XM-L Custom Cooper PCB - M@g Mod with 44500lux Hot-spot

eprom

Newly Enlightened
Joined
Apr 17, 2007
Messages
146
Location
Turkiye
Hi Friends,

I have used Kai 14.2mm XM-L PCB and I got max. perfomance with 1.75A drive current. I got 25000lux with this setup. Over 1.75A lux readings decreases. So I tried to made my costom PCB with 0.3mm thick Cooper sheet and Scissor. Result was perfect. With 4A (AMC7135) drive current First 5min lux reading was over 44000lux and after 30min test lux reading was already over 41000lux.

With Kai's 2.8A AMC7135 Driver Hot-spot was 35500lux on High(2.78A), and 3350lux on low(150mA)

PCB Schema

CustomCooperPCB.1.png


Hard Pressed with ArticAlumina to a HA22 Heatsink

CustomCooperPCB.2.JPG


Thermal Transfer to Mag body is Perfect.
 
Last edited:

Walterk

Enlightened
Joined
Jan 21, 2010
Messages
755
Location
Netherlands
Most interesting!
I assume under the Led there is just void space to make connections electrically isolated?

Could you please measure Lux @ 1m without the optics?
 

ma_sha1

Flashlight Enthusiast
Joined
Jan 10, 2009
Messages
3,042
Location
CT, USA
Eprom, are you the guy posted lot of DX driver testings on DX?
Your work is excellent & much appreciated!

DX is selling a Copper PCB version XML, I just got one, the led has a gap,
I'll have to re-heat it probably. But looks like real copper.

4A is not necessary need copper, I've driven the 20mm XML cutter version to 4.6A in my Triple XML mod. The brightness increas was huge from 2.8A to 4.6A.
Techjunkine has driven to 5Amp in alu. PCB.

your drop after 1.75A might have more to do with Kai low quality PCB & not necessarily material Alu. vs. Copper.
 
Last edited:

eprom

Newly Enlightened
Joined
Apr 17, 2007
Messages
146
Location
Turkiye
I only put a little kapton tape under the + side (which is small cooper part) for electrical insulation. There is only ArticAlumina under the big part of the cooper PCB.

With bare emitter I have read 230lux @2,78A

wish helps.
 

eprom

Newly Enlightened
Joined
Apr 17, 2007
Messages
146
Location
Turkiye
Thanks ma_sha1,

I trust your experience and I will give a try to DX Cooper Board. Kai sell the first 14mm PCB so I got it but it is useless over 1.75A.

Did you try to trim DX Cooper PCB to 14mm?
 

ergotelis

Enlightened
Joined
May 31, 2007
Messages
734
Location
Greece/Hellas/Crete
masha, i think that cutters' 14mm pcb for currents over 1,5amp is useless too.You might be seeing a difference but i am pretty sure that there is a drop in performance after a few minutes. This is what i have concluded from my tests.
Also, what do you mean that the led has a gap in DX copper pcb?THere is no optimal heat transfer?I have ordered some of them and wait to get them.
Eprom, brilliant idea for you mod, but isn't it difficult to solder the led in that way with two pieces of copper?
Also, a HAIII heatsink might be better for your tests, but i think that HAIII have worse thermal conductivity...
 

ma_sha1

Flashlight Enthusiast
Joined
Jan 10, 2009
Messages
3,042
Location
CT, USA
Thanks ma_sha1,

I trust your experience and I will give a try to DX Cooper Board. Kai sell the first 14mm PCB so I got it but it is useless over 1.75A. Did you try to trim DX Cooper PCB to 14mm?


Sorry guys, I have to make a correction. The 4.6A triple XML mag I made used cutter 20mm Star. & the 14mm Star so far I only drove to 2.8A, haven't gone above that yet.

I haven't used the DX copper star board yet. It's noticibly thinner, maybe 1/3 of the Cutter 20mm board thickness. The led has a visible gap to the board on one side, if this is not fixed, it'll perform worse than Alu. star.

I got this thinking driving to 6A, but it doesn't look that capable, I'll need to fix it, just haven't got around to do this.

I would rate this DX copper star equal to Cutter Alu. 20mm after fixing it by re-flow again, yes it's copper but too thin. If used "as is", I would imagine this one will perform worse than cutter 14mm star. The re-flow job on mine is pretty bad & the gap will trap more heat before it even gets to the star.
 
Last edited:

onetrickpony

Enlightened
Joined
Mar 10, 2011
Messages
262
Quick comment on aluminum versus copper: From everything I've read about high end heat sinks, the combination of aluminum and copper is to use the metals in their best capacities. Copper is better at soaking up heat, absorbing it from the die. This is the reasoning behind the copper star. Aluminum is better at releasing the heat to the atmosphere, which is why a nice big aluminum sink under the copper is ideal.

So, if you can fix that copper star so that the led is properly mounted to it, you'll suck the heat away from the led faster than an aluminum star. The thickness of the star is practically irrelevant, you just need to make sure that it is super flat so that all of the surface can be used to transfer the heat to your aluminum sink.

Look up heatsink designs on wikipedia or whatever source you find more trustworthy for a more in depth explanation. Keep in mind too that every time you have two parts that need to be connected you lose some of the efficiency of the thermal path. I like stars for their ease of installation and somewhat more predictable results, but I'm sure an led mounted directly to a big slug of copper that is threaded or otherwise bonded to the body of a flashlight will be able to run much cooler due to the lack of bottlenecked heat transfer at connection points between different components of the heat sinking system.
 

eprom

Newly Enlightened
Joined
Apr 17, 2007
Messages
146
Location
Turkiye
Aluminium PCB's Anatomy is like this

PCB_Layer.png


Problems are,

* On cooper layer if the cooper is too thin, it can not transfer the heat efficiently
* On cooper layer if the layout is too small, it can not transfer the heat efficiently
* Insulater layer is always problem. It has a high termal resistance but this layer must be there for electrical insulation and bonding the cooper layer to aluminium board
 

mudman cj

Flashlight Enthusiast
Joined
Dec 14, 2005
Messages
1,827
Location
Where corn and pigs are grown unimpeded by trees
Quick comment on aluminum versus copper: From everything I've read about high end heat sinks, the combination of aluminum and copper is to use the metals in their best capacities. Copper is better at soaking up heat, absorbing it from the die. This is the reasoning behind the copper star. Aluminum is better at releasing the heat to the atmosphere, which is why a nice big aluminum sink under the copper is ideal.

That does not apply to this case because this type of heat sink does not dissipate heat through convection. Copper beats aluminum for heat transfer when it comes to conducting the heat through the solid material.
 

StefanFS

Flashlight Enthusiast
Joined
Oct 12, 2006
Messages
1,262
Location
Silicon Road 1, Sweden
Great news on overdriving, must test!

About the standard pcb's. I did a test with a Mag 1D on one D-size LiIon with the charger connected to the charging port to get constant power. XM-L T6 1D from Ledtech.de reflowed to a trimmed KaiD 14 mm pcb driven by an AMD 7135 2.8A driver. Starts out with 33000 lux @ 1m. The led and pcb is covered with a thick layer of glow powder & epoxy. After 30 minutes it's 31000 lux.
 

eprom

Newly Enlightened
Joined
Apr 17, 2007
Messages
146
Location
Turkiye
Stefan, I specially put the LED 1mm off center on one axis to get an "oval" and "bigger hotspot" so comparing lux values wil not be possible. I know your builds are error free.

Edit: And I am using XM-L T5 2B from my Dist.
 
Last edited:
Top