I was, of course, referring to swapping emitters as a mod… company prices are typically lower and the offset of the replaced emitters comes into play in that scenario.
Pricey upgrade, nonetheless…
preordered with 4000K 95 CRI
would like to buy 3×18350 tube drill a hole for external power supply.
If any Chinese light manufacturer offered the LH351D as an option other than 1 zebralight….a lot more people would……I will replace a whole bunch of my SST and XP-L leds once I get some single emitter practice swaps done. If not just for the fact (I got some from that W6 tint bin on Digikey that you recommended) that betters even the XP-L in lumens! & with more flood….whats not to like at $2 something a piece!
(Buying in bulk they are around $0.80 a piece so price ain’t the issue - it’s that they are made in South Korea)
I have no experience with most of the newer LEDs under discussion (self imposed buying freeze here for over a year) but keen to learn more.
Naturally when 18 LEDs are required, unit cost must be a consideration to meet a price-point. $100 will be the most expensive torch, by far, that I have bought. Plus the extra tube.
My MF01 (Nichia) has been a great success, particularly as a video/photo fill-in light as well as for tracking furry things that may be leaking red fluid, this should be smaller, neater, lighter, more performant, more sophisticated, well built.
18 LEDs behind TIRs should give a great smooth uniform spread beam. The MF01 does. I’ve tried to get my Q8 somewhere close, but given up. Managed it with some heavy diffuser foil at longer focal lengths, but too much loss incurred, it’s really much more of a thrower with a distinct beam pattern.
I’m not obsessed with “the BBL”, being an Applied Physicist, and understand why the highest (on paper) emitters might sometimes seem a bit green. This is understandable when you know that lumen output is not measured as a straight line on a spectrum, it is weighted by the human eye response, which peaks in the green. So if you want the best (on paper) lumen output, don’t be surprised that the manufacturers may tweak the phosphor balance on these to make the bigger numbers. Which is actually justifiable and generally useful.
Many people (particularly male, and there are also ethnographic differences) do not have eye responses that match the standard CIE profile and see things subjectively differently. Their opinions are valid, for them, but do not match mine, and I can get on with a variety of quite different LEDs (confident that my eyes are boringly typical).
Also making torches even with a healthy green output is more use to me than lighting up white walls or bedroom ceilings. They get used outdoors, where the world is mostly green, and where my eyes are most sensitive.
But I also need good output in the red region. Which usually means one of the lower colour temperatures, with the lower headline lumen numbers.
And an overall tint that I can use with digicams, without too much tweaking with white balance, never mind the full X-rite colour checker treatment.
Professional light panels usually have two different LED types, because it can’t be done properly with just one, when blending with available light.
I suppose this could also be done here, with major changes
Digicams mostly use RGB filter arrays over a monochrome sensor. When the spectral profile of the LEDs roughly matches the digicam colour filter array (Nikon for me), results can be very good indeed. This is unexplored here. And beamshots taken without carefully controlled calibration, misleading.
So, the answer is, I don’t know, but am very interested in real-world experience. Together with overall efficiency (i.e. run time when running in linear mode). Max. lumens for a few seconds before overheating, much less so.
Meanwhile Hank has to get these out, and I trust his judgement for the initial choices.
Just placed an order. Thank you, Hank for making warm white emitters available for each of your lights. I think only you and Simon with Convoy give customers this option. And for using anduril as the UI!
I own two Meteors and four Emisars in various emitters and tints.
I couldn’t resist ordering a 4000K D18, especially after reading the glowing reviews of the SST-20 high CRI emitters.
Thank you, Hank! Keep the new Emisar designs coming! 
I’m going to hold off on the pre order in hopes an XPL HI 5000k option comes up in the future fingers crossed
Thanks Hank, Just placed an order.
Ordered , thanks Hank
Hank, do you plan to sell this FL with 95 CRI 5000K led emitter?
4000K is too yellow for me and I would like to have 5000K or 6000K 95 CRI version.
If your answer will be yes, I will wait. 
Thank you
Kind regards
Unfortunately, there are no SST-20 5.000K 95CRI LEDs. They only exist for 2.700K - 4.000K. 5.000K and 6.500K is CRI65 - CRI70.
Where is the link for preorder?
https://intl-outdoor.com/emisar-d18-318650-14000lm-flashlight-p-939.html
Not good news regarding 95CRI 5000K leds. Is there any other led available with has 5000K 95CRI in will be good match for this FL?
Thank you
I think your best bet would be the Nichia 219C which has 5000k and 83CRI. I don’t think there’s anything with better CRI at 5000k?
Even the 219B at 4500k is 90CRI but once they eclipse 90+CRI they drop to 4000k
Does anybody know what the CRI is for the SST-20 5000K?
R9080 quality E21As are only rated for 90+ CRI, but they usually reach or surpass 95cri on measurements. Good luck using them on a FET driver though. There are 18 however, so that may help their cause.
65-70 I think. SST20 only comes in 95 cri up to 4000k.
To give people an idea what to expect for sustained output, here’s a runtime/output test. Full turbo is crazy-bright, and when it gets hot it does a smooth ramp down to about 30% power. After that it regulates to whatever temperature the user configured, which in my test worked out to around 2000-ish lumens. After the initial ramp-down, the changes aren’t really perceptible:
On a longer time scale, there was a bit of slow wobble… but not too bad. As it keeps searching for the ideal level, the fastest it moved was a few percent in almost a minute, which I wouldn’t even notice without a light meter.
Full power is definitely a “turbo” mode, not a sustainable “high” mode. It’ll generally ramp down in a minute or less, if you’re not using water cooling.
So it looks like only about 30 seconds (or even less) at full power?