Nichia 219 lumen sphere testing

[foxtrot824]

The Nichia NVSL219AT-H1 otherwise known as the 219 92+ CRI emitter has been a treat to the flashlight community so I decided to do a quick test in the small sphere

mA	v	w	lm	lm/w	CCT	x	y	CRI
100	2.806	0.2806	35.09	125.07	4502	0.3609	0.3646	89.2
200	2.948	0.5895	68.27	115.80	4526	0.3599	0.3631	89.1
300	3.056	0.917	99.31	108.30	4536	0.3593	0.3618	89.1
350	3.099	1.084	114	105.10	4543	0.359	0.3612	89.2
400	3.14	1.256	128.2	102.07	4550	0.3588	0.3604	89.2
500	3.216	1.608	155.5	96.70	4559	0.3583	0.3593	89.4
600	3.28	1.968	180.9	91.92	4566	0.3579	0.358	89.6
700	3.338	2.3366	204.6	87.56	4573	0.3574	0.3568	89.8
800	3.394	2.7152	226.7	83.49	4583	0.357	0.3556	90
900	3.44	3.096	247.4	79.91	4588	0.3566	0.3543	90.4
1000	3.486	3.486	266.2	76.36	4597	0.3562	0.3531	90.7
1100	3.536	3.8896	284.8	73.22	4604	0.3557	0.3518	91
1200	3.578	4.2936	300.4	69.96	4613	0.3553	0.3506	91.4
1300	3.617	4.7021	314.3	66.84	4622	0.3549	0.3492	91.8
1400	3.656	5.1184	326.3	63.75	4631	0.3544	0.3479	92.2
1500	3.694	5.541	336.4	60.71	4642	0.3539	0.3466	92.7
2000	3.886	7.772	360.2	46.35	4718	0.3509	0.3395	94.9

I was curious to see how the emitter performed through out it's recommended current range. While it has a lovely tint and great CRI it's, however performance drops below 100 lumens per watt pretty fast.

Below is a chart of the CCT shift as current increases.

 

[Illum]

The Nichia 219 safely operating at 2A is news to me. On their datasheet the maximum threshold for continuous operation is 1.5A. Huge dip on the 2A efficacy though. How are you cooling this little bugger?

 

[foxtrot824]

The Nichia 219 safely operating at 2A is news to me. On their datasheet the maximum threshold for continuous operation is 1.5A. Huge dip on the 2A efficacy though. How are you cooling this little bugger?

Not sure about safely If I remebered correctly the data sheet says it can be pulsed at 2A but not continuous. The jig that holds the LED on a star in the port of the sphere has some mass to it so the LED being tested stays cool for a short time. It doesn't stay cool for long so all of these numbers were pretty much just an instant on test.

 

[THE_dAY]

Thank you for the tests, really great info there!

Have you done any burn-in on the 219. I was wondering if this LED can have the Vf drop a bit.

 

[romteb]

Fantastic info, very usefull.

Thanks foxtrot824

 

[jtr1962]

Thank you for doing this! I've been curious about the Nichia 219's lumen output myself but haven't had the time to do a lumens test. Vf does seem a bit high although I suspect that will drop to 2.9V or so @ 350 mA after a few thousand hours. Yes, these aren't efficiency champs but then again no high-CRI LED is by its nature. In general illumination applications running the 219 at 350 to 500 mA seems to be the sweet spot which will deliver reasonable efficiency while still keeping the number of LEDs relatively small. It might be interesting to compare the efficiency of Nichia 219s versus a setup using standard whites plus red/cyan to deliver similar CRI. The latter might have a 10% or 15% edge in efficiency, but at the expense of circuitry to actively balance the output.

 

[foxtrot824]

Thank you for the tests, really great info there!

Have you done any burn-in on the 219. I was wondering if this LED can have the Vf drop a bit.

I have not, but if I remember I'll retest that LED after some use.

 

[Hoop]

Perhaps you could run the led at a constant 1 amp for a while and see how the voltage fluctuates as the heatsink temp increases? I am not sure that instant on readings are an accurate reflection of the operating voltages?

 

[neutralwhite]

thanks. by the way what is the Lumen output on the basic Eiger pocket torch neutral white/medium optic?.
they use the Nicha 219 LED right?. or am I lost ?.
thanks.

 

[foxtrot824]

thanks. by the way what is the Lumen output on the basic Eiger pocket torch neutral white/medium optic?.
they use the Nicha 219 LED right?. or am I lost ?.
thanks.

I am not sure what their stock drive current is and or optic losses are. This was strictly an at the emitter test, sorry.

Perhaps you could run the led at a constant 1 amp for a while and see how the voltage fluctuates as the heatsink temp increases? I am not sure that instant on readings are an accurate reflection of the operating voltages?

Next chance I have I'll do some longer test. The LED is currently just running at low current to get some hours on it.

 

[THE_dAY]

The LED is currently just running at low current to get some hours on it.

Hi,

If thats for burn-in, I'd recommend high current and the Vf usually settles after about 24hrs.
There was also mention in a couple of threads that heat helps with the burn-in process.

Check out this thread:

http://www.candlepowerforums.com/vb/showthread.php?130622-burn-in-data-from-UX1L-after-24-hours

Thanks for all your work! :thumbsup:

 

[gunga]

Any chance for lower current tests too?

 

[foxtrot824]

Hi,

If thats for burn-in, I'd recommend high current and the Vf usually settles after about 24hrs.
There was also mention in a couple of threads that heat helps with the burn-in process.

Check out this thread:

http://www.candlepowerforums.com/vb/showthread.php?130622-burn-in-data-from-UX1L-after-24-hours

Thanks for all your work! :thumbsup:

Noted!

Any chance for lower current tests too?

How low are you looking for?

 

[gunga]

I'd find it useful to go as low as 5-10 mA. I am setting up levels for custom circuits (eg. gdup etc) and would like reasonably accurate estimates instead of trying to read very vague charts. This information is much appreciated!

 

[Beacon of Light]

+1

I'd find it useful to go as low as 5-10 mA. I am setting up levels for custom circuits (eg. gdup etc) and would like reasonably accurate estimates instead of trying to read very vague charts. This information is much appreciated!

 

[BenChiew]

Thanks for publishing this.

 

[PCC]

Nice that you publish the measurements at 100mA intervals. It's increasing by about 30 lumens per 100mA at the low end and it increased by 10 lumens from 1400mA to 1500mA. Looks like there's not much to gain after about 1300mA.

 

[fnj]

I'd find it useful to go as low as 5-10 mA. I am setting up levels for custom circuits (eg. gdup etc) and would like reasonably accurate estimates instead of trying to read very vague charts. This information is much appreciated!

I would be VERY grateful for testing down to at least as low as 5 mA. I would be ECSTATIC if you would go down to 0.5 mA or even 0.1 mA. At least intensity if not CCT and CRI down there. Don't need a whole lot of data points; just enough to see if trends continue, or perhaps some reverse. I suppose measurements might get dicey at extremely low levels, but the information is of great interest to me. I am a fan of extra low; I know you couldn't tell that.

 

[netprince]

Great info. Thanks for doing this.

 

[Hoop]

A belated: Thanks for the post! Very useful.