Very nice , need to make me one of those too 
Comfychair 
Show off.
lol nice. What springs are you using?
The plate used to be 5/8" copper pipe, slit lengthwise then hammered flat (I think it's .040" wall?). Springs are the 'A' version from IOS, the big mean ones.
Your just using them as pictured or you put braid on them as well?
Last night melting those two was a pretty big scare, smoke and all… Especially since one of these is for a buddy I want to do all I can.
Cereal_killer, I normally buy those springs in 10-packs from CNQ/fancyflashlights.. Im not sure about max amps.. But I think they go soft/melt/overheat at around 7amp... I have melted some of these springs when testing them out.. Copper braid and I have not been able destroy them at 12A (highest current I have had through a single spring). Cooper braiding is a must mod on lights like this anyway if you want best possible performance out of the batteries. I normally prefer stock springs that have been copper braided on SRK lights. Those springs are quite stiff are are not as suitable for different lengths of batteries..
braid??? BRAID!! We don need no steenkin’ braid! 
Hello! I’ve been mod my SkyRay Clone, now that I have much free time. I changed the wires of the LEDs for a silicone 22awg. Now just need to increase the output of driver.
this would be a solution?:
or would not help? :~
No, you do NOT need to increase the output from the driver, not until you replace those aluminum MCPCBs with copper ones.
thank you comfychair! with this would be enough? http://www.ebay.com/itm/5PCS-20mm-Star-Copper-Base-Board-for-Cree-XM-L2-XML2-XML-XM-L-LEDs-LED-/221243386040?pt=LH_DefaultDomain_0&hash=item33832228b8
I need some more component to increase driver output? or with the bypass is OK?
Get sinkpads or noctigon, they have no dielectric (isolating) layer in the heat path.
I have been looking for information on the dielectric layer and do not understand that is necessary for this mod. I can not use copper base without dielectric layer?
wanted it to be a simple and cheap mod and” I read this thread”: 6x 20mm XML-ledboard comparison and I think the little difference in performance between copper base and sinkpad is not necessary for my. I can do the bypass resistors to check only the luminous difference? I will not use the flashlight until I put copper base plate on leds
+1 - those copper boards are pretty useless over standard aluminum ones. Usually we refer to copper boards as either SinkPAD or Noctigon. There's been some tests done lately that show aluminum boards that have no blocking dielectric layer perform very close to copper SinkPAD's or Noctigons, but the SinkPAD aluminum ones are the only ones I know of.
Those no-name copper boards from eBay can be dremel'd down to the copper at the center pad, then filled in with solder - probably not as good at SinkPAD or Noctigons, but it will work better, but I've found them arkward to reflow the LED's back on.
The problem is when you cut down the resistance, you will bump up amps, a lot potentially, and those LED's may not last very long because the heat runoff is getting blocked by the dielectric layer. Also you will get heat sag, which means the output will start dropping right away, sometimes turning the tint bluish.
Or you can do this and have similar results:
Crap - I read this thread way back, and totally forgot about it. Dunno if I can do this the comfy way - don't understand it fully, or how to do it - tools, etc.
Tom, I'm guessing you've got all of the necessary tools. All you really need is a drill, a few bits, hammer, file/sandpaper, and some time. It really is an ingenious idea. Unless I'm mistaken, these are the condensed steps:
- Drill hole(s) in thermal pad that are a pretty tight fit for whatever size copper wire you are going to use
- Chamfer/slightly enlarge the backside of the drilled hole so that the copper will have a "stop" to press into when it is peened into the hole
- Smash/peen the backside of the copper wire, leaving a small gap at the LED side so that it doesn't need to be sanded
- Sand/file the backside of the MCPCB/copper flush
- Reflow
- Enjoy!
Comfychair has come up with some of the most innovative "budget" solutions I've seen to date.
If you add a small chamfer on the LED side as well, the copper pin can expand on both sides and possibly lock in more securely. Probably doesn't make a bit of difference since once the LED is soldered back on top it's not going anywhere anyhow, but you know how these things are...
The copper pin should end up flush on the top if you hammer it from the backside, with the front of the board laid down on a smooth surface (and make sure to clean off all the solder left behind on the LED pads).
Can be used with much larger copper bar stock to join pieces together as well...
Ok - I may be lacking one tool though... A Brain 
It really is pretty simple I suppose and this is really incredible the more I think bout it. It would be great to have some emperical data to back this technique up, but really can't think of any down side to it. Based on djozz's bench tests of MCPCB's, aluminum does nearly as well as copper, so in theory taking a dielectric blocking alum star and converting it this way, should get it incredibly close to a SinkPAD's result. What I've seen when you use proper bonding techniques of LED to PCMCB and MCPCB to pill top mount, not only do you get more initial lumens, but it holds the output better over those first 30 secs, then going into first couple of minutes. I've seen improvements just from adding screws to lock down the MCPCB.
Big advantages will be for those 25 mm and 30 mm size stock aluminum stars, or sometimes copper ones with a blocking dielectric layer. Sometimes swapping in a SinkPAD or Noctigon creates more issues of fitting, wiring, clearance, etc. Also for those multi-emitter lights with the custom single piece MCPCB, as comfy did.
I have to see what I have, or can find as far as drilling holes for flat head screws - that's the first challenge. I got some thick copper wire, not sure though if it's thick enough though.
I mis-drilled the center hole and had to knock out the first piece I used. Good opportunity to show how it expands radially when squashed.
Wow, that is a beautiful process and result.