Introduction: Blue LED Light Box

Picture of Blue LED Light Box

Build your own blue LED light box to combat SAD and winter blues. Sleep better and wake up with more energy while exploring the world of high-power LEDs. If 18 watts of LED power isn't enough to wake you, just add more LEDs!

Also, check out Blue LED dawn simulator for Soleil Sun Alarm.

Step 1: Why Blue Light?

Picture of Why Blue Light?

According to light box manufacturers 446 to 477 nm wavelength light is supposed to be the most effective for treating seasonal affective disorder (second to natural sunshine, of course). Instead of buying a commercial product that lacks external control and might be based on psuedo-science, I decided to take advantage of the rapidly falling price of high-power LEDs to build an 18-watt blue light.

Step 2: Materials

Here's what I used and where I got it.

Luxeon III Stars, (Lambertian) 3 blue (470 nm), 3 royal blue (455 nm) -
BuckPuck 700 mA driver (03021-D-E-700) -
30 volt DC power supply - RadioShack
1/8 Mono phone jack and plugs - Radioshack

lying around the shop but available at RadioShack or Jameco:
small breadboard
heat grease
1/8 thick aluminum sheet approximately 8 x 10
#4-40 nylon screws and nuts

Step 3: Rough Cut and Drill Holes in Aluminum Base

Picture of Rough Cut and Drill Holes in Aluminum Base

The LEDs generate a lot of heat, so I mounted them in an aluminum base that could also act as a heatsink.

Cut the aluminum to size and mark where you want to mount the LEDs. I printed a template on paper and used it to mark where to drill mounting holes.

Step 4: Round Edges and Sand

Picture of Round Edges and Sand

Round the edges of the base and then sand it. I sanded with 100, 200, and 320 grit paper.

Step 5: Bend the Aluminum Base

Picture of Bend the Aluminum Base

Clamp pieces of wood to the front face of the base to keep it flat. Bend it into an upside-down "7" shape.

Step 6: Cut Mica

Picture of Cut Mica

The back side of the LEDs is not electrically neutral, so they need to be isolated.

Cut thin pieces of mica, which is thermally but not electrically conductive, into the shape of the LEDs. I used a laser cutter, but you can also use the mica's existing shape and carefully drill holes to fit the LED's screws by starting with very small drills and working your way up.

Step 7: Mount LEDs

Picture of Mount LEDs

Put heat grease on both sides of the mica and sandwich it between the LEDs and the aluminum. Attach the LEDs with the nylon screws.

Step 8: Solder the LEDs Together

Picture of Solder the LEDs Together

Solder the LEDs together in series and connect them to the LED power supply. The Buckpuck power supply will drive 700 mA up to its maximum voltage. Connect the DC power supply to the Buckpuck. You can solder all these connections, or if you want to do fancy external control, putting everything on a breadboard is easier.

Step 9: External Control

Picture of External Control

If you want to control the light from an external source, like a dawn simulator, bend up the backside corner and drill a hole to mount the 1/8 connector, or whatever connector is appropriate. If you are not going to do external control, you should at least put an on/off switch on the light.

Step 10: Mount Power Supply

Picture of Mount Power Supply

Hot glue the DC power supply and bread board to the base.

Step 11: Felt Feet

Picture of Felt Feet

If you're going to put the light on a nicely finished wood surface, put some felt feet on the bottom.

Step 12: Enjoy the Blue Glow

Picture of Enjoy the Blue Glow

Bask in the blue glow of the LEDs. Some people use them in the morning to wake up, while others use them in the evening to help go to sleep. Run a few experiments and see what works best for you.


lawren_0607 (author)2015-12-05

Hello, will this diy light work for acne?

I am considering purchase a blue light as reviewed on, if this light works for acne so I will make it myself.

1403 (author)2010-01-18

this is a great insperation to my school work

larsmunck (author)14032010-03-03

LED in ears cures SAD!!

Listen guys. This is going to sound like SCI-FI i know, but bear with me because this is awesome news! :)

Scientist in finland have found out that if you insert a pair of very strong LED lights in the ears, it cures SAD.... Simple. Yes. Weird. Yes. Effective. Yes, because they have allready launched a product thats on the market now caled "Valkee":

Its a portable lighting device, looks like a walkman, but it costs like 200 euros!

Now I am really curious to make a DIY set for myself, but a bit unsure how strong the LED should be, what size, color, and if there is risks like burns/ shock. I asked my local electrician and what he could produce was a white 20 MA LED.

What do you think!!?

hanlin_y (author)larsmunck2012-02-14

According to the study, 1 lumen was effective at treating SAD. It looks like 20 mA white LEDs are more than bright enough.You can dim it further with resistors if you prefer.

Blue and cyan may work too but the longer the wavelength, the safer it is. I have never used the Valkee before. If you have, please let me know about it. There is little if any risk of getting shocked from 3.6V but shorts can burn at any voltage.

Here's an article about my opinion

larsmunck (author)14032010-03-03

LED in ears cures SAD!!

Listen guys. This is going to sound like SCI-FI i know, but bear with me because this is awesome news! :)

Scientist in finland have found out that if you insert a pair of very strong LED lights in the ears, it cures SAD.... Simple. Yes. Weird. Yes. Effective. Yes, because they have allready launched a product thats on the market now caled "Valkee":

Its a portable lighting device, looks like a walkman, but it costs like 200 euros!

Now I am really curious to make a DIY set for myself, but a bit unsure how strong the LED should be, what size, color, and if there is risks like burns/ shock. I asked my local electrician and what he could produce was a white 20 MA LED.

What do you think!!?

calskin (author)2011-10-21

Do you remember an approximate cost of this project?

ABaruwal (author)2011-01-25

Im making a wake up light for my final engineering project and and unable to find useful information :(
Can anyone help me with the circuit for the wake up light, please.
Help much appericiated

Ragasteady (author)2009-10-31
This design involves a risk for the eyes. According to Lumileds documentation :

"Blue and RoyalBlue power light sources represented here are IEC825 class 2 for eye safety."

This means that LED can cause damage is the user stares at them. HB LEDs should always have a diffusor in front of them.

jkollin (author)Ragasteady2011-01-16

Yes, all commercial LED light boxes have diffusers on them.
I am looking into the blue light hazard and while it is not completely settled at what point it starts to be dangerous, it is clear that small point sources will be much worse than diffuse ones. Given how powerful the Luxeons are, I would STRONGLY urge you to use a diffuser. Even a thick piece of vellum an inch (3cm) away would help a lot.

If the mechanism for BLH is as expected, even 480nm light is likely to be hazardous, and even when diffused. 505nm light ("Cyan" for Luxeons) would be < 10% as dangerous as blue while being 60% as effective for reducing melatonin levels (the presumed mechanism for reducing SAD and restoring circadian rhythm).

So I am looking at building a box with 505nm to replace my Philips unit.

Finally, make sure the light box is in your upper peripheral vision - it appears the receptors for melatonin/circadian are there, and the area of most concern for BLH is the macula (center point of vision).

Treknology (author)2010-06-08

They've started flogging these in Oz too. How confident are you regarding 660nm and 880nm?

larsmunck (author)2010-03-03

Ok guys!... This will sound like Sci-fi i know, so bear with me!

Scientists in finland have found out that if you insert a pair of LED lights... in the ears (!) it cures SAD. Simple. Yes. Weird. Yes. Effective. Yes, because they have launched a product called the "Valkee". A moblile light theraphy gadget, that looks like a ipod with earplugs (i even think there is a mp3 player in it).

So, now I am very curious of how to make a DIY version!
I am uncertain about size, color, power and if there is any risks of heating/ burns of the earcanal. But, please if you too are willing to test this, put the results up and lets hear it! My local electric store suggested a pair of 20 MA LED 5mm lights. What do you think?!!

larsmunck (author)larsmunck2010-03-03

Here is a pic of the Valkee

Thav (author)2010-01-11

I haven't looked at LED prices lately, but maybe something like this would be easier for some people to deal with.

I think all you would have to do is then provide a power supply and mount that enclosure to a heatsink. It certainly doesn't have the aesthetic appeal of your design, but it's an option.

Farles (author)2007-08-07

Why use those Luxeon Leds, as you could simply use raw LEDs... anyways, looks great, love the aluminum shaping... gives me ideas :P

ewilhelm (author)Farles2007-08-07

I used the Luxeon's because they are much brighter.

moep (author)ewilhelm2009-08-17

very professional :-) thanks for sharing. We are in the Led lighting field also, manufactuer in Ningbo city, China I am Annie from Paneralighting Co., , hope we have a change to talk about the lighting details:-)

captdecoy (author)2009-01-16

Have you considered a hat version of a lightbox? I've been thinking of modding a Luxion LED headlamp with a blue LED and attaching it to a hat visor facing inward.

Xharhoo (author)2008-06-26

Blue light is actually proven to keep you awake.... But that is beside the point. Nice Instructable!

awang8 (author)Xharhoo2009-01-13

If it keeps you awake, turn it on in the morning.

sensoryhouse (author)2007-10-14

I suggest heating the aluminum before bending any angles tighter than the ones used in this project. I wanted a true 90 degree angle for my project but the aluminum broke without heating with a torch. It made the bending easier too!

sunlike422 (author)sensoryhouse2008-09-19

Just think my company make the similar products Morris Shenzhen aglare lighting co.,ltd

ewilhelm (author)sensoryhouse2007-10-15

That's a nice looking light!

sensoryhouse (author)ewilhelm2007-11-01

why thank you, Mr. CEO.

robin67 (author)2008-02-27

just curious, why the controller and not just the power supply directly.

robin67 (author)ewilhelm2008-03-03

thanx, now it makes sense! great idea.any more cool instructs?

MatrixEagle15 (author)2008-02-02

would you seriously want something that bright in your face as you sleep? i would have covered it with a translucent piece of plastic

ewilhelm (author)MatrixEagle152008-02-02

It's only on while I'm waking up - it's not on while I sleep.

nobby_sk8 (author)2008-01-27

thats reli cool but i would of used light reflective acrylic insted that would be easyer to bend aswell :P

ewilhelm (author)nobby_sk82008-01-27

Post an Instructable when you finish!

mori (author)2007-10-21

Umm... Is there a circuit diagram?

mori (author)ewilhelm2007-10-21

Thanks ewil. Your circuit kicks'a. I've got some questions, which I'll put on the other thread. Brian

oconnorm (author)2006-02-21

Thank you for sharing this project. You beat me to it!

You might be interested in reading the following article, which will be published in Biological Psychiatry this summer:

Light Therapy for Seasonal Affective Disorder with Blue Narrow-Band Light-Emitting Diodes (LEDs) Glickman G, Byrne B, Pineda C, Hauck WW, Brainard GC.

You can find a preprint of this article on PubMed.

One of the problems with your implementation is that there is a huge tolerance on the light output of the devices you've chosen. If you want to replicate the device in the paper, you need 607 uW/cm2 at 50 cm.

Entertainingly enough, Apollo LIght is trying to patent the use of LED arrays to treat SAD. If you're American, and you care about this, please be sure to voice your opposition to patent application # 20060009822 "Hand-held programmable ocular light therapy apparatus and methods"

dan (author)oconnorm2006-02-22

i can only read the abstract of the paper without paying, but even there it sounds like this study only tested one light intensity setting, so there is no data to support the idea that the intensity is highly critical. the distance from the lamp to your body will affect the received radiant energy tremendously.

i am not impressed with the quality of this study anyway, because they only tested red vs. blue, and did not test intensity effects. far more interesting would be to test against white light, which includes blue. why bother building a blue-only light when a bright white light has just as much blue in it anyway? oh right, so you can sell $500 blue lamps.

refer to the bottom of this page:
one of the studies listed tested white vs. blue and found little difference. all the studies agree that red has no effect. others found little effect from bright light at all. the results are all over the map.

mori (author)dan2007-10-21

I'm pretty impressed with this.. basically takes the possibility of melatonin suppression as a given and researches different wavelengths.

oconnorm (author)dan2006-02-23

>this study only tested one light intensity setting, so there is no data to >support the idea that the intensity is highly critical. Yah, it being an inverse square law and all, I see what you're saying. The study attempted to control for this - they showed people what 50 cm looks like at the beginning of the study, and asked them to reproduce that viewing distance at home. I would gently suggest that the idea that has no experimental support is your idea that the intensity doesn't matter. This study gives evidence of an effect at a particular intensity, and that is the only evidence that exists (at least, in my limited, nonspecialist survey of the literature.) >i am not impressed with the quality of this study anyway, because they > only tested red vs. blue The reason they tested against red is that it has been previously established that red has no effect. It's an experimental control thing. If you want to do an experiment where you tell the people that you're trying to see if light has an effect, you have to give the controls some colour of light since it would be a little bit of a tipoff if you gave them a box that doesn't light up at all.

dan (author)oconnorm2006-02-23

ok, you can send this paper to i'm familiar with experimental methods. i'm not suggesting that intensity of light does not matter, of course it does. its just that we have no reason to think that the value selected by the experiment is anywhere near the useful minimum or maxium because they didn't test that variable (at least, according to the abstract). similarly, their study of red light vs. blue light is i'm sure completely valid scientifically, the problem is that it is not a very interesting question to be studying. answering this question sheds little light on what we are really interested in, questions like: can SAD be reduced through environmental therapy? what is an optimal strategy for doing this? what is the scientific basis (proposed mechanism) for any observed effect? could an understanding of the scientific basis lead to development of enhanced SAD treatment methods? there are many variables involved in answering those questions, and such a specific test as 'red vs. blue' gets very little further data towards an answer. some of the earlier studies seemed to have been more logical in terms of the variables tested in the studies. from a brief look at the abstracts of a number of papers it seems that a couple studies showed (unsurprisingly) that white light is just as effective as blue - which of course is completely logical since white includes blue. i'm not aware of any detrimental effect of the non-blue part of the white light (except the UV), so why is anyone bothering with making blue-only lights? it seems very silly. if you could show that blue-only was somehow more effective than white, now that would be interesting.

oconnorm (author)dan2006-02-23

Yah, I agree that there are lots of questions still to be answered here, and that the study isn't perfect. For one thing, the study only used 20 people or so; it would be nice to have larger numbers. They used a pretty big array - wouldn't it be great if you only needed one (possibly high brighness) LED to get an effect? My light is going to try to reproduce the irradiance they used, since that's the only evidence-based treatment I've found. It would be really great if it didn't matter though - I have to drive 3 hours and beg shamelessly for access to a suitable radiometer.

dan (author)oconnorm2006-02-23

if you know the rated output of the LED's (which is always on the spec sheet), you can calculate the irradiance. you don't need a meter.

oconnorm (author)dan2006-02-24

Hee. Have you ever actually looked at the data sheets on the high-brightness LEDs? The tolerance on the light output is like 50%, and nobody is binning these parts for brightness. I suspect that's why Apollo is using old-style 5mm LEDs.

ewilhelm (author)dan2006-02-23

I think manufacturers are making blue-only light products because blue-only can be smaller, more portable, and perhaps more power efficient than white light products that include the same amount of power in the specific blue frequencies. For the time being, take as given that they are correct about the specific frequencies; the amount of blue power from the GoLite could be replicated by a certain size incandescent bulb (for example a halogen bulb that would generate "waste" power in other visual frequencies and make a lot more heat than the LEDs). If this size halogen bulb could be housed in something smaller and more portable than the LED-based GoLite or this Instructable it wouldn't make sense to use blue-only lights. Off the top of my head, I don't know what size incandescent would be required, but I do know where Dan's next Instructable should be focused!

oconnorm (author)dan2006-02-23

If someone were to post a (munged) email address, someone might receive a copy of the paper from an anonymous donor.

a11en (author)oconnorm2007-05-10

Ok, I tried to find that pat application- can't seem to find it. :( Is there an inventor name to go along with it? Has the application been accepted already, or perhaps I'm searching for the number in the wrong way? Thanks!! [yes, patenting a number of blue led's in a row is definitely not a good patent]

a11en (author)a11en2007-05-10

that's not a link- sorry!!

oconnorm (author)oconnorm2006-02-21

That's Apollo Health, not Apollo Light.

shimself (author)2006-11-06

If you look at an undiffused LED for just a minute you see spots which remain when you shut your eyes. The standards are not yet written, but I cannot believe that this is safe. (Interest declared but I want to stay anonymous)

ewilhelm (author)shimself2006-11-06

True, but I don't typically stare at the LEDs. However, the same thing happens when you stare at an incandescent light bulb or the sun, for that matter. Perhaps someone should write a standard for use of the sun? I, for one, welcome the government in to regulate all aspects of my life!

i_make_stuff (author)ewilhelm2007-06-27

look, the same thing happens with nearly any object. it is used in optical illusions books. what happens is that the image is temporarly burned into retina. those "dots", although annoying, are completely safe, with the exception of the sun.

Saintkosmosis (author)shimself2006-11-26

If you look at any LED for more then a few seconds it will leave dots. You need a diffuser and or reflector, by placing the led against the cone you just get light and not the harsh point. Get used to these lights they are starting to take over conventional lighting. If you want to buy LED's in bulk or normal feel free to email me and ill send you onto to suppliers that know what they are doing and have great prices. Cheers

About This Instructable




Bio: Eric J. Wilhelm is the founder of Instructables. He has a Ph.D. from MIT in Mechanical Engineering. Eric believes in making technology accessible through ... More »
More by ewilhelm:LEGO table with integrated parts binCustom Wooden Train Track X-crossingMad Max and War Boy Nux father son costume
Add instructable to: