I hoped for illumn.com because I like 4000k leds way better than 5000k. (apart from doing output tests I like building flashlights )
oops, I didn’t even notice his were 5000k. 4500k is my preference. I can order some from Illumn then? Shoot me a PM if you’re interested.
Hi djozz, regarding your 219B in the graph I see it hitting 600 lumens around 2.7A.
In your older test of the 219B here it is maxed out just under 600 lumens around 4A.
Is that because older graph shows lumen values with reflector losses and the graph on in this thread is lumen value without reflector?
Just to be clear these were the same binned HCRI 219B?
Thanks for your awesome work!
Hi djozz, regarding your 219B in the graph I see it hitting 600 lumens around 2.7A. In your older test of the 219B “here”:testing two of KD's new Nichia 219b leds (4500K'92CRI' and 5500K) it is maxed out just under 600 lumens around 4A. Is that because older graph shows lumen values with reflector losses and the graph on in this thread is lumen value without reflector? Just to be clear these were the same binned HCRI 219B? Thanks for your awesome work!
I'm not absolutely sure of the binning of the 219B from the graph of the OP actually, the numbers are from a test together with the new Oslon square, at a later date than my first 219B test. I think that the led came from the same batch that I used in the first 219B test, from Kaidomain so uncertain, but I ordered one or two from intl-outdoor as well later.
The method indeed changed since my first 219B test, that first test uses a ceiling bounce method that measures the output from a reflector placed over the led, which can lead to (up to 20%?) light loss, this is on top of calibration inaccuracies that come with the ceiling bounce method (my calibration light -D40A- is much more throwy than the small OP-reflector that is used over the led that was being measured, and the ceiling bounce method is sensitive to source distance and incoming light angle, I expect the test-led to come out a bit too low compared to the calibration light). The later test that is used in the graph in the OP above (and all other tests since then) measures bare led lumens in an integrating sphere.
Any idea if they are going to come out with 90 CRI 219C?
I’m really liking the current batch of 83 CRI 219C. I’ve installed 10 of them in various lights over the last few months. Beautiful tint. Nice beam pattern in zoomies. Low forward voltage makes for easy regulation especially in small 1x14500 EDC lights, etc.
Just last night I swapped the XPL HI emitters in one of my DQG Tiny III triples with 219C. Not bad. Though with a 3.8 amp 10x7135 driver at 100% power the triple 219C doesn’t throw as well as my other DQG equipped with triple XPL HI and a FET emitter…. running at 30% power.
Any good sources for the 4000K version of this LED? I like the 5000K, but would like to see the other as well… I found the Osram SSL’s on Mouser in 4000K, but cannot seem to locate these anywhere in anything other than 5000K
I haven’t seen any 4000K 219C but there is 4000K 219B V1 HCRI at Illumn.com here
The beamshots I’ve seen of the HDS lights with these 4000K HCRI 219B look really good.
Thanks for the great test, I need to get more familiar with Nichia emitters, and the graph made things much clearer.
I’m not absolutely sure of the binning of the 219B from the graph of the OP actually, the numbers are from a test together with the new Oslon square, at a later date than my first 219B test. I think that the led came from the same batch that I used in the first 219B test, from Kaidomain so uncertain, but I ordered one or two from intl-outdoor as well later.
The method indeed changed since my first 219B test, that first test uses a ceiling bounce method that measures the output from a reflector placed over the led, which can lead to (up to 20%?) light loss, this is on top of calibration inaccuracies that come with the ceiling bounce method (my calibration light
D40Ais much more throwy than the small OP-reflector that is used over the led that was being measured, and the ceiling bounce method is sensitive to source distance and incoming light angle, I expect the test-led to come out a bit too low compared to the calibration light). The later test that is used in the graph in the OP above (and all other tests since then) measures bare led lumens in an integrating sphere.
Thanks for the explanation. I’ll keep in mind that your current tests are bare LED lumens.
Any good sources for the 4000K version of this LED? I like the 5000K, but would like to see the other as well… I found the Osram SSL’s on Mouser in 4000K, but cannot seem to locate these anywhere in anything other than 5000K
I got my 219C 4000K's at Cutter, here: http://www.cutter.com.au/proddetail.php?prod=cut2744 - good price except for the shipping.
The Nichia 219C is the newest upgrade in the 219 series. The Nichia 219 leds have become famous for their 90+ CRI variants, giving better colour reproduction than any Cree or other led, but in the performance section they were way behind Cree at start. But since back then the 219 has come a long way: the 219B performed way better than the 219A, and this test shows another huge leap with the 219C.
The 90+ CRI versions of the 219C are not on the market yet, but I was able to purchase this Nichia 5000K 83CRI B280-300 led from Lumitronics in Germany (the picture on the site is wrong btw). Voltage and current performances from this test will be comparable for the upcoming high CRI variants (when time comes, I want to test them as well for output and colour performance), but for now let’s see how this 5000K 83CRI led compares to a latest bin XP-G2
The Nichia was reflowed on a 16mm Noctigon board and tested for voltage and output using my usual method, described in the XP-L test-thread, except that for this test my 46cm integrating sphere was used (should not make a difference for the results). Also I tested a Cree XP-G2 S4 2B led, for comparison and also because I have thusfar only tested lower binned XP-G2’s and the newer XP-G2’s are said to have higher voltages (and higher output of course).
For the graphically challenged, here are the raw collected data from the tests, also because I find the actual numbers always very useful for reference (in blueish green the Nichia numbers, in greenish green the XP-G2 numbers):
For comparison, in the graph I added some already posted test data (collected with the exact same method): for the Nichia219B 4500K 90+CRI, the new Osram Oslon Square 4000K 90+CRI, and an old batch cool white XP-G2 R5. Note that output differences are partly a result of varying colour temprature and CRI of the different leds in the graph.
So what can be seen?
*I have put way to many lines in this graph, I hope you can make things out
*the output of the 219C is very very good, close to the (cooler tinted and lower CRI) latest XP-G2! The maximum at 6.2A is the same as the XP-G2, so Nichia has (silently) managed to achieve the same low thermal resistance as Cree. And look at the difference with the 219B (maximum at 3.8V), this is a huge leap again.
*the 219C has the lowest voltage of the bunch, and so much lower than the XP-G2 that it is more efficient at all currents. The voltage is even significantly lower than the Cree XM-L2 (0.25V ! lower at 6A). This makes the 219C very easy to use in single li-ion flashlights, it will even require current regulation (polish up your Qlites!), with FET-drivers it may draw more current than you would like!
*I really tried, but could not kill the 219C, I stopped at 10A. After that I looked for decreased performance, checked at 4.4A, and found just 2.5% less output. This could be caused by what looks from the side like a semi-detached dome (led at .5mA). Having been at 10A for 2 minutes, I can not blame this led :
*the 92CRI new Oslon Square that I tested recently, outperformed the Nichia 219B, but compared to the 219C the Oslon will probably loose it (but a definite answer is there when the 92CRI 219C becomes available).
*this XP-G2 S4 2B has a fairly constant 0.2V higher voltage than this old batch XP-G2 R5 over most of the current range. This is consistent with the general observation on the newer Cree leds.
With these results I am curious about the dedomed performance of the 219C. The XP-G2 still has the advantage of a bit better output and of a bit smaller die size (1.45x1.45mm, compared to 1.6x1.6mm for the 219), but I really like the uniformity of the 219C die for use in aspherics:
(here at 0.5mA, and after the torture)
Conclusion
The Nichia 219C seems a fantastic upgrade over the 219B, in performance, voltage and thermal resistance. I’m also really looking forward to the 90+ CRI variants. As already said in the new Oslon Square test: it is high time that Cree launches a new generation XP-G2.
Thanks for reading!
(stereo picture)
You should add a bunch of other emitters like the XPL and XML, need moar data!!!
Check his signature block at the bottom…. Djozz is a very prolific tester….
Hint: It says DJOZZ Tests in blue…
Camo5 - please read this opening post thoroughly: https://budgetlightforum.com/t/-/23020, report back . You will (may) find buried in there a whole section dedicated to djozz's tests, all the thread links - I'm guessing it's bout 1/4 the way into it.
The linked thread's OP is an incredible amount of work to put together and does a great job summarizing things.
For some reason I had been having a very hard time finding this data. thanks!!
You should add a bunch of other emitters like the XPL and XML, need moar data!!!
Check his signature block at the bottom… Djozz is a very prolific tester… Hint: It says DJOZZ Tests in blue…
just added another one: Nichia 219B V1 4000K 93CRI :party:
So……in a light with one cell only it’s probably best to use this new Nichia Led, due to the low Vf (3,5V @ 6A) because with one cell the XP G2 S4 would probably reach no more then 4A @ 3,9V……?
Grtz
Nico
I dedomed the XP-G2 S4 2B also (this one worked fine except for a small piece of silicone on one corner of the die) and made pictures of the two emitters under the stereo microscope at the exact same magnification, then Photoshopped the dies next to each other (I did not match the sizes, both leds are indeed the exact same size), the dies look the same size too (so is the S4 bigger than before? I don’t know). I have to go to bed now, so no further measurements, but last thing I did was measuring the XP-G2 emitter with caliphers and I measure exactly 3.5mm (if we did not already know that, but I never measured it myself), so you can work out the real size of the dies from the picture below (good luck )
EDIT: Oh shoot, took out the ruler and I measured it all the same, I measure the XP-G2 die size at 1.48x1.48mm. Apparantly the 219C die is smaller than the 219B die (and now for the succesful dedome, who can do it?)
Hi djozz,
I’ve just found this picture and it could be the answer to a question that was bugging me for a while. Assuming this is a XP-G2 on the right:
Are there some revisions of the XP-G2 with a different number of “dots” at the die? I’ve swapped some Latticebright with XP-G2 in those violet AA Zoomies (“Ultrafire Q5” or SK72XPE/SK72) and the older one of the XP-G2 has rows of 4/5 dots. The one I’ve just recently bought from Fasttech has more dots per row (6/7) and they are less noticeable, what is a good thing when the light is focused. Another difference is the lack of the small “+” sign on top beside the dome at the newer XP-G2. I was afraid they are fake, but the colour of the base and other characteristics match the known CREE’s I have. Brightness is also on the same level.
I’m wondering if I should use 2 or 3 of this nichias for a mountainbike-helmet-light with a LD-2(M) driver set to 6A. Is it safe to use them constantly for 20 minutes on 2 or 3 amps, if well cooled?
I red that they survive much higher currents for seconds, but has any of you tried them with higher currents for a longer time?
I’m wondering if I should use 2 or 3 of this nichias for a mountainbike-helmet-light with a LD-2(M) driver set to 6A. Is it safe to use them constantly for 20 minutes on 2 or 3 amps, if well cooled? I red that they survive much higher currents for seconds, but has any of you tried them with higher currents for a longer time?
The test in the OP was on a DTP copper board. Assuming such boards, my gut feeling tells me if they survive 10 amps for a short period without damage, they will survive 3 amps, which is well below its maximum output, without problems for a prolonged period.