Samsung LH351D 5000k LED output test by Texas_Ace - Amazing output, low thermal resistance but high Vf

kiriba-ru sent me these LED’s to test. I had never even heard of them before so I didn’t hold out a lot of hope for them but they really surprised me! They have one of the lowest apparent thermal resistance I have seen so far. This can be seen in how the output hardly jumps from a cold start, that says it is moving the heat away very efficiently.

The output is also very impressive at over 2200 lumens! This is more then any other 3V LED I have hot tested. The XP-L2 will make a bit more from a cold start.

The catch? It has a pretty high forward voltage compared to modern LED’s by Cree or nichia. Max amps you could expect in a single cell light would be around ~5-6A if I had to guess. That is still just under 2000 lumens though, so pretty respectable.

This also further confirms my hypothesis that the max amount of heat that the 3535 footprint can handle is around ~35-40W as all the 3535 LED’s I have tested have maxed out around there.

The dome started to burn around 5A but after scraping off the burnt area it completed the test just fine and a retest after it was over showed only a few lumen loss at 1A even after all that abuse.

I tried sticking an S2+ reflector over the LED and the beam looked pretty good, no tint shift.

Here are the results, click the pictures for larger versions.

Can’t find anything about this LED with a quick search. Could anyone share some pointers?
I absolutely LOVE the efficacy of this LED. Smells of really high sustained output.
How much can Q8 sustain? Because these with efficient boost driver would do 5500 lm with 55W including all losses.

I see why I couldn’t find it…it’s LH, in IH.

TA, I guess you don’t know the bin?

I’m very curious about luminance…

The high Vf is a blessing for functional lights with our direct drivers, like the BLF-A6 or the Q8, those modern low Vf leds are driven far into the inefficient top of the output curve in direct drive, which is interesting for output addicts but it makes it a less useful flashlight. This could be competition for the XP-L, especially if it comes in good tints.

Btw, both the looks and performance are very close to the LG H35F0, I’d be surprised if these two korean leds are not very close related. Expect unbearable green-ness in the high CRI versions!

(I tested the LG H35F0 4000K 90 CRI led but did not post it, the led that I tested maxed out at 8.5A at a Vf of 4.09, with 1730lm output. Considering it was a 90CRI led, the 21% lower output compared to this Samsung led is perfectly consistent)

Edit: forgot my manners for a moment: thanks TA for another insightful test!

Nice work TA. :+1:

Available in 90+CRI. :+1:

Thanks for the update, that would be my bad. I updated the charts if you refresh the page with CTRL + F5 it should show them. It also includes the bin now.

Yeah, it will limit the current to make the light more useful but on the flip side if you already have a regulated driver it will drop out of regulation much faster then other LED’s.

For example I generally use the TA drivers with the all 7135 mode but it is nice to have the turbo mode for when you want some extra.

Sad to hear that the high CRI versions could be green, I have had enough green LED’s for my lifetime.

Wow! 1000lm - 137lm/w (XP-G3 - 100lm/w). Way be very usefull for small capacity lights.
LEDs from same reel were tested by maukka: Samsung LH351D color testing

Yeah, these blow the G3 out of the water. The tint is much better then the G3 or L2 as well, no ugly tint shift that I saw.

Good to see that they also have good tested tint as well.

These samsungs are not using the flip chip setup, which in part explains the higher Vf. It could also mean they can be dedomed, although I have never really liked or been very good at dedoming.

Looks like an XP-L(2) sized die. So it’s no suprise that it’s brighter and more efficient than smaller LEDs like the XP-G3.

Sure, but cost-wise the LH351D is a competitor to the XP-G3 instead of XP-L2.

Nice test, TA. It looks like DigiKey has these at a pretty decent price and with good price breaks, too.

I’m wondering… would you be able to do a combination chart with this vs the XP-L V5 data you have? Looks like a real close competitor and a graph might help.

Also, what (if any) reflectors have you tried this in? How’s it look (especially with regards to tint-shift)? Maukka mentioned it looked good in his initial findings, then said it looked ugly when put in an EE X7.

Not necessarily, that is only for buck drivers with the same amount of input cell stages as in series output emitter stages. Besides all the theorizing, this for example is no problem with 2+S input for a single emitter, the down-side of this approach may be finding good 18350/26350/whatever short cells, formats which are not developed/manufactured by the big players in the li-ion industry.

This is the reason many times I've wished for a buck/boost driver which, on the other hand, would be necessary for driving violet/UV based emitters properly, I believe.

Cheers :-)

Is there still nobody who sliced it and can estimate die size? Or throw?

I almost did but didn’t :stuck_out_tongue:

But looking at the domed led the die size is at least very close to XP-L/XM-L if not exactly the same.

what is latest thread on these samsung?

Stupid question here: are these led “good” to be used with the MTN DD drivers in triples or quads?

Since I started using the Wuben TO50R with these Leds, much has changed in my perception of neutral white and high CRI, so I wanted to build another light with these. But I want it to be triple or quad, floody!

Thanks in advance for the answers!

These LED’s seem to be better suited to reflectors vs TIR optics in my experience. I think that the large die is just bigger then most TIR optics are designed for.

In reflectors they actually look pretty good.

I recently built a few AAA lights with some of these that a member on here sent me for my family members “purse lights” and they look really good in those.

Not a big fan of them in my S42 though.

That said, if you’re pleased with how they look in the TO50R optic and you can find a suitable optic for your build, I doubt you’d have any issue with a triple/quad and a FET driver. Between these having a higher vF and you running multiple (presumably in parallel), I don’t foresee any problems.