspot at low setting:
A lot of work but the results seem worth it, nice job.
Edit; no clear switch boot in parts drawer?
congrats on your Awesome Mods!
but, the low mode of 6.5 lumens with the new driver is not much different than stockā¦ is that as low as the driver can go, or just your personal preference does not include a sublumen level? (I have a BLF A6 driven S41, and it certainly can go sublumen)
Moon is only in the 7 mode group. The old A6 UI is not full bistro. I assume moon in 7 group would be sublumen.
thanks
yes, I have the S41 set to 7 mode. The moon level is 0.1 lumens on my meter.
The Tool AA has a 12mm boot cap, in this size there is no clear one available, I only ever found orange. And I kind of like the black stock cap that does transmit some light although not very efficient.
Edit: I did manage to squeeze a 14mm clear cap in my aluminium Tool AA V2, with 6 slow colour changing leds on a ring pcb, the effect is cool but I think the cap is a bit too bulgy to look really nice:
It has the same led and driver as my ti version btw, but the driver is faulty (it has messed up lower modes but the 7135 is not the cause) either the BG driver was broken already or something went wrong with the transfer of the components to the new board (some Oshpark boards are less than perfectly fabricated). A new driver for that one is my next project.
Thanks. I set it on the group with 4-modes because that is how I use it most, but if set to 7 modes it goes much lower yes, I just measured it, exact 0.300 lumen.