Wow thank you so much for trying this and supplying the measurement data!
It looks like CCT went up 800K (which is a lot!), while duv slightly increased, and CRI did not change meaningfully. The very high initial+final CCT is likely due to lack of a mixing optic.
I’m super excited for your future measurements with an optic, which is representative of the real-life spectrum. (The Convoy 15deg beaded TIR would be my choice too!) I’d also like to hear your subjective impressions of the CCT/tint/CRI!
I couldn’t wait. Hehe.
I installed the optic very carefully using fine hemostats, paying very close attention to not touch the new dome.
Most of the measurements seem to have maintained a similar % ratio with and without optic, aside from cct. It went way down with optic installed.
Before left, after right.
I’ll test again after another day or 2 of curing, at which point I’ll blast it with turbo a bit.
These emitters are really cheap and great for the price. $0.97 each on superbuy for bare emitters.
They are very yellow in person.
Similar to an lhp531 3000k.
Adding the dome really cooled it down a bit, and I don’t hate it.
Seems similar to lhp531 4000k, but higher cri.
I’ll report back after some curing time and beamshots to compare before and after.
That’s how I felt when I did it! Thank you for the new measurements!
The difference in CCT is a bit less now at +450K. Tint gained +3mduv, and CRI went down a bit. R9 dropped 9 points, and similarly did R12 (though not too surprising given the higher CCT).
In general, I suspect that this mod is more desirable for emitters that are either pink, or red-oversaturated, to begin with, and conjecture that doming these emitters could increase CRI/R9. For an emitter that already has too little red to begin with, doming makes the color quality slightly worse.
Thank you for these observations! I can’t wait for the high CRI LHP531 to drop…
Thank you for the beamshots, they look fantastic! The CCT has become very noticeably cooler, and subjectively it feels like it’s putting out a bit more light too.
No problem.
This emitter was the perfect candidate for this mod.
It seems the offensive yellow has been lessened quite a bit.
I think it’s just about right for me now.
If you look at the brick wall, you can really see the difference in color rendering.
Post dome looks to be much closer to what I see IRL.
I wish I had before pictutes with reflectors though.
I think I’ll try SMO and OP reflectors at this point, for science.
The red paint on the wall stands out much more in the domed beam! I also find green tint more tolerable at higher CCTs.
I suspect that the dome would introduce substantial angular tint shift in a reflector, which doesn’t mix all of the emission like a well-designed TIR does.
This is an interesting topic. It’s surprising to see adding a dome to a flat emitter is so simple.
Avoiding an overly rosie light is the reason I chose the FFL5009D 4000K in my Z1. It doesnt appear, to my newbie eyes, to have the rosie hews Foyfly is known for..
So I knocked the dome off the 519A in an m31 dropin, and the issue I have is that the whole dropin is pre-set, including the reflector to a particular focal height. That gives me some wiggle as to the dome height, but I wanted to ask: How viscous is the kafuter? Will it spill off the glass plate on the 519 when applying? If not, would it be possible to place a mini-hemisphere bead (like half scale to the original) directly on the glass plate?
Picture for reference…it’s not QUITE so bad to my eyes, but I tweaked it to highlight the issues in the beam now
As you can see in the upper right quadrant of the hotspot that blooms to the corona, the corona is definitely darker than the rest. I figure since I can’t push in the reflector, I’d re-raise the apparent image of the LED to emulate the extreme angles of the stock dome, i.e. 120º cone, to thicken the dome to the base level of the reflector, which would be roughly 0.3mm difference. I estimate the CCT to go up to around 3200k, but if it can clean up the beam a bit and put more lumens into the hotspot, that’d be great… what do you think?
The kafuter is significantly more viscous than water, but has poorer cohesion, which introduces a tendency to spread out rather than bead up when applied too thickly. I would start with a 2mm thick layer, which is not tall enough to spread under its own weight.
Alternatively, UV resin can also be used for the dome–it’s closer to water in physical behavior, and in particular has good cohesion.
Thanks for the photo! I think I see some silicone residue, which can certainly mess up the beam profile. You don’t even need to remove it–the Kafuter will smooth out the jagged surface.
We must have converged on the same Aliexpress shopping strategy of sorting by number of orders and then rating. This strategy has never let me down over dozens of orders.
Yep, that’s how I use AliExpress as well.
The only issues I ever have with AliEx is when they use Caniao shipping.
It’s 50/50 I get my package with that carrier.
Could be a local issue, but seems systemic.
Actually now I remember the old days when they used cainiao, and tracking would just randomly stop, and I’d get a random package 2 months after ordering. Fortunately nowadays they use SpeedX and GoFo for my address (New York City), which have been reliably fast.
SpeedX and Gofo are great.
I have a Cainiao package that went out for delivery on the 23rd.
It’ll probably never show up.
I’ve had that happen several times.
