Good advice, I’ll wait for Richard.
I ordered a D4 last Friday from him. I had to settle for the xpg2 version because all others were gone already. At that time he said they would be getting a large number of units in various emitter versions in a couple of weeks. This is now a week later so I assume surely in 1 to 2 weeks his supply will be in. Go to his site and register for an email alert when they get in stock. I am planning on getting both a version with both xpl hi and a nichia model. Hope this helps
Also, Intl-Outdoor made the D4, and I think they actually build the lights themselves, so they are effectively the origin point. Everyone else will be a bit lagged because they’re not the origin.
This is pretty much the same thing is a traditional heat sink with active cooling. Did you ever see the one Old Lumens made?
Or LinusHofmann started one, but I don’t know if he ever finished it:
The heat sink design needs to strike a balance between surface area, thermal mass, and thermal conductivity. So, the fins or wires should be thick enough to conduct heat through the entire heat sink, thick enough to hold a useful amount of heat, and thin enough to offer lots of surface area. A scrubber might be too thin to hold or move heat, while a solid block doesn’t have enough surface area. CPU heat sinks seem to have a good balance.
I keep hoping for some technology crossover — all the best ray guns have used big cooling fins for years. Surely they’re applicable to flashlights.
How good is titanium at dispersing heat? Here are some:
http://signupsale.com/sale/gghuxs0f3iqu2xd3wglvzc7zvhk0lxox
Ooh, copper!
And the architect-type LEDs are coming with active cooling:
Though I admit, what I’d really like is water-cooled-with-steam-whistle ….
Did you ever see Rufusbduck’s 2013 handmade ray gun light?
I tried to buy it but he decided it wasn’t for sale. 
Uh, those aren’t cooling fins on the ray guns. They are high voltage insulators.
If so it comes off without leaving any adhesive. I have some on my R50 and it looks pretty decent.
TK wrote: quality matters
Yep. That message should be conveyed to every manufacturer who gets approached about building for a BLF group buy, too.
Good to see a clear illustration pointed out demonstrating why it matters.
Thank you.
Yup very true.
BLFers are flashlight enthusiasts. On average our knowledge about flashlights is much great than the non-flashaholic. We have high expectations for our custom BLF builds and any manufacturer who fails to live up could get roasted!
CRI is a measure of color quality. The higher the CRI (an average) and R9 (a test specific to deep red), the more vivid the colors of illuminated objects will look. This isn’t just aesthetic, as lower CRI can mean it’s impossible to see subtle variations in color that help your brain recognize and differentiate objects (is that a snake or a stick?). It can be more useful to have better light than more light.
R9050 means CRI over 90 and R9 over 50. R8000 means CRI over 80 and R9 over 0 (both measures can be negative, and R9 often is with LEDs). 100 is the maximum.
Thanks Zac for that explanation. I had been scratching my head in the dark 
I was interested in why this especially high CRI is only mentioned when talking about Nichia LEDs and never with Cree?
I see CRI ratings regarding Nichia in light manufacturers websites but never see any Cree led related CRI info.
I have S41 with Nichia and like the tint.
I also have Noctigon Meteor with S4 3D leds and tried to ask its CRI rating in the Meteor thread but I think that never got an answer.
Doesn’t any other of these mainly used leds other than Nichia have any good CRI?
Does anyone know how long does Hank process orders?
anyone tried Yuji LED they sell high CRI COB LED upto 100w i think. Most COB leds are just too pale.
Nichia was the first to offer a useful LED in high-CRI at 4500K or above, and produced some with unusually good tint. Cree’s high-CRI offerings have mostly been in warmer tints, and typically don’t look as good. Meanwhile, Cree was making the brightest and most powerful LEDs while Nichia lagged behind.
That may be a bit outdated now, perhaps, but the history still plays a big role. Now we have a 219c which has both good tint and high output, and Cree has increased their variety of high-CRI emitters somewhat.
However, Cree’s latest LEDs like XP-G3 and XP-L2 don’t seem as nice for flashlight use. Cree always had kind of a tint rainbow where a single emitter would have different tints at different angles, and it sounds like the latest models might be even worse about that. OTOH, Nichia has had consistently good performance on that count, with relatively consistent tint at all angles.
It’s kind of an arms race, really.
My favorites are the Nichia 219B 4750K 92 CRI and Cree 3D (~4850K) tint at 80+ CRI. They hit what I consider the sweet spot for neutral white. The D4’s 5000K 219c is pretty nice too, though a bit cooler than I’d prefer.
Thanks!
So there is much more than just the LEDs tint that makes the CRI.
Now I just have to wait for more stock and then try to decide what model to buy.
I understood that D4 comes with 18650 tube and you can add extra tubes? I have some 18350 high drain batteries.
It’s on top of the cost of almost $40 in buying the boards and emitters I would want. An out of the box solution would be much nicer and cleaner.
According to the datasheet the output difference between the R8000 sold at Intl Outdoor and the R9050 is about 15. The 219B R9050’s sold there are about 5 back from the 219C.
R9 (red) is one of the color patches not used in calculating CRI average (R1-R8). It’s useful for skin tones and meat but Cree emitters, particularly the higher color temperature ones, tend to score poorly and far behind Nichia’s because their spectrum output is very uneven. Nichia’s R9050 LEDs in the 5000K and 5700K versions have beautifully balanced SPD against Crees and are about as close to natural sunlight as you can get which is why some people like me love them.
Datasheets again?
Nope. Get out.
We don’t use that kind of language around here.
…
…
But seriously, have you seen djozz’s test data?
