NEW Nichia emitters NV4(W/L)144A(R/M)

Copied from my thread in CPF:

Hi guys,

I just knew Nichia launched the 5 x 5mm 6V and 12V array emitters. And in my opinion they're VERY NICE!
Look at these links:

Compared to XHP50 this NV4W144 doesn't have the "X" die, better beam with secondary optics. The trend is going towards colour quality not just raw power. At its finest colour grade, this NV4W144 beats XHP50 in both efficacy and quality. Available in min 73CRI, 83CRI, and 93CRI & R9 min. 50!!!

I overlayed the the chart so you can see what I'm talking about. It's also better than Nichia 219B in colour richness especially in 480nm region. Nichia really cares about colors judged by it's detailed spectral chart details available for public. You can plot detailed head to head emitter comparison using their spectral datas, unlike Cree. I don't know where I can get detailed relative spectral distribution chart for Cree emitters. Cree never publish the complete chart for all its emitters CRI and CCT range.

[IMG]nichia vs cree by Clemence Ronald, on Flickr[/IMG]

Tailored with color gel filter to reduce the blue spike, the R9050's curve can be made almost perfect.

Below is the Excel spreadsheet I made to compare recent favorite emitters for flashlights. Cree datasheets combined with it's PTC made comparing very easy. That's not the case with Nichia. So, I decided to use Nichia's standard in the comparison spreadsheet. To make it more objective I used wattage rather than amperage.

[IMG]emitters comparison by Clemence Ronald, on Flickr[/IMG]

But wait, there's a problem.....

[IMG]NV4W144 footprint by Clemence Ronald, on Flickr[/IMG]

I’m more Nichia than I ever Cree before :smiley:

check out this thread ” Nichia 144A and 229A series: bigger dies, more output, 90CRI included,…….but no thermal slug :-(“:Nichia 144A and 229A series: bigger dies, more output, 90CRI included,.......but no thermal slug :-(

Djozz did it first! Didn’t know about it. Well, I did it first too in other forum hahaha.


These are just preview. I’ll let the honour to Maukka and Djozz.

So are both anode and cathode doing the heat transfer?

90% through the cathode. The 3mm thick copper base is connected to the cathode. And yes, it works better than Sinkpad (at this huge 30mm x 30mm board size)…
Depends on applications, you might use dielectric material down the base.

Okay, i see.
I just posted some of that thought in the other thread, isolating between star and pill, not between LED and star.

Already replied to that too :wink:

That looks fantastic. If you don’t mind my asking, where did you get these? I’ve looked everywhere and nobody seems to have them. I would love to swap an xhp35 zebralight.

Direct, from Nichia Japan. Those are samples for us to test. None are for sale until Q1 2017

Any hints of models that will be compatible with Cree boards? A 6V version of the 219B/C would be so nice if it could be dropped directly for XHP.

There is no 6V version for 219 series for now.
The XHP50 equivalent are just NV4(W/L)144AM 12V and NV4(W/L)144AR 6V. And you have to buy/make custom board.
I already have DTP MCPCB specifically made for them though.

I will buy those LED in bulk quantity for my own project. But I have the option to order different colour temperature with the same LED. Just in case any of you interested, please vote for your favourite colour and voltage.

Note: I will order the Ra9050 version only

For the CRI and colour test result you can check Maukka’s Tests

If people wanted to put these in a light, I think it might be easiest to go body positive and load the batteries in backwards. Wouldn’t be hard to stack an existing driver that way, or I could draw one up specifically for that purpose.

clemens, are the 90 % due to soldering the cathode to the larger copper part of the board, or is it the LED itself that transfers 90 % of the heat through the cathode?

In other words: Can I solder the anode to the larger copper part of the board and still have the same 90 % heat transfer?

The reason is that I possibly will build a light with reversed battery where the host is the positive conductor.


OK, I have to clarify my new findings here….
I realized something when I was thinking why such a big chunk of VirEnce board resulted only in moderate boost in performance compared to Djozz board.
The 90% cathode(-) : 10% anode (+) proportion applies ONLY to single die LED without thermal pad such as 119A/B/C, E21A, etc….
I made this board and totally FORGOT that 144AM/AR is a totally different animal. This is a multi die LED consist of 4 x 119C sized small dies configured in 2xSeries - 2xParallel.

- For 144AM this means each pad shared the same thermal load from cathode and anode of the two small dies.

  • As for 144AR, it’s a 4S - all wired in series. Each pad shared the same thermal load from cathode and anode of the four small dies.

I came to this conclusion since both pads are of the same size and symmetry. The correct proportion for this configuration is 50:50!
Thus if I made the VirEnce board with Anode/Cathode 50:50 the performance COULD be higher. I haven’t done any test yet.

So to answer your question:

- Yes, 90% of the heat generated by most single die LED will be transferred to the cathode. (confirmed by Nichia’s test results)

- Some multi dies LED such as 3V Luxeon MX does behave like a single die since it’s wired in all parallel.

  • You can try to reverse the connection in the VirEnce S3030 knowing it’s wired 2S 2P. Should performed the same either way.

Interesting test result (“90:10” examples):

- Clemence

EDITED: 144AR description added

Very interesting, clemens. Thanks a lot for clarifying!

Thanks for that update, Clemence.
Good to know it’s 50-50.