For work I still fav 5600-5700K. Outdoor 3500K. General purpose 5000K
But this new 6500K R9080 is so amazingly clean, whites are brilliant while the rest of the colors pops. Perhaps I can tolerate this one for my office space. Never liked anything above 5700K before.
- Clemence
OK guys, here’s the beamshots:
I used Texas_Ace’s order for showing purpose here. I’m sure he doesn’t mind 
With the new VR16SP4 board, cleaning the flux residue has never been easier.
Using commonly available 20mm TIR optic for Cree XML, the footprint loosely fits the holder. Although it works OK, for extreme power application this holder base has to be removed (unless you can carefully lower it down to slightly below the phosphor height). With faceted optic like this slight off focus won’t be too noticeable but the most important thing is to prevent the opaque holder from getting burnt. At 4A and beyond the side spill was hot enough to burn the neighboring LEDs (in previous gapless prototype).
Texas_Ace’s favourite CCT: 4500K. The combination of sm403 and sm503 should be close enough for his taste.
Control shot: Cloudy morning in Bali
15cm from the wall. If there’s any artifacts or weird beam shape, it should be obvious at this distance
Driven at 1,4A
Vertical direction
- Clemence
Yes it will fit most 20mm based TIR optics. But to get the most OTF lumens, the optics’s aperture should be 7mm (the diagonal of 4,8mm x 4,8mm is ~6,7mm).
Any optics with 6mm will work OK since the dies in this quadtrix setup don’t extend all the way to the edges. The losses would be minimal since E21A has narrower emission angle than Nichia 144Ax or Cree XHP50, thanks to the reflective side barrier.
Light Blue = Cool white E21A
Orange = Warm white E21A
Red = 219C
Blue = 144AM/R
- Clemence
I can’t wait to try out those LED’s, I have an optic very close to the one you are showing with the dimpled surface that I plan to use. Although it doesn’t use the white base, so it should not have any issues with burning. Although I might have to glue it in place to keep it from damaging the LED’s.
I wish I could find some optics with that dimpled surface but with a 20-30 degree beam angle, right now the best I have with those dimples are 60 degrees. The dimples work fantastic at mixing the tints in my experience.
Bought this, they’re on the way.
https://www.aliexpress.com/item/High-quality-High-power-LED-Lens-CREE-XML-T5-T6-U2-Lens-Diameter-20mm-60-degrees/728091430.html?spm=a2g0s.9042311.0.0.MYJ5co
Not optimal but a little aperture widening should works.
- Clemence
Do you think that mixing sm303, sm403, sm503, sm653 on VR16SP4 would be nice? I’d like to have something between 4000K and 5000K. Maybe 2*sm403, 2*sm503 is better idea?
Thanks for the heads up, that is what my 60 degree looks like now and it is really nice, just a little too floody for my tastes. If the 20 degree has the dimpmes that should be perfect, even the 38 degree would work.
I am thinking about using this with an XHP50.2 90CRI in my new EDC, or maybe one of the nicha 6V emitters if they are give enough output with these new mcpcb’s. I forgot about that actually, PM inbound.
Please scroll up Damian, those example were 2xsm403 + 2xsm503
I’ve seen that, but no 1*sm303 + 1*sm403 + 1*sm503 + 1*sm653 configuration 
:person_facepalming: how could I see only 403 and 503 in your post?
Anyway, CCT mixing can be predicted by simple mathematics. But you have to put output into the equation too. In an ideal case where all the emitters have exactly the same output and forward voltage, the resulting CCT would be very close to the average number. But unfortunately in your combo, each have different output. Sm303(D200), sm403(D200), sm503(D220), sm653(D240). Average them first then use the 200:200:220:240 ratio to approximate the resulting CCT
- Clemence
Thank you. So 3000*200/860+4000*200/860+5000*220/860+6500*240/860=4721K. Ideal Order 10146 
Hi Damian,
Please specify the configuration as well. Use the LED config diagram in the product image
For example:
-LED1=303
-LED2=653
-LED3=403
-LED4=503
Wiring:
3V/6V/9V/12V
Thanks,
Clemence
Wiring would be 3V for both boards. About the led config… I don’t know, your example looks good enough. For the second - all 503. Thanks.
Got the 3K 4K and 5K nichias premounted on the quad boards. Had just a bit of time to light one of each, and definitely need more time because I am a bit confused about what I observe.
Just a quick question: Clemence what solder did you use on the contact plates? I had spent half an hour trying to solder wires to these plates and only half of them were eventually soldered securely enough. The rest just keep snaping off. The generic solder I’d used for years just would not adhere to the plates, I am really confused.
Are you using a high-wattage soldering iron with wide tip? The low thermal resistance of the dialectric means that the pcb is soaking up most of the heat from your soldering iron.
Correct, you have to apply a LOT of heat to mcpcb’s to get them to solder properly in most cases.
I usually solder around 380c but for mcpcb’s I crank it up to 450-500c depending on how large the mcpcb is.
I meant this specifcally for this pcb. Soldering the wires to it is much more difficult compared to normal DTP PCBs.
yeah, I was too, although it depends on which mcpcb you are talking about. The GT mcpcb is a special kind of difficult to solder wires to lol.
This is the first thing that came to my mind. I use a regulated iron which I set at 300C and when I kept it attached to the PCB for some time, the blob of solder did attach to the plate and looked like a normal blob. But then it just snapped off as if there was dirt underneath. Which was not since I used soldering acid that had worked every time before that.
I am not sure what wattage is this iron, but I used it with an attached aluminum plate for reflowing an LED on another board and it did it easily.
But OK, I’ll try higher temp. I think the iron can go up to 450C.
So far I am very happy with that single 3K. It does have some tint shift, but for a warm tint this is irrelevant.
The specific 5K I tried was a benchmark mule emitter. On par with my trusty 083 but with the potential for a much higher output. Just need to confirm the observation with the other three 5K LEDs on the board.
The one 4K I tried (out of four on the board) had bad tint shift, that nasty green corona which I hate in 219B/C. Not that much of it, but still unacceptable. This was so strikingly different from the 5K that I was really puzzled. But it looked like the tint shift was different in different directions, which may have to do with the remaining three unlit LEDs, reprocessing the light of the lit LED.
Regardless of the tint shift, the on-axis tint was excellent in all the CCTs. No green, no magenta, but I’ll need to double check this in a well illuminated ambient light conditions. My little workshop was not like that today.
So next thing will be to test the consistency of the tint shift patterns across the samples.
Yep, loads and loads of heat is needed. I even bought a new tip for my 80W iron especially for these virence boards.