2D Maglite Tri XM-L U2 1C **PIC HEAVY**

I been planning this build for quite a while now, but never got around to actually building it til now. I got the idea for a triple XM-L from threads over at CPF and decided it would be a cool build to make. I originally planned on using a sandwiched AMC7135 driver to reach 3A and just adding a resistor inline to the switch to drop off the extra voltage so the drivers wouldn't heat up so much. Luckily, kaidomain had some new drivers on their website that would better suite my purposes. Thumbs up to Angelos for showing me those drivers!


Tail cap readings:
Freshly charged 2 x 26650s after 30 minute cool down.

High: 3.75 A
Med: 2.06 A
Low: 599 mA

Light output testing:
Testing was done in a dark room. Lumen output was done using the light box method. There is still a degree of error using this method so take it as an approximation.

  • Light box lumens:
    • 1923 lm on high @ 1 sec
    • 1792 lm on high (stable)
  • Ceiling light bounce test:
    • 212 lux on high
  • Lux @ 1 meter:
    • 12000 lux on high

Flashlight pics:

LEDs wired in series

Beam shots:
I'll have some outdoor beam shots up soon! Hopefully these will suffice for now. :)

These shots were taken about 7-8 feet away

Final thoughts:
Definitely one bright light! It's hard to tell from the pictures and in person, but it beats the DRY on high. The hot spot isn't as nicely focused using the LEDIL Boom reflectors, might do another build using the extra DRY driver I have and a triple head reflector like the DRY's. The 1C tint is pretty nice, but for some reason it shows up kinda greenish on some pics. Gotta love these Maglites! They're so moddable, they should be called Modlites! :beer:

Thanks for looking!

Nice build.

How do you fit 2x26650's into a 2D?

nice light.

where did you get the heatsink, and could you link the driver?

Awesome! Would also love to hear about all your parts, though I suspect "I own a lathe" will come up.

Aside to oldbobk: if you hit the 'post reply' button at the top left of the original post, then you won't end up quoting all the pics (which is hard on the bandwidth).

Very nice build! I really like it and I'm interested in a part list too.

It probably has even higher actual lumen counts. The TK70 runs at 3 amps each and hits 2200 otf lumens. This one approaches DRY levels (likely with better efficiency due to nicer parts) so I'd think it would hit 2500ish.

Great job! One day I will build one of those.

I find it interesting that it's brighter than your DRY. Your light is pulling 2.5A/LED at the tail while the DRY are pulling anywhere from 3.2A-4.2A/LED. It has to be the reflector shape focusing more light into the center making it appear brighter.

@Hill: Thanks! I used a fiberglass sleeve made by LPF member jayrob

@dthrckt: Thanks! I got the heatsink from download at CPF last year, this build has been pending for a while heh. Here is the link to the driver: http://kaidomain.com/product/details.S020077

@oldbobk: Thank you! It's actually a more neutral color to my eyes, but the camera turns it kind of greenish. Your right most likely medium is all I need. The high setting is more of a wow, that's bright mode, heh.

@ruffles: I wish I owned a lathe! I would love to make my own heat sinks, but I'm kind of out of budget for that right now =[ I got the heat sink from download @ CPF last year.

@N.shock: Thanks! I'll update the parts in the specs section with links!

@joe1512: I would like to think so, that would be awesome!

@JohnnyMac: Thanks! On high it beats the DRY, but on turbo the DRY is the clear winner. Running the ceiling bounce and light box test while both on high, the Mag got the higher numbers. Probably better efficiency since it is a U2 bin vs the T6 bin in the DRY.

thanks jufran

any chance of you trying the light w/ one real and one dummy cell? I know it should go into low voltage protection mode - but I'm wondering if cycling on/off twice will disable that.

I just tested this out and it does work! When you turn it on it will go into protection mode and flash, but if you cycle through it will bypass the protection and I was able to get it to work normally.

that's awesome - and means someone smarter than I am might be able to figure out how to disable the protection.

any idea what kind of output (watts) you were getting?

I Estimate around 22 Watts. I took apart the light to do this and get the Vf and I. :bigsmile:

wow, that's great - I hope someone will figure out how to permanently bypass the protection - though it is kind of big to fit in single cell lights...

edit - wait...so you were measuring output too LEDs and not input to driver? did you measure input, too? because if it has decent efficiency w/ one cell.....even better! thanks!

Wow! nice work!

@dthrckt: Wait, I think I'm lost.. Yes, I was measuring output to LEDs with 2 cells. Measuring input wattage to the driver? Hmm, that would be batteries voltage (under load) x tailcap current = 7.78 x 3.75 = 29-30 Watts?

Calculating for efficiency using Vbatt x Ibatt x Efficiency = Vf x If

Efficiency = (Vf x If) / (Vbatt x Ibatt)

Efficiency= (9.2 x 2.4) / (7.78 x 3.75) x 100 = 75.68% efficient

Not calculating other losses such as internal resistance on the switch and tail cap.

@atbglenn: Thank you!

that's what I meant (output, input, efficiency)

but i thought you were getting 22 watts to emitters with ONE cell :) which is why i was pretty excited lol

anyway, 75% beats the pants off the last driver I tested efficiency on

Ahah, sorry about that. The driver could be around 80% efficient if you factor out the resistance from the tail cap, switch, wires.

if you don't mind I'll fwd a link for this thread to ian mander. he maintains this driver db http://www.videofoundry.co.nz/ianman/laboratory/research/driverlist.php

so, would you say (at least visually) you get decent output w/ the single cell?

I don't mind! I say there is a bit of decreased output with only 1 cell, but it does work. It draws about 2A on a single cell.

ok, i'll email him.

thanks, for the info.