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

I'd love to hear about emitter current after you upgrade those bulb filaments 26ga wires.

Oh, and is it in fact an XM-L2? Hard to tell in your closeup pic.

-Garry

Mine is completely unbranded since I bought it from Shenzhen Wholesale. Shenzhen Wholesale
The light is now also available from Kaidomain under the brand name Courui 1600 xml2 or something vague like that. Kaidomain link
We haven’t had reports yet of this version but I suspect it’s the exact same light just with a name on it.

Also I’ve seen some pop up on ebay recently with the same brand on them, also available from willbuying.com. Check the original thread linked above to get some more details about buying one.

I can’t measure lumen output or even lux at the moment so I wouldn’t be able to give you any specifics yet. But I can tell you that in it’s completely stock form the light certainly doesn’t out throw my modded HD2010 running 5.5A. And I wouldn’t expect it to, (even when running at the same current) until that led centering ring has been shortened or replaced and the led is better focused. As I said there’s lots to improve on this light before it’s performing at it’s potential.

I will measure some tailcap currents when I have the light reassembled though and post them here.

The tailcap doesn’t really do anything except cover the back end of the light. It’s not part of the electrical pathway at all. The front threads are not anodized so locking out is probably not possible by simply unscrewing the battery tube a turn or two.

Also forgot to mention that the contact between the battery tube and the ground contact ring of the driver could be improved. I had some flickering and wavering of power when testing, think that’s because the contact board isn’t in direct pressure contact with the edges of the battery tube. Instead the board is simply press fit into the head of the light and makes contact ground contact on ly through those solder blobs you can see on the top side of the driver board ground ring. It’s really not ideal either, again I’d hate for someone to buy this light and expect it to great out of the box, based on the light I have here it probably won’t and will need some considerable tweaking and modding to get it to a good state.
Be warned, you’re in for a project with this one…but if you’re up for it it should be a fun project with a great final result.

Yes the light is definitely equipped with an XM-L2, sorry my photos were a bit rushed. No idea what output bin it is though, no way to tell and it wasn’t advertised as being a specific output bin emitter so I would think XM-L2 T6 probably. Tint is hard to say too since a fair bit of the warmer light from the edges of the emitter is being swallowed up by the centering ring but it looks fairly close to my XML U2-1C emitter. Would guess it’s either 1B or 1C.

Thanks for the clarification.
Of course lockout isn’t possible, I am a bit tired or whatever…

Yeah could give that a go, don’t know much about it though. I’d need some suggestions from the driver gurus to find the sense resistor and figure out the conversion values and whatnot. I’ve got some smd resistors of various values laying around that I could use.

Here’s a bigger shot of the driver board.

There’s actually not all that much on the board.
Any idea which the sense resistor is? I know there’s only two on there but I haven’t done this before :stuck_out_tongue:

The R200 resistor is the sensing/limiting resistor. So after you have measured the stock current we will see further.
Do you have a experimental power supply? You should test if it is regulated at all.

I am a bit unsure because the r200 Is just in the input path of the little transistor and I can’t see a trace from the source side going to the controller…
But it could be a sensing resistor as through the source of the small FET flows all the current, you should check if there is a trace from this point to anywhere to actually measure something…
Interesting that there is so much empty space left on this driver board, maybe this is just a cheapo variant of a more complex driver. Looks like if there could be another identical circuit on the empty space…

Ok cool thanks :slight_smile:

There's lots of promise if someone does a redesigned copper pill for this. Thanks for the teardown.

Thanks Linus - subscribed.

Hopefully my KD one will show up soon.

Yeah absolutely, plenty of space inside the driver cavity for some extra heatsinking mass and the threads in the head are considerably longer than the thickness of the shelf. Currently those threads make very poor contact to the emitter shelf.
I would definitely be interested in a copper pill for this light!

Ok I reassembled the light and I’m reading 2.58A across the tailcap “filament” attachment points on high.
Fully charged senybor 18650s reading around 4.16v after the test and it stayed stable at that reading for a good few minutes.

I didn’t take note of the reading on low because my meter doesn’t like the low frequency PWM current signal. It gives inaccurate averaging results.

I have 1 loose r200 (0.2 Ohm?) resistor of the same size as the one on the board and a roll of r500 and r100s but they are smaller, like half the size.
If I added the extra r200 in parallel to the sense resistor should I be seeing around 5.2A at the tailcap?

Cheers

Subscribed :slight_smile:

-Jamie M.

Ok another small update, just replaced the original tall centering ring with a nice flat one and as expected this has made the world of difference to the hotspot and beam profile.

The centering ring I used is one of those white plastic ones with the wings on both sides and not much more than a slightly raised section around the emitter square. It won’t work if the emitter is not already pretty much in the center and can take a bit of coaxing to lock the reflector down right in the center, but this is the type I’ve had the best luck with centering HD2010 reflectors and it’s worked superbly again here.

The emitter is now sitting almost completely flush with the flat shelf of the reflector and it’s now pretty much perfectly focused. The spot is really nice, even and tight with fairly sharp defined edges. I’m sure that’s the perfectly smooth finish of this reflector that’s providing the definition.
It’s much more defined in fact than a nicely focused HD2010 and the spot is about 20% smaller.
So yes this light will out throw a HD2010 if given the right treatment :slight_smile:

Here is a direct side by side comparison between my best HD2010 running at over 5A on copper and the Big Head running at the measured tailcap current of 2.6A. I think I need to measure emitter current because this thing is keeping up with and maybe already surpassing the HD2010 in perceived lux! :stuck_out_tongue:


Distance to wall is 2.5m, Bighead on left and HD2010 on right.

Such a shame it didn’t come like this, would be so much easier to recommend if the beam profile was nice right out of the box…

Wow, talk about flawless beam!!

Subscribed for outdoor beamshots :slight_smile:

-Jamie M.

Thanks for the tear down. Great pics and summary of the light. Now I don’t have to do it myself - just the build log. :slight_smile:

I am slightly disappointed with the visible PWM, as I would like to use the existing driver. It is good to know that there is enough space in there to use my own driver though down the track.

Regarding the resistor mod, from the LD-29 thread , if you were to simply replace the R200 resistor with an R100 resistor, you should see an immediate double in current going into the LED. I plan to do this for the XM-L2 version light I will build. For the XP-G2 version, I plan to stack three R120 resistors and put them on top of the existing R200 to get around 4A from the tail cap. Both lights I am building will be de-domed.

Note that, the above figures are using your own numbers. To reduce resistance throughout the light, I intend to use higher gauge wire in place of the existing wires, along with copper braid the springs, as you have suggested.

edit: I am hoping that this light can dissipate the heat coming from XP-G2/XM-L2 being driven at >4A.

edit 2: just looked at your pics of the pill again. Man, is there a lot of space in there. Thinking out loud, me thinks soldering copper on the bottom of the pill should help with increasing the heatsinking mass, therefore, help with dissipating heat.

http://www.taskled.com/h6flex.shtml

$40 driver in a $35 light? Sure, why not. Since when did any of this make sense anyway? ;)

Being a parallel setup, doesn’t it mean that the light is only seeing 4.2V? If that is the case, the linked driver would not work as it requires a minimum voltage of 7V.

It is a massively nicer beam than when I started and the HD2010 should be a good comparison if you know what that beam can look like.
But I should state that beam shots like these can be a bit deceiving. They don’t look quite so flawless in person unfortunately.

In person you will see more shall we say “artefacts” around the center hotspot. This is usually in the form of like a flower petal type halo around the center hotspot, as well as sometimes a faint ring just at the far extent of that flower petal. This stuff is comparatively so much dimmer than the hotspot that the camera with it’s limited dynamic range fails to show it accurately. Our eyes pick it up though.

It all depends on how well focused and centered the emitter is in the reflector as well as the reflector type and quality but if you combine a smooth reflector with a square emitter those artifacts are always going to be there to a greater or lesser extent.

So both these lights exhibit these effects when looking at the beam on a white wall, on the HD2010 with it’s softer hotspot transition it’s only really noticeable when spinning the light and watching the center of the hotspot. On the Bighead the flowerpetal is a little more obvious, I suspect simply as a side-effect of the enhanced definition of the hotspot itself. The ring is also slightly more defined. But this is all with just a quick 5minute focusing so that may still be improvable by further tweaking.

The HD2010 took considerably longer than this to get to the “perfect” focus and alignment…like hours… :stuck_out_tongue: :slight_smile:

Outside you’ll never notice either of these artifacts so ultimately it doesn’t really matter.
I just want to say that if you expect a buttery smooth hotspot when white-wall hunting, you’ll have to settle for less throw and go with an orange peel reflector!

Cheers

I wonder if the 5a ramping lck led driver could be peesuaded to fit?

from the quoted size, I’d say so but it depends on the positive pcd. comfychair has one so perhaps he could give us a measurement to compare to the stock board?

Yep plenty of space in the driver cavity. You’re going to struggle soldering anything to the aluminium shelf. But you could simply bolt a copper or aluminium slug to the bottom of the emitter shelf/pill and try to size it (or alu/copper foil wrap it) so it’s a press fit into the driver cavity. Then you have some more heatsink mass on the bottom of the emitter shelf and an extra thermal path from there into the walls of the driver cavity/head. Kinda like I did here with my Apex 5T6 here that light had a similar issue with a “floating” driver shelf/pcb.

I’m looking forward to seeing what people come up with to rectify the problems with this light! :slight_smile:

I measured the inside diameter of the stock driver board ground ring at 32mm. That’s the gold ring on the top of the board, the inside diameter of the driver cavity is probably ~34mm/35mm going from that measurement. I’ll get accurate measurements for this when I have the light apart again tomorrow.
So it might just work…

Personally I’m not going to go for multi cell setups on this light, it’s parallel and I like it like that :slight_smile:

1.504"/38.2mm OD, .562"/14.3mm thick. This one also needs more than 4.2v, and also won't run anything but a "3 volt" LED - it's not a buck driver despite the input voltage range.

Of course. Sorry I didn't add all the required health and safety disclaimers! NOTE: SOME ASSEMBLY REQUIRED, TOOLS NOT INCLUDED, WARRANTY VOID IF REMOVED