A Perfect Dedome?

I did them dry.
It was a while ago so I don’t think the gas trick had been invented :slight_smile:

Thanks guys. Think I’ll try the soak first, then pull if needed.

i’m going to have to try gas on an xr-e. I’d LOVE to get a clean/no-ring de-dome.

2x the current of xp-e…should be quite a thrower

I have used the heat method with near 100% success. Fairly sure given 100 XML or XM-L2s I could do every single one of them without failure now that I have the technique down.

I have done approx. 2 dozen XMLs (+ two more this morning), an XP-G2 and a XP-E.

I find it easiest to do if the LED is already mounted on a MCPCB in a torch. From there I remove any screws or glue etc and lift the board so the board is “floating” in the air with the leads still attached. I then run the torch on high until the board is too hot to hold.

Placing my figure as far down the dome on the opposite side of the bond wires I peel towards the wires slowly: you will notice if the heat is not right the dome is harder to peel off. If doing this for the first time have a play with peeling the dome at different temperatures and try get a feel for how hot it has to be. You can reheat the LED at any time during this process.

The result is almost exactly 110% increase in lux for reflector lights.

No solvents are needed. I just let my led heat up a bit by running it away from a heatsink until it gets good and warm, not burning up hot just hot.

I then stick the razor into the dome a bit adjascent to the bond wires and open it like tupperware. Basically on an angle and peel the dome off. Had good luck with this method and if any goop is left, the clump stays by the bond wires!

Well, it turned out a not too great test, actually just two old xml-leds and two solvents: petrol and petroleum-ether. Not much news here, but I do have the pictures! . Have a look at my first 'chemical' dedoming session (powered by this thread):

So here are the victims, two spare cool white xml's:

The first thing to do was cutting the silicone away around the cone. can you see the difference?:

So now the soaking begins, the left bottle is petrol, the right is petroleum-ether:

Suddenly it is 24 hours later , something has happened there under those domes:

So now the dedoming can take off, I even made a crappy video of the first one, it is the petrol one. It is a somewhat messy business this dedoming, no fun. I use a binocular microscope to see what I am doing, the video stops after the dome was removed from the die, it does not show the cleaning up of debris around the die later:

The second dedoming went better, the petroleum-ether seemed to have loosened the dome a bit better than the petrol did. After some cleaning up this is how the leds came out, left the petrol-treated (dirty stuff!), right the petroleum-ether one:

And they both still work, just tested them up to 50mA, but I guess they would be ok in a flashlight. For fun I made pictures of the right led above enlarged and at 1mA current:

So the chemical way worked for my first two tries! I believe I am confident trying this on brand new leds as well .

Got a “This video is private” error… Please show us your privates!

sorry, fixed that, there they go

The few times I've tried it, that's what mine look like too except not at 1mA, but at full power |(

I am the Dedome King ! Watch this (in HD ):

Not bad for a very first attempt at hot dedoming , I am a happy modder.

This is what came out, no further cleaning needed:

And it works (at least up to 50mA, the picture shows 1 mA):

That's amazin stuff! I guess you could reflow, then de-dome in one fell swoop! I assume that's an XP-G2 - you are a better man than I -- I never would have attempted a de-dome technique the first time on a XP-G2.

“This video is unavailable” error…

I did reflow and with the led still hot dedomed it, the video is one shot (and perhaps a lucky one ). The led is a XP-E2 R3, 1D tint.

It is going into my 'budget test-thrower' (49mm reflector, led @2A), just to see what an extreme pencil beam will look like . I may do a beamshot and some lux-readings if it gets finished this weekend .

I am not sure what happened, I made it public this time. Anyone else can't see the video from post #87 ?

I can see it

ok, I was inspired by the post of zizo and quickly finished the mod mentioned above. Just two pictures, hope you don't mind, Tom? (I do not want to derail this thread):

reflector 49mm, dedomed XP-E2 R3 1D-tint on copper, screwed on pill, arctic silver in between , driver LD25 2.0A, tailcap:2.0A. Ceiling bounce lumen estimate: 290 lumen. Throw (measured at 2 meter,3 meter and 5 meter, same results): 170klux@1m (165klux after 5 minutes), functionality: zero, this is an utterly useless flashlight.

beamshot:

I don’t know about that. Looks like it would be handy trying to find buoys out on the water. In that situation, flood reflects off the water and boat and blocks your far view. That light you built is sweet. Looks like you could see buoys way, way out there.

Funny you mention that, I have done that many times during sailing boat races at night here in Holland, but that was 10 years ago. At the time we had big maglites for that (very useless), and lead-acid battery powered big halogen lights, those were great, but lasted 7 minutes on a battery. And now we have this in our pocket . But still an aspheric would be even better for buoy-searching because of the total lack of spill

djozz - I don't mind at all and it's not off track, that's the next phase, what are the results and what do we do with these de-dedomed emitters?

This is probably the thread to note that

I’ve wasted >$100 trying to dedome mtg2s

You CAN shave the dome off with a razor, but the phosphor is thicker than you think, and the throw gains are not as high as with other emitters (neither are the output losses).

Petrol - no go, all the phosphor will come off the center of the emitter

Same with acetone.

If you try the ‘off heatsink heat method’, the dome will bubble at the emitter and pull the phosphor off…and it takes a @#$% load of power to do that…these things can handle some serious current…