PDA

View Full Version : OT: LED lighting, part 37...



Evan
12-03-2012, 04:34 PM
I installed some new LED lighting in my living room. I need to control the colour of light to which I am exposed in an attempt to try to regulate my diurnal sleep cycle. Bright bluish light in the am and sunset colours with no blue at all in the later pm. I have set up a colour scheme on my computer to help with that but I also needed variable colour lighting for the living room, which is where I live and sleep the vast majority of my time. I seem to suffer from what is called "non-24" which is a diurnal sleep cycle that is not 24 hours but usually longer. I cycle around the clock about every 2-3 weeks, sleeping during the day and awake at night about half that time.

I installed a variable colour LED light strip from DX that has about 30 watts worth of RGB LED emitters, 300 of them over 16 feet. It runs on 12 volts and draws about 2.5 amps. The remote that comes with it is OK but I bought a compatible remote with control module for $7.00. It is the same type remote but has full independent dimming of every colour from full bright to off and also has memory for the dimming setting. The entire setup cost less than $40, not including the 20' strip of aluminum angle. That I got at the local job shop by trading an LED bulb to put in the light fixture of their brand new Haas CNC lathe. I am looking forward to playing with it.

A few pictures: The strip light can be set to any possible colour and brightness using the dimmer controls. That includes any colour temperature you prefer as "white" light. It does not come with a power supply. The shown night colour is "pure" white. I normally set it to sunset yellow.

BTW, I am going into town so I won't be answering questions for a few hours.

http://ixian.ca/pics10/ledlr1.jpg

http://ixian.ca/pics10/ledlr2.jpg

http://ixian.ca/pics10/ledlr3.jpg

http://ixian.ca/pics10/ledlr4.jpg

outlawspeeder
12-03-2012, 04:42 PM
I run the same on my Harley. Call them chicken lights. They hold up well. I was thinking the same thing but not sure how to do it. I like your setup and may copy it.

Thanks for sharing

beanbag
12-03-2012, 08:09 PM
Very interesting project!
Recently I have been playing around with fluorescent tubes in my living room to try to get some more diffuse lighting.

How is the Color Rendition Index? (which is a measure of how close to blackbody a light source is) With only 3 colors I can't image that it can be that great. I have read that if the CRI is low, things look unnatural due to gaps in the color spectrum. This is the main reason fluorescent lighting doesn't look as good as incandescent or halogen lighting.

Evan
12-03-2012, 08:30 PM
These are RGB LEDs. The colour rendition index is whatever you set it to be. RGB is sufficient to create the entire spectrum closely. There aren't "gaps in the spectrum". The primary additive colours combine to create all the other colours in the spectrum. To start with, green plus red creates yellow, blue plus green creates cyan and red plus blue creates magenta. Vary the proportions and they create any other colour. Keep all three the same luminance and it creates white which is all of the colours. The spectrum of visible light is nearly exactly a full octave in frequency. That makes it possible to create any frequency between the bottom and top of that octave by interference and reinforcement depending on the mix of frequencies. It requires a frequency close to the bottom, the middle and the top which is what RGB is.

A.K. Boomer
12-03-2012, 08:37 PM
Is there a black light setting?
nothing like roasting up a good bowl of red bud and breaking out the black sabbath paranoid album...
neighbors usually cant here cuz i'll be in the basement where I now live...
Cops are no issue cuz it just got legal in my state...


that is allot of light for the wattage for sure... im going to be looking into them.

Evan
12-03-2012, 08:43 PM
UV LEDs are available but I don't recall seeing them in a light strip. Could well be that I missed them since I wasn't looking for them. The RGB strip I bought is currently by far the best deal at DX. The prices vary all over the place depending on how recent the product is. DX doesn't change the prices when product goes out of date. It's up to you to find the best bang for the buck. That requires knowing something about the actual LEDs used to make the product. The newer they are the more lumens you get per dollar. It pays to look up the data sheets and figure out the lumens per dollar. That is how I determine how good a deal LED lighting is.

A.K. Boomer
12-03-2012, 09:24 PM
what about the name "cree" --- are they the best to get?
my bro only buys flashlights with the cree led's in them and says their the best...


Edit; just thinking about all the stuff Evan builds and have come to the conclusion that he must have one very nice and patient Wife,
I can just see him handing over weekly updated diagrams on how to operate the house in which they share lol

Honey, how come there's a micro-switch on your snooze button? --- ughh - don't press that sweetie - it operates the toaster,,, lol

beanbag
12-03-2012, 09:26 PM
I looked into the CRI issue a little more, and I think my suspicions were correct. RGB clusters can produce any (most?) visible light colors, as appears to the human eye, but because the lights lack certain frequencies, the color rendition of objects will look funny.

For example, green and red led light can appear to look yellow to the eye, but if there is a real object that only reflects yellow wavelength light, it will look funny under rgb illumination because there is no actual yellow light frequency.

Here is an article I found written by some artsy French guys talking about how RGB clusters are horrible at making certain colors look good:

http://www.crcc.cnrs.fr/IMG/pdf/DiodesElectroLuminescentes_AIC2005_Vienot-Mahler-Ezrati_en.pdf

Adding amber to the RGB cluster makes it look less bad.

However, I am also interested in how you subjectively think things look under this illumination.

Oh, and I think that's the reason white led's (blue + phosphor) are used for lighting and not rgb clusters. Because the phosphors have a broad spectrum.

beanbag
12-03-2012, 09:28 PM
what about the name "cree" --- are they the best to get?
my bro only buys flashlights with the cree led's in them and says their the best...

yes

Disclaimer: cree shareholder

Willy
12-03-2012, 09:32 PM
Very timely thread as I've ordered 4 5m rolls of LED light strips for some testing and evaluation purposes around the yard, shop, and house. It's been a couple of weeks so they shouldn't be long.
Looks as though I won't be disappointed.
Thanks for a sneak preview of what to expect.

AK, prices are all over the map for these strips but I have seen ultraviolet LED strips go for about $55 for a 5M strip and approx. $15 for a 1M strip.
Definitely more than the single color or RGB strips but they're out there at any rate.

Evan
12-03-2012, 10:47 PM
For example, green and red led light can appear to look yellow to the eye, but if there is a real object that only reflects yellow wavelength light, it will look funny under rgb illumination because there is no actual yellow light frequency.

Not true. Photons have wave properties. When electromagnetic waves of differing frequencies are mixed they produce a series of harmonics both above and below the fundamental frequencies of the original waves. Photons exhibit this property called constructive interference. This can be precisely simulated with audio wave forms, which I have done to illustrate the principle.

Red is accepted to be a wavelength of 650 nm. Green is 510 nm. I created two sine waves of those audio frequencies, 650 hz and 510 hz. Then I mixed them and did a log spectral analysis to see where the peak spectral frequency lies. It is at precisely 590 hz. Not at all coincidentally, 590 nanometres is yellow light. When red and green light are mixed the product is a range of colours with the peak power exactly at yellow. Other colours are also present but the light is seen by the eye as yellow since that is where the greatest spectral power lies. There is no "gap" when RGB colours are mixed. There is however a great deal of misunderstanding by a lot of people online that write on this subject.

http://ixian.ca/pics10/freq1.gif

beanbag
12-03-2012, 11:28 PM
hmmm...
The principles of constructive interference and etc have to do with phase, and don't change the frequency of light at all. The same way lasers bounce around and interfere all over the place and keep the same color.

In think in normal life, light only mixes linearly, i.e. the waves are additive so there is no frequency change either. No sum of sine(a*t) and sine(b*t) will give you any fourier components at sine(c*t).

The mixing you are referring to must be some kind of non-linear mixing, where you start getting sine(A*t) * sine(B*t), and then you can start getting new frequencies. Or some kind of non-additive distortion of the waveforms. Also, you did windowing on your fourier transform, so that will also throw frequencies all over the place. Try the same analysis with two signals lasting a much longer time (more periods) and use a more "gentle" window.

In short, I don't think reg and green leds can make pure yellow light, even if it looks yellow to the eye.

Evan
12-03-2012, 11:36 PM
Increasing the resolution on the window just picks out more little peaks that correspond to the various higher harmonics of the waveform. Same thing happens at light frequencies. They aren't "imaginary" resultants. Mixing frequencies results in new frequencies that didn't exist in either original wave form. It does it with sound and it does it with electromagnetic waves like long wave radio, microwaves and with light.


The same way lasers and bounce around and interfere all over the place and keep the same color.
Of course they keep the same colour if it only interferes with itself. Try it with more than one colour.

beanbag
12-03-2012, 11:59 PM
wave mechanics aside, I am mostly interested in your opinion on how things look under this lighting. Whether it is similar to natural/incandescent lighting, or whether things, especially skin tones, look "kind of funny".

jnissen
12-04-2012, 12:06 AM
I used the same type of RGB lights for an outdoor kitchen. Installed them under the counter top ledge to provide some useful accent lighting. Wonderfully adjustable.

tmc_31
12-04-2012, 01:30 AM
Don't want to confuse the issue guys, but I think you are confusing Color Temperature with Color Rendering Index.

Color temperature is the color of a light source as it relates to a blackbody.

Color Rendering Index (CRI) is an expression of a light source's ability to render the colors of objects near the light source compared to what an ideal light source (typically bright sunlight in a blue sky or an incandescent light source) would do.

Into the fray with both eyes closed!!

Tim

Evan
12-04-2012, 02:18 AM
I am very happy with the subjective light quality. I may not be a good reference to go by and neither is my wife. We both happen to like blue white daylight white which is a colour temperature 7000 to 10000K. As for colour rendering index, I don't pay much attention to that except on my monitors. What you see around you is extremely subjective and depends on a huge number of factors. Your age, your genetics, the light sources of course, the types of materials that surround you and others. There are many ways for the effect of colour to be produced. My only standard is 6000K sunlight white. Sunlight white and daylight white are very different. Daylight white includes the large amount of very blue sky light as well as sunlight.

Colour as you see it can be the result of subtractive reflection, subtractive transmission, dichroism; transmission and reflection, interference effects such as oil on water, emission of several types and a few others. What colours you perceive also depend on what colours you last looked at and that changes as you continue to look at something. You brain adjusts constantly for these effects so you really don't know what colour you are looking at. The only way to make an accurate judgement of any colour is to use a properly calibrated colorimeter.

I forgot to mention that the cones in the retina do not have a sharp colour sensitivity band. In fact, each type of cone is sensitive to the entire spectrum to some degree. The curve of spectral sensitivity is very broad with no particular peak, just a taller hump at the centre of the spectral sensitivity for each of the three types of cone. There is a very large overlap and it is the brain that sorts it out.

batt-man
12-04-2012, 02:48 AM
Thanks Evan for sharing this - it's very interesting to see this working as it just so hapens that i ordered two of those led strips last week.

I plan to use them for under-cupboard illumination as part of my kitchen refurb / xmas project (as the wife calls it!!).

Question : how did you attach the led strip to the aluminium angle? double-sided tape?


Cheers
Batt....

Evan
12-04-2012, 04:48 AM
They come with already applied 3M peel and stick tape. Make sure the surface that you stick it to is clean of any trace of grease or oil. These high density LED tapes need a heat sink so it needs to be applied to some sort of aluminum strip or trim. The aluminum strip should be at least twice the width of the tape although it doesn't need a flange. It really does make a difference. Mine ran hot without the aluminum angle. With it it runs just barely warm. That will make the LEDs last much longer and run brighter. It will also prevent the sticky tape from loosening.

macona
12-04-2012, 05:58 AM
UV LEDs are available but I don't recall seeing them in a light strip. Could well be that I missed them since I wasn't looking for them.

They do make them. I looked them up for someone some time back. The only thing is the LEDs are not what I would really call UV. They are in the 400nm, really good ones are in the 395nm range. A fluorescent blacklight is about 365nm. There are LEDs at 365 but they are pretty expensive.

Barrington
12-04-2012, 07:19 AM
Beanbag is absolutely correct.

Additional frequencies are NOT generated in linear systems by the simple addition of frequencies !!!

Here is a somewhat more realistic spectrum analysis of a signal created by adding 510Hz and 650Hz at equal amplitudes.

http://i564.photobucket.com/albums/ss82/MrBarrington/mix.png


Similarly mixing two wavelengths of light does NOT produce any light at an intermediate wavelength !!

Cheers

.

Evan
12-04-2012, 10:49 AM
Would you care to explain where the harmonic frequencies come from when you mix electromagnetic waves?

ed_h
12-04-2012, 11:01 AM
I'm pretty familiar with radio waves, and the only time we see new frequencies occur when mixing is if there is a non linearity in the system.

beanbag
12-04-2012, 11:25 AM
Try reflecting the light from this RGB strip off a CD, and let me know if you see any yellow :)

Evan
12-04-2012, 11:27 AM
Light is inherently non-linear since it also has particle nature. Regardless, when you mix two frequencies in a linear system you will get the sum and difference as the first order harmonics along with the fundamentals.

If I mix red and green what will I see with a yellow filter? Yellow light of course. The luminance of that light will not be equal to the sum of the luminance of each colour source because the some of the luminance will be distributed along a range of values other than yellow. That is determined by probability of any particular frequency being produced. The graph I produce shows what is equivalent to the probability distribution of the colours produce by mixing those particular frequencies. Note that the graph produced by Barrington is linear and not log. It shows a much smaller range and is not comparable to the one I made.

ed_h
12-04-2012, 11:46 AM
Regardless, when you mix two frequencies in a linear system you will get the sum and difference as the first order harmonics along with the fundamentals.


Evan--

I'm having trouble envisioning the math that would support this.

If we apply a two-frequency mix to a non-linear transfer function (often approximated by the first term or two of a Taylor series expansion of the function), the harmonics and the sum and difference frequencies readily appear. If the transfer function is linear, no new frequencies are produced.

danlb
12-04-2012, 11:56 AM
It looks like we are talking of several different things here. Sure, you can blend red and green light to end up with something that our eyes perceive as yellow (when shined on a white surface). That's because of the receptors in our eyes.

But the next step is that certain objects are really yellow, and will only reflect certain wavelengths of light. If you don't have a yellow wavelength in the the light source, it does not reflect much back at you. This is really apparent when using a white LED light to look at green leaves. The leaves look almost black. The white LED shines a blue light through a phosphor that emits a yellow light. The blue and yellow combine to look white, since the yellow satisfies the eyes receptors for red and green too.

The high CRI lights are trying to match a standard set long ago. You can look it up, but standard is basically the output of a tungsten filament bulb as seen by a Scandinavian ( i think).

The problem with RGB LEDs is that they create light with peaks at three frequencies and valleys between them as they overlap, so the colors between red, green and blue are not as vivid.

BTW, Nichia, Luxeon (Phillips), Cree, Seoul and Osram are all big names in the LED game. Each has made some good stuff. I'm partial to the Luxeons myself. :)

Dan

Barrington
12-04-2012, 12:04 PM
Regardless, when you mix two frequencies in a linear system you will get the sum and difference as the first order harmonics along with the fundamentals.This is not correct.

A perfect multiplicative mixer produces sum and difference frequencies, harmonic terms arise from non-linearities.

If we could mix and create new light wavelengths as suggested - could we add some IR to some UV and get visible light ? (There are non-linear optical effects in some materials at extremely high intensities, but nothing relevant to domestic lighting!)

Imagine the havoc if every RF transmission mixed with every other one to produce yet another at some other frequency...

Confusion can arise from the appearance of the time domain waveform. Beating (adding) two frequencies produces an amplitude envelope at the difference frequency, BUT this does not exist in the frequency domain.


Note that the graph produced by Barrington is linear and not log. ???? Both axes are quite obviously logarithmic ????

Cheers

.

Evan
12-04-2012, 12:08 PM
This is really apparent when using a white LED light to look at green leaves.

You must be using some very odd white LEDs. The white LEDs I use make green plants in my living room look just as green as with sunlight. I only have LED lighting in our house. There are only a few 4' tube fluorescents, the rest are LED.

EVguru
12-04-2012, 12:11 PM
http://www.jands.com.au/support/product_support/lighting_technical_materials/what_to_look_for_when_judging_an_led_fixture146s_c olour_mixing_capabilities133

danlb
12-04-2012, 12:25 PM
You must be using some very odd white LEDs. The white LEDs I use make green plants in my living room look just as green as with sunlight. I only have LED lighting in our house. There are only a few 4' tube fluorescents, the rest are LED.

Nope, not odd at all. There are, of course, many shades of plants. Some render well while others don't. It all depends on your LED at the moment and the plant in question.

Then there is the question of what you perceive to be the correct color. Sitting at my desk with a tungsten light, I see a white piece of paper. If I turn that off and turn on an LED lamp the paper looks blueish. Ten minutes later the paper looks like a crisp white. Why? Because my brain tells me that's a white paper and I do the internal white balance thing. When I turn off the LED and turn the tungsten lamp back on the paper looks yellow for a whiile. Are you really seeing all the hues of green in your plants, or are you just used to them that way. :)

The link from EVguru is right on target. Thanks.

Dan

Evan
12-04-2012, 03:46 PM
Actually the link is very misleading. They use as an example the sRGB colour gamut which has nothing to do with RGB LEDs. That is a very restricted gamut originally designed to accommodate the colour space of colour monitor phosphors. RGB LEDs can produce a wider gamut than any other light source that uses only RGB.

From the link: The sRGB colour gamut

http://ixian.ca/pics10/gamutsrgb.jpg

This is the true LED RGB gamut.

http://ixian.ca/pics10/gamutledrgb.jpg

I am very acquainted with colour reproduction. I was trained on it by Xerox.

Incidentally, the NTSC (Never Twice the Same Colour) gamut is not the gamut of the monitors. It is the colour gamut of the TV signal.

Barrington
12-04-2012, 04:23 PM
Actually the link is very misleading. They use as an example the sRGB colour gamut which has nothing to do with RGB LEDs. That is a very restricted gamut originally designed to accommodate the colour space of colour monitor phosphors. RGB LEDs can produce a wider gamut than any other light source that uses only RGB. Hardly fair comment. sRGB appears in the flow of explanation quite legitimately.

From the article, in the section "Colour Spaces in the Lighting World" which follows the discussion of sRGB:-
"We can see the colour space in a standard 3-colour LED fixture on the left is already larger than the sRGB colour space we saw earlier, but still has large areas missing."


Returning to the previous matter. Are you maintaining your position that light of two wavelengths of light added together results in a third (or more) wavelength(s) being generated ?

Cheers

.

danlb
12-04-2012, 04:23 PM
You are still looking at it from the standpoint of generating light as opposed to reflecting light. A CRT or LCD can do a good job of mixing light emitters so that they shine in millions of shades. Try displaying a white screen on your monitor and use it to illuminate things around the room. It's quite disappointing.

A half ripe apple can't reflect the myriad of hues of greens, yellows and reds if those wavelengths are not present in the light source. The same goes for a well cooked steak. You can't see the brown reflected as brown unless those wavelengths are in the light. While you can mix RGB to get various hues, they will not be in proportion to the full spectrum we see in natural light, or even in incandescent lights.

Don Mcleash did a nice writeup on this web page. http://www.candlepowerforums.com/vb/showthread.php?199054-High-CRI-and-its-significance
He actually has a spectrometer and integrating sphere and shows some nice graphs.

Dan

Evan
12-04-2012, 04:36 PM
Most things around us create the appearance of colour by subtractive reflection. That works just fine with RGB illumination so the wide colour gamut will illuminate what we see around us in accordance with the colours we think we should see. It works for me.

For example, a yellow coloured piece of paper reflects all but blue, which it absorbs. Same with most things coloured yellow in your living room. Under RGB light they will all look yellow and if the colour balance is correct they will look the correct hue of yellow too. The correct balance is determined by the colour temperature of white produced, which is adjustable with this illumination. The RGB illumination can reproduce all of the colours within the gamut in the diagram above and it can do it all at the same time.

danlb
12-04-2012, 05:56 PM
I don't see quite how you get all colors in the gamut in equal amounts when you have 3 LEDs that each have a very narrow (25 nm or so) band for the majority of it's luminosity.

I like the graph at http://www.jands.com.au/__data/assets/image/0014/27131/luxeon_vs_cones_RGB_LED.png. It clearly shows that each LED has a very narrow emission band.

If the RGB LEDs sold by DX had a high CRI, that would probably have been noted somewhere. :)

Dan

beanbag
12-04-2012, 09:30 PM
I am very happy with the subjective light quality. I may not be a good reference to go by and neither is my wife. We both happen to like blue white daylight white which is a colour temperature 7000 to 10000K. As for colour rendering index, I don't pay much attention to that except on my monitors. What you see around you is extremely subjective and depends on a huge number of factors. Your age, your genetics, the light sources of course, the types of materials that surround you and others. There are many ways for the effect of colour to be produced. My only standard is 6000K sunlight white. Sunlight white and daylight white are very different. Daylight white includes the large amount of very blue sky light as well as sunlight.


yes, now that you mention it, the sky IS blue, although things around me outside don't look bluish or harsh, as opposed to "daylight" indoor lighting. In my living room, I think 3000K works out well, and 3500K is too "harsh". But then again, my living room has brown wood walls. In my kitchen with white walls, 3500K looks a lot better. I think a general rule of thumb is the brighter the light, the higher colour temperature you prefer.

dp
12-05-2012, 12:38 AM
I think the perception of color using RGB light sources is entirely in the brain. The frequencies don't mix outside our heads - they are as discrete and isolated as radio frequencies. They don't mix when they bump into each other in free space otherwise we'd see ghost images where there is no light source. In the case of light our eye receptors are sensitive in different degrees to frequencies of light. Our brain mixes the output from the cones to produce the colors based on light frequency and intensity. Our brain is quite clever at back filling missing parts of the spectrum as well as making sense of three-color light. The brain knows what the color of tree leaves should be. By example - the movie Oh Brother Where Art Thou had much of the outdoor scene backgrounds translated from spring/summer green to the ambers of fall. It looked very unnatural because we know leaves with characteristics of those should also be green. Same with the grass. Healthy blades of grass in a field are not parched brown. It all looked fake to me when I saw the movie and it later was shown to be just that.

Sound is a similar thing except our ears are mechanical and non-linear. Sounds will mix in our ears mechanically and we will hear beat frequencies. The non-linear characteristic of our ears gives us under and overtones that are products of mixing frequencies, as well, and our brains learn what the nuances of such sounds mean. That is part of how we know the direction a sound comes from - particularly in the vertical plane.

beanbag
12-08-2012, 05:11 AM
I looked into the concept of LED lighting some more, and it turns out that many companies make LED emitters with 90+ CRI, such as Cree, Nichia, Phillips, etc. The spectrum is fairly broad, except for a lump and dip right near blue. These are available as individual emitters for things like flashlights. There are LED strips too, although I have only seen 80+ CRI. I learned a lot reading around on candlepowerforums.com

Ian B
12-08-2012, 12:38 PM
Evan,

There seem to be 2 types of RGB LED's commonly available on strip mounts - 3528 and 5050, both in 60 per meter, waterproof variants. Which did you pick, and any reason for the choice?

Good discussion!

Thanks,

Ian

Evan
12-08-2012, 04:14 PM
I think the perception of color using RGB light sources is entirely in the brain. The frequencies don't mix outside our heads - they are as discrete and isolated as radio frequencies. They don't mix when they bump into each other in free space otherwise we'd see ghost images where there is no light source.

When looking at this from a quantum viewpoint it changes a lot. Then probability comes into play. The reflection from non specular surfaces is subject to the probability functions and that is where just red and green can produce all the colours which contain no blue. That is what the colour gamut I posted above is based on and it isn't an illusion of the brain. If an object is a particular colour that is produced by mixing red and green it has a certain probability of reflecting those colours in the precise proportions that create the colour of that object. There is no spectral "gap" that fails to reproduce the right colour. It is very easy to verify. Just illuminate an area with red, green and blue LEDs and hold different objects under that light. The white produced by mixing red, green and blue illuminates objects in the same way as sunlight given that the colour balance is the same. No colours are "missing".


There seem to be 2 types of RGB LED's commonly available on strip mounts - 3528 and 5050, both in 60 per meter, waterproof variants. Which did you pick, and any reason for the choice?

The RGB strip is made with 5050 LEDs as they are the current best available in terms of lumens per watt and $. I also have another white LED strip that I will post about soon made with the newest available white LEDs for this application. It uses 5630 white LEDs and they put out about 120 lumens per watt. That is about double the efficiency of fluorescent lamps. They emit up to 50 lumens each and the strip has 300 of them. In theory that means if run at max the strip can emit 15,000 lumens. That will require excellent heat sinking. I am running them at about half power which means the strip will emit about the same as a two tube fluorescent lamp. The strip is $40 but the lumens per $ are excellent. I will be playing with them to see how they do. DX only has one light strip using those LEDs and it has a very high colour temperature which won't suit many people.

The 3528 LEDs are not a good deal. They are a previous generation LED and generate a lot less light for the $.

Ian B
12-08-2012, 04:36 PM
Evan,

Would these do it?

http://www.ebay.co.uk/itm/5M-300LEDs-SMD-5050-RGB-Light-Strip-IP65-Waterproof-24Key-IR-Remote-Controller-/261129235273?pt=UK_Home_Garden_Night_Lights_Fairy_ Lights&hash=item3ccc83eb49&autorefresh=true

About $20 for 5m of them, waterproof, free postage. Thanks for the info on the difference. I saw somewhere that the numbers refer to the dimensions - a 3528 is 3.5mm x 2.8mm, and - unsurprisingly - a 5050 is 5mm square.

Ian

Evan
12-08-2012, 04:50 PM
That seems to be very similar to the one I ordered. That looks like a DX reseller to me or maybe it is DX themselves selling out of the UK warehouse they recently created. It mentions a remote which strongly implies it includes the actual power controller although it is not mentioned. It does not include the 12 volt supply. You will need a supply with a well regulated 12 volts and able to supply at least 3 amps without overheating. I am using an 18 amp hour gel cell with a trickle charger. That way it is independent of the mains power for a few days.

beanbag
12-08-2012, 05:48 PM
There is no spectral "gap" that fails to reproduce the right colour. It is very easy to verify. Just illuminate an area with red, green and blue LEDs and hold different objects under that light. The white produced by mixing red, green and blue illuminates objects in the same way as sunlight given that the colour balance is the same. No colours are "missing".

This test is precisely what they do in that pdf I linked with the french art guys. The RGB light strip received the worst rating of all the "white" light sources.

danlb
12-08-2012, 07:33 PM
When looking at this from a quantum viewpoint it changes a lot. Then probability comes into play. The reflection from non specular surfaces is subject to the probability functions and that is where just red and green can produce all the colours which contain no blue. That is what the colour gamut I posted above is based on and it isn't an illusion of the brain. If an object is a particular colour that is produced by mixing red and green it has a certain probability of reflecting those colours in the precise proportions that create the colour of that object. There is no spectral "gap" that fails to reproduce the right colour. It is very easy to verify. Just illuminate an area with red, green and blue LEDs and hold different objects under that light.



That's a good SUBJECTIVE test, but by it's nature we will think that it looks better than it really is. Without a simultaneous reference most people can not tell the if a shade of green is correct or not. We see that it is green and that is enough. Cameras, TVs and magazines would be disappointing if we did not subconsciously ignore the subtle changes in hue with different lights.

The Macbeth color checker is used for checking colors in photography. The idea is to ensure that the colors you see when printed match the originals. It's only a sample of colors with no built in way to validate it. See http://www.ae5d.com/macbeth.html for an example.

I used to have a chart that was to be used to validate my printer colors. It was similar to the IT8.7 chart used to check the accuracy of scanners. ( http://upload.wikimedia.org/wikipedia/commons/1/11/IT8target.jpg ) The chart was also good for evaluating light sources. It had a graduated set of color blocks that were progressed from one hue to the next. A poor light source would make several adjoining areas look about the same. A full spectrum light source would show each color as unique. That

Dan

Barrington
12-08-2012, 08:14 PM
That is what the colour gamut I posted above is based on and it isn't an illusion of the brain.No - the colour gamut you posted is ALL about human perception, and nothing to do with basic physics.

'Colours' not on the spectral line (i.e. involving more than one wavelength) do not exist outside ourselves.

Two 'colours' of light which appear identical to the human (or to be exact, the average human) eye may be made up of completely different sets of wavelengths and amplitudes which happen to result in the same tristimulus values.

A species with different visual system would not recognise them as the same colour.

Cheers

.

Weekend_Scientist
12-08-2012, 08:14 PM
Hi Evan,

Just wanted to chime in and add my appreciation for you detailing your latest LED lighting endeavor. I'm no stranger to LEDs but I had been procrastinating using them as an actual usable light source. This thread kicked me over the edge and I've got a few strings on order.

Mark

Mcgyver
12-08-2012, 08:39 PM
Those pictures make me want to buy a small car and install a bunch of those strips under it :D

Weekend_Scientist
12-08-2012, 09:05 PM
McGyver,

Any "small car" with a bunch of LEDs under it has got to have a muffler/exhaust the size of the Hubble Space Telescope. Better factor in the cost of the muffler when getting the car/LEDs.

Willy
12-08-2012, 09:10 PM
Those pictures make me want to buy a small car and install a bunch of those strips under it :D

And they just happen to be 12v.
It's as if it were preordained.

ed_h
12-09-2012, 12:34 AM
When looking at this from a quantum viewpoint it changes a lot. Then probability comes into play.

Evan--

I don't think it is necessary (or even useful) to invoke quantum probabilistic effects here. A color source can have any mixture of spectral wavelengths, and this color space can be mathematically modeled as one of many (even infinite) dimensions. Human color perception is often modeled as a space of three dimensions: the responses from the three types of color receptors in the retina. This alone suggests that human perception of a color cannot be unique to a single spectral or mix of spectral colors.

So while two sources might look identical to the human eye/brain, the spectral mix could be quite different, and this would show up when the sources illuminate objects or pass through filters.

Evan
12-09-2012, 12:51 AM
QED describes how light is reflected from surfaces according to probability. The same applies to dichroism and polarisation. As for the varying perception of the human brain, I noted all that earlier and if my brain can make it all look right then there is nothing to complain about.

I would take some pictures under the light strip to illustrate but it is impossible to reproduce what I see or even what actually exists since it would of necessity be greatly altered by the process. The only way to judge is to try it yourself and see what you think. I find it a very satisfying light source that has capabilities that were entirely unavailable just a few years ago for an incredibly small amount of money. It suits my needs almost perfectly. The only thing that will make it better is to be able to program it from my computer to follow a preset lighting pattern synchronized to the season and time of day.

It would be interesting to have it detect the outdoor light chromacity and use that to control the indoor lighting. A spectacular sunrise or sunset would then be replicated inside.

darryl
12-09-2012, 01:16 AM
Hey- just what I need to help me get up in the morning- a simulated sunrise in the bedroom.

I have a really crude form of that right now- some halogen bulbs that come on with a timer. INSTANT sunrise :)

ed_h
12-09-2012, 11:02 AM
. The only way to judge is to try it yourself and see what you think. I find it a very satisfying light source

I don't think anyone was suggesting that you shouldn't be happy with the lights if they suit your needs. It was just some of the physics points that had some of us rattled.


I think this is a great find. I'd like to play with it myself

.

Peter.
12-09-2012, 03:32 PM
Evan,

Would these do it?

http://www.ebay.co.uk/itm/5M-300LEDs-SMD-5050-RGB-Light-Strip-IP65-Waterproof-24Key-IR-Remote-Controller-/261129235273?pt=UK_Home_Garden_Night_Lights_Fairy_ Lights&hash=item3ccc83eb49&autorefresh=true

About $20 for 5m of them, waterproof, free postage. Thanks for the info on the difference. I saw somewhere that the numbers refer to the dimensions - a 3528 is 3.5mm x 2.8mm, and - unsurprisingly - a 5050 is 5mm square.

Ian

How cheap?

I just ordered a strip of them and put up a link to a place where people often ask about lighting their workshop with no mains power available.

Ian B
12-10-2012, 02:37 AM
Evan,

You mention computer controlling the lights, to give variations throughout the day. Would this be a nice little job for an Arduino, running each colour off a PWM output pin? It might take a power transistor on each pin to deliver the current (especially if you end up daisychaining more of the strips), but that shouldn't be too difficult.

For even brighter lighting, it would be possible to buy 3 strips, one red, one green and one blue, each of 60 5050's per metre, and control them as if they were combined RGB LED's. Maybe also add a strip of yellows?

At full brightness, how much current does each strip take per colour?

Ian

TRX
12-10-2012, 09:06 PM
yes, now that you mention it, the sky IS blue

Nitrogen is almost colorless, oxygen is pale blue. Easiest to see in their liquid states.

Evan
12-10-2012, 09:25 PM
The sky is blue due to Rayleigh scattering, not the inherent colour of the element.

Evan
12-10-2012, 09:39 PM
Using separate RGB light strips is not a good idea. Because of the separation shadows will show severe colour fringing. The LEDs used in the strip I have have very small separation between emitters and act like a single light source unless viewed very closely.

Each colour draws about 1 amp maximum at 12 volts. However, that is one at a time using the included controller. If two are on at full power it draws about 2 amps as expected. If all three are turned on it draws a little more than 2 amps, about 2.3 at most. Since the controller is rated for 2 amps per colour I assume it is intentionally limited to prevent overheating. Without a heat sink strip does become very hot at full power. With the aluminum angle heat sink I used it only becomes mildly warm. I am sure it could be run at higher voltage for greater brightness when used with a heat sink.

Ian B
12-11-2012, 04:05 AM
Evan,

Yes, I can see the issue of colour fringeing in the case of direct lighting. I was thinking of doing the same as you did - let the LED's shine on a white wall, and then use the scattered light for illumination. In this case, i guess the fringeing wouldn't be an issue?

Good information on the current draw, especially for cable sizing - around 0.2 amps per colour per meter. A power supply of an amp per meter sounds about right.

Thanks,

Ian

Evan
12-11-2012, 04:24 AM
BTW, there is no possibility that my power supply was sagging. Its an 18 amp hour battery that can supply up to 300 amps. My strip does throw direct light on the floor. Using three different colour strips might still be an issue with all reflected light since it would be difficult to ensure balanced reflection at different angles. Better would be to just use a couple of RGB strips.

Ian B
12-11-2012, 05:10 AM
Yes, I can see that. I just noticed that discrete colour strips were cheaper per meter, thought it could be a cheaper way of getting more light intensity.

On PWM - this might be of interest:

http://research.ijcaonline.org/volume58/number22/pxc3883827.pdf

Ian