Built: The Everyman's Triple XM-L 3-mode NiMH Optic

Techjunkie

Enlightened
Joined
Nov 16, 2007
Messages
943
Location
in the brightly lit suburbs of NYC (Long Island)
I haven't built or modded a light in a good long while... until today. I had parts laying around for three new builds since ~January, but couldn't muster up the energy to start any of them. Now that I've sold a few from my collection, and I'm trying to sell a few more (trying to finance a new bicycle purchase), I figured it was now or never to start these builds. This one I'll probably keep for myself - for a little while anyway.

Specs:

3 XM-L cool white stars
Cu Perfect Sink-D machined flat.
Triple Optic from DX sku 1916
Blue Mag 2D host
3 AccuPower Evolution LSD NiMH C Cells
3 2.8A AMC7135 regulators, each regulating 1 LED
Primary regulator on which the other two are slaved, is set for Lo-Mid-Hi

TripleXM-Lwithoptic-Headinnards.jpg


TripleXM-Lwithoptic-Assembledhead.jpg


What's amazing about this "everyman" flashlight (uses NiMH cells, no need for fancy Li cells), is that on three cells that were in my drawer since January and haven't been charged since then, it's every bit as bright as the three mode triple SST-50 light I built that requires 3xAW IMR26500 cells to run.

Beamshots...

What follows is an indoor shootout of my collection of multi-super LED lights - aka, The Floody Buddies.

All shots taken at ISO200 F3.5 1/10sec WB=5200K

The Floody Buddies (angle)
Floodybuddiesangle.jpg


From left to right:
1) The Lil Giant - Mag 1C, 3 Neutral White MC-E, MOP Triflector, 1 AW IMR26500, DD
2) The LED USL - Mag 3D, 5 NW MC-E, SMO Pentaflector, 16 Duraloops, 0.45 Ohm resistor
3) Mag 2D, Triple 4500K SST-50, SMO Triflector, 3 AW IMR26500, 4.2A 3-mode linear regulator
4) Mag 2D, Triple CW XM-L, DX Tri-optic 1916, 3 NiMH LSD C cells, 3-mode 2.8A linear regulator 3-stack

The Floody Buddies (faces)
Floodybuddiesfaces.jpg



Control, lights off:
01Controllightsoff.jpg


Control, lights on:
00Controllightson.jpg


1) 3x NW MC-E, MOP Triflector
02LilGiant3NWMCEMOPTriflector.jpg


2) 5x NW MCE, SMO Pentaflector
03LEDUSL5NWMCESMOPentaflector.jpg


3) 3x 4500K SST-50, SMO Triflector
043NWSST-50SMOtriflector.jpg


4) 3x CW XM-L, DX Tri-optic
053CWXMLDXtri-optic.jpg
 
Last edited:
Very nice Techjunkie! This build looks very clean. Definitely want to see the beamshots. How is runtime in regards to heat buildup? Is the heatsink contained entirely in the head or does it protrude into the body as well?
 
Awesome build!!!!

Close to what I am working on currently! What are the advantages to running those 3 drivers vs. just 1 like most do? Also, any details on this part (Cu Perfect Sink-D machined flat.)? What the source was or where ya got it?
 
Nice simple build Techjunkie! Are you getting ~8.4A at the tailcap? I'm suprised those Accupower C LSD cells can handle that kind of load without sagging too much.
 
Very nice, especially the heatsink :thumbsup:

Are you heatsinking the drivers? I read in your Yard Sweeper-Thread (again, very nice :naughty:) that the drivers need to do quite some work when powering the XM-Ls because its Vf is so low.
I'd very much like to do something similar to a 2C (though with just 2 XM-L, but it will be a flooder as well), but I am still thinking about how to power the beast. 3 4/5-Sub-Cs + LEDs in parallel + resistor(s) (low mode achieved by using 2 alkies :rolleyes:) or like you, 2 linear drivers, what do you think? The former sounds a bit dangerous to me, considering the strength of these cells. I don't want to torture the LEDs ...
I'm still curious about the heatsink, too ... copper would be nice, but aluminium should do the trick as well for just 2 LEDs, right?. However, this not much of a problem.
 
I want to do one of these!! No machine shop of course... would you sell a heatsink and optic alone, along with a wiring diagram? Did you use Shiningbeam drivers?

Edit: I PM'ed you.

Rich
 
Last edited:
Very nice Techjunkie! This build looks very clean. Definitely want to see the beamshots. How is runtime in regards to heat buildup? Is the heatsink contained entirely in the head or does it protrude into the body as well?

The PES (perfect emitter sink) threads directly into the head and the lip mates with the ledge inside the head so heat transfer is very good and it has excellent mass and thermals. At full power (2.8A per LED), it heats up quickly in the hand. I haven't done a full runtime test yet to see if it can go a full discharge cycle on high, but I expect it to perform better than the triple SST light I referenced above, which has hotter 5A emitters and less sophisticated heatsinking. The bottom of the PES protrudes into the neck, but doesn't make contact except where the bottom of the PES lip bottoms out on very top of the tube when the head is screwed down fully.

I owe beamshots, I know. I'm actually very pleased with this optic. I tried the same triple SMO tri-flector on the XM-L as I had used with the triple SST-50, but I felt that the output of the optic was superior. That was fortunate, because the metal triflector was too tall for the bezel to screw down when paired with the PES D-Sink. I would have had to trim the ledge down inside the head to allow the PES to screw down even further and that would be difficult and messy. I'll use the tri-flector I modded to fit the mag head in another build some day.

Awesome build!!!!

Close to what I am working on currently! What are the advantages to running those 3 drivers vs. just 1 like most do? Also, any details on this part (Cu Perfect Sink-D machined flat.)? What the source was or where ya got it?

The three separate regulators are required because the LEDs are not in series. Basically, there's three sets of regulator+LED in parallel here. I provided some links to the PES D-Sink info at the very bottom of this post. See below.

Nice simple build Techjunkie! Are you getting ~8.4A at the tailcap? I'm suprised those Accupower C LSD cells can handle that kind of load without sagging too much.

Those are the best NiMH cells I've ever used, and the XM-L has very low Vf. I only measured 7.5A at the tail, but the cells were rested 7 months before I tried 'em. Also, experience with the SST-90 testing showed me that the tailcap provides less resistance and higher current than the conduit I use in my clamp tests at the tail. Unfortunately, I didn't clamp test at the LED lead wires this time - I was in too much of a rush to finish assembly.

Very nice, especially the heatsink :thumbsup:

Are you heatsinking the drivers? I read in your Yard Sweeper-Thread (again, very nice :naughty:) that the drivers need to do quite some work when powering the XM-Ls because its Vf is so low.
I'd very much like to do something similar to a 2C (though with just 2 XM-L, but it will be a flooder as well), but I am still thinking about how to power the beast. 3 4/5-Sub-Cs + LEDs in parallel + resistor(s) (low mode achieved by using 2 alkies :rolleyes:) or like you, 2 linear drivers, what do you think? The former sounds a bit dangerous to me, considering the strength of these cells. I don't want to torture the LEDs ...
I'm still curious about the heatsink, too ... copper would be nice, but aluminium should do the trick as well for just 2 LEDs, right?. However, this not much of a problem.

The regulators are bonded together with high temp "PC Farenheight" epoxy putty and heavy copper wire was used on the outside edges of them to aid in thermal transfer, but the truth is that they do not need to be cooled nearly as much as with lithium solutions because there is very little voltage overhead to burn off. Remember, unlike the serial build with three 4.2v cells in series and three 3.2 LEDs in series (huge potential difference), here we have three LEDs in parallel and three 1.2v cells in series. The 3.6v pack requires very little work of the linear regulators. Basically, they're only there so they can be PWM'd for multi modes. I could have as easily used a single PWM only circuit, but it would have to be one that could handle 8.4A. This was easier and cheaper and more proven.

I want to do one of these!! No machine shop of course... would you sell a heatsink and optic alone, along with a wiring diagram? Did you use Shiningbeam drivers?

Edit: I PM'ed you.

Rich

The driver is exactly like the Shiningbeam driver and the one that Download sells, but I got 'em from KD for ~$5 each. The optic is from DX sku 1916. The machined PES D-sink, I bought from this thread here, but the Aluminum version is still available from the Sandwich Shoppe, here. (They also sell a C sized version.) I don't have a machine shop either. The only grinding involved was a little beveling of the bottom edge of the optic to allow its standoffs to contact the top of the PES with the PES screwed down all the way. It was floating by a fraction of a mm before that and I didn't want to sacrifice heatsinking at all by even slightly unscrewing the PES.

As far as wiring goes, you simply remove all the non AMC7135 components from two of the regulators, and strap the VDD from the first to the other two with very thin wire. Then you wire a ground to all three regulator grounds, and a positive from after the switch to the one positive input on the in-tact regulator that you didn't remove any components from. All three LEDs get a positive in, and each has it's negative paired with the LED- of its own dedicated regulator. In that way, each regulator regulates a single LED, but all three are pulsed at the same time for Hi-med-low by the PWM circuit on the first regulator. (Solder the third star to ground to set the mode selection to H-M-L.) I don't have a diagram, sorry.
 
You got an extra optic you'd wanna sell?

Thanks for the info-I had to get whatever the SS had left of heatsinks, and it ain't much. So I hope mine'll work.

So with the SB driver I: Extract all components on two boards except the 7135's and parallel wire battery + and ground? Where do I pick up the VDD?

Rich
 
You got an extra optic you'd wanna sell?

Thanks for the info-I had to get whatever the SS had left of heatsinks, and it ain't much. So I hope mine'll work.

So with the SB driver I: Extract all components on two boards except the 7135's and parallel wire battery + and ground? Where do I pick up the VDD?

Rich

No, you just parallel the grounds. Only the first board with all the components gets a positive input. You then solder wires from the VDD signal input leg of one of the 7135 chips on that board to the same leg of one of the 7135 chips on the other two boards (you can use the solder pad exposed by removing the IC as your contact point on the secondary boards). When you're looking at the 7135, and the three legs are pointing up, it's the leg on the left.
 
OK, cool-last thing-LED+ and - gets wired independently from each board, correct? Also, where can I find the Accupower cells?

Thanks for your patience! Providentially I just found my 3D host that I thought was lost for good-so I guess it's meant to be!

Rich
 
OK, cool-last thing-LED+ and - gets wired independently from each board, correct? Also, where can I find the Accupower cells?

Thanks for your patience! Providentially I just found my 3D host that I thought was lost for good-so I guess it's meant to be!

Rich

LED- gets wired from board 1 to LED1, from board 2 to LED2 and from board 3 to LED3. All three LED+ go directly to + from mag switch.
 
OK, someone make a drawing before I go sledge this goddamned thing... because this ain't working. Already burned up one driver. I need this laid out in front of me-I don't learn by words-gotta see it in pictures. It's 0330 and I'm totally pissed. No wonder I've stayed away from complicated setups. Never again.

I'm thinking I'm gonna toss the whole thing, I can't get how the positives are hooked up, and I'm totally confused trying to put it together from all these posts. One says each LED gets a +, another says they all go to battery +. So I'm supposed to have four wires, one from each emitter and one from the first driver +?

Or should I just say eff it and use one driver stacked with multiple 7135's, and have done? Better that than the trash.

Rich
 
Rich,
We've all been at the breaking point with a project at one time or another. Just walk away from it for now till someone can step in and lend a hand. Eventually you'll get it finished and it will be great. Hang in there.
 
OK, someone make a drawing before I go sledge this goddamned thing... because this ain't working. Already burned up one driver. I need this laid out in front of me-I don't learn by words-gotta see it in pictures. It's 0330 and I'm totally pissed. No wonder I've stayed away from complicated setups. Never again.

I'm thinking I'm gonna toss the whole thing, I can't get how the positives are hooked up, and I'm totally confused trying to put it together from all these posts. One says each LED gets a +, another says they all go to battery +. So I'm supposed to have four wires, one from each emitter and one from the first driver +?

Or should I just say eff it and use one driver stacked with multiple 7135's, and have done? Better that than the trash.

Rich

A picture's worth a thousand words:

TJs3-mode3NiMHCrecipie.jpg
 
Last edited:
Beautiful! Thank you very much! I'm gonna go back to it later on when I'm in a good mood.

Very much appreciated!

Rich
 
Well, I rewired everything exactly like the drawing. I'm using the Tenergy NIMH D cells from my TK-70, 3 of them. I checked everything, and I don't get any 7 amps. Hardly 3A. So this strapping/slaving thing does not work. Pending an answer, I'm gonna tear the whole damn thing apart and piggyback one driver and have it done, I've really had it.

So much for my sigline...

Rich
 
Last edited:
It appears that's not your fault though :p. You are only powering the main board (2.8A) if you have connected it like the schematic above.

Thats because I think the schematic is incomplete :). The schematic shows two wires connected to the slave drivers. The PWM wire from the master board and the current wire to the led. What is feeding the slave boards with power? They still need to be connected to a power source as the master driver board.

If I'm wrong I'll edit this post, because it will only confuse people :x.
 
Well, I said the heck with it and paralleled the drivers the way I know how, along with the LED's, same way I wire my XP-G triples. Everything worked but I still only got 3A. So I tried 4 cells. With NIMH that still falls within spec for 7135's, and now I get insane brightness... draws ~1A on low, 2.8 on mid, and a good 7A on high with no angry blue or tint shift. So I guess it works! <g> A hack job, but it works. Wanted to use 3 cells though-4 is getting heavy for night walks.

Thanks everyone!

Rich
 
Congratulations Rich, now post some photos in a new thread with beamshots!
 
Top