LH351D 90CRI Spectrum Tests

I have now purchased most variations of the 90CRI LEDs from Digi-Key. I also reached out to a few other vendors to try to get info on some 4000K 90CRI bins for sale in larger quantities but so far nothing has panned out with that.

I created one thread already for the 2700K SPHWHTL3DA0GF4WPN6 part but I will also be doing some tests on the 3000K, 3500K, 4000K, 5000K. Any undesirable ones I will (probably) attempt to slice them to see how it changes the tint.

I’m a bit sick of SST-20 for it’s variations in tint across bins and power levels and I think we are all ready for something to replace 219B for high CRI. The LH351D doesn’t have the R9 of either but it does have some advantages and disadvantages of its own:

Pro:

  • Small deviation across power levels
  • Fairly consistent at various angles under TIR/Reflector
  • Higher output and should survive direct drive in most cases
  • Slices easily with good results
  • Cheap price

Con:

  • With dome on it’s very floody so may require slicing for certain uses
  • It’s too big for some ancient XPG2 type optics
  • R9 isn’t as high as 219B 9080 or SST-20 but it surpasses 219C I believe
  • Availability may vary for international buyers

Stay tuned for testing in the coming days…

Datasheet: https://cdn.samsung.com/led/file/resource/2019/03/Data_Sheet_LH351D_Rev.9.1.pdf

Digi-Key: Index of available LEDs

SPHWHTL3DA0GF4 (2700K) - from Sofirn
SPHWHTL3DA0GF4WPN6 (2700K, probably bin F2W7N2)
SPHWHTL3DA0GF4VPP6 (3000K)
SPHWHTL3DA0GF4UPN6 (3500K)
SPHWHTL3DA0GF4UPP6 (3500K, probably bin F2U7Q2)
SPHWHTL3DA0GF4TPP6 (4000K)
SPHWHTL3DA0GF4TPQ6 (4000K, probably bin F2T6Q2)
SPHWHTL3DA0GF4RTR6 (5000K)
SPHWHTL3DA0GF4RTS6 (5000K) - from Arrow
SPHWHTL3DA0GF4RTS6 (5000K, bin F2R2S2) - from Convoy thx user thefreeman
SPHWHTL3DA0TV4QTS6 (5700K, bin V2Q1S2) - from Convoy

Reserved

Very interested to see the results. My biggest problem with LH351D is how variable it can be even from the same distributor. I got some of the very first small batch of 4000K that AEDe sold, and they were the best I’ve tested so far. The next batch from him was a bit greener. From Digi-Key they were greener still. The SPHWHTL3DA0GF4RTS6 I got from Arrow when they were really cheap were great sliced, but the batch I ordered from Digi-Key weren’t as nice. Going back to Arrow a further batch was somewhere in between, and now of course Arrow has made their shipping excessively expensive. It’s hard to rely on LEDs when any given order could be greener than I’m comfortable with and there is no way to purchase a known good bin.

That said, they do slice up pretty nicely. The die is the same size as XP-L HI so I’ve never found them excessively floody with a tight trim. Unfortunately the more you take off, the closer you get to a point where the silicone isn’t a single piece, so it’s easy for it to tear and take off the edge of the phosphor.

I agree with everything you said. The AEDe first batch is the best (4000K of this LED) i’ve seen and will be the benchmark for any other 4000K I test.

I can already tell you the highest bin TPQ6 from DigiKey is pretty high Duv. However, it slices to a very close to the BBL 3500K (even with a SMO reflector). The goal is to find one that doesn’t need slicing though and i’m optimistic that the lower flux bin may be better in this regard, we’ll see. From what I have read I think the highest flux bins tend to be more green for most types of LEDs in general and is a reason why I purchased multiple flux bins of each where possible.

Watching this one, I am a huge fan of 4000k and an even bigger fan of the 219b R9080 sw40 and sw45.

Wow! Thank you for doing this.

Subbed!

Thanks for doing this! For me the best quality of the LH351D is how good it just looks to the eye (plus efficiency of course). Not 219b good, but miles better than SST20 or CRI90 219C regardless of CCT and duv. LH351D is fine even a bit above BBL, just like the venerable 219C 5000K CRI80.

SPHWHTL3DA0GF4RTS6 (5000K)

Am I the only one that reads this really quickly and thinks “DOG FARTS”? I’ll see myself out now, thanks.

No :smiley:

https://www.reddit.com/r/flashlight/comments/98yk8c/highcri_samsung_leds_now_available_from_arrow/

:smiley: hahaha!

This is probably not the final format but i’m going to try to refine it a bit before I test every LED. Feedback welcome if you see something missing.

The goal of my testing is to figure out what LEDs are worth checking out and which ones should just be avoided. I’m not trying to take loads of beam shots or do death tests on the emitters, there is enough of that info out there.

I limited the measurements to the hot spot because these LEDs generally have an even beam, I assume that’s OK.

I can include voltage at each current if it’s helpful. I chose 5 currents (100mA, 300mA, 600mA, 1.0A, 1.5A). Seems like a good mix to me but open to feedback.

Here’s the lower flux binned 5K 90CRI (the small dog fart if you will):

I may need to come up with a good way to line up the S2+ SMO cause it looks like I have an outlier. In any case the Duv was a bit high with it as expected. The TIR’s were both pretty acceptable but without the other bins measured I wont make any final conclusions yet.

This CCT was between ~4750-4950K so on the ‘lower’ half of the 5K bins. CRI was between 91-93 (R9 54-67) depending on current and lens. Lower current = higher CRI and Frosted Carclo also gave the higher end of the readings with SMO reflector giving the lower end.

I really like the LH351D. My favorite domed is the 4000K 90cri that I purchased from arrow. The real magic happens for me, when then dome is sliced.
I sliced a 4000k for the first time a few days ago and I was really impressed with the tint. My favorite led was the 5000k 90cri sliced, also purchased from arrow.
But now I’m starting to lean more towards the sliced 4000k.
I like the higher light output of the LH351D, even sliced it still makes a very decent amount of lumens. I modded a TCR1 with a triple LH351D sliced 5000k 90cri with a KD optic and it made 1022 lumens OTF with the lens installed over the optic.
.
Thanks for doing this Contactcr, I would be interest in seeing the results with the domed sliced also.

Did I miss Arrow carrying 4K 90CRI? I don’t remember them ever stocking small quantities.

I’ll post some results for shaved 4K, only done those on one so far but probably a dozen or more 5K and they all turned out great.

Any specific one you want to see shaved?

They carried both a good bin of the 4K 90CRI AFAIK, and definitely the “Dogfarts” as I ordered some from them.

Oops, I was mistaken. The first 4000k’s I purchased were from AEDe, then a second batch from AEDe, then a third batch from Digi-key which I haven’t used yet.
My mistake. I got several of 5000k’s from arrow when they were $0.74. I liked the 4000k better because the 5000k looked to white to me. But dedomed is a different story, I like the tint.
Dedomed testing on the 4000k would be interesting, that’s the only two color temps I use anyway. :+1:
.
Kind of off topic for this thread, but I was curious if anyone had dedomed or tested a dedomed a 219b SW45K?

I keep thinking I ordered some 4000k from arrow but I went back and checked and I order twice from AEDe and once recently from Digi-key.
I logged in at arrow to check my order history (which I have placed many orders) but it said I didn’t have any order history. :person_facepalming:

This was from a few nights ago, the first time slicing a 4K for me. It’s the TPQ6 from DigiKey that is (spoiler alert) not that great compared to AEDe 1st batch. I never sliced a 4K before now because the AEDe 1st batch was too nice by itself. I still need to test it formally like the above but in an actual S2+ SMO with AR lens (which both tend to increase Duv) it is spot on BBL and a little under with FET modes, around ~3500K CCT in hot spot. Once I test with no glass and try the TIRs I suspect it will be even better. To be continued…

Love the LH351D High CRI series. its like having the high CRI abilities & tints of the past Nicha 219, but with much more deficiency & output capability of a XP-L.

Nice beheading job contactcr :+1: .