Nichia 219C, testing a 5000K 83CRI emitter, comparing with a XP-G2 S4 2B and other leds

Thanks for sharing that Dale, that’s good to know. After the tint changed, did it return back to normal or is it permanently altered?

I just put together a A6 with a Nichia 219C (on Noctigon) with a LD-1. I like the results but is gets super hot pretty quickly. I’m not sure how practical it is in a light with so little mass to draw heat away. Then again, I don’t think it’s any worse than a BLF A6.

The hot spot is smaller and more defined than the XP-L and it is more intense. It seems to throw a little further but I don’t have a way to measure. Tint is pretty nice as well, but not as nice as the XP-L 3D.

I’m going to remove the LED and try it in either a S8 with a SMO to see if I can get a little more throw.

So far the tint appears normal with a lesser cell. I just can’t use a 30Q or 35A or LG HE-4 or….

Put together an A6 with a 5000k 219C and I have to say I’m pretty disappointed. Compared to my other A6 with a V2 3C XPL-HI the 219C seems inferior in every way. On a NCR18650GA it pull 5.5 amps, 1 more than the XPL-HI, to produce ~10% less light, ~35% less lux at 5m, and has a cooler tint.

The lower output isn’t unexpected; it’s the lack of throw and tint that got me.

I just did this combination in my Tool build, my multimeter says 2.8A, however I know it’s more as it’s visibly brighter on shorting the leads. My guess is 3.5-4A. The 219C can take it but requires a DTP star as it goes blue if fed that much on a “normal” star. The efest gets quite warm if left for 30+s tho, but seems OK.

I have both the 4000k and the 5000k 219C now. I prefer the tint of the 5000k as it’s more neutral to my eyes, whereas the 4000k is leaning towards warm. The 5000k is quite similar in tint to the famous 92 typical CRI 4500k 219B, however the reds (in both the 4000k and 5000k 219Cs) are visibly more subdued than the 219B.

Hence Nichia’s guaranteed minimums (lower than typical values) of:
219B: Ra 90, R9 50
219C: Ra 80, R9 0
Ra being the aggregate score of the 8 colours used to calculate CRI and R9 being the colour “Natural Red” which is sadly not used in the CRI calculation. R9 values of 50+ are extremely useful in surgical lighting as many tissues and blood are shades of red.

EDIT (source): Why the LED R9 value isn’t important – LeapFrog Lighting

^

Wow, really good info in the above post t&b. Thank you :)

In case you didn't see them, I posted some comparison beamshots in "My Bike Light Thread" post #46 & $47. I modded a Nitefighter BT40S that came stock with NW XP-G2's (which I had compared tint with other lights and found to be 4C or very very close) over to Nichia 219C's. I then compared these Nichia's to my favorite XM-L2 5B1 tint.

(So I did get around to using them without letting them sit on my workbench for months!)

-Garry

Hi,

I was able to get in on RMM’s recent 219C special offer, with three 219Cs mounted on a Noctigon along with Carclo optics.

My order from RMM also included the MTN-17DD FET driver loaded with Toykeeper’s excellent firmware / UI, the Convoy S2 host, the S2 copper spacer/heatsink, and Samsung INR18650-30Q battery.

Last weekend, I assembled the above S2 with the triple nichia 219C noctigon, the S2 copper heatsink/spacer, and did a switch and spring bypass (with 18 gauge wire) on the S2’s clicky. The MTN 17DD driver already had a spring bypass done by Richard with copper braid.

I kept the stock 17DD driver wires provided by Richard to attach the driver to the nichia LEDs on the noctigon. I think these are 22 gauge wires, and I am contemplating replacing these wires with something heavier, given that the projected amount of current draw with this setup (over 10 amps?) is in excess of 22 gauge wire’s rated maximum current carrying capacity. I’d like to use 18 gauge wires to attach the driver to the LEDs, but these may be too thick to solder onto the driver’s + and - pads. Additionally, the thinner 22 gauge wire, however, may be acting as a resistive “safety valve” to limit current to the nichias, and it may be better to retain the smaller gauge wires provided they don’t eventually act like fuses when running on turbo.

So far, I have run this light configuration on turbo for no longer than 15~20 seconds or so - I’m not yet brave enough to run it on turbo much longer. :wink: . I’ve done this quite a few times over the last 3 days, and the 219Cs seem to be taking it just fine so far.

It’s output is truly stunning on turbo!

To my eyes, I don’t see any tint anomalies when running on turbo, just that pleasant nichia whiteness.

I don’t have a way of measuring the light output, but doing a ceiling bounce with the above S2 configuration running on turbo lights up an entire bedroom! I have a Trustfire X100 (9 LEDs) that appears, to my eyes, to have similar ceiling bounce performance. The x100 light is supposed to emit roughly 3000 OTF Lumens.

On turbo, the S2 copper spacer/heat sink and S2 flashlight body get warm pretty fast (in about 7 seconds or so), but the heat sink/spacer appears to be doing a good job at removing what must be significant amounts of heat density. This seems to help keep the nichias from burning up (at least for short turbo bursts for roughly 15 seconds).

The 30Q battery seems to be one hell of cell when coupled with the 17DD+FET driver!

I don’t think I have a way to measure the amount of current that must be flowing in the setup. My meter has a 10 amp maximum capacity. I hope to investigate means to measure the current in this setup a bit more this weekend. Don’t know that I’ll be successful, but will post any findings or observations.

I just tested this with a KeepPower 18350 in my S2+ with one of RMM’s triple 219C’s from the recent sale.

I used a clamp on ammeter and some 12ga stranded copper (5 inches long) from negative on battery to side of tube.

I got 11.2 A initially, from an 18350…

I held it for about 10 seconds until it got too hot, it was down to 10.7 A

Holy smokes these 219C’s are awesome for pulling amps… That is about 3.5A for each emitter and according to Djozz’s chart in the OP, about 3000 LED lumen.

As for the heat, ten seconds was all I could handle holding it in one hand. I will have to adjust the firmware. With the XP-G2 S4 2B triple prior, I had it start to step down at 15 seconds, taking about 35 more seconds to ramp down to 50%…. That will not do for these little heat generators.

The tint is outstanding….

Hallo

I’m new on this Forum and i read a lot of interesting things.
The new Nichia 291c with that low Vf maybe is good for LiFePo in Direct Drive ??

Yes, it will never be a pocket-rocket but it is possible. You will have to remove some resitances in your flashlight, mainly bypassing the spring(s), then the 219C should run at up to 1.5A direct drive on a LiFePo. I like LiFePo's because the chemistry is relatively safe. :-)

I just put one in a cheap SK68 clone. The 219C on copper draws 3A DD from a protected 14500. Not much runtime, but it sure is nice.

Yes the chemistry is nice.
1,5A only ??

I looked at HKJ's test of a 14500 LiFePo cell and compared the currents/voltages with the 219C current/voltage and that comes down to a bit over 1.5A, ignoring any resistances in the system. A 18650 size LiFePo cell should do a bit better.

Well this is interesting, I had not considered that LiFePO4 cells would be good for DD until reading this. Would be interesting in the Noctigon Meteor M43 that Richard had with the 219C… Plenty of light, pretty cool for the amount of light being put out and safer chemistry…

Aahhhh……ok, that for the 14500. I try to build a “High Beam” for the Bike so i can use bigger cells.

Thanks for information

An old Illumina Ti LED/MCPCB upgrade mod to a 219C 4000K (Cutter):

Stock light did 241 lumens @30secs, 219C/10mm SinkPAD does 296 @30 secs in a much better tint and better beam pattern as well. Both tests were done on a EFEST 10440.

Nice upgrade, Tom!

Tom, do you have any of the 10mm FET drivers? Those will get you up between 800 and 1000 lumens on an Efest IMR10440.

Nice job indeed, I didn’t realize those have a pill in em. Gave mine to my daughter, wonder if she’s lost it yet… 20 yrs old and 4 mos pregnant.

Yea, was think'n bout replacing the driver for sure. Actually I got another one I want to mod as well - might do with the 5000K from Richard. Since it's a contact board/driver design, should be perfect for our 2 sided FET based 10mm driver... Hhhmmm... I know I got a couple.

Also need to do something better with that alignment piece, and the contact board is not sitting down flat, so it's a little more flaky switching modes that it used to be. I left the tiny stock LED wires - might be ok with the low Vf of the 219C, but definitely could get more power and heat for sure out of a FET with heavier wires. Didn't have much time of course.

Nice mod Tom :slight_smile:

I am working on a Olight I3S & that Illumina Ti led holder & the pill looks identical, i wonder if Olight is the OEM for the Illumina Ti.
The biggest difference seem to be Olight used lots of threadlock + glued the driver well in place unfortunately, i have gotten the threadlock lose but haven’t attacked the glued in driver yet i want to test an XP-L on it before i have to probably destroy it to get it out :frowning:
Good point about contact board & the 2 sided 10mm fet driver, the pill is very deep for something so small, i will have to see if the I3S also has that setup also, the pill should be deep enough for it anyway.

If you are going to test an 10mm fet driver in the Ti wouldn’t the alu pill very quick get overheated due to the bad heatsinking of Ti?
I guess you could use an very short turbo timer to save it destroying it self though.

I hope you try it, it is always interesting to see more hotrodded 10440 lights pushed to the edge :slight_smile: