[Reference] Nichia E17A/E21A (2000K - 6500K, R9050/R9080, color) CCT and tint shots

SKV89 · 2019-04-19T23:00:44.000Z

Thanks Maukka for providing those extremely useful test data for us. I have a few questions.

1) Why does tint mixing drop R9 so much on the 2000k+ 6500k

2) Could the low R values for 2200k be a fluke?

3) If I remember correctly the batch of E21A you tested last year was much rosier. Is that correct? This batch seems to hug the BBL similar to the Optisolis.

Also a request if it doesnt require too much effort is if you can test 2000k+2700k, 2000k+4000k, and 2000k+4500k. Of it require much effort to setup the test then nevermind.

Thanks again for your hard work and contribution.

clemence · 2019-04-19T23:07:34.000Z

SKV89:

Thanks Maukka for providing those extremely useful test data for us. I have a few questions.

1) Why does tint mixing drop R9 so much on the 2000k+ 6500k

2) Could the low R values for 2200k be a fluke?

3) If I remember correctly the batch of E21A you tested last year was much rosier. Is that correct? This batch seems to hug the BBL similar to the Optisolis.

Also a request if it doesnt require too much effort is if you can test 2000k+2700k, 2000k+4000k, and 2000k+4500k. Of it require much effort to setup the test then nevermind.

Thanks again for your hard work and contribution.

Let me try to answer those.
2000K has quantitatively more red than 6500K. 6500K measured with D65 standard while 2000K with incandescent baseline. Combined, the resultant R9 will result in less R9.
Early test done at 100mA current. Some LED produces less CRI at low current, and more R9. Different behaviour for different phosphor chemistries.
2000K and 2200K rated at R9050 not R9080.

[Clemence]

goshdogit · 2019-04-19T23:11:54.000Z

These two scenes are amazingly similar!

What’s that game called, Clemence?

clemence · 2019-04-19T23:28:24.000Z

goshdogit:

These two scenes are amazingly similar!

What’s that game called, Clemence?

AFAIK it starts with something “…find hidden objects…”

goshdogit · 2019-04-20T00:14:49.000Z

clemence:

AFAIK it starts with something “…find hidden objects…”

Thanks, Clemence! :+1: I hadn’t seen this ‘hidden object’ game genre before.

It seems your image is a screenshot from a game called ‘Criminal Case.’

maukka · 2019-04-20T08:24:20.000Z

clemence:

Amber looks like a low CRI extremely warm LED. Mixing red and amber can produce something interesting…I request a test for this Maukka.

Amber+Red

maukka · 2019-04-20T08:33:00.000Z

RGBA. Bad Rf, huge Rg. Very magenta.

iamlucky13 · 2019-04-20T15:13:12.000Z

maukka:

RGBA. Bad Rf, huge Rg. Very magenta.

Thanks for making that measurement. I suspected it would end up being unbalanced. Trying to build anything around this mix would result in a freakshow light if current to each channel can’t be managed separately. One the other hand, being able to adjust each channel individually would be what would make an RGBA light fun.

It seems like blue would need to be dialed back quite a bit, and perhaps the amber increased a little in order to get a more natural result.

maukka · 2019-04-20T15:16:09.000Z

Sure, by individually addressing the emitters, you can tune it to anywhere within the color space quadrangle they cover.

It’s trippy.

clemence · 2019-04-20T15:23:08.000Z

iamlucky13:

maukka:

RGBA. Bad Rf, huge Rg. Very magenta.

Thanks for making that measurement. I suspected it would end up being unbalanced. Trying to build anything around this mix would result in a freakshow light if current to each channel can’t be managed separately. One the other hand, being able to adjust each channel individually would be what would make an RGBA light fun.

It seems like blue would need to be dialed back quite a bit, and perhaps the amber increased a little in order to get a more natural result.

To save you time:
http://www.nichia.co.jp/specification/products/led/ApplicationNote_SE-AP00042-E.pdf

[Clemence]

clemence · 2019-04-21T08:25:14.000Z

Normally, the middle rank is the safest for good output at or below BBL tint. The higher the output the farther up from BBL the tint.

[Clemence]

clemence · 2019-04-21T09:40:10.000Z

Red E17A is perfect for turtles!
https://conserveturtles.org/beachfront-lighting-turtle-friendly-fixtures-lights/

[Clemence]

maukka · 2019-04-21T12:43:20.000Z

Added lumens and efficacy numbers to the Google Sheet.

Individual Current-Flux-Voltage curves

2000K
2200K
2700K
3000K
3500K
4000K
4500K
5000K
5700K
6500K

Tint in an integrating sphere without any optics, i.e. how these measure diffused and how the binning is done (much closer to advertised CCT values)

Individual CRI data

2000K
2200K
2700K
3000K
3500K
4000K
4500K
5000K
5700K
6500K
2000K+6500K

clemence · 2019-04-21T13:55:19.000Z

GREAT job Maukka . These integrated measurements is extremely close to OEM rated, a good example of what to expect from 3 steps binned Nichias. Many thanks for all your efforts.
FYI, this also closer match to my tint shots in the OP (no optics, matte white reflector)

[Clemence]

Ryzbor · 2019-04-22T10:40:02.000Z

Great job as usual Maukka, thank’s for measuring the requested mixes :+1:

maukka:

Ryzbor:

4000K+4500K as a neutral tinted (?) 4250K

Not much reason to mix such close color temperatures. The difference between 4000K and 4000+4500K will be practically indistinguishable.

The thought behind this is to get the tint of a d240 4000K Wizard Pro below the bbl on the medium modes since the current setup is above it even on M3 ~250lm as bob_mcbob measured. And I see 4000K+4500K is slightly below the BBL at around 450lm.

Speaking of, Maukka could you add info about which flux binning you used in cases more than one is available? Eg. your previous the 4000K you measured abut a year or more ago had the d220 bin but now you’ve measured the d240 bin.

I find the results a bit optimistic besides the 5000K (which bin?). At 700mA/led, thus around 990lm it’s pretty obvious the tint is going to be below BBL and the E21A is well known for that. Even the XHP50 WW in the Wizard is below BBL on turbo. 990lm is more a turbo mode than a regular used and sustainable mode in a typical host. What really matters is how every setup performs a medium mode and a low mode which are going te be used by fat the most in real world usage. IMO it would be more interesting to know how they perform at lets say 200lm or even less. That’s where the “big table” helps.

clemence:

Normally, the middle rank is the safest for good output at or below BBL tint. The higher the output the farther up from BBL the tint.
[Clemence]

I couldn’t agree more…

But I’m actually open for a positive duv setup for special tasks. For walking in the woods I actually prefer positive duv above a pink tint, maybe because trees aren’t pink and leaves look less green under a pink tint

Or maybe a 4500K (rosy)+5000K(greenish) mix to create a true neutral 4750K…

maukka · 2019-04-22T10:47:33.000Z

Ryzbor:

Speaking of, Maukka could you add info about which flux binning you used in cases more than one is available?

Clemence probably knows the flux bins. I just got the bare leds on mcpcbs.

clemence · 2019-04-22T11:12:31.000Z

maukka:

Ryzbor:

Speaking of, Maukka could you add info about which flux binning you used in cases more than one is available?

Clemence probably knows the flux bins. I just got the bare leds on mcpcbs.

New E21A flux bins measured by Maukka can be checked via my store. Sm503 = D240, sm403 = D240.
No LED can satisfy your needs for consistent tint. You asks for below BBL from the very low current yet stays there at very high current. This is almost impossible with such wide operating current, not with current technology. That’s why PWM control is highly desired in CCT and tint critical lighting.
If you want to get the most consistent tint and CCT, go with Optisolis. Or any mid power LED. There’s no magic here, with such low current range you wont get tint/CCT swings as those in high power LED.

[Clemence]

clemence · 2019-04-22T11:14:20.000Z

FYI, we tested all those LEDs at 700ma because that’s the current used for all E21A binning standard (Nichia)

[Clemence]

Ryzbor · 2019-04-24T12:48:08.000Z

clemence:

No LED can satisfy your needs for consistent tint.
[Clemence]

I don’t have needs for consistent tint. But I want them to have a decent tint at 100lm and above and my 3000K+4000K(d220) Wizard and 3000K d220 non-mixed Wizard give me that. That’s one of the reasons I’m extremely satisfied with it. I use low and medium modes the most and this mix gives a good tint even on low modes.

Eg. a 2x3500K+2x4500K mix at 0.1A/led ~140lm/otl will give a duv of (0,0007−0,0014):2= −0.0007.

malkoff_junkie · 2019-05-21T21:49:50.000Z

are these possibly still an idea to make?. I need a 16mm MCPCB for Carlco 10507. E21A 503.

clemence:

Tally-ho:

Will you make a triple MCPCB for carclo 105xx or a quad MCPCB for carclo 106xx if the E21A makes a decent beam with those TIR series ?

Triple yes, quad not yet

[Clemence]