Edit: It looks like what happened is the ‘k’ in ‘sw45k’. It changes the tint quite a bit.
—— original post below ——
The popular Nichia 219B 4500K LED was everywhere, and then it was gone for a while, and then it came back… but now the new ones look different. Anyone know what happened?
New models look like a colder CCT, with less green and more blue. For example, here are four 219B 4500K lights. Left to right…
L3 L10 219B (~2013)
Custom brass 219B triple (~2017)
Custom 5×7135 219B single (~2015)
Noctigon KR4 Ti w/ 219B 4500K (2021)
(if pics don’t show up, it’s probably an issue with browsers trying to enforce https even on older http-only sites; try clicking on it to bypass that)
I noticed the same pattern with Fireflies lights. Here’s an original ROT66 219B (left) next to a newer Gen2 model (right):
I don’t know if the old ones were sw45 or sw45k, but the new ones are rated as sw45k. The ‘k’ is supposed to imply lower duv, meaning less green and more red. But it looks like there is also significantly more blue, so the new ones look like they’re not 4500K any more… looks more like they’re between 5000K and 5500K.
Has anyone been able to properly measure these? I don’t have the right equipment for that.
Is this a sw45 vs sw45k thing? Did Nichia change how the LEDs are made? Is it something else? I’m hoping someone has more information.
I also think it’s k Vs non-k related. I personally find the k versions to purple. Rather have sw45 which are not green or purple, but maybe look a bit green next to sw45k.
Yes, tints of my 219B sw45k D220 lights vary. Not extreme, but in comparison, the green/rosy perception phenomenon is clearly visible. D200’s also differ somewhat from each other. Thought that would be normal.
From my understanding, 219B was never available in sw45 in R9050 or R9080. The original datasheet for NVSL219B-V1 (warm white) shows R70 and R8000 with CCTs from sw27 to sw45. At some point it was available in R85 according to Cutter, and the 2015 Nichia catalogue shows R9050 available in sw27 to sw40, with sw45 “under development”. When R9080 became available (early 2017?), sw45 was replaced with the special sw45k chromaticity rank, which is guaranteed to fall the below the BBL. My theory is this was necessary to produce the required R9 since sw50 was only ever available in R9050.
In theory, sw45k can be anything from 4260K to 4752K and Duv –0.0098 to 0.0000 at the rated 700mA, even ignoring other factors like reflector or optic choice, plus CCT and tint shift from low to high current. I have measured various batches at both extremes of CCT, though Duv tends to follow the flux rank, with D200 generally being much rosier than D220. I mod my own lights, so it’s hard for me to comment on any kind of general shift over the years.
They feel different to me as well. The Convoy T2’s I just got in with the sw-45k are very, very rosy with no blue or green. I really like the ones from Convoy, I have a decent stock of the old ones and have many installed. To me the new ones feel cooler.
Ok, so broke out some of my older builds with 219b sw-45k emitters… I could not have been more wrong! Side by side the new ones appear to have a hint of either green or blue! They still are not bad, just not the same.
Could this be a difference created by the increase in the brightness bin? My old ones were D180 and D200, I think the new ones are D220
Comparison, camera on 4500k had to almost over expose because of PWM on the left/old light. Very similar, but yet the left is a little different. Difference is more obvious on my camo shorts.
The camera’s white balance was set to match the hotspot of the original ROT66, since it has the nicest-looking beam. Any apparent green is relative to that. The manual white balance puts “white” at about 4500K with negative duv, so the blackbody pure-white line would look green in comparison. If I had set the white balance to match the spill instead, the hotspot would look pink. Or if I set the white balance to a standard “daylight 5000K”, the old 219B beams all look a bit like pinkish amber, while the new 219B looks more blue-purple.
In person, both old and new varieties look really good… but the new ones look colder.
Yeah. They are different. It’s not just my imagination, since the camera can see it too. I’m just not sure exactly how they’re different, or why.
That was the R85 version released in 2013, which was CRI 92 typical and available in sw45. Those LEDs were widely used for years, especially since the 2015 2014 “V1” update that added R90/R9050 removed the sw45 option (again, listed as “under development” in the catalogue). 4500K was added back as sw45k in the R9080 release. It’s hard to determine exactly what was available at any specific time because many of the datasheets are gone, and the general LED catalogues weren’t always very accurate.
I traded my d4v2 with 219b sw45k (bought July 2021) because it appeared to be more of a 5500k light, i liked the tint but it was wayyy to cool. Not worth the tint for me, i unfortunately have no experience with 219b's before this summer's round of them.
They probably changed the “recipe” of phosphors for whatever reason†, and ’though they still technically meet spec, they’re still a bit “off” as far as old/new appearance.
I actually suspected both my L3 L11C and my first generation ROT66 were SW45, but didn’t have any way to verify. Both look very neutral to me. They only look rosy if comparing to lights with worse tint. It looks like Maukka tested his ROT66 at moderately rosy on middle outputs, and very rosy on turbo:
I was not aware SW45 (non-K) was not actually available in R9080, but I’ll take Bob McBob’s word for it. I guess one or the other of my lights might have actually had R85.
Maukka’s ROT66 measurement technically meets R9080, but is consistent with Bob McBob’s statement about the typical performance of R85 219B’s.
Also, here is the chromaticity diagram showing the range of DUV possible within SW45K that Bob mentioned. It looks like the theoretical range is even bigger than he stated, but from tint measurements I’ve seen shared, the numbers he mentioned seem realistic.
Yeah, that’s the same reason I haven’t used the newer ROT66 much. It looks like it’s 5500K. Granted, it’s the best-looking “5500K” I’ve ever encountered, but I was hoping for more a neutral CCT instead of a colder shade.
I am pretty sure that led is used in tk’s L3 L10.
