Teardown and Mod Thread of Courui XML2 Aka "Big Head 3x18650 Side-Switch Thrower"

Yeah should be possible if you do it like that, although you may run into issues with battery clearance once the contact ring is high enough to press against the battery tube.
I tried adding some tall solder blobs onto the battery side of the contact ring where the little ears are. Just to try and increase the height and make contact with the battery tube edge, but after a certain blob height the batteries started hitting them when screwing on the tube…and I still wasn’t making contact with the battery tube…the gap on my light is pretty big it seems but ymmv.

The emitter shelf is the main bottleneck really, I wouldn’t drive the light above 4A without taking some serious time to beef up the thermal transfer of that thin emitter shelf. Even at the roughly 3A stock it’s really not adequate…

The heatsink/driver section of the light isn’t particularly massive or well finned as you point out but it does have a decent thermal path from there into the fairly massive battery tube through the raw aluminium threads. Thermal transfer forward into the rest of the head isn’t ideal because the threads on the two part head are anodized, but the combination should still suck a fair bit of heat out of the pill and dissipate it pretty quickly.

I don’t see any real problem with the kind of power you need to dissipate from a single XP-G2 or XML even when driven at up to 6A it should do fine.
Provided the pill is improved anyway.

I just retested this one with 3x 20Rs in parallel. It, well, sorta works. It's not entirely happy with the voltage, it makes a truly wicked squeal from around 40% up to 80%. It's fairly quiet at 100%, but still making some noise. It only sounds not too bad because of how loud it is before it gets to 100%.

Output is 4.10-4.15A, shorting the big sense resistor on the top board only bumps it up to 4.30A. It is much happier at 8.4v input and above, though I don't think it would be unsafe/unreliable at 4.2v. The squeal is from the driver itself, not any of the connections or other components. No squeal at 8.4v.

How wide is the top circuit board? That may be the biggest hurdle to getting this to fit.

1.31"/33.something mm.

The inside surface of the lower board has components out to within .070" from the outer edge, so the ledge the driver sits on can't be any wider than that.

I love this beast :slight_smile:
on the left is the big head with dedomed stock emitter flowed onto noctigon, replaced wires, braided springs, r100 sensing resistor and smaller centering plastic
on the right is t08 dedomed xm-l u2 1c on noctigon, replaced wires, some resistors stacked up…
and damn! this thing throws :slight_smile:

also I turned the emitter plate upside down to be able to focus better (get the led deeper into the reflector and create some pressure on the board) and pasted it a little - then tightened it really well
but I screwed up the stock driver a bit, it doesn’t turn completely off :frowning: measured ~16ohms between L- and ground and couldn’t find where was it coming from :frowning: oh well, it’s a nice nightstand light when “off” :smiley:

good to see some more of these lights in the hands of modders!
The hotspot it produces really is very nice once focused, dedomed it must be a beast. Nice one!

Did you manage to get a reading of tailcap current after the r100 resistor change?
Cheers

not yet, I’ll try to get a good reading this weekend across the tailcap
the hotspot is really nice, with a bit of flares
it got a brighter ring around the hotspot but that doesn’t bother me that much and I will not mess with the focus no more

Pretty much what I am planning to do with mine. The XP-G2 version will be using either an R120 or R150 resistor. I am not entirely too sure yet.

How are you finding it handling the added heat? Do you think it’s necessary to add some more heatsinking mass to the LED plate/shelf?

[quote=Rod911]

On my light I would say yes it’s pretty much necessary for higher output. Definitely for overdriving an XML to the max. Maybe not for XPG but the stock solution is really not ideal.
janko.hrasko how was the tolerance of the pill shelf threads on your unit?
Mine was very loose, as in I could wiggle the shelf easily and by quite a large amount when it was partially threaded on and in the focus position, that’s the worst thing about the thermal transfer I think. Very few threads for a small contact surface, and loose as well. Putting thermal grease/paste on the threads helps but not enough really.

the shelf was very loose on mine in the focus point too… that’s why I flipped it upside down to tightly screw it against the rim underneath it
in stock there was a recess in shelf in which the led was sitting… now it’s higher on the flat part of the shelf and even the reflector is pressing on it from the front
the heat transfer seems fine to me, it’s getting pretty warm but not hot even when I’m not holding it - the whole massive body is working as a heatsink

Yeah nice idea to flip it over, that’s probably going to result in the best solution using the stock components! :slight_smile:

Is there any chance you can take measurements between the bottom of the LED shelf once it is screwed in and the top most driver components? It looks like the clearance is around 30mm, but that is a total guestimate.

Sure, I’m stripping the light down again in a bit so I’ll take a bunch of internal measurements of everything

Ok here is what I came up with.

top of emitter shelf from top of heatsink when in focus ~2.7mm
thickness of emitter shelf 3mm (3mm-0.85mm for pcb cutout)
-
depth from top of driver board to emitter shelf rim 27.5mm
depth from emitter shelf rim to electronic switch port ~12mm
driver cavity inner diameter 34mm (33.98mm)
-
emitter shelf thread O/D 41.75mm
I/D of emitter shelf threads in body 41mm
depth of emitter shelf threads in body 7mm
-
thickness of driver board 1.6mm
height of driver at highest component (including board) ~12mm
-
diameter of flat base section on reflector 27mm

So Rod, if you want to know exactly the distance between the bottom of the emitter shelf when it’s focused and the top of the driver board.

It’s…

distance from top of the driver board to emitter shelf rim (27.5mm) + depth of emitter shelf threads (7mm) - gap from top of threads when emitter shelf in focus (2.7mm) - thickness of emitter shelf itself (3mm).

And if I didn’t mess any of that up it comes out as 28.8mm, so your guesstimation was pretty damn close :slight_smile:
If you’re wondering about the size of the copper slug you could fit in there though, you’d have to take into account the height of the toroid on the driver (around 11mm) and the electronic switch port which is only 12mm down from the edge of the emitter shelf rim.

-

Easiest heatsink enhancement I think would be a 34mm X ~13.5mm slug (copper/aluminium) lapped and bolted to the bottom of the emitter shelf. Then drilled to accomodate the two emitter wires.

Starting with a 35mm bar it should be fairly easy to sand the O/D down a bit and also remove the anodizing on the inside of the driver cavity walls (and open it up a bit as well) to get the combination close to being a press fit.

At least that’s what I’m going to be attempting I think, I have 35mm aluminium bar stock at hand and it should be doable without heavy machinery…I hope :stuck_out_tongue:

Better would be to machine a whole new pill out of copper with threads and everything but that’s definitely out of my league, and I’m pretty sure a lathe is mandatory for something like that… :wink:

Many thanks for the info. I am hoping the heatsinks which I ordered from FT arrive within a couple of weeks. A copper slug of around 10mm in height might do the trick if I can find said slug in a hardware store.

Thinking out loud, the plan is to place the LED shelf upside down. With it upside down, the area where the LED is supposed to occupy, there will be a 20mm copper slug in its place. The surrounding area, I will place heatsinks which are 5mm in height. I will have to rely on the threads to transfer the heat collected to the body.

Also, with the shelf upside down, it should focus the LED better as suggested by janko.hrasko as it is sitting up closer to the reflector rather than the default position.

The problem with this approach is you’re not really doing much to improve the thermal path away from the shelf and to the outside. It will work well initially to suck and store heat from the led but the real weakness of the shelf is not really it’s thinness but that it can only transfer heat to the body through it’s small and loosely machined threads with minimal contact surface. Upside down or not the contact area is still pretty minimal and that will be the main bottleneck.

The idea with adding a 34mm slug that fits down into the driver cavity is that this will both collect and transfer the heat out along it’s own contact points to the driver cavity walls. Even if the press fit isn’t optimal you can increase/improve the contact area very easily with a few layers of copper or alu foil.

I’m pretty sure that will do a much better job at keeping the led cool than relying solely on those threads on the emitter shelf. You’ll see when you get the light, they’re pretty suboptimal…to put it kindly. :bigsmile:

Incidentally how easy/hard is it to remove anodizing on a specific part of a light without destroying the rest. Like could I paint on the degreaser/greased lightning only to the inside of the driver cavity and let it do it’s thing or would just sanding the coating off be a much better idea?

flipped shelf isn relying solely on those poor threads… it sits tightly on the rim and with some thermal paste there it isn’t that bad :slight_smile:
sure copper pressed really tight would be better but I don’t think that it matters to me that much to have someone machine it for me

Hi Linus,

Thanks for the review.

Can you tell me what the outermost diameter of that heat sink plate the LED sits on is?

Sure, that’s what I’ve called the “emitter shelf thread O/D”, probably could have named it better. It’s 41.75mm.

I got tailcap lockout!
just unscrew the tailcap, unscrew the b- pcb, take it out
screw the three screws that were holding it to the body

place the pcb OVER the screws

get something that can create pressure on the pcb from the tailcap (I used a lid from film canister) and place it over the pcb

screw the tailcap on and BAM! :slight_smile:
when you tighten the tailcap, it pushes the pcb on the screws and completes the circuit
loosen it a bit and the pressure from the springs breaks the contact

Nice and simple, good job! :slight_smile:

Update: I ordered a custom firmware driver from DrJones (mokkadrv - 4 mode) that I’ll be using in this light.
Probably won’t do much work on it till I get this driver, too many projects on the go and I really need to get my BTU-MTG2 monster project finished up. :slight_smile:

But I’m keen to see what others are coming up with in the meantime so feel free to post any progress or ideas in this thread so we can have a collection of useful info on this light.

If anyone has any questions or wants any measurements let me know, the light will stay disassembled for now.