I thought about doing the same but replacing them all or slicing the 5000k
There were several here who said that you could simply pry the dome off the 519A when the led is hot and that you are left with a clean glass surface. I tried and the dome came off but the split was not nicely at a silicone-glass surface, it left me with still a messy bumpy layer of silicon on top of the phosfor. So I went back to the slice method to remove more silicon (edit) and get a flat surface, then all was good (it is in my SP10 Pro now). So apparently YMMV with dedoming these.
I imagine their must be a “Goldilocks” zone for that to work consistently. How much “heat” for how long before it affects the state of silicone directly above the phosphor/glass opposed to the peripheral portion? I’m no physicist, but I would think it all would have to be tempered equally to get a clean tear off.
glad it cleaned up easily…
Ive been getting very clean, down to the glass, dome removals, from just warming the light up very slightly by running it for a few seconds. I use tweezers to pinch the dome, which causes one edge to lift up, and then I just lift and peel it off with the tip of the tweezers.
I still use a plastic tooth pic to remove any stray bumps of dome left near the edge of the die… but over 90% of the dedome is clean, down to the glass.
.
whatever works, counts
my first attempt, with a cold LED, had the most stray bits of dome remaining adhered… but they clean off easily…
Euhm… pretty sure it’s silicone you were removing 
Fwiw, I’ve never been able to remove the dome with the glass completely clean. Always have to do some touching up to get it squeeky clean. Unheated btw.
Since they’re on completely different channels, there is no concern with mixing LED’s for the LT1. In contrast, LED’s with different forward voltages on the same channel in parallel would cause the lower voltage LED to run brighter.
So it really depends what you want on each channel.
If you stick with the 5000K LH351D, you have a good balance between brightness and CRI, but probably a slightly green tint (per Maukka’s measurements of the LT1). If you swap to 5000K 519A, you take a very slight (imperceptible?) brightness reduction, get a slight increase in CRI, and get neutral tint.
If you slice the 5000K LH351D, you’ll probably a little under 4500K with close to neutral tint (per thefreeman’s measurements of various slicing heights) and a slight decrease in brightness. A sliced 5000K LH351D or a dome-on 4500K 519A will likely be very similar all-around. I’d personally favor the 519A in this case - slightly more neutral, slightly better CRI, but the the LH351D doesn’t become worse simply because you have 519A on the other channel.
If you want rosy tint at all color temperature mixes, a dedomed 519A would be the way to go, but because the color temperature drops so much for this LED when dedomed, even starting with a 5700K would likely end up below 4500K. I think if I did this, the result would be that mixes of the warm and cool channels would end up too rosy for my taste.
The slicing shouldn’t change the light distribution from the LT1 noticeably. If you look at the datasheet for an LED that has both domed and dedomed options from the factory like the XP-L (datasheet page 16), you will see a slight difference in the relative spatial distribution at moderate angles, but the more important difference in the distribution you get from an optic is the difference in the apparent visual size when there is a dome changing the apparent source of the light rays.
I saw this 519A dedoming video shared on Reddit yesterday:
Cheule says the method recommended to him by Clemence is to first vertically section the dome into quarters with a knife (the glass protects the phosphor), then pry it off laterally.
He tries a couple variations on the technique in the video, and still had a little cleanup to do on some of them, but it seems like something that a person might be able to do very consistently with a little practice.
My impression is the goal is a mix of minimizing setup (no washer/shim), reducing cleanup of residual dome material, and avoiding risk of pulling up the glass cover and any phosphor.
Good info in there, appreciated!
Personally I don’t do tint mixing with the LT1. It’s either warm on lowest brightness as a nightlight or cold high brightness when I want to see as much as possible. So any rosyness overkill is probably not a problem for me. What worries me is that the narrower beam after dedoming is too distracting projected on the diffuser. It’s not just the LED specs, the distance to the diffuser also comes into play. Even with the stock LH351Ds you can see the difference in distribution quite good. I just took a photo of both CCTs to illustrate what I mean. And that’s just the distance affecting it. Only way to be sure is actually try it though.
dedomed 519a is more pink than e21a
.
Nice! Was hoping for this. Have a bunch of 3000k on order.
Keep them coming jon_slider, I love seeing these shots with measurements. Big thanks!
Jon this is fantastic comparison! Great work keeping us up to date here!
thank you all for the encouraging comments
stock 219b 3000k is more Rosy than stock 519a 3000k:
.
Just gorgeous, Jon! Excellent work. Thank you for sharing. Makes me so glad I have some of these 3000K on the way. Artie is going to make the RRT-01 how it should have been… you turned me on to it, also! Thank you for that. It’s been great to carry… Ready to get rid of the 6000/6500K… 
Great work Jon! However, all the E21A 2000k I tested are pretty much spot on 2000K with DUV at pretty much 0, could it be the optic or reflector of the light you put it in shift up the CCT?
My Opple is probably different than whatever instrument you are using. I dont expect my values to match other peoples, which is why I do side by side measurements for two lights, so my Opple bias is consistent.
glad to hear my Opple agrees with your duv measurement… Ive never seen 0.0000 before this
the light I tested is using an aspheric lens, I do not know if it is affecting CCT… it could be the output level that I chose, that is influencing the difference in CCT of the E21a compared to your tests…
I think your data is based on a more expensive tool than an opple… so it is probably more accurate. These variations in numbers, is why I like to include the two side by side beams, so we can see any CCT and Tint differences visually, in addition to numerically.
You can remove the lens and test the bare emitter. E21A is very close to 2000k
That is a very informative video. I watched it last week and cringed when he sliced down on the dome with an x-acto knife.
I was able to pull off the dome with my fingernail at room temperature. I was surprised as to how easy it was to do! The glass over the phosphor was nice and clean (no clean-up necessary). I just removed one of the triple domes to give it some rosiness, but not too much. It brought the overall CCT down to around 4000K.
What about dedomed 519A efficiency? Any changes? If dedomed more warm and rosy, so efficiency should drop?
This also isn’t quite apples to apples, as it’s comparing different CCTs. Lower CCTs are always less efficient in lm/W.