I dont know anything about the cells but when I look at replacement ones on Amazon it says 3V nominal
Huh. Yeah. I just looked. You’re right. Hmm. Curious thing.

I just now am realizing it had 2 stacked CR2016 batteries so I am probably giving this thing way too much voltage? Maybe a 2032 battery would have been more appropriate, who knows. I guess if it dies i’m out 55 cents…
I have used and gifted dozens of these small lights, all the white ones use 2xCR2016 and i have never seen a single one with a fried led.
These coin batteries have a high internal resistance so the voltage drops when you load them with a led.

…
Huh. Yeah. I just looked. You’re right. Hmm. Curious thing.
CR2016 cells are “3V”, but their internal resistance is very high. Thus, 2x in series CR2016 cells drive a 5mm led well.
That and the fact that you don’t turn that on for a sustained period. Even if you overdrive these LEDs and degrade them way faster than their rated lifespan, you probably won’t notice it.
CR2016 cells are indeed 3V. The “CR” code denotes a lithium manganese dioxide chemistry, nominally 3V.
Alkaline button cells are 1.5V and are given “LR” code. The typical form factor is also fatter and taller than lithium coin cells.
A pair of CR2016 cells shouldn’t be any concern, but because of the low forward voltage on the Yuji’s, they run pretty well on a single CR2032 cell, in my experience.
I haven’t done a controlled test, but I’d expect you would get more stable output and significantly longer runtime out of a single CR2032 compared to a pair of CR2016’s, although you might get higher initial output out of the latter. The massive voltage drop the pair of smaller batteries experience is in part due to a lot of energy being wasted on internal resistance.
Either cell does better at these kinds of loads with intermittent use (a few minutes at a time) than continuous use. It’s surprising how much the voltage recovers when a load was removed in a test HKJ did:
Low current discharge of batteries [image] My usual battery tests uses a fairly high discharge current, but in many applications batteries will last for weeks or months with the device on. To get some better data for this I decided to do some low current discharges, i.e. with discharge times up to a few weeks. For this I I will show capacity down to a couple of different voltages. Due to the time involved I can only do it with a few batteries, other brands may be better or worse. …
Shipped
slowtechstef: RR146459731TH
Noctiluco: RR278756610TH
mitsuki08: RO247700037TH
Thank you so much, rngwn!!!
Mine arrived in Denmark today.
My wife have some tee-lights she bought in Ikea :
The tint in the original on the left is perhaps more candle like, but it also has a hint of green. I like the “upgraded one on the right better.
Thanks again for making these available.
LEDs arrived today, thanks!
Got mine yesterday! Thank you!
rngwn, thanks for making these LEDS in good bins available in small quantities for the BLF community
In my experience replacing a stock 5mm LED with a warm-white high-cri one can turn almost any shit flashlight into a hit flashlight. Well at least in my humble opinion
Action Halloween projector, 0.69 Euro
Kudos to iamlucky13 for the suggestion of using 2AA direct drive for a 5mm Yuji.
No PWM, no standby drain, no soldering required for 5mm LED swap
Current somewhere around 25mA with 2 “newish” 1.5V alkaline AAs.
Mod with Yuji 3200K LED.
Q.E.D.
Mine arrived today, thanks!
I did a voltage and output test on one of these 3400K 95CRI leds, and it does not disappoint, output is better than the Yuji, voltage is lower, and it can handle more current. The max is about 22 lumen at 100mA, going higher up to 120mA gets you a bit more light for a moment but it settles back to 22 lumen again within a couple of seconds. My test rig for 5mm leds is not great, the current can not be set very well and the graph shows it, but the output numbers are correct. I also re-tested a 3200K Yuji for direct comparison. This new tested Yuji has about 1 lumen less maximum output as the one I tested before, but apart from that the voltage and output curves are pretty much the same.
Also I have been running one of the 3400K 95CRI rngwn leds all day at 100mA, and it does not seem to bother the led much, output does not degrade, steady at 22 lumen. It is not quite a 50,000 hour test but it does tell something.
Edit: there seems an effect on tint though. I should do another test for confirmation but the led seemed more yellow and when I measured it again, the duv, that was –0.0038, after 12 hours at 100mA it went to –0.0007. CRI is invariably good.
Here is the tint data at 50mA, I also measured at 5mA and that is a bit better even but only marginally
I also measured tint in the warm edge of the beam, CRI is the same but in 2800K. The CCT has some variation apparently, the first led last week measured 3320K, and now 2 leds measure both at 3200K.
(the bottom leds are two of the 3400K 95CRI leds from rngwn, top left is a Nichia GS, top right are a cold and warm Yuji)
My LEDs have arrived - thanks, rngwn!
Now I’ve seen those test results, I’m really glad I bought some Thanks for the testing work, djozz!
Thanks for making measurements for us djozz. I’m impressed how well this LED performs - not only more output and able to be driven a little harder, but also a lower forward voltage than the Yuji’s. Taken together, that means 20% better luminous efficacy, which would be great in lights that use boost circuits like the C01.
And very similar color. I looked up your spectrometer test of the Yuji’s for comparison:

Djozz, did you do a spectrometer test on these? Sorry if I missed it. I have the E01vn with the 3200k Yuji and I like the tint on the SST-20 3000k 95CRI much better. So I just want to see how it compares. Wondering if Sofirn can offer a more expensive version with the SST-20. That would be a different flashlight. The C01 uses 5mm DIP leds and that makes the construction very simple and robust. Using SMD leds requires a construction with lens, optic, ledboard, shelf etc., it may be that Sofi…
I took the liberty of summarizing your two samples:
Yuji 3200K | rngwn 3400K | Yuji 5600K | rngwn 5800K | |
---|---|---|---|---|
CCT | 3162K | 3237K | 5390K | 5666K |
Duv | -0.0037 | -0.0038 | -0.0004 | 0.0010 |
CRI | 96 | 96 | 95 | 98 |
R9 | 84 | 84 | 94 | 93 |
Output, 20mA | 6 lm | 7 lm | 7.5 lm | 9 lm |
Vf, 20 mA | ~2.9V | ~2.8V | ~2.8V | ~2.9V |
Peak Output | ~14.5lm | 22lm | 22lm | 25lm |
Peak Output If | 80mA | 100mA | 90mA | 120mA |
Peak Output Vf | 3.6V | 3.3V | 3.25V | 3.45V |
(1) Yuji color tested at 20mA? rngwn sample tested at 50mA
(2) Maukka C01 average of 2 samples
* Edit - added Duv.
** Edit 3/24/23 - Reformatted table to appear properly on new forum software.
Thanks rngwn for finding these and setting up the thread to sell them.
Great to see this test results Djozz. This LED looks like a fine replacement for the Yuji.
The Yuji color is very special, probably second to none. Still rngwns LED is much more useful imo. Very bright in contrast with an evenly distributed beam (beamshot added in posting #102). Really nice. In the Fenix E01, it clearly outperforms the C01, but maybe the Fenix has a higher current driver.
unheard, nice work making your E01 much friendlier. It’s the perfect form factor for this time type of emitter.
(edit - corrected typo)

Action Halloween projector, 0.69 Euro
Kudos to iamlucky13 for the suggestion of using 2AA direct drive for a 5mm Yuji.
No PWM, no standby drain, no soldering required for 5mm LED swap
Current somewhere around 25mA with 2 “newish” 1.5V alkaline AAs.
Mod with Yuji 3200K LED.
Enjoy casting silhouettes on the wall in warm, high CRI!