Why does a perfect dedome work well?
If I understand DrJones correctly, even a flat dome should reflect less light back towards the phosphor than no dome at all. Reflection of perpendicular beam is down from 18% to 10% (6+4). TIR angle is the same. So total reflection with a shaved dome should be lower, producing better output and smaller tint shift. But numbers here, here, here indicate otherwise.
Can someone explain why is it so?
I have never seen a “perfect” shaved dome. So there is more lumen loss with a shaved dome because the top is no longer as smooth and perfect as with the dome on. I imagine there is still some scattering of the light because of the imperfections in the “top” of the emitter. Not that I think a “close shave” de-dome is bad, because done properly it is still a good way to de-dome.
That was my first thought too. But @relic38has sanded the shave with 2000 grit paper. I think it should be OK…
Another thought is that in a perfect dedome there’s still a thin layer of silicone left.
Or maybe there’s some more fundamental reason.
Dunno…
Just killed a LatticeBright “XM-L” by driving it directly from a 3.5+V battery while submerged in white spirit. Gave it a few good zaps a few hours ago, can't say for sure if there was some apparent damage mostly due to my unwillingness to look at the thing while driving it.
Gave it another zap a few minutes ago, and it just died.
I have an old genuine XM-L next in line, but now I am in doubt with regards to submerging it with wires attached.
2013 manufactured cool white XM-L is enjoying a white spirit bath, with attached wires. I've given it some current, but the alu board doesn't even seems to get hot to my finger. Maybe the mia LatticeBright “XM-L” was already “cooked”, it came from a friend's headlamp whose emitter I swapped.
Going to raise the battery voltage to increase my diver's driving current. :-D
0K, at this point I believe the LatticeBright XK/XL I was bathing maybe died from overcurrent.
Estimating:
Test battery: two worn out King Kong INR26650E cells in parallel; let's say ≈40mΩ of internal resistance.
Multimeter and its leads, ≈200±50mΩ of resistance.
Emitter wires: ≈½ a meter both positive and negative, AWG24; 86.56mΩ.
This at least ≈277mΩ (≈277mV/A) from the measuring setup. Checking the current flow to the submerged XM-L revealed from 1.5A to nearly 2A, with a no-load battery voltage of only 3.54V! Boldsheesh fellows!
Test XM-L may be getting ≈3A when the multimeter is removed. Remember my testing is being done with my led's bathing fluid at room temperature, this means Vf remains higher also (excellent cooling).
I can see many of you may have killed a good deal of emitters this way because of hot bath (less Vf) plus high starting battery voltage. Seriously.
My old XM-L was left in the white spirit bath, waiting for nature to take its course. I extracted the board a few days ago and the dome nearly fell by itself, leaving just an unimportant tiny leftover where the bond wires are, yet not over the die which is… completely clear! (ᵔᴥᵔ)
I know have an XM-L2 in the bath. The good thing with white spirit is that it does not seem to completely evaporate at room temperature, of course because of the longer chain hydrocarbons.
It is fun to observe the white spirit slowly doing its magic and giving it some stirring just for the sake of it. I can clearly see a curious optical effect in the emitters while immersed: the die looks unscaled, like if it were dedomed. As time slowly goes by it can be observed how the white spirit diffuses under the dome, progressively soaking the die under.
If anyone is interested I believe this slow cooking method may be worth a try with the newer generation Cree emitters.
Yeah, it’s apparently the toluene that does the job best.
I.i.r.c. the nitro thinner contains toluene, unlike regular cellulose thinner.
But toluene is hard to come by (and i think it’s rather toxic too).
That’s why nitro thinner is presently the go to solvent, but it’s not easily available in every country.
I can order a liter online, but it’s some $15 including shipping cost.
Home Hardware also sells a lacquer thinner made up of toluene, methyl alcohol and methyl ethyl ketone, that seems like a pretty nasty combination.
Mem talked about his method of dedoming being a two step process, attack the glue first then swell the dome off. I can’t figure out if he was using different chemicals or wetting the base of the led for a while and then submerging it.
@Tom I also couldn’t find Nitro thinner. I read somewhere that MEK worked, so I started using it. The SST-40 dedomes nicely with it. The couple of XML2 U4-1A’s I did in it had the usual tint shift.
Readily available at HD. I have only tried it hot, so not sure how well it works cold.