X-ML de-doming method with 100% success?


Only attempt to de-dome an LED if you are willing to do an emitter swap or scrap the light! Expect failure on your first de-dome procedure and treat it as a practice run. To maximize chances of success stick to the methods and the appropriate tools suggested

I tried to make a very brief summary of the main methods of de-doming that have been contributed to this thread. Sorry for not linking to the contributors as yet, this is a preliminary summary that I will flesh out later this week when I have time.

The three main methods of de-doming are as follows:

Heating the LED:
The LED PCB can be heated by an iron, stove, or the LED driven at a high current, to the point where the dome softens and can be pried off. Sealing the bare LED afterwards is recommended. Risks involve damaging the LED with heat, dislodging the LED from the PCB when the solder melts, damaging the LED bond wires as well as removing phosphor from the emitter surface. This is the “original” method of de-doming and the most popular.

Heating the LED dome:
An iron or other hot implement is used to heat the dome uniformly until it is soft enough to be pushed off, in the direction of the LED bondwires. Sealing the bare LED afterwards is recommended. Risks involve damaging the LED bond wires as well as removing phosphor from the emitter surface. This method has not received much testing as yet.

Slicing off the LED dome:
A very thin, very sharp blade, such as a razor, is used to slice off the LED dome a millimeter or less above the phosphor layer. A spacer/shim for example a washer of appropriate thickness can be placed around the LED to act as a guide for the blade. The motion of cutting should be a single push or slice to allow for a smooth clean cut. A dremel or fine grit sandpaper can be used to polish of the surface afterwards IF needed. Risks involve nicking the LED with the blade, getting too close to the LED and destroying it during polishing. This method provides seemingly less reliable results than removal of the entire dome provides.


I’ve stumbled across an easier way to de-dome an XM-L that should yield close to 100% success if one is careful (maybe its been discussed before but I didn’t see it). I just dont know how much light is lost in the process since I didn’t have the good sense to take before and after beamshots.

Anyway The method is to take a razor, like this (the sharpest, thinnest thing I could think of) and just slice the dome off as close to the surface and possible, leaving maybe 1mm or less of material. Mine actually came out quite untidy but the beam still had no artifacts, though this probably affects brightness a bit.

I did it to a UF 3mode dropin. The beam is about 50% smaller and more intense, with a slightly warmer hotspot. Another pro of this method is that the thin layer remaining above the phosphor means that there will be no degradation of the LED over time through oxidation etc. My previous attempt at de-doming ended up ripping off some of the phosphor, which resulted in a purple bulls-eye in the hotspot.

It would be nice if someone with a luxmeter could replicate this and check the results.

Sounds very interesting. I might try this later this week with one of my lights. If I break the emitter it gives me a good reason for buying a new, brighter one.. ;)

This sounds interesting but honestly? I am just not sure I have the b@lls to do it.

Interesting. I’m tempted to try this. Do you have any pics?

Just tested it with my FF STL-V6 and the results were (arbitrary units):

Original Dedomed
Ceiling bounce 47 40
Throw 5000 6170

=> 23.4% more throw with a simple 2 minute operation.

(could have shaved it closer the emitter surface and cleaner and..)

That was quick!!! results look really promising…

Ledsmoke, MattSPL

I can post pics of the emitter tomorrow if you want,but there’s nothing special really. Cutting off the dome is a lot easier than it sounds, you don’t need to cut it off in a perfect single pass, you can cut it off a bit higher then “carve” off pieces till you get to the desired level.

edit to add: One thing that surprises me is that the ceiling bounce indicates only a 15% light loss from this method of de-doming… I was lead to believe that de-doming caused something closer to a 25% light loss, great news.

Nice! I've been thinking of sanding down and polishing to dedome one but this way is FAR simpler. I can't believe I didn't think of it myself. :p

I wonder what a difference it would make in an HD2010?

Ok thanks.
Is the dome solid, or is it hollow once you cut the top off it?

So will actual output decrease, but throw should increase?

yep 15% drop in lumens, increase in throw. What remains of the dome is just a thin sliver acting as a flat lens.

JohnnyMac, well after cutting the dome down its left quite rough, so maybe polishing might help after that?

Ok thanks :slight_smile:

It's solid. And yes output will decrease when dedoming an emitter.

Ok, thanks. I might give this a try on my spare Trustfire Mini-01 and see if i can turn it into a pocket rocket thrower :smiley:

I wouldn't be so quick to say output decreases. Identical drivers and emitters in different hosts, flood & throw will yield different ceiling bounce levels. Flooders score higher in bounce tests because they scatter more light than a thrower.

A stock domed xm-l spreads light 125 degrees. Cutting off the dome focuses the light reducing ceiling bounce numbers but increasing lux numbers for throw. The lens on an emitter doesn't increase the lumens produced, it just refocuses them. If I'm wrong I'm sure someone will be quick to correct me and explain why.

Well from what I understand, the dome on an LED not only refocuses the light, but allows more light to escape from the emitter that would otherwise be reflected backwards. This is also the reason why the tint becomes slightly warmer after de-doming.

Ok, just cut most of the dome off the xm-l in my trustfire mini-01.
Any suggestions for polishing it a bit?

Preliminary results are: hotspot is smaller and slightly brighter to the eye.

Anyone up to de-doming a TN31? :bigsmile:

How about someone with spare XR-E’s and a lux meter measure the lux from a cheap thrower (Smallsun ZY-C10-S for example) and then cut the dome off and remeasure? :slight_smile:

I’ve gotten rid of most of my cheap experimental lights or I’d do it.

I was planning a 4A driver and an aspheric mod on my STL-V2. I might add the de-doming to the To Do List.

What are the side effects of this procedure, except the (said) lower lumen output and warmer spot? Will it suffer more from the 4A current than it would without de-doming?

Looks like a lot of people have successfully dedomed their lights. I might have to dedome my stl v2 soon :smiley:

I think dedoming isnt good for an aspheric light, because when dedomed more light hits the reflector or in this case the inside of the head and not the lens which is what you want.