Following with interest Kourtinoksylo. All the best for the build.
Update, see first post. 3 refelctors tested, 2 more on the way.
Interested! 
The project is progressing pretty good. I tested 5 reflectors and decided to go with the 73mm from kaidomain.(ordered 4 of them). I managed to focus it really good on my sbt90.2. (3 more one their way from Hank). Delays in delivery are unfortunately unavoidable so I am concentrated on the heatsink for now.
I have drilled the 280holes (2.9mm) in the 6mm copper disc and will later add the 3mm thick copper pins. These will vary in length from 46mm to 16mm depending on the space avaliable in the head.
Question I was thinking and researching a little about how to increase the surface area of my pins. I came up with the idea to cut threads to each of the copper pins that will run their whole length. I found some rough calculations that a screw has around 50-65% and maybe up to 100% more surface area than the same diameter cylinder/rod. This sounds pretty amazing to me.
What do you think about that? Will the threads possibly improve cooling? The airflow will be parallel to the pins and perpenducular to the threads. My main concern is to not make things worse, because the airflow will be perpendicular and wont hit directly the valleys of the threads. Any other ideas to increase the surface area?
UPDATE:
The heatsink is finally ready, here are some photos of it 
It weighs 907gr/1.99lbs 
and has and an actively cooled surface area of aproximately 720cm2/111square inches. This is only the copper, didnt include the head of the flashlight.
(If I remeber corectly) it has 278, 3mm copper pins that reach through the 6mm thick copper to the leds. It took ages to drill all the holes and press fit/hammer all the pins but I had enough time during the holidays and enjoyed it. The other side doesnt look so nice for now as some pins are longer and some shorter, it needs some more work… 
I managed to calculate the surface area of a BLF GT94 from pictures and found out that it has around 1450cm2 of fins and probably some more surface area if you take into account the rest of the head. This is twice as much as my heatsink but it is not actively cooled. The heatpath is also shorter with my heatsink. Honestly I have no idea what is better and from which point is a passively cooled heatsink equal to an actively coolled one. Material wise copper is better than aluminium, but I dont know if I managed to build somehing superior to the the GT94. probably yes. (But also extremely inpractical due to the weight).
The battery tube and the contact point to the driver are also ready. I visited the local machining shop again and had the tubes done there. Then JB welder them together. The contact points are just 1mm thick copper sheet and I made them with a dremel, didnt use a lathe for these.
And a small preview. The upper part of the head is missing of course. I am waiting for the reflectors to come.
This is looking great so far. Im looking forward to seeing how well the copper heat sink dissipates the huge amount of heat that will be conducted to it. If anything, it should dramatically increase run times at the higher power levels. Thanks for sharing. 
That is a lot of work ’ Kourtinoksylo ’. And you have our attention. 
I do believe this light may become epic 
UPDATE:
The heatsink is completly ready
Also screwed onto the head, already some copper patina from the constant handling…
This is the cooling concept and almost final assembly of the head:
I milled out the existing fins using a drill press, a 2mm end mill and a jig to hold the head stable to create the air intake. Then the final work with a small file. Also drilled through the tripod mount at the back side to allow slightly more air flow.
Next step is to cut the 4x 73mm reflectors and epoxy them together. After that I need to make some centering rings for the leds and finally screw them on the heatsink. The speed controller for the fan will be probably at the upper compartment where the reflectors are and not near the driver, because I dont want any extra components near my air intakes obstructing airflow to the pins. The potentiometer for the fan speed controler will be placed at a random location that I will need to figure out and consider comfortable to use. I wanted to use the tripod mount (drill through) as place to put the potentiometer but it seems the least comfortable place, plus I need this hole for extra air flow.
That’s pretty incredible. The heat sink looks like an under water creature.
Wow!! amazing build… Thanks for sharing.
I love your ambition! Can't wait to see how this light turns out as you progress! Keep it up for sure!
Ohhh wow…bookmarked this share!
.
Are you going to cut the reflectors to fit closer together ? Just checking what you are looking to build. :sunglasses:
.
Yes exaclty, from these 73mm reflectors if you remove most of the “bezel” you end up with 67mm inner diameter. The inner head diameter is 128mm(4x 64mm possible), I dremeled out an extra 1.5mm from the inner diameter of the head and ended up with 131mm. I have to cut 2 sides from each reflector to get them to fit all together and also reduce their diameter to 65mm. In the end with tolerances mistakes etc I should end up with an inner diameter reflector set of 129-130mm. The original BLF GT94/GT4 (which I am trying to outperform) have an inner reflector diameter of 118-119mm so I am ok with it.
Reflectors are done. Cut with a hacksaw in a jig from the hardware store and jb welded on a glass surface for flatness.
Almost 130mm, it is a really tight fit in the head, and 12mm wider than the BLF GT94 !
Part of the reflectors is covered by the bezel. I thought about thinning out the bezel to have more of the reflectors exposed. Measurmetns with and without the bezel showed only 1% improvement so maybe I will leave it as is.
And some numbers with the head not in one piece and not ready yet. Double 16awg silicon cable directly to the leds which are also connected with double 16awg in 4s.
I meassured 31.59A !!! with 4s3p 
with the cables being extra long (extra resistance) until final assembly. Cells were at 4.16volts. WIth 1p I got 13amps and with 2p, 25amps. With the 3rd set of MJ1s I got the 31.5amps, so cells are not the limiting factor.
My lumen meassurment box is far from perfect but is the only method I have right now.
My Convoy C8 is meassured to have 852 lumens, and if we suppose it makes 1000lumens we have a factor of 1.17
The Astrolux MF02 meassured at 1743 and should produce 2000? Factor of 1.14
So lets suppose a multiplying factor of 1.15 for my box. The 4 sbt90s produced 21100 lumens which is pretty much spot on with the reviews of the GT94. Seems realistic to me.
If you spoted this suspicious c8 there, its not on. I just used to support the conical heatsink…
I need to measure kcd in the weekend, then de-window 
and meassure again.
A quick kcd meassurment showed 4.42 times more lux than the mf02 (4.42 times 400kcd=1.768Mcd) but only at 3 meters so not very accurate. I will need to test it more accurately and report back.
Last steps are to put in the fan and speed controller. If I dont break any bond wires by de-windowing it will be an amazing light.
Thanks for reading
Wow dude, color me impressed!!! 
Wow this is incredible! I can easily see you hitting 2 million lux if you are already getting 1.7 that close. Can’t wait to see more progress pics!
Quite the monster! I have mixed feelings about active cooling - I think it’s about the only route forward for sustained lumen increases, but waterproofing is out the window.
Man, I think I need to get a few more SBT90.2s. Not cheap, though…
Finally ready!
Fully assembled fully functional.
Side by side with the Astrolux MF02 (xhp35 hi).
The de-window gave me an 11% increase in candela and around 4% in lumens, measured 21890lm. I didnt notice any changes in the tint. Also I dont know if the reflectors need refocusing after delensing, I will leave it as is. However it gave the beam some artifacts (rings) which were not there before. Maybe later I will make some cut-outs to the bezel with a circular saw to have more of the reflectors exposed.
The potentiometer for the fan controller was placed on the left side to be easily operated with the left hand while holding the light. I made a wood knob for it. Looks a bit akward but I like it.
Total weight 2700gr or 5.95lb, not very practical but it was never meant to be.
I haven’t tested and wont test how long it lasts on turbo on max fan speed which is probably the number one question one would have. It does get warm after some minutes but the fan blows hot air and seems very efficient to me. At low fan speeds its pretty quiet but at max it really blows and whistles somewhat because of the many sharp edges at the intake. Doesnt bother me as I am never gonna use it for practical purposes. Just wanted it to be there as a feature.
Question: The driver is from the BLF GT4/GT94 and has a ceiling at 9.5-10amps and then double click to 31.5amp turbo. I want to disable the ceiling and have ramping up to turbo but didnt manage to find the right instructions manual. I dont know exaclty which firmware I am running. Does anyone have a clue? 4 clicks from on (which is supposed to be configuration mode if I am right) is locking the light out.
I will add later some outdoor beamshots when its dark enough.
Thanks for reading and for the kind comments
Very, very impressive !
And fun to read and informative
Job well done!