Well…. i think we use quite different hot dedome methods here
By hot dedome i ment i use a hot battery-electrical dedome method- i usualy use a realatiely bad Great Empire 6800mah battery( that is actualy 900mah): this way even fully charged it gives max 2.5 amps when direct connected to the emitter….i have to make a video though….most of times the dedome is nearly perfect btw
Alright, so you use a crappy low amp cell to power the emitter while the MCPCB is hanging in free air, wait till it heats up and pluck the dome off while the emitter is still illuminated? If so, do you need welding goggles to see what you’re doing while you lift the dome?
Something like that, its only i power the emitter while its on the table, but unter its PCB i use a paper to issulate the heat , then i power the emitter in series, 4-5 seconds working, 1 second off, repeating this 4-5-6 times( i use to that like i drink water, i damaged many till i found the way to do it like i should)…meanwhile i touch gently the dome to see if its melted, after the 10th try one will know when the exact moment to dedome is, the dome becomes hot and soft: then i dedome the emitter with 2 cuts ( the dot in the upper left corner is the starting point):
1. One fast cut but only to the emitters core from the side that is pointed on the picture:
2. A final complete cut , from the site pointed on the picture:
If its done under 10 seconds usualy results are great
If the emitter is on an alu PCB it heats up 2-3 times faster, so series in that case should be 2 seconds work, 1 off
Every VERY crappy battery will do the job, there are some Ultrafire batteries that give 3-4 amps - those aint good, they heat up pretty fast, the main idea here is to maintain a current 2.5-3amps tops…it can be done with a led driver ofc
So you are slicing the dome before lifting, creating 2 cuts 90 degrees apart (and away from the bond wires to protect them). Then lifting from the corner where you cut the dome. Thanks Mitko! If this is the method, I can manage the heat pretty well just after reflow onto the DTP MCPCB.
Does hot dedome removes silicone under bond wires?
I use gasoline dedome method but I am using medical grade gasoline, much purer than car gasoline and doesnt leave any smell but it needs at least 24h of soaking. Anyway it is very very hard to remove particles of silicone dome under the bond wires…
Great work Mitko, 176 kcd for a C8 is very very good. Was the reflector the stock reflector that came with the C8?
Usualy it doesnt, but 25-30% of times it does, the only sillicone that is left is in that place…but i actualy prefer that for the wires to be protected at some level
Yes djozz, only the glass is an AR one( it brings aditional 10-12% in the overal candelas, tested it- tails spring bypass also brings like 8-10%)
Those Ft AR glasses have issues - from 5 ordered 2 have several non AR-ed dots, i already created a ticket about that
The funny part is that that particular reflector wasnt perfect, it wasnt dusty but it has some muddy dots here and there- usualy Convoy reflectors are fine
The new host has changes btw, i usualy order 4-5-6 monthly, lately they use thicker glass and kinda longer reflector, compared from the first samples that i got months ago
[/quote]
There is only one XP-G2. I don't understand... I have no clue where you saw 1.5A - as was said, we never restrict our amps based on CREE's specs. Please try to find your reference - it doesn't make any sense. We've been using XP-G2's well over 2A since they were released - no problems.
Mitko try the Intl Outdoor C8 AR glass, it’s way better than the FT ones. Almost Nitecore quality to my eyes.
I have a Convoy C8, original one, with a de-domed XP-G2 I built maybe 2+ years ago - can't find my notes on it, probably a 7135 stacked driver, at the time, pre-FET period, somewhere after the Jurassic age 
, but can't find my notes on it. THinking it did bout 120 kcd.
FWIW, when taking an amperage measurement at the tail it’s important to have proper polarity. One might think the numbers would be the same either way, but I haven’t found that to be the case. The 5.50 that’s showing also has a negative mark to the left, meaning the contacts are reversed. Given proper DMM probe placement that one most likely won’t be 5.5A. You might check it out and see if that’s true in your set-up as well.
I hot de-dome in the light, at high amps, usually within 30 seconds. We all develop a style that works for us, interesting how many different ways there are to do the same thing…
I have already give the link to the spec sheet in my previous post. I am listing it here (Redirect Notice) again. It clearly says maximum current is 1.5A. Please let me know what I have missed.
There is only one XP-G2. I don’t understand… I have no clue where you saw 1.5A - as was said, we never restrict our amps based on CREE’s specs. Please try to find your reference - it doesn’t make any sense. We’ve been using XP-G2’s well over 2A since they were released - no problems.
[/quote]
Ok - you didn't miss anything. We happily ignore, and throw shame on the CREE max amp specs - it's what we do, it's what makes us happy 
. We've been totally ignoring them for so long, guess I just assume that anyone here on BLF knows what we modders do.
Yes, CREE states max amps - of course they do. We, on the other hand, do not take that as gospel, so we test on our own what the little LED diode can really handle. So, we establish our own maximums., and they are all much higher than CREE's maximums. djozz has been the unofficial/official best BLF source of the test data. Does it shorten the LED's expected lifetime and MTBF specs? Sure, of course it does, but by how much? Not really too sure. There's been some testing done by Mitko on that, but not on all LED models. Do we really care? No, not really. If it last for 3 years under common usage for a flashlight - that's great! It would be obsolete in 3 years anyway. ANother contributing factor is we use high quality, expensive copper MCPCB's that have a Direct Thermal Path (DTP) that keeps the LED cool. CREE's max specs are probably based on a poorer mount and housing environment that what we build.
Is CREE correct on their stated maximums? Of course they are - for their own concerns which is mostly for the large commercial use of their products, like in street lights, home lighting systems, etc. Regular failures in the field under stress would be a disaster for them.
Hope this explanation helps....
You (TomE) just reminded mi that I need to order more XP noctigon mcpcbs :bigsmile:
And some of those S2 2b m8 
I stacked nice ammount of those….unfortunately FT shipping lately is a nightmare
Ooh… a short form of mate… m8 Back in 3dmark forum we used that alot and i got used ot it
Thanks!
Tom:
Thanks. You certainly explained that clearly. I am more inside box guy would never think about doing some thing like that on my own.
Overdriving the LEDs, just like overclocking CPUs, are certainly interesting and can be rewarding. One thing I observed from what reported on the forum is it appears the overdrived XP-2G appears less efficient in the very high current region (~ or >3A). At low and mid regine, e.g., up to 1.5A (top of the spec.) it is very efficient. I tried it on an SK68 that I have, it can produce the same or maybe more lumens than XM-L U2 with 14500.
Do you think my observations are correct?