Red LED's are Orange !!

lifeson99

Newly Enlightened
Joined
Jun 4, 2007
Messages
15
OK guys . . . I need some help here if anyone has a decent source. Let me explain. I have trouble finding "true RED" colored Water-Clear LEDs. They are almost all Red/Orange and in many cases they are completely Orange. I have tried several different brands but they all have a orange look to them. I bought 3 different packs of them from 3 Chinese eBay vendors . . . and all were orangish. Here is a photo I took of two of my typical Red LEDs - the diffused LEDs are great for use as Panel indicator lights, but for illumination I need the added brightness and the angluar focus spotlight from water-clears:

LED-Red-is-Orange-5mm.jpg

Diffused is Red - Water Clear is Orangish/Red
The actual LED is even more "orange" - the Camera
has somehow made it slightly more red than it actually is

Here is a "Red Fan" I bought . . . at least it was advertised as a Red fan. But when it arrived and I fired it up . . . Orange !! Just Google for Red Case Fans and you will see that quite a few of the vendor pics show an orange fan. This pic is from Xoxide:

LED-Red-is-Orange-fan.jpg

Even my Power LEds (Luxeon and Lamina) red die have an orange tint to them, especially when I turn up the current/brightness level.

One exception to this, is govt approved Traffic stoplights, which usually contain true, deep Red LED's. Has anyone found a good, solid Red LED on the market? I am looking for a recommendation.

Here is another exception to this problem. This is a decent sample of what I would like to find for my own Red LEDs (it is a gorgeous LED lamp, actually) :

LEDs-standard-RGB.gif

 

Citivolus

Enlightened
Joined
Apr 1, 2008
Messages
232
Location
Happily back in Sunny QLD
I think the problem is that you need to be more careful about wavelength. You are probably getting LEDs in the 600nm area, where true red is perceived as somewhat longer wavelengths. Here are a few that might be worth trying, in descending order from NIR through towards orange:

25mcd, 700nm, 30 degree:
http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail?name=67-1104-ND

2800 mcd, 660nm, 30 degree:
http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail?name=67-1612-ND

3000 mcd, 647nm, 10 degree:
http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail?name=404-1104-ND

The first one is likely much brighter than it sounds, but due to the response curve of the eye it appears dimmer.

If you are after a bit more power at the trade-off of cost, wavelength, and needing a heatsink:

23 lumen, 630nm, 120 degree:
http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail?name=67-2045-1-ND

34.9 lumen, 627nm, 100 degree:
http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail?name=492-1134-ND

44 lumen, 625nm, 140 degree:
http://www.luxeonstar.com/item.php?id=331&link_str=1385&partno=LXHL-MD1D

You may consider 625nm to be pushing it a bit towards orange, though. 660 may be an optimal trade-off of brightness vs deep red, and you may get the right amount of brightness at that level to not overwhelm whatever other lighting you have in place.

Regards,
Eric
 

jtr1962

Flashaholic
Joined
Nov 22, 2003
Messages
7,506
Location
Flushing, NY
Hebei has 640 nm reds for $0.047 each. Look for part number 530FR4C in their pricelist. 660 nm is even deeper red but less common. In the quest to get red LEDs ever brighter, apparently 625 nm is now considered "red". Eye response at 625 nm is several times greater that 660 nm, so for any given percentage conversion efficiency a 625 nm LED will appear several times as bright.
 

2xTrinity

Flashlight Enthusiast
Joined
Dec 10, 2006
Messages
2,386
Location
California
More on wavelengths color:

400px-Luminosity.png

The black curve is the human eye response, relating wavelength to overall perceived brightness. The "lumen" unit is defined in terms of this curve.

The green curve is the night vision sensitivity -- note that for your night vision to be truly insensitive to the red light, it must be 650nm or longer, which few lights are anymore. Note -- if you're interested, check out a laser pointer. It will likely say what wavelength it is on the side (usually around 650).

287px-Cones_SMJ2_E.svg.png

These are the frequency responses of the different colored cones in your eyes. heights are all normalized to one, but in reality, the M (or green) cone is the most sensitive. You can see that once you get to 650nm or longer, really only the L (red) cone is activated, so going longer than that won't change the color anymore, it will only get dimmer/less efficient.

Most red LEDs on the market are made to be around 625, due to the fact that it will appear a lot brighter. However, that is what I would call "red-orange" territory, as the green cone has some sensitivity there, too. For reference, low-pressure sodium lamps, as well as some yellow traffic lights, are around 590nm, so 610nm will appear appear to be "orange".

One phenomenon that you observed is that as you crank up the current, LEDs tend to shift to shorter wavelengths. As you can see on the graphs above, the ratio between the green adn red cone sensitivity changes a lot, so even a decrease of 5nm to around 620nm will cause a perceptible color shift. On the other hand, a similar amount of wavelength shifting in a blue LED won't really be noticeable, since there is really only one color cone in the 450nm territory.

The actual LED is even more "orange" - the Camera
has somehow made it slightly more red than it actually is
Cameras do not perfectly reproduce the frequency response of the eyes. Especially not for saturated monochromatic colors. Not all cameras are equal either. For that matter, neither are most computer monitors, so it's quite possible that everyone reading this thread is "seeing" slightly different colors in your comparison. That is why you're best off describing the LEDs in terms of wavelength.
 

IMSabbel

Enlightened
Joined
Dec 4, 2004
Messages
921
One other effect to consider is that for very bright light sources, the eye starts trying to white-balance on them.

Thats why molten steel seems to by white-hot, even though its really a deep red colour.
Bright red lights will also be "whitened" by our perception.
 

PPJ

Newly Enlightened
Joined
Feb 2, 2010
Messages
5
Hi, I myself have been asking the same question. The truly 660nm area is a very difficult one, where most brands in the market faile at best emittes red at 650nm. One tip, the ones to go on emittes a truly red colour with a pinkish tone, when you see it once, then you know the difference. The answer to your question currently in the market is this one as post #2 also describes:

SSL-LX5093SRC/E
5mm ordinary round through hole with a good strenght

SSL-LX5093SRC/F have been avaliable before but it´s not a better choise

the company site:

http://www.lumex.com/en/products/detail/standard54/high_intensity2/

Best regards PPJ
 
Last edited:

Erasmus

Flashlight Enthusiast
Joined
Dec 29, 2004
Messages
1,077
Location
Belgium
I have some red Cree XR-C LEDs. These are pure red, and perfectly fit the application I need it for (providing light in the dark while not destroying my eyes' night vision). Better than all those Chinese LEDs of which the specs aren't always correct.
 

PPJ

Newly Enlightened
Joined
Feb 2, 2010
Messages
5
(some pictures is to be added when I figure out how to do)
 
Last edited:

G12002

Newly Enlightened
Joined
May 14, 2010
Messages
9
630nm: also known as "Hi Efficiency red" or "HeNe red" only really look orange when you compare them to the purer red 660s. The same is true if you compare the 630nm LEDs to the pure orange 610s. The orange LED then begins to look like an amber which, in turn, makes the 590s look pure yellow. The context seems to change depending on which colour was viewed last.

IMSabbel said:
One other effect to consider is that for very bright light sources, the eye starts trying to white-balance on them.

Thats why molten steel seems to by white-hot, even though its really a deep red colour.
Bright red lights will also be "whitened" by our perception.
I think it's because our channels begin to saturate allowing the cones which are less sensitive to "catch up". Molten steel is white hot because it emits blackbody radiation which makes that colour at those temperatures.
 

IMSabbel

Enlightened
Joined
Dec 4, 2004
Messages
921
630nm: also known as "Hi Efficiency red" or "HeNe red" only really look orange when you compare them to the purer red 660s. The same is true if you compare the 630nm LEDs to the pure orange 610s. The orange LED then begins to look like an amber which, in turn, makes the 590s look pure yellow. The context seems to change depending on which colour was viewed last.


I think it's because our channels begin to saturate allowing the cones which are less sensitive to "catch up". Molten steel is white hot because it emits blackbody radiation which makes that colour at those temperatures.

Ah, come on. If molten steel were hot enough to be white in the blackbody spectrum, it were not molten, but plasma.
Iron evaporates way below the normal operation temperature of thungsten filaments in even normally driven incans. Look up just how "white" a 1800-2000K blackbody really is...
 

G12002

Newly Enlightened
Joined
May 14, 2010
Messages
9
Steel boils at little over 3000K (Halogen). Probably safe to assume it isn't heated to near that but BB radiation gives off a spectrum of colours so the colour has a low saturation; even more the case for lower temperatures where there is less of an emphasis for a particular peak wavelength. Standard incandescent bulbs are 2700K but the diffuse reflected, ambient light is an acceptable white. So molten steel, while unlikely to be at those temperatures, will give off a whitened amber. It wouldn't be white to the directly observing human eye if the colours were pure, you would just see orange because the blue cones are limited in the lower 500s (nm).
 

lyyyghtmaster

Newly Enlightened
Joined
May 24, 2006
Messages
148
Location
Tucson, AZ
Agreed, the black body radiation will tend to be white-balanced by the eyes, where the monochromatic colors will saturate the receptors. And it's amazing how very fast that happens; within seconds if you A/B your eyes on a colored LED.

Looking at the eye-response curves again, it is now clear why apparent color shifts so rapidly in the aqua range! Both curves (blue and green) have very little overlap there, so slight changes in wavelength would seemingly have noticeable effects, and that's exactly what I observe. I wouldn't be surprised if I could perceive a 1nm difference around 490nm, even though it's more like 5~10nm in the deep blue!

Now I would like to remind everyone of this: I red awhile back, and have seen posts discussing it here, that some percentage of women (40% I think) are more sensitive than everyone else to differences in wavelengths in the deep red area of the spectrum due to some recessive(?) cone variation. How deep and how much more sensitive were never established though. Not surprising considering the news item came from a popular press article! I would leap for joy if these specifics could be pinned down! As a man, this article bugs me. I want my deep red vision to be "all that" too!! :sigh:

You know what? When I compare 630 and 660, adjusted to equal perceived intensity, the 630 actually seems drab and brownish by comparison to the vibrant, rich, melodious 660! Yes it does look oranger on the spectrum, but the aesthetic effect to me is one of comparative brownness. Weird, huh?

I agree though the normal 625~630nm red LEDs do appear orangeish when my eyes wear my analytical hat. I greatly prefer the 660nm flavor when I can throw enough power at it to do it justice! One great LED I like to play for this musical effect is LZ1-10R205 by LedEngin. 5W, 20mm star. Similar beam angle to XR-E. They do make several other variations as well (bare emitter, 10W version, etc). I get mine from Newark.

Lately there have been appearing on eBay at least two sellers with 1W 660nm 20mm stars! These seem to be the same Vf and brightness as the LedEngin product, but are only rated to 350mA vs 1500 for the LedEngin. They do have a smaller die as well, not surprisingly. The build quality also isn't as good; there is some goopy silicone-like gunk all around the dome (similar to LuxI and III but a lot more, even on the dome itself!) Nonetheless they seem entirely adequate for any less-demanding applications. I ran one at 500mA for 4.5 hrs (somewhat poorly coupled to a 4" x 1" x 1" finned heatsink) and could detect no light output degradation at all, using my light meter! Another I accidentally cracked the dome of, using a malformed optic. LED still works fine, except for the screwy beam pattern and slight loss of output.

I have used Lumex's SSL-LX5093SRC/E for many years now to build true red LED Christmas light strings from! Works great, if you don't mind the narrow beam angles (30deg). (true they can be sanded to diffuse effect, but that's another time-consuming step in an already very lengthy process). The 50deg offering from Ledsupply.com seem to be effectively a wider-beam version of the Lumex product, and even the leadframes look surprisingly similar. These work even better for light strings; however they get a bit more pricey if you need a lot of them. The ~SRC/F variation of the Lumex was supposed to be a higher-output version (3500mcd IIRC) compared to 2800 for the /E. I almost bought some, but they were pricer and a larger number had to be ordered, and the manuf. never responded to my queries as to whether the higher mcd was just a special bin of the same package/die that happened to be throwier but with the same mW output. I think even the /E is rated higher than 2800mcd now IIRC(?)

For some reason many (but by no means all) center-high-mount stop lights on cars seem to use LEDs in the 650~660nm range, even though the main stop/tail lights (often filtered incandescent still) appear to have a dominant wavelength around 620~625nm! (Which seems to be the NHTSA standard, BTW). That's very weird to me, considering you'd think manuf. would want to optimize power efficiency/heat generation as well as consistency of appearance!!! Many a CHMSL with failed segments! Why don't they use 625s??????

Traffic-control stoplights appear to use LEDs in the 640nm area, which seems to be a reasonable compromise between oranginess and dimmer-but-purer 660 red if you're not aiming for artistic perfection.

BTW I'm not a synesthete (AFAIK!!!), I just really enjoy a good, deep wavelength! Same goes for violet, though of course I'm more careful with higher intensities in this range. :shakehead 'Course, at both ends of the spectrum, lower intensities read deeper, so that works to my advantage eye-health-wise.
 
Last edited:

lyyyghtmaster

Newly Enlightened
Joined
May 24, 2006
Messages
148
Location
Tucson, AZ
Just thought of something: Try the phosphor-converted Rebel Amber!!! Those are noticeably less saturated than direct-color Amber, yet still appear Amber to the eye. But there's a fair amount of red and green in the spectrum, and even they actually do start to approach "white" compared to the direct-color Amber LEDs.
 

PPJ

Newly Enlightened
Joined
Feb 2, 2010
Messages
5
Testing to post some pictures with the SSL-LX5093SRC/E in work.

Here in a array with 4 leds drived at maximum rating @ 30 mA.


dscf8719z.jpg


content_round.php


dscf8719w.jpg
 
Last edited:

PPJ

Newly Enlightened
Joined
Feb 2, 2010
Messages
5
Here against a shiny surface. The pinkish note it emittes is seen here as I described earlier.

dscf8721.jpg
 

PPJ

Newly Enlightened
Joined
Feb 2, 2010
Messages
5
And here against a mat surface. The SSL-LX5093SRC/E produces a pure deep red led light.

dscf8725.jpg
 
Last edited:

axd

Newly Enlightened
Joined
Apr 12, 2009
Messages
93
Location
gent, belgium
I just received a ZL 502pr, but I see no significant difference with 51r and 501r :-(

should the difference be easily seen? or is this only a subtle difference?

Here is a comparison between a 502pr and a 51r:

https://github.com/axd1967/502pr

Try to guess which one is which...
 
Last edited:

JoakimFlorence

Newly Enlightened
Joined
Jun 4, 2016
Messages
137
If you want saturated red color, you need to put a 650nm behind a red filter. The reality is that LEDs are not entirely a single wavelength and you need to filter out the traces of orange light because that results in a substantial perceived color shift in the hue of light coming from the emitter.
95% pure red light and 5% green light still appears orange.
In certain specific color ranges, the corresponding color of the peak wavelength of the LED may vary substantially from the actual color appearance of the LED's light. Red light needs to be very pure to remain red.

Or just get 660nm, but be prepared for a sacrifice in brightness. A 660nm emitter appears reasonably red (maybe ruby red, or almost).

(note that none of this applies to CRI, 635nm spectral reflectance can appear fairly well saturated against a neutral white backdrop)
 
Last edited:
Top