IKEA Ledare, type LED1708R5.
A dimmable GU10 bulb that uses very warm leds at the lowest brightness setting and gets cooler tint at higher setting by gradually mixing in a cooler set of leds and switching off the warm set at the highest setting.
Lowest settings, just the warm leds:
CCT=2060, duv=+0.0012, CRI=93.6, R9=58, R12=98, Rf=93.2, Rg=96.8
An intermediate setting, some cooler light mixed in:
CCT=2450, duv=+0.0005, CRI=93.8, R9=60, R12=92, Rf=92.6, Rg=99.4
Highest settings, just the cooler set of leds, warm leds are off:
CCT=2640, duv=+0.0008, CRI=91.7, R9=54, R12=88, Rf=90.3, Rg=100.0
The newest insights are that (“western”?) people prefer the tint well under the BBL, which these bulbs are not, but they are only slightly above it, and I do not see an ugly tint.
Since there are so many measurements being posted, I just want to remind everyone you should allow LED bulbs to warm up for a while before testing because Duv tends to drop significantly.
Are there any good tube style replacements? I got a box of Hyperikon T8 bulbs that claim >= 90 CRI. They’re better than the fading florescents they replaced, but the PWM is noticeable sometimes.
High output bulbs with mediocre cooling tend to heat up to 70-100°C within an hour. Since home lighting is used for prolonged periods, I would usually wait at least 15-30 minutes before taking the measurements. Similar shift in tint goes for flashlights too, but they’re mostly used in shorter bursts and as such with less heat generated at the emitter.
I use it with an IKEA fixture that may be especially made for this led, so I’m not sure if it is a normal voltage dimmer. But the dimming is certainly not up to BLF-standards :nerd_face: , not very smooth and at certain postions the cooler leds keep switching on and off (may I introduce “tint flicker” ? :party: )
Added in the draft new png table, in a kind of "promising chip" category.
You are right, we should build some recommandations for testing. Warming up is one, testing multiples bulbs (2/3) of the same spec (if the testers owns them of course, or can be offered them) and average the values could be another.
At any moment, if any of our dear testers wish to provide new measurements, I'd be glad to keep up to date and update our thread table, if needed.
Just took 30s to read the view of wavelightning take on CRI, and learned something about geography. According to them, in general the US way to describe CRI (Ra) is only based on R1-R8 values. But, the Chinese/European way to describe CRI is said to mostly involves the full R1-R14 scope. And sould be noted Re, with e for extended, but it might not always be the case. All our BLF testers posted here Ra values, so normally R1-R8, but do you think any your spectrometers softwares might misuse the term Ra and do a full Re computation instead?
I don't find this particular "304.084.70" 806lm version tested anywhere yet, but the 920lm version, according to this french website review, has its CRI varying between 85 @2700K & 82 @4000K. They advise the reader towards this otherTradfri kit instead, in regard to quality of light (92 in that case). But both published measurements lacks the interesting R9, R12 and Rf values.
At least my software tools (CT&A, ArgyllCMS spotread) use R1-R8 for Ra measurement, which is the CIE 13.3:1995 standard.
Here’s the data for the 950 lumen Ikea Trådfri model LED1546G12, which looks like has been discontinued. Not integrated nor warmed up for more than a couple minutes, just a quick test in front of the bulb.