Picture of 10W Retro-Futuristic LED Lamp

LEDs could be very efficient and  economical in long term use. A 10W LED lamp can replace a 100w incandescent lamp or a 30W compact fluorescent lamp. Despite an relatively high initial cost, compared with other kinds of domestic lamps, your electric bill could drop significantly if you use it instead.

Here I will show to you how to make your own stylish 10W Retro-Futuristic LED Lamp, spending around US$ 25 bucks. Let's go!

Remove these adsRemove these ads by Signing Up

Step 1: Materials

Picture of Materials

You will need:

1. One old burned compact fluorescent lamp, for the LED lamp socket  (anyone serves);

2. Two grip flat connectors;

3. One 10W LED: - color of your choice;

4. Two little screws (search in your tech scrap);

5. One 10W LED driver:

6. Thermal paste;

7. One old computer cooler (search in your tech scrap);

8. Heat shrink tubing;

9. 30 cm of wire (2 mm). 

Step 2: Tools

Picture of Tools

You will need also:

1. Soldering iron;

2. Little screwdriver;

3. Small plier cutter;

4. Long nose plier;

5. Medium plier cutter;

6. Puncture tool;

7. Flat screwdriver;

8. Grip plier;

9. Little drill;

10. Electric drilling machine.

Your will need some oil drops too.

Step 3: Disassembling the compact fluorescent lamp

Picture of Disassembling the compact fluorescent lamp

Be careful to not break the CFC lamp. The glass could cut you and there are mercury inside the lamp. 

Slowly try to open the base of the lamp, inserting the flat screwdriver and moving it up and down along the junction. These could damage the plastic a little, but it's not critical. You could after scrape the plastic with a knife or sand it to minimize the faults.

Carefully cut the two wires that join the circuit with the socket, without break the connection. Now cut about 0.5 cm of the insulation of the wire.

Step 4: Welding the LED driver to the socket

Picture of Welding the LED driver to the socket

Place two pieces of heat shrink tubing around the wires. It is important to isolate them and make a good finish.

Now solder the two wires to the AC wires of the LED driver and position and heat the heat shrink tubing for shrinking.

Step 5: Cutting the base

Picture of Cutting the base

Make a little V cut in the plastic base to pass the wires.
1-40 of 97Next »
Lindsley (author) 1 year ago
Man, why don´t you buy a cheap LED driver from DX? Problem solved, without headaches.
KDS44441 year ago
I realize this conversation is now over three years old, but it brings up some questions for me that remain unanswered with regard to current, voltage, drivers, and LEDs. I had been under the impression that any electrical device "draws" current as required, and that too much voltage is what causes them to fail. I have now let this notion go, and am accepting that high-power LEDs draw whatever they can in terms of current until they die. Got it. But this then leaves me with this question: if I have a power supply that is rated to 1A, doesn't that mean that the power supply is only "good for" one amp of power and that if you try to draw more than one amp through it, it will overhead and die on its own? And if so, doesn't this make finding a power supply for a high power LED rather difficult? Or am I not understanding the difference between a power supply and a driver? Or can a power supply contain a current-limiting driver? Or is such a driver a standard thing in most power supplies?

Here's the recent quandry: I have a series of five white power LEDs that I am running in series using an old power supply from a computer that claims to put out 18.5V of power and 3.5A of current. These bulbs seem to be doing OK (I have not run them very long, not more than 2 hours at a time over the past few days) though they do get rather hot (I have them attached to an aluminum bar, a finned aluminum heatsink, and then a fan on top of that and they still get quite hot). I am still rather disappointed with their brightness, though. They are bright, but by themselves even all five are not bright enough for my intended design (an aquarium hood). I also had a set of 3 such LEDs hooked up to a 12V 2A supply for a couple of hours— I finally got the brightness I wanted, but then the bulbs slowly burned out and are now useless. I am about to get a new 10W RGB power LED that I would really rather not destroy. The manufacturer claims the bulbs run at 30-34V and a current of 1.5A. I would rather have the bulb slightly under-driven and last forever than overdriven and burnt out in a year. Can I get some help figuring out what kind of power supply I should be using with this new lamp? I have read ever message above and the constant back and forth of it ("You are wrong about X", "No, you are totally wrong about X", etc. becomes very, very confusing!). Thanks!
Lindsley (author) 3 years ago
Well... Today, my five LED lamps was functioning whitout any problems for more than one year. I think that is a good testing time for me.
burn3173 years ago
I'm very amateur when it comes to electronics but still trying to self teach as much as I can. wouldn't you be able to use the current power supply in the cf bulb housing and run a series of leds off of that?
its my understanding that the bulbs burn quicker than the components behind them.
Lindsley (author)  burn3173 years ago
The CFL power supply can't be used for LEDs. They use high frequency AC to light the tube. You have to use current constant drivers for LEDs, voltage and current according with the LED specs. Or use resistors to limit the current, but they are inefficient.
Nice instructable !
Nice design , too , but there's one thing , add a cover like those ceiling CFL's have , or else somebody's gonna go blind looking at it , for real .
How about installing a fan too ? It might get hot , why not ;) especially if you encase the cooling fin in a case , i might be wrong , but you could correct me .
Lindsley (author)  Electronics Blurred4 years ago
The diffuser have the function to soften the light, and it's ok. These LED is really very bright and there are many other options to diffuse his light. I choose a very cheap.
The fan is not necessary, if the cooler is exposed. It reaches 53 Celsius, and the LED supports it. Don't close this thing with a case or it will fry. In this case, the fan is mandatory.
ac-dc Lindsley4 years ago
Your conclusion is premature. It is not the temperature of the heatsink that matters, it is the temperature of the LED die. Similiarly you don't know the temperature of a car engine by measuring the radiator temperature.

Further, I cannot recommend using the generic power supply linked, it appears to not be UL listed and may be unsafe. I would be glad to be wrong though, if you will provide the CE and FCC certificates for it that are implied by the stamps on its label and a proof that the components are the same as submitted for the certification.

Further, mounting the power supply above the heatsink will heat it up, I would be worried about it causing a fire.
gnawlej ac-dc4 years ago
(removed by author or community request)
sreeci gnawlej4 years ago
EGO issue !
rattyrain ac-dc4 years ago
I think people need to stop judging others' intelligence based on how well they speak English; it doesn't work too well. I judge intelligence based on whether or not one has a profile picture.
Lindsley (author)  ac-dc4 years ago
Well, I'm using five of these LED lamps in my home without any problems. The circuit inside the CFL heats a lot also (in a closed case), and don't burn out. The temperature of the air around the LED could be up to 80 degree Celsius, so 53 degree in the cooler is OK.
I have done this project about 3 years ago, with 1 watt LED.
The heat sink was from an old CPU cooler with integral fan.
without the fan , junction temperature was recorded as 140 F.
As soon as the fan kicked on, it dropped down to room temp, ie 70 F.
Thanks for the info.
Dr.Bill4 years ago
Just ordered the parts for the light. Wish there was something like this for Hot Water Heaters in apartments. Not allowed to modify the heater at all or anything else for that matter. Very good project.

Thank you Instructables for the spell checker.........
frankmiao4 years ago
Hi Lindsley, I run into this post by accident. I am Frank from SatiLED company. I would like say Thanks on behalf of all our company. My question is if we can quote this article on our blog. And as a reward, we would like to send you one 10W power LED for free. Please contact me via to leave your registered email on our website.

Best regards
Lindsley (author)  frankmiao4 years ago
Dear Frank,
Thank you for your interest in my article. Of course that you can quote it in your blog. I'm working in other projects with your LEDs and I will send the link to you when they are finished.
Thanks for the 10w LED too!

Airazz4 years ago
Would this one work? It's cheaper.
Lindsley (author)  Airazz4 years ago
Yes. It is more cheaper because don't have the case or the wires, but it works well.
nerys4 years ago
suggestions on other drivers? this one is rated for 900mah (warm white) but the driver is rated for 1000mah. I would like to run it underspec say 800mah.

IF the specs are true the lumens for this LED are actually pretty good as its getting 60 to 90 lumens per watt. 60 is pretty bad but if you get a few at 90 thats pretty good actually.

The warm CFL I have in my hands is 900lumens for 13 watts. or 69 lumens per watt so unless I get a few sucky 600lumen LED's these are superior to your average CFL.

Its superior again because the light is directional. ALL the 900 lumens go in the same direction instead of "all around" like the CFL. so if your usage is set up to take advantage of that its that much more efficient than the CFL.

I have about half the house converted to LED. This might let me convert the rest so I just might build one of these and see what happens.

yes these can get 50k hours if you under burn them a little bit. even 40k hours would mean around 40 years of usage.
Lindsley (author)  nerys4 years ago
This driver is current constant, it will deliver only the current that your LED requires, and not more, even if it rated for more. Your comment about the direction of the light is very important. This LED sprays the light in a cone of 120 degrees, concentrating then on it.
nerys Lindsley4 years ago
Yes current constant at 1000mah or 100mah more than the warm white is supposed to get. I want to under drive it - lower power and longer life. It delivers the current its designed to deliver.

how would it "know" otherwise?

another trick with LED's is multiple fixtures all crossing each other. Eliminates the hash shadows.
Lindsley (author)  nerys4 years ago
(removed by author or community request)
nerys Lindsley4 years ago
ahh maybe these work differently but its you that does not understand how this works.

LED's are Current Hungry. the whole REASON you need to use "fixed current" power supplies with an LED is because as it gets HOT (and they DO get hot 80% of the power comes out as heat) they increase in resistance. as the resistance increases they draw MORE CURRENT which makes them hotter which makes them draw MORE current which makes them hotter.

you get the idea. eventually you get a run away effect and they DIE as they "melt down"

this is the whole reason you have to RESTRICT the current to the LED and this is the whole reason so many Mass Produced LED bulbs DIE far far before they need to (over driven to make them brighter)

so running a 900mah LED at 1000mah won't make it blow up but IT WILL shorten its lifespan.

You plug a 100mah LED into a 1000mah power supply that LED IS DEAD and FAST.

Running it (the 900mah led) at 800mah will LENGTHEN its lifespan while only losing a tiny bit of lumens.

unless "something else" in the loop is restricting the current somehow if you plug a 900mah led into a 1000mah power supply IT WILL EVENTUALLY DRAW 1000mah from that power supply especially in the summer or unless you really go crazy on the cooling.

its LED 101.

9.15watt is nice. How long was it running and how hot was it when you measured that? how hot did the actual emitter get? Lumens from that?

I want LED for the low cost low power and MOST IMPORTANT to me. LONG TIME. I would rather have a dimmer product where I use TWO to get more lumens and have it outlive me than have a brighter product that dies in 10 years.
Lindsley (author)  nerys4 years ago
My lamps was running by two months, without any problem. The cooler fins temperature stabilized in 53 degree Celsius. The emitter is more hot, the specs limits the temperature in 80 degree Celsius. I measured the volts and amps at 15 min of functioning. I don't have the lumens meter, but the lamp illuminates well a 4m x 4m room. Cool white have more lumens than warm white, but cool white is more bluish. Pure white is an ideal color, when you have the choice.

The resistance of the LED actually decreases when they hot, so the current increases.

How the driver limits the current?
nerys Lindsley4 years ago
no the resistance goes up when they get hot.

as for how the driver works. I have no idea.

as fot temps ahh NO if that temp is celsius its way way way too hot.

anything over 100' F or 38' C is way too hot for a White LED. you wil start burning away the phosphorus at temps over 100' F

if your emitter is running 80C or 176" F its NOT going to last long (two months is not long LONG is 40 years) unless its a VERY different emitter than your typical white LED.
Lindsley (author)  nerys4 years ago
If the resistance goes up, the current should goes down. (I=V/R)
nerys Lindsley4 years ago
but not if the resistance increases HEAT which increased resistance and increases power demand.

hotter = more resistance. More resistance = more heat.

more heat equals yet more resistance.

your equation does not factor temperature. you equation is for when CURRENT and RESISTANCE intermingle.

LED's RESISTANCE AND TEMPERATURE mingle and a side effect of that mingling is more temp equals more resistance which means more current demand.
more resistance = less current. less current = less heat....
as an led overheats it's junction allows more current through which causes it to explode or melt down.
josephlebold the resistance is inside the led so it does not decrease current inside the led it decrease current as it move through the resisted energy is then turned to heat and the more electricity resisted the more energy turns to heat

I hope i helped a little bit
hmm so I would ask this question. what makes a resistor different?

why does a lower rated resistor get hot faster than a high rated resistor when connected to a power source? of course the resistor with less resistance draws more current. how is an LED different?

nerys says:
but not if the resistance increases HEAT which increased resistance and increases power demand.

now incandescent light bulbs have an increased resistance once the tungsten filament is heated up but there is no runaway overheating because of it because increased resistance = reduced current.

lindsley and switch62 seem to correspond to what I thought I knew about leds.
V=IR, so if V is constant, as R increases, I must decrease to satisfy the equation. Also, be careful of your terminology with resistors. The value of a resistor is in ohms (resistance.) It also has tolerance in % and power rating in watts. A resistor with a higher power rating will get hot more slowly than a similarly rated resistor with a lower power rating, simply because it's bigger, and therefore has more thermal mass and surface area to shed that heat. Some high-power resistors come in small packages, but must be heatsinked.

LED's are not resistors. As they heat up, their current increases for a fixed voltage (similar to resistance going down.) This is how most semiconductors work. Normal conductors such as metals behave the opposite, as they heat up their resistance increases. Ceramic resistors resistance usually drops a very small amount as their temperature increases. The compounds used in resistors is chosen very carefully to minimize changes in their resistance due to temperature.
Thank you
That was a perfect explanation. It corresponds Exactly to what I had in mind.
nerys please read and understand the truth in his explanation.
There is nothing really for me to read at least not as you intended.

he is saying what I said but he is using the correct terminology :-)

end result is the same

"As they heat up, their current increases for a fixed voltage"

as current increases they heat up as they heat up current increases as current increases they heat up.

We are saying the same thing I was just using the wrong words.

Point stands. if you don't properly cool and properly control the current to an LED it will get hot and as it gets hot it draws MORE CURRENT which makes it get hotter etc.. etc.. till it nukes itself.

I am sorry I used the wrong words - I got them from someone else without fully understanding them. I just understood the "process" without having the correct words to describe the process I knew.
cdousley nerys4 years ago
And it really smells when an led nukes itself too
I read somewhere that laser diodes can emit arsenic gas when they are destroyed.
good good. I understand you now. :)

Lindsley (author)  nerys4 years ago
38 degree is the temperature of our summer in some cities. The manufacturer absolute limits is 80 degree Celsius.
nerys Lindsley4 years ago
Exactly. which is why LED's are so hard to use properly if you over drive them.

once you hit 100' 38c you start to degrade the phosphor. this is why "bad" led's die and why they die faster in the summer.

NOW this is not your normal drop led its a truer emitter so it may be able to handle higher temps. but my Normal LED's I do not let them run more than 5' above ambient.
switch62 nerys4 years ago
LEDs are diodes not filament globes. Once a LED has past it's knee voltage (the voltage that it starts to conduct) it's resistance is nearly linear and constant, but very low. The reason it will consume more power is that for more voltage, more current will flow. I=V/R where R is constant But it will not keep using more current if the voltage stays the same, provided voltage or power does not exceed maximum ratings. If they do then you get thermal run away, this is where the resistance actually drops as the diode junction starts to breakdown. Then more current, heat, and resistance drops, etc until LED is destroyed.

There are 2 usual ways to run any LED.

Constant voltage and vary/limit current. This is done with lower powered LEDs where currents are from 10mA to 30mA. Usually use a resistor to limit the current and drop the voltage across the LED to 2 - 3V depending on type of LED. If a white LED needs 3V and 20mA to light at full brightness, and supply voltage is 5V. Resistor is (5V - 3V)/20mA = 100 Ohms. Resistor will dissipate 40mW of power.

Constant current and vary/limit the voltage. This is usually used for high power LEDs. Using the other method (constant voltage) with a resistor would dissipate too much power. The 10W LED needs 900mA at about 11V for full brightness. At 12V supply you would need a 1 Ohm 1Watt resistor. Since the supply may vary you might need to run at say 15V. That will be about 4.5 Ohms at 3.6 Watts. If you are using mains power a proper constant current circuit is the only practical solution. These high power LEDs are more sensitive to overload and need proper regulation with an active circuit.

A LED driver should sense both current and voltage and not exceed the 10W to protect the LED. Notice the measurments Lindsley made, they are in line with the LEDs specifications. In fact the voltage seems to be a bit low and the current slightly high. I'd say that is due to the meter/s affecting the circuit when making the measurements or the variations of components.

To get less current you would need to get an 8W driver. The 8W drivers I've found only put out 350mA and so will only give a very dim light, if at all on a 10W LED.

As for long life, from what I can find the LEDs could last 30,000 to 60,000 hours of continuous use. That's 3 to 6 years continuous or 10 to 20 years at 8 hous use per day. Even if the life is shortened by slightly incorrect driving you would still get many years of service, with huge power savings.
1-40 of 97Next »