Introduction: LED Only Apartment Lighting

Picture of LED Only Apartment Lighting

My on average power consumption is 100 watts! If I turn on ALL the lights in my apartment I will be consuming 324 watts.

Motivation: saving money on electricity, being energy efficient and doing it on a budget

Scientific data: When a light-emitting diode is forward biased (switched on), electrons are able to recombine with electron holes within the device, releasing energy in the form of photons.

This makes perfect sense to me, but if you're a bit confused read this wikipedia page on Light Emitting Diods

LEDs use even less power then Compact fluorescent lamps ( and in the recent years the price of these lights have gone done enough to be affordable.

Disadvantages to using LEDs:

   1. They do not emit as much light as regular or CFL light bulbs.
   2. If they do, they cost way to much.
   3. To achieve normal lighting conditions you must use many LEDs
   4. If you use DC power, you must use very thick and costly wiring to
       reduce resistance at low voltages

I looked online and in various stores for LED lights. The most affordable ones seemed to be flood lights with a GU5.3 plug. However, it would take very many of these "bulbs" to produce sufficient light and lots of wiring. But since these worked with 220 Vac (European standard) I would save on wiring.

 My choice LED light bulb consists of 18 small LEDs put together in a hexagonal matrix with a reflator background with a small individual power supply inside each lamp. Thus, it uses regular voltage and each light reduces 220 Vac to 12 Vdc the diods need (pic included).

I bought 10 of these lights in a hardware store and started experimenting with how much light these would give out. It turned out that I was happy with 4 lights per 10 square feet (1 sq meter).

I also had a white string of LED Christmas lights from IKEA, which I wanted to use in the bathroom.

Once I had the lights, I could start planning my lighting with LEDs.

Step 1: Plannig

Picture of Plannig
My apartment is 516 square feet (48 sq. meters - 8 x 6 meters) 26 x 20 feet (I'm including a photo with detailed dimensions, although in Russian and metric). One studio/room and a bathroom.

I figured to grid the ceiling in 1.6 x 1.6 in  (50 x 50 cm) squares and put the LEDs in intersections. I needed 135 LEDs.

I looked online and in various stores. I found the best deal in Moscow, Russia to be at LED-LAZAR – a manufactured of LEDs in small town of Serpukhov (, outside of Moscow.

For the LED - G5.3- 220 whites I paid 70 Russian Rubles which is roughly $2.50
I also bought 20 of the improved/higher luminosity LED-G5.3-6UW-C at 80 Russian Rubles per light bulb, which is roughly $2.75

I bought these at wholesale prices since I was getting over 100 light bulbs.

I paid $275 for 130 of the  LED - G5.3- 220 and $55 for the improved  LED-G5.3-6UW-C. I spent roughly $330.00 for all my lights.

I then had to buy the fixtures to install the light bulbs in a fake ceiling made out of drywall. These ran 29 Russian Rubles per fixture, which is about $1.00 I spent another $135.00 for the fixtures to install the LED lights into.

Now I had to buy the wires. Each LED uses 2.4 watts of energy. So If I turned on all 135 lights simultaneously I would be using 135 x 2.4  = 324 watts. And according to Ohm's law, this means that there would be a current running through the wires of (324 watts divided by 220 Vac) 1.5 Amps. I then looked up wire diameter tables online and got 3 times the diameter I needed, which was 1.5  square millimeters.

I purchased 700 feet (200 meters) of this cable at about $150.00

Below is a link to over 20 photos of my lighting setup (in Picasa web albums)

slideshow via picasa

Step 2: Tricky Places, Like the Shower LEDs

Picture of Tricky Places, Like the Shower LEDs

In some areas I had to use some tricky techniques like drilling holes in the drywall and placing a strand of 130 LED Christmas lights from IKEA. I had to paint the piece of drywall before I drilled the holes and installed the lights. I then put covered it with cellular poly carbonate panels and installed the tropical shower head under them.

Here is a video on youtube of my tropical shower with LED lighting in action:

Some of these photos are self explanitory of what was done.

LEDs in the shower

Step 3: Installation

Picture of Installation

I ran the fatter central light power wires to corner locations of the kitchen, main area and bathroom. I then drilled holes and ran wires from the central cable through each hole in zigzag pattern. Each LED light is connected in parallel according to city/electrical code. I have the skills and permission to work with wiring under 20 k Wt, but I called the building electrician to inspect my setup and attach the main power circuit breaker.

Below are self explanitory pics of my setup already powered up.

I've also added some photos of my LED lighting over a computer desk/treadmill, an idea I got from this wonderfull instructable by ewilhelm

More photos and slideshow via picasa

I'm in the finalists for LED CONTEST! Thank you everyone!


decc1954 (author)2012-04-03

To the author of this magnificent instructable.
Thank you for THINKING, and
thank you for sharing this.
I think all of the hoopla merely reflects that you have an impressive instructable.
It is EASY to find fault.
Finding diversity and flexibility and ingenuity would be more productive.
Thanks for this. When can you help me with my place? LOL.

rotoole (author)2013-01-22

Great write up. I'd love to see if this could be solar powered since you pull 12v from the backup batteries usually.

gauzz (author)2012-01-31

Looks a bit violet but it could be tha camera's white balance setting since a couple photos don't have the violet tone.

SDdiy (author)2011-06-19

I am really curious about the details and what products you used. I am assuming somewhere you had to use a power brick or some sort of converter to step down the power going from your power service into the lighting systems. Did you use something custom or just a generic power block that comes with an under cabinet lighting system?

I want to do something like this in our laundry room with exposed ceiling. Not being an electrical engineer I was hoping to be able to use an under cabinet modular system and tie it into the existing wiring.

pashanoid (author)SDdiy2011-06-19

No power brick was used. The leds are designed for 220Vac use. I took one apart and there is a capacitor and a diod bridge which brings down the voltage to whatever the leds use:

HiFiMan (author)pashanoid2011-06-19

It takes a transformer to raise or lower voltage.A diode rectifies AC to DC and the capacitor smooths the pulses out of the DC voltage.The diodes and capacitors have little effect on voltage manipulation other than minute parasitic voltage drops.
Usually placing 3 LED'S in series will allow them to be powered with 12volts using a current controlling resistor,12 volts seems to be the preferred voltage set up since this allows the user to switch to battery power in case of power outage.If you plan to power many LED'S then I would recommend using a current regulator such as a LM317 or larger current controlling package. LED'S can be powered by low voltage AC as well.
To all those interested in a project like this I suggest NOT using typical dome type LED'S, as these are not designed for lighting applications but rather for indicator applications.The sm5050 is a good choice for low power consumption use at low cost but you will need to use several in a series parallel configuration.
LED'S are available in many wattage's now for lighting applications from less than 1 watt to over 10watt each.

ac-dc (author)HiFiMan2012-01-02

These are set up with a current regulating switching power supply. No transformer is likely to have been used as they aren't needing to achieve a precise voltage, rather a switching pulsed high frequency output through an inductor which ends up as an average current tolerable by the LED and a frequency high enough the human eye cannot detect it.

Here is the big design choice. You can opt for modularized "bulbs" each with their own step down rectified power supplies in them, leveraging existing bulb sockets available for a mains 110VAC or 220VAC system, or for various ways to drop current on a 12VAC or DC system, OR you can do it from scratch with the emitters, custom housings, and a single or fewer (think I'd want more than one so A single point failure does not cause loss of all light everywhere) centralized power supply.

Generally the centralized supply will be more power efficient and "can" cost less in total if you are skilled enough to fabricate your own light housings, but it will be much less user/homeowner/etc friendly for someone else to deal with later which is a very good reason to avoid that kind of setup for permanent residential lighting where average people have to deal with problems later opposed to specialized industrial situations or portable and/or disposable lighting devices.

On the other hand if you were to use a centralized supply with fancier circuit including active PFC, your power factor could be very close to 1 so measurements and conclusions about power usage were more precise.

pfred2 (author)HiFiMan2011-07-27

LM317s are linear regulators and being such they dump excess voltage as heat. The further you need to drop from input to output the worse it is too. As much as I may like linear regulators efficiency isn't really their strong suit. A better choice would be a switching supply like we use in our computers.

I powered a regular dome LED once with some diodes and a resistor to use as an indicator light in a project I built and was amazed how much power I had to dump to get an LED to work off wall power. I think I used a 5 watt 2200 ohm wire wound resistor and the thing got hot enough to cook on! Calculating it I could have even gotten away using a 2 watt resistor. The 5 watt one I used was almost getting too warm for my liking though so I stuck with it.

It worked out OK replacing a burnt out neon grain bulb I didn't happen to have another on hand to replace it with. But I'm sure it is a massive waste of power.

It is on the right side of the tan box with the black stripe on it in this picture where I'm wasting another 100 watts of power load testing a linear supply I'd built:

I was so mad because I'd already cut that case out for that particular indicator lamp then the neon bulb burnt out! All the circuitry for the LED is inside that metal case. I built it on an old lug strip.

For the curious:

The insides of that linear supply. It is LM723 based adjustable current limit variable voltage with 3 TO-3 boost transistors. Sorry I never did take a picture of what is inside the tan box. Well, this is one thing in there:

The rest is a number of block bridges to rectify outputs and a custom barrier block I made to connect everything together. Mostly it looks like a rat made a nest out of wire though.

pfred2 (author)2011-07-27

So with your total watts consumed figure you've actually used an ammeter on the AC power input to measure this or did you just add up the rating of all your LED lights?

pashanoid (author)pfred22011-11-12

I use readings from my invertor

pashanoid (author)pfred22011-07-28

I have an ivertor that shows my power usage

pfred2 (author)pashanoid2011-07-28

What model? I'd like to see such a device that has a gauge that displays its input. That is rare.

pashanoid (author)pfred22011-07-28

pfred2 (author)pashanoid2011-07-29

I must say it's all Greek or Cyrillic to me? Klingon? Chicken ran though spilled white paint? The correct answer is D All of the above.

zilcho (author)pfred22011-11-11

Looks Russian

pfred2 (author)zilcho2011-11-11

Same difference.

pfred2 (author)pashanoid2011-11-12

I'm glad you have absolute faith in your instrumentation. Having been exposed somewhat to the complexities of power factoring AC current myself let me just say I've my doubts.

The only way you could test the two for me would be to use the same method for both loads. Anything else is apples and oranges I'm afraid.

I'm sure you're saving power, I'm not so sure you know exactly how much power you are saving though. Because actually figuring that out isn't trivial. Well, it isn't for all us folks who can't count every electron flowing past a point. ha ha!

pashanoid (author)2011-08-03

Today I recieved an T-shirt and badge and some stickers! Thank you!!!

angelosantoz (author)2011-07-01

Such a great project! If only more people would do this. The only thing for me is I don't think I could handle such harsh lighting.

pfred2 (author)angelosantoz2011-07-27

That was what I was thinking. While the overall project is impressive the results I mean the place looks blue! I'd have tried to get a different spectrum LED. Some LEDs do seem to have a rather bluish cast about them to me. But I've seen purer white emitting ones as well.

pashanoid (author)pfred22011-07-28

Looks blue in the picture, not so blue in real life. I'm ok with the blue, but you can buy more expensive, less blue leds and that will soften the tones I guess

pfred2 (author)pashanoid2011-07-28

Cameras can take photographs with false colors. Noted. It might have been a good idea if you tweaked your photographs digitally just so a lot of us viewers weren't fooled by them.

No I really can't, buy more expensive solutions that is. It does appear to be a growing trend that worries me though.

Actually I find it sort of funny that you even thought I could, or maybe even should. A disturbing trend indeed.

mrwo1f (author)2011-06-18

I dont understand why that guy thought this was a joke ....? anyways i have tried a similar project (including aquarium led lighting) but found that most leds out there seem to fade over the months with constant use even though they say they will last 10 million hours etc. have you had any problems like this or are there a specific type to go for what can do the real deal?

nerys (author)mrwo1f2011-06-19

Short of defective bulbs FADING LED bulbs are not defective. They are USER defective (in this case user is whoever MADE them)

the reason a white LED fades is you are over driving it and its "over heating"

as the white LED over heats it "burns" the phosphur (they are actually UV LED's with a phosphur inside that converts the UV light to WHITE light not unlike a Florescent bulb.

then they just DIE. you see while LED's produce a lot less heat than a incan or even a FL bulb they STILL produce a LOT of heat relative to power input. roughly 80% of the power you put into a typical white LED comes out as heat.

SO how 100 watt LED array is 80watts of heat 20 watts of light essentially.

here is the problems. LED CAN NOT TOLERATE HEAT. it kills them deader than a door nail.

anything over 100' is typically DEATH to LED's

problem is "WE" do not percieve 100' as "too want" since out body temp is 99.6' near external is 98.6 and skin temp is 95-97'

so 100' is only "vaguely" warm to us.

if your LED so much as feels lukewarm YOU HAVE A PROBLEM and need to provide COOLING for the LED.

typically you need 9cm/2 of PASSIVE cooling area per LED watt Consumed.

now if you have a SINGLE emitter (little 30mah led) it can typically cool itself.

but once you go to higher power emitters or you start to ARRAY them the problems begin.

many of the LED bulbs you can bulb online etc.. ARE NOT SUFFICIENTLY COOLED.

I bought these wonderful LED tube lights from someone on ebay. 192 white led's in a 3 ft tube.

wonderful amazing bright white (warm) light.

till they turned blue and puked withing 3-4 weeks tops. I had one in the summer eat itself inside a week.

I shoved a thermometer inside the tube. 160' yeah thats why it died. he was cranking them at full tile 12 watts with ZERO cooling in fact being inside a tube they were INSULATED.

So I got a VARIAC off ebay and started LOWERING the voltage to lower the wattage. once I got down to 89 volts they stopped over heating. (I won't let them run hotter than 5' above ambient so if its 80' in the room they can't run hotter than 85')

its been 7 years and they are still running. even the ones that started to "blue" and fail ceased and work fine (though still a bit blue of course can not reverse that) to this day.

they are about half as bright but consumed 1/3 the wattage (4 to 4.5 watts each) instead of 12 watts.

I just USE MORE. the way to make LED's brighter is NOT with more power its with MORE LED's just keep adding them till you have sufficient light.

you should "UNDERPOWER" LED's as much as possible to extend their lifespan.

to me if it does not last 50,000 hours MINIMUM 80,000 preferred ITS DEFECTIVE in my book.

Many of the chinese designers run the LED's AT their rated voltage and power values.

but these values are MAXES and ASSUME proper cooling.

so your aquarium bulbs likely dimmed because of bad design or WAY to much power consumption.

Here is my typical "WELL DESIGNED" led bulb I got off ebay

NOTE at least or more than half the bulb is HEAT SINK. if your bulbs is not like this you either need to dramatically lower power consumption or MAKE it look like this. (this is a higher power 5 watt emitter array)

its about as bright as a 30watt bulb. a wee less than a 40watt bulb.

HEAT HEAT HEAT - HEAT is everything. you MUST control heat.

bulbs turn color or fade for 99% of the time one reason. TOO MUCH HEAT from the bulb itself and no way to get rid of it properly or sufficiently.

I have boxes of dead LED bulbs over the last 10 years and almost 100% of them died "heat death" by being over powered or insufficiently cooled.

ideally you want to use (if room allows) 1 watt or less LED's and ARRAY a lot of them like the author did for his apartment.

THAT is how you do LED lighting. you have to change your concept of lighting from massive bright POINT SOURCE (ie one bright bulb) to massively ARRAYED lighting.

the result is a lot more work BUT much much nicer even "day like" lighting and incredibly long life and lower power.

I have about 60% of my house converted to LED. when I manage to get to 100% I will remove the lighting from the grid all together and goto solar/battery since at that time 100% of my lights all burning should be under 400 watts meaning typically less than 60-70 watts being used "at any one moment" in time.

at that level suddenly solar battery becomes practical and affordable.

guitarman63mm (author)nerys2011-06-19

I'm curious as to what makes you an expert. What are you talking about, claiming that LEDs can only be lukewarm? You would need to be under-driving the emitter so heavily that it would be completely ineffective in cost.

Take a look at this data sheet. Operating temperature: -40 ˚C to 110 ˚C. That's a maximum of 230 ˚F, not lukewarm.

I currently have a 50w emitter ( sitting in my garage on this heatsink:

Now, assuming 1 square inch per watt (a proven recommendation, I might add), that gives me a heatsink dimension of 66.5 cubic inches, and a surface area of 99.7 inches squared, not to mention the fan.

Are you trying to tell me that I don't have sufficient heatsinking, that you know more about LEDs than the electrical engineers who actually make the data sheets? Than the manufacturers who rely on those data sheets to make quality products?


stoobers (author)guitarman63mm2011-06-19

This fellow "nerys" has wired most of his house with LEDs. You, "guitarman63mm" have something "sitting" in your garage. Nerys knows more than you.

Also, since he has actual data gathered through genuine use, Nerys does know more than the "experts" for the application he described.

Those of us who have built products from electronic components - most will agree with me, you need to TEST the components. Even a resistor can be WAY off what the manufacturer claims.

And it doesn't make any difference what the manufacturer says. If what the manufacturer says DOESN'T WORK, you can go by why the manufacture says. An ounce practical application will always trump a pound of theory.

Practical application does not include sitting an LED on a counter then spouting off numbers pulled from a spec sheet.

pfred2 (author)stoobers2011-07-27

I beg to differ. Practical applications do indeed begin with carefully studying device data sheets then calculating appropriate values.

I'm curious as to just how many resistors you've actually tested and why you seem to think that many are off their rated values? Why just yesterday I tested about 100 assorted 100 ohm resistors in vain hope of finding one 96 ohms and the best I could come up with was a silver 10% band that measured 98 ohms. None of the gold 5% bands were off by more than an ohm. In other words everything was exactly as it should have been.

So excuse me if I call BS on your fabrications as marketed electronic components are far beyond the realm of theory.

stoobers (author)pfred22011-07-28

Practical applications begin with inspiration and end with trial and error. The data sheets are details for the worker bees to save some time.

10% deviation is the spec. You found the deviation less. Therefore, the manufacturer spec was wrong, which is what I have been saying - you can't rely on them as they are frequently wildly inaccurate. Install the resistor in a circuit, change the temp and you will change the resistance.

Regarding the "realm of theory", an electron is just that: a theoretical particle. It happens to be a useful theory, but a theory just the same. No practical application of electronics is going to "subject" itself to theoretical compliance for the sake of manufacturers data sheet.

guitarman63mm (author)stoobers2011-06-19

That doesn't make him more knowledgeable at all, quite frankly. You can do something thirty years and still be awful at it. Go ask a politician!

What data has he collected? What does he know about it? Where did he come up with that crap about "20 watts of light, 80 watts into heat"? There is nothing supporting such an outrageous claim. In fact, the following link says just the opposite!

The 50w LED was nothing more than an example. Talk about a straw man! I have been building flashlights and continuous lighting rigs for the past year from CFLs, quartz halogen, and LEDs, so no, after that short amount of time, I am not an expert. I do happen to have plenty of practical knowledge, and I've learned quite a bit from experience, candlepower forums, and the DX forums.

Spouting off data sheet numbers? Are you serious? Hell, what does Toyota know about cars...who needs oil in a car anyway? This guy is just one more person on the internet spreading misinformation.

pfred2 (author)guitarman63mm2011-07-27

I agree I've seen lots of folks do things wrong for their entire lives! Each of us to our own abilities I guess.

nerys (author)guitarman63mm2011-06-20

lets get something straight guitarman. maybe you should stick with guitars?

you see we went from discussing QUESTIONABLE DATA sheets from CHINESE knock off manufacturer's of question quality and knowledge and WELL KNOW FLAKY PRODUCTS

to you posting a data sheet FROM a reputable manufacturer who designed the LED's they are selling.

can you say apples and oranges?

how much ARE those luminous LED's ? I would LOVE to be surprised but I am guess they are NOT $35 a pop.

in fact I could be reading it wrong but it seems you even went so far as to post an INVALID data sheet a data sheet for a product that does not even appear to be available yet.

did you LOOK at the data sheets for there EXISTING products? no where NEAR the power efficiency of this cherry picked product you point out

there lowest power emitter besides this one is 30watts. at 30 watts there is no POINT for a "home" application for a user trying to SAVE power.

guitarman63mm (author)nerys2011-06-20

The data sheet I posted ( is for a 30w DX emitter. It's the closest I could find to the 50w emitter.

I also posted the data sheet for the Luminus emitter simply to show that not ALL manufacturers lie about their data.

nerys (author)guitarman63mm2011-06-21

I don't think any of them "lie" (emitter designers) I think some are just low quality and shoddy (IE the DX emitter you posted) but they are CHEAP so worth it if used right.

pfred2 (author)nerys2011-07-27

If you want to use an electronic component right an excellent place to start is by reading and understanding the ratings on the device's data sheet.

nerys (author)guitarman63mm2011-06-19

First a 50watt emitter is a 100% different kind of beast. I said very clearly in my statement YOUR TYPICAL LED

so let me clarify further.

we are talking about THESE

kind of LED's once you get into massive array SMT emitters like on your 50watt emitter my experience stops. I have never really played with those kinds.

as for your electrical engineers. I don't care what they think. I care what YOU can read and interpret.

did your EE tell you how much heat sink to use? did he talk about duty cycle? did he talk about ambient temps and humidity? material for the heat sink? an aluminum heatsink of the same size as a copper heatsink will not dissipate heat NEARLY as well as the copper sink,

what about lifespan? did he specify projected LIFESPAN at stated wattage and cooling potential?

let me give you an example of an EE or manufacturer being CLUELESS

Phillips. they have this nice new bulb out. it puts out 60+ watts equiv of LED light from a single bulb. this bad boy has a massive heatsink (and I fan iIRC) and it gets HOT.

works as designed right? yes. for 15,000 hours.

so let me get this straight. your ok with spending $60 on a bulb that will last 15,000 hours or in other words produce the SAME amount of light as a $1 13watt CFL and only last 3,000 hours longer and only save you ONE WATT of power.

you would have to be stupid to buy that bulb but it works "AS DESIGNED" by the phillips EE's

Here is MY criteria. consume as LITTLE power as possible with the maximum number of lumens per watt.

Last longer than the lifespan of my future children (ie I want 100 years) I would take 60 years (my protected remaining lifespan)

this is WELL within the capabilities of LED technology but NOT at 240' and 50watts in a single emitter with inadequate cooling (FOR THE DESIRED TASK)

You also seem to have a READING problem. I was very clear.

9cm/2 PASSIVE cooling area. now maybe your just unfamiliar with the terms. PASSIVE means "NOT ACTIVE"

see that FAN on your heat sink? well THAT IS ACTIVE not passive.

you would need a heat sink 6-10 times that to PASSIVELY cool a 50watt emitter. Personally? your design has a massive flaw. that fan both CONSUMED POWER (reducing your lumens per watt since you now need to add the watts consumed by the fan) and if that fan fails? I give your LED 5 minutes before it cooks itself (I could be wrong on that but not by much)

you going to notice when that fan dies and turn it off before you eatup 50% or 100% of your LED's projected lifespan?

50 1 watt emitters will produce FAR FAR more light than your 50watt emitter and produce LESS HEAT doing it and consume LESS POWER doing it (remember you have to add the power of that fan as WELL as the lower efficiency of the high power emitter)

if your getting less than 70 lumens per watt your WASTING your time (in most applications) as the CFL will produce near that at far far less cost.

I just remember reading a technical doc on LED's it said over 100' reduces the lifespan of an LED (white phosphur) and over 110' (F) KILLS them pretty quickly.

ever since I started ADHERING to this concept I stopped burning up LED bulbs. its that simple. after HUNDREDS of LED bulbs and THOUSANDS of dollars over the last decade I have accrued a lot of practical experience?

does that make me an expert? I would say no. does that invalidate my information? NO.

My information after all is FROM experts. My experience simply confirmed that "expert" data.

its is a proven fact (go check out some papers on CREE lights) that the higher the current the LOWER the efficacy of an LED bulb. as you ramp up power you make more and more heat but less and less light.

LED's have hit as high as 160 lumens per watt at VERY LOW POWERS.

when you can get LED's as CHEAP as you can today its no longer a valid argument to use EXPENSE to justify higher power.

how many LUMENS does that 50watt emitter net you?

DX claimes 4000 lumens. first HOGWASH. I would be VERY surprised if your getting 2500-3000 lumens. (they massively over estimate the lumens of their LED's ie the manufacturers flat out lie or don't even TEST they just figure 80 per watt and use that with no data to back it.

no way in HE double hockey sticks is a current tech cheap chinese emitter getting 4000 lumens out of a 50watt emitter. not even CLOSE.

if your getting 60 lumens per watt I would consider that an IMPRESSIVE emitter. so about 3000 lumens.

a 1 watt emitter "CAN" easily get 90 lumens per watt (the discrepency is SMALLER the lower the power consumption)

I would lower the power even MORE and easily hit 100 lumens per watt.

so lets figure 70 lumens for .7 watts (that is what I would target my power supply for)

I would need 43 1 watt emitters consuming a total of 30.1 watts of power and getting just a hair over 300 lumens. not only 3000 lumens but those lumens would be passively cooled (no fans) last 5-10 times longer than your 50 watt emitter and provide BETTER LIGHT that is evenly distributed throughout the area.

Cost? $258 for the emitters. (deal extreme $6 a pop)

cost benefit. you paid $35 for the emitter and $15 for the fan PLUS your power supply (mine come with power supplies) figure $40 for that.

$90. I have a $168 deficit to make up for. lets see if I can do it.

I figure your emitter will last 5 years (assume 4 hours per day usage averaged out) 10 if you used it less or got really lucky.

that fan? a year if your lucky. so each year your GOING to be replacing that fan even with maintenance (more cost but lets not factor that in)

so what? $10 for another fan?

lets use 10 years.

your spend $90 for extra fans (plus shipping or gas to go get them) $70 for emitters $40 for power supply and $15 for the 1st fan and heatsink.

$210 now my deficit is $48

now lets look at power. YOUR consuming at least 55-60 watts of power (remember the fan) lets assume low end 55 watts.

I pay 17c/kw summer 13c/kw winter so lets use 15c/kw

I will consume $49.44 in electricity over the next 10 years

YOU will consume $90.34 in electricity.

so after 10 years YOU will have saved a wee over $7 compared to me.

but you will have burned up 2 emitters 10 fans and now after 10 years you will need to buy yet another emitter meaning afte 10 years and 1 day "I" will beat you buy $28 and will from now on continue to save money by larger and large amounts since in theory my emitters should last 50+ years. (remember lower power passive cooling AND intentionally under powered) reducing load on ALL the components.

in theory of course. no idea if they will last 50 years but i Have some approaching 10 years with NO degradation so far so its encouraging.

also lets not forget the shipping and or FUEL you will consume maintaining your emitter the labor we can ignore since on day one my install is a lot more labor than yours so lets call that a wash.

Manufacturers like you. you have to keep going back to "BUY MORE" which is why phillips made that stupid 12watt $60 LED after all. its profitable for them.

you want to play with some nice LED's to get started?

goto sams club. GE 2 pack 2 watt LED globes $15 or $20.

Walmart .9watt mini globes $7

very high efficiencies plug and play. so how do you light a room with a 2 watt led?

you don't. you use an ARRAY of 2 watt LED's I light the kitchen with 6 of these. later I will install 12 total and cut the power in half.

I will end up with MORE lumens for the same 12 watts of power and "under load" the components so less heat and less wear on the internal components IE longer life.

cost a wee bit more up front but saves a TON of money longer term as you use dramatically less power (this 12watts of LED replaced an 80watt FL fixture) and last essentially forever for all intents and purposes.

SO to answer your questions again.

do you have sufficient cooling? to ME not even close even with the fan. I would want to use a copper heat sink at least TWICE that size WITH the fan.

personally I would reduce the power to the LED to 20-25 watts use 2-3 copper heat-sinks the size of your heatsink each and no fan.

and then use 2 emitters in 2 locations in the garage. MORE light less power longer life and better more even illumination.

Do I know more about LED than the EE's who actually made the data sheets.

I can not easily answer this as you have a MASSIVE number of assumptions that are massively flawed.

the assumptions start with you clearly thinking what the EE things is anything close to what the MAN did and that they might be the same person.

the EE writes the data based on a set of assumptions. you ASSUME the manufacturer (man) who BOUGHT those emitters ADHERED to the assumptions of the EE (but the man just GAVE you the data sheet he got from the EE)

so you have two people here. the EE who designed it and the MAN who built it into his design application.

I can assure you that chinese maker probable has a data sheet for ONE of the itty bitty tiny emitters in that 50watt emitter.

they slapped a bunch of them crudely into an array with some conductors and just multiplied the EE data by the number of emitters they used and said HERE YOU GO. have fun.

you are ASSUMING that the MANUFACTURER adhered to the design criteria that the EE Used to PRODUCE that theoretical data. hah funny :-) (this is CHINA afterall)

Quality Products. god make me gag. your buying a 50watt emitter for $35.

quality belongs no where within an AU of that statement. PREY its quality is high enough to allow it to even turn on. I have had plenty of DOA's from china :-) there quality control is out of this world pathetic but the prices are SOOO STINKING CHEAP its worth it to "blindly shoot" with an order and hope to hit the side of a barn now and then since you will STILL save money.

SO do I know more than the EE who "theoretically" designed the emitter? (after all its very like the original EE Is not even the EE who DESIGNED your emitter it might be a chinese EE who COPIED something from another EE to replicate it and sell it cheap)

don't get me wrong. The chinese are good people offering a good service. but they cut a LOT OF CORNERS to GET you those cheap products.

do I know more than the manufacturer of that 50watt emitter?

HELL YES. I probably know a LOT more than them about their own products and I am NO EE.

Here is what I know.

SO FAR in my experience high power low lumen per watt emitters tend to make a lot of light lot of heat and then DIE.

low power high lumen per watt WELL COOLED LED's tend to produce a good usable light for a VERY long time.

guitarman63mm (author)nerys2011-06-19

First: Can you please construct a coherent sentence without jumping all over the place? It's honestly impossible to reply to a slew of CAPS LOCK statements with no relevance to each other. I suspect English is not your first language, and if it is, I "PREY" for you.

I also think the current LED bulbs being offered are pretty much crap, as well as overly optimistic. That is completely irrelevant. I did not say a single thing about the Philips bulb.

Yes, DX is also overly optimistic, but I've taken EV readings from the emitter and it seems to match up. I do not have a lux-meter, so my numbers are estimations at best. Quite frankly, 60 lm/w would be extremely inefficient, and I believe it is closer to 70. 70 x 50 = 3500 lm. That's still not the point of this. My statement on the " 1 square inch for watt" heatsink is from the DX forums from people with much more experience than I, and it's passive. The fan was an added bonus when I was looking for a heatsink.

Those small LEDs you have linked to are not even worth my time. I build flashlights or lighting rigs, both of which are generally compact. Putting in a ton of dinky little LEDs take up time and valuable space. Furthermore, depending on the bin, I can get differing color temperatures and tints. There is more to go wrong using many smaller LEDs.

This has nothing to do with me or China. This has to do with you simply thinking that you are omnipotent. Here is an american company, Luminus. Their SSM-80 (1 has a maximum temperature of 302 degrees Fahrenheit. You telling me they're lying too? Page 8.

No one will debate that it's smarter to run your emitter at a lower temperature for better output and lifespan, but to get it to run at under 100 degrees fahrenheit would be extremely expensive with the amount of active and passive heatsinking you suggest.
If those little emitters can't handle more than 100 degrees Fahrenheit, fine. I know nothing about them. But the fact is that that is not true for all emitters, and if you were going to wire your house with something, it makes more sense to use a product with a larger safe zone, such as 302 degrees Fahrenheit. On page 9, Luminus has a chart showing efficiency as it heats up. Yes, at 25 Celsius, at is at 100%, but by the time it reaches 135 Celsius, it's only at 80% efficiency! Hardly a destroyed emitter.

nerys (author)guitarman63mm2011-06-20

I am downright poor and I seem to have NO problem getting 100% of my emitters to run below 100'

in fact I won't USE an emitter that runs over 100' (BARE EMITTER!!!)

the point of the heatsink is to remove the heat FROM the emitter so while the heatsink will get hot the EMITTERS will stay cool.

thats kind of the point.

and if your going to go insulting people and calling them names can you at least get your insults right. its omniscient not omnipotent.

at least I assume you meant to call me a "know it all" and not "all powerful"

I will say it again and this is PROVEN factual.

with existing tech 9 1 watt emitters will produce significantly more light than a single 9 watt emitter.

this is a fact that has yet to be shown otherwise.

if you want insanely bright rooms well you have work to do. if you want "ENOUGH" light to work with ie JUST enough you go with the minimum.

10 of those 1 watt emitters will light up pretty much any normal sized room with PLENTY of light.

as for the heat values.

any years ago when I was researching what was killing my bulbs I cam across a document on the web (university site can't find the link any longer it was 7+ years ago afterall)

this is where I got my data bout phosphur based white led's not surviving long over 100' and led's being 20% efficient (compared to 2% for incandescent bulbs)

the university document said 80% of every watt you put into an LED comes out as heat 20% of each watt comes out as light.

this was for your typical encapsulated LED. (which is what was in my TUBES which is what were giving me a problem so that was where my focus was)

I have no idea if this also applies to SMT led or array emitters like your 50w led but I will ASSUME it does until I am shown otherwise.

I can't afford to throw away dollars assuming otherwise.

plus I don't want to run the LED's as designed I want to run them MY WAY that means I want 50 years out of them. to me the bulb is a failure if it can't last 50 years.

lowing the power and temperature and spreading the load over many emitters as a result WILL get you significantly longer lifespan and more lumens per watt efficiency.

your 302' is probably LINKED to a lifespan AT that temperature. I bet its a lot less than 20,000 hours.or my desired 80,000 hours.

if your willing to live with 15,000 hours FINE run them at 302'

I am not. to me they are a WASTE OF MONEY at 15,000 hours.

SO let me get this straight. you CHERRY PICKED a specific LED that meets the criteria you argue about and you call me names based on that one emitter?

I was pretty clear which emitters I was talking about. I was pretty clear MY EXPERIENCE STOPS at the emitters I was talking about.

yet you still me names? sound like your a moron picking for a fight.

I do thank you for pointing that LED Out to me btw I LOVE its specs. 3000k (a little on the warm side) over 100lm/w 1600 lumens 70% lumen retention after 60000 hours.

I bet if I ran it at 8 watts I would keep 80% of those lumens and double the lifespan. I just might have to try one of those out and see how it goes.

so how much do these bad boys cost?

guitarman63mm (author)nerys2011-06-20

Woops. I did mean omniscient.

I can't measure the temperature of the emitter. I don't know what my emitter is running at. I suspect an infrared thermometer would help with that.

Your way is underpowered in my opinion. Let's see. You want to use 10w of LEDs at around 80 lm/w. That gives you 800 lumens, or the equivalent of a single 60w incandescent. That's really not much light at all. I mean, light is light, at the end of the day, but in a 10x10 foot bedroom (around 9.2 square meters I believe). 800 / 9.2 = 87 lux. That's around 5.3 EV at best, not allowing for spill or the inverse square law. Double the distance, and you get a quarter of the light. 2 meters away from the floor, and the LEDs give 21.75 lux, a bit over 3 EV. With 8.6 foot ceilings, you'll get 15.5 lux, a bit over 2.5 EV. Most interior homes are lit to around 5-7 EV. That is going to be a very dim bedroom.

Forget the document. If it applies to encapsulated LEDs, that's fine. I don't think it applies to metal plate emitters. Of course the lifespan is shortened by running it at that temperature. I linked to it to merely show a point.

I'm not looking for a fight, but making blanket statements about LEDs is just downright incorrect.

The emitters from Luminus are well-established emitters being used in many top-quality lights. Their SST-90 is used in the Olight Intimidator (2200 lumens @ 9A if my memory serves me correctly).

Depending on your country, they can be purchased in many places online. Click on the 'where to buy' link on Luminus's site. has them here in the states. See the following link.

I cannot find the one I linked to. I suspect that it is not in production. That being said, a CST-90 chip is $45. Their biggest chip, the CSM-360, is $137. To be honest, I trust them a lot more than DX.

If you're in for a laugh, look at the following link. Even I recognize the woefully inadequate heatsinking (although it is quite bright!).

nerys (author)guitarman63mm2011-06-21

But thats just it. I DID NOT MAKE blanket statements about LED's in fact I was VERY clear about the limitation of my experience. the fact that you IGNORED this quid pro quo and chose to assume I was making a blanket statement is your problem not mine. I can not "make" you understand what I typed.

We were talking about DX EMITTERS which was the limit of my statement. again the lack of understanding is not something I can fix I was pretty clear. if you read my statements IN CONTEXT its clear.

you elected to assume I meant ALL data sheets when in fact I NEVER had a problem with ANY data sheet. my issue was in assuming that some Cheap chinese dx made emitter was remotely close to the data sheet provided with it.

but again you elected intentionally to ignore that.

I have no doubt those emitters are fine emitters. what are there lumens per watt?

you see I don't care about lumens. I care about lumens per watt.

why? because if I need more lumens I can just ADD MORE EMITTERS.

whats critical to me lumens per watt and lifespan.

I want as close to 100 lumens per watt as I can get and I want as close to or more than 80,000 hours as I can get.

lumens per watt GOES UP as you reduce power to the LED, NOT DOWN as you imply. (to a point of course)

so I would rather UNDER-POWER the led and use TWO of them (giving me the SAME lumens as your getting with one at full power) but at LOWER wattage and longer lifespan and no need to use active cooling.

I am sorry but I can not accept that the heatsink you listed is sufficient to cool a 50watt emitter without the fan for any appreciable amount of time.

my experience say no way is that going to work so your going to have to prove it to convince me.

MOST people use those emitters for FLASH LIGHT type projects where that heatsink might work for a LIMITED RUN duty cycle.

but a continuous duty cycle like illuminating a room well I just don't see it lasting very long.

I personally would not be happy without at least 4 times that surface area for cooling and something better than aluminum.

guitarman63mm (author)nerys2011-06-21

I happen to believe their data sheet to a degree.

I posted the can calculate the lumens/watt very easily...

The CSM-360 at 13.14v and 3.15A puts out between 3000-4300 lumens, depending on the bin. 13.14 x 3.15 = 41.39W. 4300 / 41.39 = 104 lumens/watt at the high end. Of course, this is an extremely expensive emitter. I'm sure it would get even better efficiency powered down to the levels you like though.

The CST-90 at 3.25v and 3.15A (10.2w) puts out between 850-1200 lumens. That's about 117 lumens per watt from a $40 dollar emitter at its peak performance. It can be pushed to 13.5A giving 2750 lumens, and a dismal 62.5 lumens/watt.

I would love to run many of them on lower levels if it were cheaper to do while offering the same bin characteristics. For my purposes (video lighting and flashlights), yes, the duty cycle is much lower than one running nonstop.

I'm sorry for creating a large argument based on unfounded accusations. I just meant to offer a counter-argument to some of the claims you put forth, including the "80% heat / 20% light" statement.

nerys (author)guitarman63mm2011-06-21

well remember the 80/20 value is for NORMAL phosphor encapsulated white LED's

I have NO idea if the same applies to these other emitters (and I was clear on that)

I know for example that gallium ?? based white LED's are far more tolerant to heat than normal white led's but are also much more expensive.

mrwo1f (author)nerys2011-07-02

thanks for all the info, i never realised leds were so heat sensitive. i will also turn down the voltage to my aquarium lighting to make that last longer.

you would think designers would run led's below there rated voltage etc but i suppose they would last too long then and you wouldnt buy another ha ha

pashanoid (author)nerys2011-06-19

You know, there is a gap between the leds and the actual ceiling so they are getting some passive cooling in my case. Thank you for a thorough writup on the subj

pashanoid (author)mrwo1f2011-06-19

I haven't used them on a regular basis for a long time yet. They seem to be thes same kind as the ones in led flashlights. I've taken several aprat and used individual leds. So I don't know if they will fade or not. I've had them for about a year now and they seem to be as bright as when I first installed them, however, they haven't been used much...

nerys (author)pashanoid2011-06-19

put 4 of 5 of them together let them run 30 minutes then wrap your hand around them (don't short them)

is it WARM? if its WARM you might want to lower the power consumption just a bit. even a 10% drop in power consumption will result in a nice drop in heat production and a LESS than 10% drop in produced light. its not linear.

you should always "UNDER" run LED's at lower than rated power consumption. it reduces heat and DRAMATICALLY extends overall lifespan.

in theory if you design it right those LED's should OUTLIVE YOU (literally) and outlive your kids (literally)

if your running them off a single power source lower the voltage THEIR to reduce power consumption and therefore HEAT production.

you area bit shy on the details. you seem to be using individual emitters wired yourself (shower ??) but CLEARLY are not using that style in your main room.

you appear to be using read made flush mount 1 to 3 watt emitters ??? or did you wire up your own and put them IN a flush mount fixture ??

I am looking to get a few of these to torture test for a few months and see how they hold up

they have them in warm and cold white (3500k ish and 6500k ish.) they are 1 watt.

the lower the power the more efficient (lumens per watt) they are and the harder it is for the designer to screw them up :-) (heat dissipation wise)

Here is another nice one if you REALLY need more lumens from less bulbs

let us know how you made the primary ones your using in your kitchen etc..

nerys (author)mrwo1f2011-06-19

sorry for the typo's (just got up) it won't let me edit.

NOTE those bulbs I showed pictures of? they still work 100% and they are (I have 4 of them) over 7 years old. still working flawlessly. ie some of them have over 25,000 hours on them.

stackerjack (author)2011-07-11

You can now buy lamps which contain 60 LED elements, and consume less than 5 watts of power.
Search the net for them.

Bernie_R (author)2011-06-23

This is fantastic. I studied "Energy Management" but I live in Chile where energy, until recently was dead cheap. Nobody would listen to my "Crazy" ideas of switching lights off when not in use. Everybody purchased 100Watt incandescent light bulbs as they gave more light! Energy saving bulbs have only recently arrived.

I will try and install your LED idea in my pad and I hope others will follow suit.

Many thanks for a brilliant (no pun intended) idea.
Bernie Robinson

pashanoid (author)Bernie_R2011-06-23

Thank you! It's in our nature to take things for granted I guess, so people are ignorant about things untill problems or lack of resources hits them. Same situation here in Russia, people think electric companies will never raise prices. But they do, they always have. I's up to us to keep an outlook on new technologies and use them to be one step ahead or at least cover our losses. Best of everything to you, thank you!

marcintosh (author)2011-06-19

Thanks for it all. This is just exactly what needs to be done. Someone has to assume responsibility for "Trying New Things" and you've done so admirably. The slide show is HUGE thanks for documenting that so well.

To see if your LED's are wearing or how they're wearing, use a light meter. Good brands are Gossen, Westin, Sekonic, You can find them on eBay really cheaply. You don't need to be worried about how accurate it is as you are only concerned with consistency of the readings.

Thanks again for your work and, for sharing it with the rest of us.

Rich Editor not working right now so it's not the way I'd format it.

About This Instructable




Bio: I'm a Linux enthusiast, avid diy-er. I work for a company that manufactures "smart" invertors (an Uninterruptable Power Supply for the entire house). I ... More »
More by pashanoid:Hand crank LED flashlight hacked - now also USB chargerLED only apartment lightingEco friendly business cards from cereal carton
Add instructable to: