The Samsung LED still has better cri and r9 than the natural light coming in my office today:
Two measurements taken in Texas, today at 2:45PM, overcast day. First is indoors with no artificial light, only ambient light from window – the next measurement is when I step outside.
Very similar to the DigiKey one I tested but with a better price.
1Amp with Frosted TIR:
Would be interesting measure after slicing for know if DUV and CCT decrease at very low current.Cheers
Here it is sliced in a Sofirn C01S AAA small TIR on the “100 lumen” setting. Not exactly apples to apples but you can see lower DUV but not much change in CCT. I cant explain why that is, maybe LED tint variation but probably just the optic doing strange things. It’s not really meant for such a large die. There is also a little ambient light in room.
The 2700k and 3500k were always the best ones so I’m not too surprised by this.
We have seen T6 be above bbl from other vendors so it seems like these squares if meaningful generally tend to be at the very top of the square/bin
Side note; CONGRATS on getting one of these mega distributors to answer this. Didn’t think it was possible.
If I can muster up the energy I’ll test my Convoy sourced 5700K 351D tonight or tomorrow. Don’t think we have seen one of these in the “wild” yet. I think these were from some stadium lighting so I suspect mega green but we’ll see…
The 2700k and 3500k were always the best ones so I’m not too surprised by this.
We have seen T6 be above bbl from other vendors so it seems like these squares if meaningful generally tend to be at the very top of the square/bin
Side note; CONGRATS on getting one of these mega distributors to answer this. Didn’t think it was possible.
Yeah, plus with 4000K and up the middle of the quadrangles is raised above the BBL, probably to merge with daylight tint (~ 0.0030 duv), on the other hand, integrated duv should be slightly lower than measured at the spot (but maybe not that much with the frosted TIR you used).
I was doubtful they would bother to check for me yes, I got no reply for ~10 days so I though they ignored me.
Quote:
If I can muster up the energy I’ll test my Convoy sourced 5700K 351D tonight or tomorrow. Don’t think we have seen one of these in the “wild” yet. I think these were from some stadium lighting so I suspect mega green but we’ll see…
Simon wrote the bins he got and it’s Q2 (5700K) and R2 (5000K) so the lower duv bins. I have some 5000K coming (not interested in 5700K) so I’ll see how they test. By the way I measured my FC11 and it’s actually worse than yours :
5170K (Duv 0.0086)
(Ra) = 87.6 [ R9 = 30.8 ]
Frosted TIR, same as before. The results are about what I expected. I purchased these mostly for two reasons: slicing and tint mixing. For both purposes they should be acceptable.
Actualy no. I was just curious how are their 5700K compared to the one you have. I had one installed in Novatac and it looks great. Not so “green and yellow”...
Is this the color we saw on the first power LEDs? We all had those Ledlensers and fakes on our keychains. Stable fluorophores had yet to be found I guess. I remember, when I was in microscopy, fluorochroms had really little photon staticstics like 60000 for this green substance whose name I forgot. You could watch them fading. Or is this a completely different story?
Is this the color we saw on the first power LEDs? We all had those Ledlensers and fakes on our keychains. Stable fluorophores had yet to be found I guess. I remember, when I was in microscopy, fluorochroms had really little photon staticstics like 60000 for this green substance whose name I forgot. You could watch them fading. Or is this a completely different story?
No it’s distinctly solid blue. More like a fake UV or “blacklight” I guess but very “visible”
A diy monochromatic blue LED. How did you get the phosphor off?
I think the old Ledlenser Photonenpumpe also was just blue.
You know, the usual. Force it into a too small TIR with a mediocre centering ring and keep tightening. It ripped it off nicely in one piece and I scraped with tweezers what tiny little bit was left.
Here it is sliced in a Sofirn C01S AAA small TIR on the “100 lumen” setting. Not exactly apples to apples but you can see lower DUV but not much change in CCT. I cant explain why that is, maybe LED tint variation but probably just the optic doing strange things. It’s not really meant for such a large die. There is also a little ambient light in room.
This is in a Sofirn IF25 at lowish output (2 LEDs at 50~100lm or something).
This is great because this is exactly what I wanted, a very warm 3535 LED with high CRI, which apparently doesn’t exist.
In the same flashlight (it’s a tint ramping light) I got Convoy sourced 5000K, also sliced :
CCT = 3877K (Duv -0.0004) 93.5 [ R9 = 60.2 ]
I’m surprised of how much the CCT was reduced compared to yours, in your FC11 it reduced to 4500K. In this case 4500K kind of was what I was aiming at for the max CCT in this light, I should probably have gone with the 5700K instead.
The TIR in IF25 is quite a bit larger I think than the C01S and C01S is very low output so I suspect those 2 things can cause different results.
As far as the Convoy 5K goes did you measure it before the slice? I assume it’s under 5000K slightly since it’s R2 bin.
This was measured at a pretty low output, here is turbo (should be around 3~4A per LED) : CCT = 2437K (Duv -0.0021) (Ra) = 94.0 [ R9 = 65.5 ], and lowest : CCT = 2357K (Duv -0.0013) (Ra) = 96.0 [ R9 = 73.3 ]
I didn’t measure the 5000Ks before slicing them but I’ll put one in another light so I’ll do it now.
The duv increase at higher current is strange, same behavior in the IF25 I modded with the 5000K leds (the 2700Ks have the classic current up -> duv down), so there is a middle tint point where the duv doesn’t move at all with current changes. The 5000K LH351D in my Wurkkos FC 11 behaves normally ( current up -> duv down)
I suspect that “mid point” correlates to the non-linear shape of the BBL?
Edit; and the Convoy LED being closer to 4700K than 5000K makes total sense. It still only changed by ~1000K CCT it just started out much lower than you probably thought.
(disregards the black dot, I don’t know how to display several measurements in Osram colorcalculator without doing the tint mix calculation )
“mid point” wasn’t really the right word, but at around 2800K the duv doesn’t change with current, under that it goes down with current (behavior of the 2700K lh351d from Sofirn), above it goes up with current (behavior of the 5000K lh351d from Convoy)
contactcr wrote:
Edit; and the Convoy LED being closer to 4700K than 5000K makes total sense. It still only changed by ~1000K CCT it just started out much lower than you probably thought.
I was under the impression that 1000 CCT was still quite a lot ? Maybe it’s because I sliced it very low ?
Here is the point I was trying to make and it’s just a hunch I do not know enough to be sure this is even true:
Lets say as we increase current it always makes this same slant (yes I oversimplified it). At some point our slant starts to become parallel with BBL at higher CCT so we will no longer have the change in Duv since Duv is fixed to BBL while at lower temps it is almost perpendicular to BBL.
Ah ok I understand, here are the 2 leds in the IF25, at lowish, midish and turbo :
(forgot to add the output on it, but higher CCT means higher output, and again please ignore the black dots and lines)
So indeed the inclination between the two is similar, and for the higher CCT one it crosses the BBL, since the tint shift is not very angled in the first place for LH351Ds. (I’m not sure those are the right words, English isn’t my first language)
Aside from that, I was curious about the effect on tint depending how low the slicing is, this is with the LED from the wurkkos FC11, with the same beaded TIR as previously :
0.85mm is the thickness I used previously, on this one it shaved a tiny amount of phosphor on one side so the results might be skewed a bit, although looking at the led at very low output I don’t see any irregularities, it hasn’t happened before, maybe this led was soldered slightly higher.
So the 1000K reduction was because it was a close shave.
The Samsung LED still has better cri and r9 than the natural light coming in my office today:
Two measurements taken in Texas, today at 2:45PM, overcast day. First is indoors with no artificial light, only ambient light from window – the next measurement is when I step outside.
like of your anDroid ColorMeter app
good time for a dogfart
can you also show the ColorMeter data screen for the LH351d 5000k please?
Here is the Sofirn sourced 2700K LH351D
Very similar to the DigiKey one I tested but with a better price.
1Amp with Frosted TIR:
Would be interesting measure after slicing for know if DUV and CCT decrease at very low current.Cheers
Here it is sliced in a Sofirn C01S AAA small TIR on the “100 lumen” setting. Not exactly apples to apples but you can see lower DUV but not much change in CCT. I cant explain why that is, maybe LED tint variation but probably just the optic doing strange things. It’s not really meant for such a large die. There is also a little ambient light in room.
I emailed Digikey the other day to ask about some color bins, I didn’t ask for a lot because I doubted they would reply, I got the reply today :
SPHWHTL3DA0GF4WPN6 (2700K) : 798 pcs F2 W7 N2
SPHWHTL3DA0GF4UPP6 (3500K) : 759 pcs F2 U7 Q2
SPHWHTL3DA0GF4TPQ6 (4000K) : 481 pcs F2 T6 Q2
Not sure how this info is valuable though because we don’t know if the bin you tested last year were the same.
The 2700k and 3500k were always the best ones so I’m not too surprised by this.
We have seen T6 be above bbl from other vendors so it seems like these squares if meaningful generally tend to be at the very top of the square/bin
Side note; CONGRATS on getting one of these mega distributors to answer this. Didn’t think it was possible.
If I can muster up the energy I’ll test my Convoy sourced 5700K 351D tonight or tomorrow. Don’t think we have seen one of these in the “wild” yet. I think these were from some stadium lighting so I suspect mega green but we’ll see…
Thanks for all your testing contactcr.
My favorite is the 3500k UPP6 from digikey sliced.
V11R and V10R Current Mod
Yeah, plus with 4000K and up the middle of the quadrangles is raised above the BBL, probably to merge with daylight tint (~ 0.0030 duv), on the other hand, integrated duv should be slightly lower than measured at the spot (but maybe not that much with the frosted TIR you used).
I was doubtful they would bother to check for me yes, I got no reply for ~10 days so I though they ignored me.
Simon wrote the bins he got and it’s Q2 (5700K) and R2 (5000K) so the lower duv bins. I have some 5000K coming (not interested in 5700K) so I’ll see how they test. By the way I measured my FC11 and it’s actually worse than yours :
5170K (Duv 0.0086)
(Ra) = 87.6 [ R9 = 30.8 ]
It’s quite strange, not 90 Ra and low R9.
Convoy Sourced LH351D 5700K 90CRI (bin: V2Q2S1)
Frosted TIR, same as before. The results are about what I expected. I purchased these mostly for two reasons: slicing and tint mixing. For both purposes they should be acceptable.
Thanks for getting those tests done! No surprises there but at least some Lee filters could help with the slight green tint.
I’ve got some incoming to see what slicing will do. Hopefully it’ll produce similar results to sliced 5000k dogfarts.
@contactr
Have you got a chance to grab KD’s 5700K emitters?
WTB Titanium 4sevens 2xAA tube
Nope, I can grab one though. Any other stuff at KD I should be getting?
Actualy no. I was just curious how are their 5700K compared to the one you have. I had one installed in Novatac and it looks great. Not so “green and yellow”...
That’s why am puzzled.
WTB Titanium 4sevens 2xAA tube
Here’s the Convoy 5700K sliced in Wurkkos FC11 (shallow light OP reflector, plain lens, on high):
Good for a day time use light since your eyes are adjusted to >5000K ambient above BBL light already.
Not bad at all.
WTB Titanium 4sevens 2xAA tube
If anyone is curious what these look like under a TIR with all the phosphor removed:
450nm and 25000K CCT (if that is even a thing, I guess it is technically.)
Is this the color we saw on the first power LEDs? We all had those Ledlensers and fakes on our keychains. Stable fluorophores had yet to be found I guess. I remember, when I was in microscopy, fluorochroms had really little photon staticstics like 60000 for this green substance whose name I forgot. You could watch them fading. Or is this a completely different story?
Spitzbube.
No it’s distinctly solid blue. More like a fake UV or “blacklight” I guess but very “visible”
UltraTac A7P w/ no phosphor LHD and Sofirn C01R
A diy monochromatic blue LED. How did you get the phosphor off?
I think the old Ledlenser Photonenpumpe also was just blue.
Spitzbube.
You know, the usual. Force it into a too small TIR with a mediocre centering ring and keep tightening. It ripped it off nicely in one piece and I scraped with tweezers what tiny little bit was left.
SST-20 JA3 2700K tested with differents type lens
I sliced two Sofirn sourced 2700K and it lowered the CCT quite a bit :
CCT = 2369K (Duv -0.0014) (Ra) = 95.9 [ R9 = 72.9 ]This is in a Sofirn IF25 at lowish output (2 LEDs at 50~100lm or something).
This is great because this is exactly what I wanted, a very warm 3535 LED with high CRI, which apparently doesn’t exist.
In the same flashlight (it’s a tint ramping light) I got Convoy sourced 5000K, also sliced :
CCT = 3877K (Duv -0.0004) 93.5 [ R9 = 60.2 ]I’m surprised of how much the CCT was reduced compared to yours, in your FC11 it reduced to 4500K. In this case 4500K kind of was what I was aiming at for the max CCT in this light, I should probably have gone with the 5700K instead.
The TIR in IF25 is quite a bit larger I think than the C01S and C01S is very low output so I suspect those 2 things can cause different results.
As far as the Convoy 5K goes did you measure it before the slice? I assume it’s under 5000K slightly since it’s R2 bin.
This was measured at a pretty low output, here is turbo (should be around 3~4A per LED) : CCT = 2437K (Duv -0.0021) (Ra) = 94.0 [ R9 = 65.5 ], and lowest : CCT = 2357K (Duv -0.0013) (Ra) = 96.0 [ R9 = 73.3 ]
I didn’t measure the 5000Ks before slicing them but I’ll put one in another light so I’ll do it now.
5000K from Convoy (bin F2R2S2) with beaded TIR :
100 mA CCT = 4693K (Duv 0.0019) (Ra) = 93.4 [ R9 = 62.4 ] IES TM-30-15 Rf = 90.42 Rg = 98.89
1A CCT = 4729K (Duv 0.0026) (Ra) = 92.0 [ R9 = 56.2 ] IES TM-30-15 Rf = 90.17 Rg = 98.87
sliced with beaded TIR:
100mA CCT = 3651K (Duv -0.0016) (Ra) = 93.8 [ R9 = 60.9 ] IES TM-30-15 Rf = 90.88 Rg = 100.11
1A CCT = 3777K (Duv -0.0012) (Ra) = 92.6 [ R9 = 55.9 ] IES TM-30-15 Rf = 90.28 Rg = 100.13
The duv increase at higher current is strange, same behavior in the IF25 I modded with the 5000K leds (the 2700Ks have the classic current up -> duv down), so there is a middle tint point where the duv doesn’t move at all with current changes. The 5000K LH351D in my Wurkkos FC 11 behaves normally ( current up -> duv down)
I suspect that “mid point” correlates to the non-linear shape of the BBL?
Edit; and the Convoy LED being closer to 4700K than 5000K makes total sense. It still only changed by ~1000K CCT it just started out much lower than you probably thought.
The tint mixes in a straight line :
(disregards the black dot, I don’t know how to display several measurements in Osram colorcalculator without doing the tint mix calculation
)
“mid point” wasn’t really the right word, but at around 2800K the duv doesn’t change with current, under that it goes down with current (behavior of the 2700K lh351d from Sofirn), above it goes up with current (behavior of the 5000K lh351d from Convoy)
I was under the impression that 1000 CCT was still quite a lot ? Maybe it’s because I sliced it very low ?
Here is the point I was trying to make and it’s just a hunch I do not know enough to be sure this is even true:
Lets say as we increase current it always makes this same slant (yes I oversimplified it). At some point our slant starts to become parallel with BBL at higher CCT so we will no longer have the change in Duv since Duv is fixed to BBL while at lower temps it is almost perpendicular to BBL.
Ah ok I understand, here are the 2 leds in the IF25, at lowish, midish and turbo :
(forgot to add the output on it, but higher CCT means higher output, and again please ignore the black dots and lines)
So indeed the inclination between the two is similar, and for the higher CCT one it crosses the BBL, since the tint shift is not very angled in the first place for LH351Ds. (I’m not sure those are the right words, English isn’t my first language)
Aside from that, I was curious about the effect on tint depending how low the slicing is, this is with the LED from the wurkkos FC11, with the same beaded TIR as previously :
Original :
0.1A CCT = 4769K (Duv 0.0058) (Ra) = 91.2 [ R9 = 51.0 ]
0.5A CCT = 4826K (Duv 0.0061) (Ra) = 90.5 [ R9 = 48.2 ]
1A CCT = 4875K (Duv 0.0059) (Ra) = 90.2 [ R9 = 46.4 ]
2A CCT = 4917K (Duv 0.0055) (Ra) = 89.7 [ R9 = 43.9 ]
(so actually even with this bin the duv doesn’t decrease a lot, even goes up a bit at 0.5A
Washer thickness of 1.52mm :
0.1A CCT = 4544K (Duv 0.0050) (Ra) = 91.4 [ R9 = 51.6 ]
1.20mm :0.5A CCT = 4605K (Duv 0.0051) (Ra) = 90.8 [ R9 = 49.0 ]
1A CCT = 4680K (Duv 0.0046) (Ra) = 90.5 [ R9 = 47.6 ]
2A CCT = 4706K (Duv 0.0044) (Ra) = 90.0 [ R9 = 44.6 ]
0.1A CCT = 4343K (Duv 0.0031) (Ra) = 92.0 [ R9 = 54.3 ]
1.00mm :0.5A CCT = 4405K (Duv 0.0034) (Ra) = 91.4 [ R9 = 51.6 ]
1A CCT = 4462K (Duv 0.0032) (Ra) = 91.0 [ R9 = 49.7 ]
2A CCT = 4525K (Duv 0.0027) (Ra) = 90.5 [ R9 = 47.2 ]
0.1A CCT = 4083K (Duv 0.0019) Index (Ra) = 92.3 [ R9 = 54.9 ]
0.85mm :0.5A CCT = 4147K (Duv 0.0023) (Ra) = 91.7 [ R9 = 52.2 ]
1A CCT = 4202K (Duv 0.0022) (Ra) = 91.2 [ R9 = 50.1 ]
2A CCT = 4293K (Duv 0.0017) (Ra) = 90.7 [ R9 = 47.4 ]
0.1A CCT = 3798K (Duv 0.0023) (Ra) = 92.2 [ R9 = 52.6 ]
0.5A CCT = 3859K (Duv 0.0026) (Ra) = 91.5 [ R9 = 50.1 ]
1A CCT = 3918K (Duv 0.0025) (Ra) = 91.0 [ R9 = 48.0 ]
2A CCT = 4017K (Duv 0.0020) (Ra) = 90.4 [ R9 = 45.2 ]
0.85mm is the thickness I used previously, on this one it shaved a tiny amount of phosphor on one side so the results might be skewed a bit, although looking at the led at very low output I don’t see any irregularities, it hasn’t happened before, maybe this led was soldered slightly higher.
So the 1000K reduction was because it was a close shave.
Pages