Best path to the 10,000 lumen club? [Planning a project/ideas]

Flashlight Modding and DIY Parts
1

As much as I’d like to say that they are, my BTU Shocker and MT-G2’ed Trustfire X6 aren’t bright enough, 3000 lumens just doesn’t cut it anymore. I need…more. J)

To that end, my new goal is five digits, ten thousand lumens. It doesn’t have to be something immediate but I want to start amassing parts and a host in the interim. I know this is going to cost more than $200 to do.

My current idea is this:

Trustfire AK-90
*4s-26650 configuration so there’s plenty of voltage for this 100W+ light. also decent capacity too. 12up XM-L configuration means it will be a flooder, but I’m not spending $350 on an Acebeam so that’s a compromise I’m willing to accept.

*12x XM-L2 emitters on noctigons. Self explanatory. Driving each emitter close to 3 amps should get me my 10,000 lumen otf goal.

*A direct-drive driver with a second slave board and FET to minimize the effect of the PCB traces adding resistance to the circuit. Emitters wired in a 4S3P configuration.

Here’s where my concern begins. I have a 3up-XML light powered by a 26700 King Kong, and it won’t even pull the 9 amps that the 105C and slave board are chipped to drive. And that’s with older, lower-resistance XM-L emitters, surely driving current will be less with XM-L2’s. Or, can XM-L’s on copper put out enough light to get me to my goal? If that’s the case it would save me about $40-50. Alternatively, are there less expensive plunger lights that allow a 4s-26650, than the Trustfire AK-90? Is my goal of a 10,000 lumen light impractical? should I buy a used TK-75, a battery extension, and three MT-G2s and rewire the battery cradles?

2

There’s so many ways! I’ll recommend two.

I’m a bit partial to LarryK14’s, which are a 600 watt airplane landing bulb stuck into a spotlight housing. I built one and it’s nice to see what high wattage incans can really do. If you want to build a regulated one, I do have an extra “JimmyM JM-PhD-X1 voltage regulator” that I am not using. I believe he is sold out and no more will be made but check with him first.

Beyond that suggestion, you can reach 10,000 lumens by maxing out a single HBFlex driver from Taskled, in whatever host. The drivers are said to be good at least to 100 watts but it is not really known where their maximum output limit is. I probably wouldn’t attempt more than 12 XM-L2’s for around 133 watts. The driver would need good heatsinking.

I have done all the relevant calculations regarding this:

HBFLEX CALCULATIONS:
XM-L2 @ 3 amps = 3.4 volts according to Cree PCT. Total wattage and minimum allowed input voltage to stay below 7A current input:
3.4 x 12 LED’s = 40.8v x 3.0A = 122.4w / .92 eff = 133.0 / 7A = 19.00v min input (14,160L)
3.4 x 11 LED’s = 37.4v x 3.0A = 112.2w / .92 eff = 122.0 / 7A = 17.43v min input (12,980L)
3.4 x 10 LED’s = 34.0v x 3.0A = 102.0w / .92 eff = 110.9 / 7A = 15.84v min input (11,800L)
3.4 x 9 LED’s = 30.6v x 3.0A = 91.8w / .92 eff = 99.8 / 7A = 14.26v min input (10,620L)
3.4 x 8 LED’s = 27.2v x 3.0A = 81.6w / .92 eff = 88.7 / 7A = 12.67v min input (9,440L)

25v max input divided by cell count = max voltage per cell
25v / 7 = 3.571v (lifepo4)
25v / 6 = 4.167v (6x4.2 = 25.2 which just exceeds the maximum allowable input voltage)

Lifepo4 voltages:
(x2.9) (x3.2v) (x3.4)
7s 20.3 22.4 23.8
6s 17.4 19.2 20.4
5s 14.5 16.0 17.0
4s 11.6 12.8 13.6

Li-Ion voltages:
(x4.2) (x3.0 )
6x 25.2 18.0
5s 21.0 15.0
4s 16.8 12.0

48v max driver output voltage divided by # of LED’s in the string:
48v / 14 = 3.428v max per led
48v / 13 = 3.692v max per led

Minimum acceptable pack voltage before reaching 7 amps input for a given wattage:
140w / 7 = 20.0v
130w / 7 = 18.57v
120w / 7 = 17.14v
110w / 7 = 15.71v
100w / 7 = 14.29v
90w / 7 = 12.86v
80w / 7 = 11.43v

3

I see what you did there

4

TK75 may not be the best idea as the stock driver wont get you anywhere near you want. If you are willing to spent a lot of money and time, then you can get it above 10000 lumen.

If you just want to peak above 10000 lumen, no special requirements to range, flood, tint, CRI, or much else. Just pick up something like a 15 emitter light with XM-L2s for 66$. Add a zener modded DD driver for bout 6-15$. Wire the emitters in a suitable way 4s3p (for the first 12 emitters, and then take each of the last three emitters in parallel with three of the other emitters). Not sure if that light have a proper emitter shelf. If so you need to make one. Should not add more than 10$ if you got some various parts and tools. If you have 4 purple Efest cells you should see well above 10000 OTF for a limited time from a light costing less than 100$. If that host is bad, find another one. Lots of suitable lights with 12 XM-L2s or more for well below 100$. The one I linked too may not be the best budget option, especially not if you need to make a pill and are not well equipped to do that. Its probably wise to buy a light that have a proper emitter shelf. You can get that verified by some members who have bought similar lights, or maybe the seller if you buy from someone reputable.

Cant remember if zener mod is suitable for 4cells in series, but I believe so. Someone will have to verify that.

You may end up replacing the switch too. Ideally you want to have all emitters on copper.

For the "retard light", I used some OEM/Kinfire light. I liked the form factor and reflector size. On top of that I was going to make a light with good tint and extra range for a more useful beam so stock emitters did not matter. I just wanted to make a likable light with good tint, beam and sick output that was not too large and heavy, yet had some runtime potential on more reasonable output. Being so short with only 2 cells in series, high amps would become an issue (over 30 amp on the tail/over 60amp divided on the emitters at startup), so I made it into a twisty. If you want a more beefy light that can handle more heat you might want to look into the Trustfire TR-J20. At the end of the day, having over 10000+ lumen without active cooling will mostly just be suitable for "bursts" anyway. Unless you have a really efficient setup, which either will just give you a ton of flood, or be really huge.

If you`re not in a hurry you might want to wait around (1-3 months?) for some XHP70 emitters and proper stars. (Or you can buy today and start customizing) Three of those and well driven will take you above 10000 lumen. Having to only use a triple reflector adds a lot of possibilities. When that is said. Cheapest option is to just get a suitable light with 12 or more XM-L2 emitters, a BLF DD, and low resistance cells.

Like Hoop. I really like the Taskled drivers coupled with E-switch. Breaching 10000 OTF may be a bit on the limit, but you will much easier get those lumens when you want it, instead of just at startup with fresh cells. Its gonna cost you though.

There is a bunch of ways to easily breach 10000, really just depends on your budget, imagination, host and ability to mod.

5

UltraFire LW-12L2 for $69 might be an option for a host. The stock driver only supports 3*26650, but you can use extensions for more if you’re going to be using your own driver.

http://www.fasttech.com/products/0/10006644/1519900-ultrafire-lw-12l2-12-cree-xm-l2-t6-5-mode-8800lm

I think XM-L2 is definitely the easiest way to go for this kind of light. MT-G2 is floodier and requires way too much voltage.

6

10,000 lumens? Pfffft! You should aim for 20,000!!!

Buy 4 of RMM’s modded Supfire M6s … and a large rubber band :smiley:

7

Nice. I don’t think I’d seen the LarryK14 builds before you posted about it here. Crazy stuff and kind of tempting.

8

Three MT-G2 are about the easiest way right now. Use something like a Solarstorm SP03. Use three of the boards I use for the Maglites with 9 amps each, master slave. It's an idea, but seriously, it will get so hot in one minute, you will have to turn it off. I think you really need to talk to machinists and have one of them make a host with a huge heat transfer area, meaning tons of fins and lots of aluminum in the head. Then you could do something like 12 high drain cells 2S 6P and possibly be able to power it. More money, but that way you could decide on larger reflectors so it would throw a little as well.

9

I actually did consider the SP03 when I thought about 3up MT-G2s, but the problem with that light, and why I think the TK-75 is the only such style host that can work, is that it only has four 18650’s to draw from. That’s 2S2P to run three MT-G2s, 5000mAh at the most, to power a combined 27 amps of draw. And the voltage curve for a Efest looks pretty bad that steep. The TK-75 at least can have additional battery carriers stacked on, and in that case one would have a theoretical 20 minutes worth of runtime on high before the cells were depleted.

10

As a professional machinist I can tell you it would be prohibitively expensive to have a custom host made. It would be pretty cheap though to use a spotlight host and mount all of the emitters to a large heatsink with a fan, mounted in the host. Use a 12v DC/DC converter such as the R-7812-0.5 to efficiently regulate the voltage to the fan. Secret: it is actually possible to control the DC/DC converter with a Taskled driver via the STAT output so that when the driver is on the fan is on, but without getting the firmware tweaked by George you would lose low voltage warnings, but still retain the emergency cutoff. :wink: With a lifepo4 battery the low voltage warning isn’t really of much use anyway since the discharge curve is so flat. The stat output is capable of 200mA stock which is fine for 24v fans and some 12v fans.

11

Hoop, did you edit your post to have less information?!? :open_mouth: I filed the LiFe brand you mentioned away in my head as something to maybe check out in the future…

(and I also mentally filed away the MRTdiver DIWdiver linear driver reference as something to look at for inspiration/reference 0:) … and I see that the forum does not correctly parse the smiley with the flushed cheeks?)

EDIT: corrected username

12

Personally I would not consider 3x MT-G2`s in a budget light a very good path to 10000+ OTF.

That is over 3333 lumen OTF from each emitter. Which in a budget light, with a budget reflector and lens, is going to mean 4000++ at the emitter. Which is basically the absolute max you can get out of an MT-G2 when its highly overdriven. That is going to result in a ridiculous amount of heat since its well beyond whats efficient. At 10000+ heat is already an issue if you have a fairly efficient setup. With the heatsag you`ll be looking at. Achieving 10000+ will probably not be doable if you are using ANSI standards and well calibrated equipment. There is a slight chance you might breach it, but why not aim for something that will do it better?

Even X60Mvn only have about 7800 lumen.

12 or more XM-L2`s will just be more efficient, and they are practically free with the right light/host. Ideally you want more emitters for better efficiency, but that results in smaller reflectors, less throw and more light close up, where you already have enough light. If you are going to stick with 3 emitters then XHP70 is probably a better solution for reaching or properly breaching the lumen goal.

13

Yea haha.

The driver I had mentioned, DIWdivers’s 10A linear driver is only good to 12 volts. I had laid out a scenario of use with 7x Headway lifepo4 cells but then I discovered the max voltage was 12v so I removed it. But there it is once again as another option.

DIWdiver also has a 13.5A linear driver that is good for up to 15 volts input. A dimmer can be used with either driver. I have three of these 13.5A drivers that I am not using.

With these drivers you could run a lot of MT-G2’s in parallel and with pretty good efficiency with lifepo4 cells.

Yea, 4x MT-G2’s would be better. Not the cheapest option but not crazy expensive either.

14

To be fair, that test didn’t use high drains, which are a requirement even for the regular X60M. Even with high drains, the X60M kills batteries. If you are going to push MTG2 that hard, I think you want at least 3 cells/emitter.

15

Thanks Hoop! I edited to correct the username mistake I made above, oops!

16

You also could jump to the 15000 and just buy an off the shelf 4 7’s Maelstorm XM-18.

17

That will go perfect with his yacht.

18

F-f-five s-s-cells in s-series… [string of drool drips out of the corner of his mouth.] S-side switch… [fingers trembling.] It…it even comes with its own c-carry strap. [puts hand on the mouse].

NO! [left hand wrenches right hand off the mouse.] No! It is madness! There must be a limit!

There is no limit! [heaves up on the desk] The only limit to the size of my flashlights shall be the strength of my back! [roars at the ceiling. Dog jumps off the bed and starts barking at me. I pause and take stock of the situation. Smoke rises from the broken computers sparking on the floor.]

*Its past 1am and I have a fever. That’s my story and I’m sticking to it.

19

I’m sticking with exactly what race r86 is saying. Try the new xhp70. Or 12x xml’s.
That’s what I’m doin. It’s in the process of being built right now by 18sixfifty. Host is
J20 and u3’s

We tried the triple mtg2 route with a j19. Didn’t quite get there. Stopped short at around 7000 lumens.

That host you linked above with side switch is super nice. But I already know the heat sink is not gonna make it. In all honesty, the j20 seems to have one or the best heat sink I have ever seen in any multi emitter light

20

3xMT-G2 90w
&
15*XM-L2 262w