Mixing red and white LEDs to get a truer white

So has anyone tried this? I was thinking about maybe 6 white 5mm and one red 5mm dialed back a bit. Would it have the desired effect of making the white seen more true?

TIA,
LT

So no one has tried this? I saw a desk lamp that used colored and white LEDs to produce "natural light" but haven't seen anyone try it in a flashlight. I'd be interested if it could work.

OK then, I guess it's up to me to try this out. And since nobody is interested I'll keep the results sealed under lock and key. /ubbthreads/images/graemlins/nana.gif

-LT

White light, (well at least what we perceive as white), is the combination of the spectral colors, look at the LED binning information for further info.

Adding a red might smooth some of the blue out that is present in most "white" LED's making it a bit warmer, but as for "truer" I can't see how it would do that.

An interesting experiment though!

Bill

bwaites,

All the graphs I've seen have a spike at blue and a dip at red. I've also noticed that reds don't show true when lit with a LED so I figured sticking a little red in it would help.

The desk lamp I mentioned above uses reds and oranges I believe but I've not seen it in person and the web pic doesn't show the head well.

I'm guessing that the "warmer" Luxes are more dialed down blue than cranked up red so I'm not certain the effect would be the same.

Do appreciate the response. I was beginning to hear an echo.
-LT

Mixing red and cool white will just add a red spike to the emission graph. No idea what that will do to the perceived output.

During an episode of CSI, there was a scene where they were using a "sewer cam". It appeared to be a camera surrounded by sevral concentric rows of LEDs. I noticed that maybe 1 in 20 of the LEDs were red. The rest were white. Assuming this was a real-world device, I would think it was exactly what you're proposing - an improvement in color rendition by filling in the "hole" in white LED's output.

I was at the Strategies in Light conference today. As it turns out, Lamina was showing a "soon" to be released 5 watt array which exactly does this. Most of the cavities were white, and then they had a dial which allowed turning on variable amounts of red.

Made a significant difference in the appearance of red objects on paper, as expected.

Good idea.

Since your eyeball has red, green, blue "sensors" in it, and your brain makes white from responses of these (photopic response) (scotopic is a different ball of wax), why not make the light peaked in the center of R, G, B as your eye senses it?

I have ben using Luxeons in bike lights for a couple of years now and have noticed the red rear reflectors of cars do not show up as well as when I am using halogen lights.

I have been tempted to add some separate red leds, or use 'warm white' Luxeons which have some red-emitting phosphor.

BTW1 Foliage looks a little odd too.

BTW2 - I sympathise with Lunal Tic with the lack of interest shown in the initial post. When I posted the lack-of-reflection-from-red-reflectors issue on CPF, no one replied.
I suspect there are few interested in anything other than maximum apparent brightness.

[quote

BTW2 - I sympathise with Lunal Tic with the lack of interest shown in the initial post. When I posted the lack-of-reflection-from-red-reflectors issue on CPF, no one replied.

I suspect there are few interested in anything other than maximum apparent brightness.

[/ QUOTE ]

I read his post earlier, and had to think about it a little before responding. A couple of replies over a few day period for a new and interesting idea is not so bad IMHO. I suspect that he might have gotton more response if he posted this in the LED forum which is more oriented toward LED light output / light mixing discussions.

This forum area tends to be more of a "I modded / made this nifty light, come take a look".

Good idea. Move this thread to led. I too have noticed the differences in folage colors at night, when using led and incandescent lights. There is more contrast in colors using incandescent lights. Led light tends to wash out the brown, red, and green colors.

Bill

Guess I figured the LED forum was more for stuff that's been done and this one is for stuff that hasn't.

-LT

Depending on the LED that you are using, you might actually want to use amber LEDs to balance the blue.

Your eye has red, blue and green receptors. Yellow light is a mixture of green and red light, and when yellow light is added blue light you get white light.

If your LED is putting out a pale blue, the addition of yellow may balance it. If it's purple, some green may balance it. It you have cyan (blue green), add red light.

Daniel

Hi LT, There is no perfect place to post an idea / question like you have - My suggestion was more based on the typical readers that post in LED vs the custom light area, not correct vs incorrect. Playing with LEDs, colors, and vision is very fun and intellectually interesting.

As you get more into it, you will quickly find that color vision is quite a bit more complex than the 2 D CIE curve or the pigment mixing charts we used as kids. There are some significant age / genetic / hunan factors that also affect how we see colors.

IMHO, these charts provide a useful simplistic way to tell an engineer how to "fool the eye under certain conditions", but really do not tell you how you actually see color and objects.

A good example of this is producing psuedo white light using blue / green + orange / red. If you draw the line through the CIE chart, you can make a binary system which passes through the white plankian, and the light produced is actually quite efficient due to the LED efficiencies at those points. Some colors come out fine / acceptable, others are completely missing.

Flashlights made from LEDs are a good example of balancing brightness, ability to see colors well, efficiency, and complexity. There isn't a "correct answer", just personal use / taste.

What modern "bulbs" have enabled, really for the first time, is the ability to have a choice in your lighting tools. There are some very interesting proposals on why 5 - 8 LED colors should be mixed to make better light. NIST has an interesting program (carried to the extreme of course) to callibrate spectrometers using MANY LEDs covering the human vision range. (I think it was every 3 - 5mn)

I was originally thinking of light that would render a truer skin tones or colors of flowers for photography. Also figured doctors would appreciate the better color for quick exams. I was also wondering why no one had tried to use a RGB LED and fire them all to get white. I've got Sauce lights that use multiple LEDs to do that but haven't seen a single LED that does. I may have something like this put together in the next few months and if it works I'll post here.

Thanks for the input.
-LT

Years ago, I made a LED headlight using this concept...

Today I did a few experiments: as you can see in all the spectral graphs for LEDs, there is a large spike arround 470 nm. This is the blue LED used to excite the YAG phosphor. There is absolutely no light from wavelengths lower then 440 nm. At 500 nm there's a large gap and the reds above 630 nm are also defficient.

If we can fill these gaps in the spectrum the color rendition *should* be a lot higher.

In the pictures below I used an array with normal 5 mm white LED's (china made, not Nichia). The LEDs used for the color correction are near UV 395 nm, cyan 505 nm and red 635 nm (I think 660 would be better...didn't have those).



-> White LED's, no correction



-> White LED's + cyan correction



-> White LED's + red correction



-> White LED's + violet correction



-> White LED's + full correction (violet+cyan+red)

Also note that the CRI (color rendering index) of nichia white LED's ranges from 65-95...95 is almost perfect (my arc aaa has a perfect color rendering...almost exactly the same as sunlight).

I hope this is the info you were looking for

Wim

All fluorescent bulbs have these spikes as well, and munufacturers publish the CRI charts. They may look perfectly normal to humans with the spikes but your pets may see, and be bothered by, the fun-house colors.

Don't forget too that the phosphors in white LEDs have a finite life, just like the phosphors in fluorescent bulbs and televisions. Near the end of their useful life the output will have gradually faded to less than half of new.

Over time, the different phosphors also degrade at different rates and you get a color shift. If you make a LED light using only red, green, and blue LEDs (no phosphors) with similar outputs, the color and intensity will not change over time.

[/ QUOTE ]
All fluorescent bulbs have these spikes as well, and munufacturers publish the CRI charts. They may look perfectly normal to humans with the spikes but your pets may see, and be bothered by, the fun-house colors.

[/ QUOTE ]

True, but the more you "fill" the spectrum between the spikes, the better the CRI will be.



[/ QUOTE ]
Don't forget too that the phosphors in white LEDs have a finite life, just like the phosphors in fluorescent bulbs and televisions. Near the end of their useful life the output will have gradually faded to less than half of new.

Over time, the different phosphors also degrade at different rates and you get a color shift. If you make a LED light using only red, green, and blue LEDs (no phosphors) with similar outputs, the color and intensity will not change over time.

[/ QUOTE ]

This is not true: the red, green and blue dice age differently. The blue die will age faster than the green and red, so the output will change for sure.
This aging process can be controlled by software...

White LEDs only use 1 yellow phosphor...as it degrades, the output will become more blueish.

[ QUOTE ]
NewBie said:
Since your eyeball has red, green, blue "sensors" in it, and your brain makes white from responses of these (photopic response) (scotopic is a different ball of wax), why not make the light peaked in the center of R, G, B as your eye senses it?

[/ QUOTE ]

There will still be differences in what happens when you shine that light on objects. While you can make "yellow" light, for example, by combining red and green, a mixture of red and green photons is still physically different from a beam of yellow photons. So you can get different reflections from objects if those objects preferentially respond to different parts of the spectrum.

Consider a surface that absorbs red light but not yellow or green. A beam of yellow photons will be reflected as yellow, but a mixture of red and green photons will be reflected as green since the red photons will be absorbed.