Ok, you guys convinced me to go ahead & peel the rest of the dome off my T6 3C XM-L (in the craptastic SMO P60) I put the solder iron tip next to the LED, and used magnifier, tweezers, & X-acto knife to carefully peel it off. It never got hot enough to melt the solder. There is still some spooge around the bonding wires Iām not messing with, but the chip is entirely clear.
I was finally able to focus a decent hot spot with this reflector, though still ringy.
I am quite sure I left all the phosphor intact as this thing is really warm now. It looks to be 3500-4000K now, I donāt know, itās the warmest LED I have. If I do another, Iāll start with a U2 1A or such, and hope to end up at 5000K. This color shift is fascinating; somehow the average light must be traveling through more of the phosphor(?) And this warm shift continued when removing the last .35mm of the dome too.
I heat the board directly and its firmly attached to a heat sink. I heat the board quickly with high
heat enough to melt the adhesive holding the dom and i use tweezers to pull it off. Now your using heat as well, only to a different location and completely different approach. I cant say i would suggest this but perhaps if you try this method before putting your iron to the dome itself.
I havenāt killed or scratched one LED yet ( by this method) but i have killed plenty of diodesā¦
I fiddled on with that reflector positioning and now I get a steady 168 kCd after 30 sec and initially when turned on it gives about 183 kCd.
Now I wait for xml 2 led with even more impatience! Muahahaha!
EDIT: Corrected some strange spelling....
I think the XML-2 is very close: I read today in an electronics magazine that Cree stated the are expecting to release a 200 lm per watt LED before the end of 2012ā¦ call me skeptical (XP-G2 is something like 160 lm / W), but ever if they manage 180 lm / Watt it is a decent step up. I think the XML-2 would be this LED. 
dedomed sky ray A0 clone 85klux initial J)
Not bad from a little light.
I think the dedome method was from Andi: I heated up the led and the pill at 4A for a few minutes, and the dome came down very easily.
200lm/W is probably assumed to be achieved at pretty low drive current and ?CRI of color renderition 
Wow! I actually would have expected more out of that little bastage. Considering how well it throws normally I would have guessed at least 120k+ considering the numbers other lights are pulling. Iām betting the wide base reflector reduces lux when you dedome.
If this is even remotely true, it should be easy enough to test. Without the dome of maybe RI ~1.5, the output should drop in half, which is hard to believe given the results here.
Also the RI = 3.7 is of the semi material itself (transparent to the emitted wavelengths), not phosphor layer.
The focus of the dome is evident enough from its shape and >1 RI, and comparison between CREE app note and COS curve as already noted.
Well, de-domed a XML headlamp ( ~38mm reflector), while it still goes I removed a fair bit of phosphor from two of the cornersā¦ the pictures can explain better then I can:
Before:
After:
Left to right: P60 XPG at 1A, OP reflector, XML at 800mA in 38mm SMO reflector and Trustfire X9 at 2.4A ( batteries needs recharging)
Smaller hotspot and marginally more intense, still good for all flood as the colour is even (if slightly blue) without the reflector.
I pressed the LED dome lightly against a hotplate for about 10 seconds before pushing it upwards with a knife bladeā¦. I donāt think I let it get hot enough or perhaps I should have heated it from the underside? Going to replace the LED anyway now.
Anyone have specs for the XM-L U3?
Donāt want too much from this little light. Other lights have a (much) larger reflector in this topic where has been measured 120-180-200kcd. So, itās not fair to compare with them. With an original xm-l it can 45kcd maximum, and this increased to 85kcd, itās not so bad.
Not so bad??? Itās an outstanding value from such a little light.
85kcd is really really good result for a single battery light. Not to mention the $15 pricetag (without emitter swap yet of course).
Anyone doing this on Small Sun T13? Nice big reflector there, standard T6 XM-L ready to peel, and with the resistor mod it should really give good result.
I plan to do T13, but mine has not arrived yet. And even if it did, I don't have a way to measure lux anyway to publish result.
EDIT: Sorry, price was $22 shipped.
Except at 1 meter the hotspot is around 5cmā¦ Itās absolutely tiny. I did it to mine before accidentally breaking the bond wires trying to fix it.
My X9 is killing my STL-V2 now. I keep grabbing it when I head outside. Hard not to try de-doming the V2, but I like the beamā¦.
outdoor beamshots Skyray A0 clone, distance 200m
mouse out with dome
mouse over dedome
Nice mouseoverā¦ that pretty much nails the results I had as well.
X9 de-domed
Old lux 38,120kcd
New lux 61,920kcd
Old lumens 786
New lumens 575
Nice beamshot viffer750. Amazing the warm (er yellow) shift on these de-domed. Mine too. I est from 3C 5000K to 3500K (er yellow). IF I do another, Iāll start with a U2 1A and hope to end up at 5000K. Craig at Illumination Supply has some U2 1A on 16mm boards on sale for $6.50ā¦ā¦
You're absolutely right, Viffer. That little A0 clone is so small I forget that the reflector isn't as big as it's output makes you think it is. That is an amazing difference!
The C8s arrived!!! (5 of them :bigsmile: )
They cost me >$15US each and include aluminum reflectors, glass lenses and an overall very good fit and finish considering the price.
Three of the five have quite cold tints so I decided one of those needed de-doming. I used the Dremel method this time: ground the LED dome down to <1mm and stuck the felt wheel in before finishing with a optics cloth.
So far it looks very promising: When at 2.5A (battery at 3.8V) it throws as well as a X9 at 2A (3.6V) however the hotspot is 1/2 the sizeā¦. just recharging the batteries to put them both on a even footingā¦ photos to follow.