Power LED's - simplest light with constant-current circuit

Here's a really simple and inexpensive ($1) LED driver circuit. The circuit is a "constant current source", which means that it keeps the LED brightness constant no matter what power supply you use or surrounding environmental conditions you subject the LED's to.

Or to put in another way: "this is better than using a resistor". It's more consistent, more efficient, and more flexible. It's ideal for High-power LED's especially, and can be used for any number and configuration of normal or high-power LED's with any type of power supply.

As a simple project, i've built the driver circuit and connected it to a high-power LED and a power-brick, making a plug-in light. Power LED's are now around $3, so this is a very inexpensive project with many uses, and you can easily change it to use more LED's, batteries, etc.

i've got several other power-LED instructables too, check those out for other notes & ideas

This article is brought to you by MonkeyLectric and the Monkey Light bike light.

Step 1: What you need

Circuit parts (refer to the schematic diagram)

R1: approximately 100k-ohm resistor (such as: Yageo CFR-25JB series)
R3: current set resistor - see below
Q1: small NPN transistor (such as: Fairchild 2N5088BU)
Q2: large N-channel FET (such as: Fairchild FQP50N06L)
LED: power LED (such as: Luxeon 1-watt white star LXHL-MWEC)

Other parts:

power source: I used an old "wall wart" transformer, or you could use batteries. to power a single LED anything between 4 and 6 volts with enough current will be fine. that's why this circuit is convenient! you can use a wide variety of power sources and it will always light up exactly the same.

heat sinks: here i'm building a simple light with no heatsink at all. that limits us to about 200mA LED current. for more current you need to put the LED and Q2 on a heatsink (see my notes in other power-led instructables i've done).

prototyping-boards: i didn't use a proto-board initially, but i built a second one after on a proto-board, there's some photos of that at the end if you want to use a proto-board.

---

selecting R3:

The circuit is a constant-current source, the value of R3 sets the current.

Calculations:
- LED current is set by R3, it is approximately equal to: 0.5 / R3
- R3 power: the power dissipated by the resistor is approximately: 0.25 / R3

I set the LED current to 225mA by using R3 of 2.2 ohms. R3 power is 0.1 watt, so a standard 1/4 watt resistor is fine.

---

where to get the parts:
all the parts except the LED's are available from http://www.digikey.com, you can search for the part numbers given. the LED's are from Future electronics, their pricing ($3 per LED) is far better than anyone else currently.

Comments

[gopugopinath]

i need to power up 3-5x 1W leds with below conditions......... Plz help me

INPUT = 12V DC
output = 350mA / 9-16V DC

[siddarth.ghaste]

Thank you very much......

[wahwah1111]

Dan, Im wanting to use set of 3 Cree XML T6 1000 Lumen LED Emitter with 20MM Aluminum Base with 12vdc marine battery, want maximum lumins.
This needs to be submersible and only used in water. Like to mount them on 6"x6"x.125" Alum by heat sink grease, then encase them in 2 part marine epoxy to edge of bubble on one side with back side of plate exposed to water for heat distribution. Thoughts, Ideas, Problems??

[urgthrash]

Hi dan

first off i would liek to thank you for this circuit.
i have made a monster of a aquarium LED fixture rockin 44 of your drivers powering 132 ,1 watt LEDs divided into 3 arrays running on a 12 volt 350 w P/S .

link : http://www.facebook.com/photo.php?fbid=468073513207432&l=30056293f5
Not sure if you answered this question previously, i did look, but i might have missed it .
My question is can i mix a different LED into the circuit? lets say i have 3 whites running at 3.25 volts , and i have a red led that calls for 2.5 volts at the same current ( ma), can i swap out one of the whites for a red and it not fry the lower voltage led when power is supplied? or do i need to run each color led in its own driver group ?

thanks for you time and thanks again for this circuit

[hexalm]

You can swap it out if it's rated for the same current--remember this is a constant current power source. That means changing the load (the LEDs, especially by a small amount) won't change the current going through that part of this circuit--and current's what will fry things

A too-high voltage will only do that if it can force more current through than the component can handle, for example putting a 3 volt LED on a 24 volt power source by itself will burn it out because the high voltage causes too much current to flow, hence the need for a resistor or other circuit to limit current.

Now the voltage drop across your LEDs is as follows:
3 whites:
3.25+3.25+3.25 = 9.75
2 whites, 1 red:
3.25+3.25+2.5 = 9

So the total change in load voltage (voltage drop across all three LEDs) isn't very big--I suspect this circuit will be able to compensate no problem.

Someone stop me if I'm wrong about any of thsi in general, or with regards to this circuit, which I haven't used--my circuit skills are a bit rusty.
(And I am assuming that the circuit is able to handle a 9V+ load already.)

[MichaelRGroves]

Hi: My application is for lighting buildings on my HO trainset using LED's and fiber optics to deliver the light. My question is, why use a Fairchild FQP50N06L which allows 37Amps of current? Most LED's run in the 0.2 to 0.7A range - is it that the Fairchild FQP50N06L is a good bye, or is there a FET that is less expensive with lower current rating that would be just as suitable?
Many thanks

[bjcryss]

Hoping someone is kind enough to help...I've built this LED driver circuit. I have a 4W led, with 3.6V dropout and 1000mA max current, which i would like to use with 4xAA rechargeable batteries in series. I need the driver because from fully charged at 5.6V, the voltage of those 4 batteries drops to 4.4V when almost empty. My problem with your driver is that for me, it works well but only when the current is limited with the "sense" resistor at aprox 500mA and even if i use 6xAA batteries the current does not change. This is great. But if i try the same thing with a resistor to limit the current to 1000mA, when adding even as little as 0.2V extra, the current increases. Is this normal?? With 5xAA batteries, the current goes up to 1400mA, with the same "sensing" resistor (2W power resistor and it doesn't heat up at all @ 1000mA). I haven't tried with more voltage because i'm afraid to burn-out the LED. I couldn't find the transistor and MOSFET you used, but i got a IRFZ44N, and a BC237. Please tell me if you see something that i'm doing wrong...i just can't figure it out

[arij]

Hello. I am new user.
Is this circuit with different resistor, suitable for 5W ir power led?

Forward voltage: 1.5-1.7V Forward current: 2400mA Wave length: 940nm
Power source 12-14V dc

Thank you
Ari

[bjcryss]

Hello! Help with this please... if i go to the store and they don't have the Fairchild 2N5088BU or the Fairchild FQP50N06L, what replacement should i ask for? Better than that, what characteristics should i ask for? Thanks alot!

[ohnoezitasploded]

Does this driver work with PWM?

Thanks for taking the time to put this together!

[naeemfarrukh]

Hi Den,

Great work. i am a mechanical engineer by profession and not good in electronics.
so that i need your help to make the constant current drivers for the following 2 different power LED s. please check and confirm the calculations for R3

1. 10 watts with 850mA/10Volts. R3= .5/.850=.58 ohms in 1watt
2. 7 watts with 700mA/7volts. R3= .5/.700=.71 ohms in 1watt.

and most important of all i have added a 10Volt Zanier Diode in the circuit (as per diagram #3) and with heat sinks to LED and FET (50N06) want to use it with 36Volt DC power source,
will it work ? please guide.

Thanks.

[surf4point0]

This circuit is awesome! Does anyone know if it will work starting at about 30 volts? I'm trying to get a constant current of 1/4 milliamp through a resistance that gradually decreases, so we start out with 30v and probably end with 5v or so.

[09khanyu]

Somebody on this page asked for the components for the 3 watt led. I am also interested in that. Could you please post the values of those components in the comments section in this step.

[Bednarz]

Quick question...I know that I=V/R but I'm curious how you arrived at LED Current=0.5/R3. Is 0.5V the remaining voltage after the drops through the LEDs and Q2?

[TheNuclearWatermelon]

Can I use my IRF510A MOSFET for Q2, and an MJE3055T NPN transistor for Q1? Also, what value resistor should be used to get approx. 1.2 amps on the LEDs?

[rufusthedog]

The FET and Transistor should work and figuring out the resistor is easy.

The transistor will start to turn the FET off once the base voltage starts to get above 0.7V so the resistor you would need would = .7V/1.2A = ~0.58 ohms, the resistor need to be above 1 Watt though to handle the current.

Rufus

[quigath]

I was also wondering about using something like an IRF510PBF, but I don't know enough about this type of circuit to decide. Any help here?

[applejuice1998]

Good job! I love this circuit!
I am currently using it for to charge some Lithium Ion cells that I pulled from a laptop battery.
I have it currently setup for 300milliamps, it appears to be charging the batteries. This is also powered from the +5V on a USB plug, I have a teensy 2 board on a bread board witch I built that on, but just replaced the LED(s) with a lithium cell.
If you decide to do that, keep an eye on the cell! Constantly check the temperature, only draw back, is it will not finish the 2nd stage of charging, witch is Constant Voltage, but that's fine with me.
I take no responsibilities what you do with the battery.

[JohnPS]

Hello,
I need the Low Cost Constant Current Source for Power LED's to supply 3.4 to 3.8 v to a LED circuit. As shown in the pencil drawing on this site, this is a series circuit which concerns me because I plan on running about 20 .5w led's which will be a light source for a small painting booth. What will need to modified for 20 led's?
Thank you,
John

[hanlin_y]

R3=0.25/I
Are you using batteries or a wall adapter? Your LED's would be using 10 watts

If all 20 LED's are in series, you will need at least 68 to 76 volts. Because this method has the least current (same as the mA rating of your LED's), this is the most efficient method and produces the least heat but there is a risk of electrocution. This is not practical as we don't see batteries in this voltage range.

Instead, you may prefer something like 4 strings of 5 LED's so your voltage source is at least 17 to 19 volts. A 24 V battery works. The resistor should be calculated as for four times the current rating.

The least efficient method would be connecting all 20 LED's in parallel. The resistor and transistor could get very hot.

You can also build more than one current regulator

[hanlin_y]

I apologize. It may NOT be a good idea to connect LED's in parallel in this circuit similarly to a resistor circuit. I blew my 20 mA LED's this way and they were hot. With multimeter readings, the mA ratings and the LED voltage drop increased slightly with higher voltage batteries. Looks like one circuit per string is the only way. Try using cheaper LED's first. 1W+ LED's are expensive

In an ideal world, you can use the parallel method but the resistance can vary in LED's so those with less resistance are overpowered and those with higher resistance are underpowered. This means some wear our faster than others. When one LED dies there is a chain reaction of other LED's dying.

I spent the whole day figuring out this circuit today because the pins of the FET are not in the same order as ordinary transistors. Check the specs.

[ashayana]

Hi Dan

I am thinking of a modification to your circuit to run on AC supply. Wanna build AC run light bulbs for my home using Power LED's and a good reliable driver is essential. The circuit which you have runs best if the voltage of the LED string matches with the input voltage else most of it is dissipated in Q2.

Can convert 230V AC to DC using a Bridge, have protection devices like MOV, spike resistor & fuse in the AC input end and limit the voltage using Zener diodes to 12V. which can be then fed to to the LED string. Does the current setting still work if these modifcation are done ?

Insights on this would be very helpful !

Thanks

[zami]

Thnx for the gr8 article.It really helps.I have a question.
I m making a 90W Led panel,Using 1W led(rating 350mA,3.2V)
In which design shud i make it??

And for Led lights panel,Constant current source is better or Constant Voltage source??????


Please Email me the design on [e-mail]

[MattATC]

Hey Dan. I have a 5w led I am needing to use. Using this set up what resistors will I need to change out and at what values? I will need to be running 750mA, and the battery will be a 9 volt. Thanks for your help!

[yohanc777]

Hello Dan.I like your circuit its quite simple and neat.However im using a 3 watt led from luxeon.which needs 700 mA. What do i do to Q2 to power my led to 700 mA instaed of 200 mA. Kindly let me know [e-mail]

[slithytove2]

I constantly reference your High Power LED schematics, usually this one. I was wondering if you (or somebody else) could check my math:

I am using http://www.fairchildsemi.com/ds/FQ/FQP50N06.pdf which has a maximum junction temp of 175C and a Juction-to-Ambient Thermal Resistance of 62.5 C/W. Assuming a maximum temperature rise of 100C, this means the NFET can handle (dissipate?) 1.6 watts without a heatsink?

My particular LED is running with a 3.31 V drop at 339 ma, thereby dissipating (?) 1.12 watts. Are this and the above 1.6 watts representing the same value (since there is nothing else on the circuit)? As long as my LED stays below 1.6 watts, I should be fine without a heatsink?

Or am I completely off my rocker? If so, can anybody point me to a site with more information in this area? My searches have been mostly fruitless..

[dan]

power on the FET is based on the voltage across it, which is the difference between the LED voltage and the power supply voltage. ie, not the same as the LED power.

[slithytove2]

Oh. So I did the right math, just in the wrong place? To check the power on my FET, just measure the voltage across it, then multiply by the current? Is the first part, about the 1.6 watt max w/o heatsink correct?

[leaualorin]

Hello Dan!
I've been "a bit of a fool" in buying some A123Systems LiFePo4 cells (26650, 3.3v, 2300mAh) and "forgot" about the recharging part...
I tried some schematics but didn't worked out as I expected!
Could I use your schematic to recharge these cells(I've got 5 of them...) considering the manufacturer recomends the CCCV procedure and a minimum of 3A for recharging them?
They could be charged at up to 10 Amps for fast charging but the current must be lowered in the final part of the charging for topping off the charge in the cell!
So the voltage could be fixed to 3.6v but I have to be able to lower the current in the final part of the charging process!
I hope you can help me and thank you in advance!
Gabriel!

[krawczuk]

how do you work out :

LED current is set by R3, it is approximately equal to: 0.5 / R3
.5 /r3

how can you divide it by r3 when your trying to find out the value of r3 by this equation ?

[aisvo]

in other word, R3 = 0.5 / Current
If you want 200mA, then your R3 will be 0.5/0.200 = 2.5 Ohm

[bobelon]

It's a good question, I can't figger it out either but it looks like the author is no longer responging.

[tdawber-mandeno]

to work out resistiance
resistiance=(suplie voltage-led voltage)/0.02(a good standed number for standed white leds if using high power you divide by recomended mA)
r=(V2-V1)/I
(9v-3v)/0.02=300ohm
or
(5.6v-3v)/0.03=86and 2/3 ohms (86.666')

[skushwaha1]

i m making a torch and i need 6v input and out put should be 3.5v with 700mA can u plz help me my id is [e-mail] hope for reply

[ngwanhoe]

Hi,

I have read your article on Leds and thought maybe you can help me.

I am trying to make a high power Led torch with small overall space, simple and cost effective.

Leds
4 x K2 3W Leds (in put voltage 4.5 - 5, current I think is around 350 -700ma each not so sure)

Batteries
2 X AAA Alkeline or rechargeble batteries ( Normal 3v, rechargeble 2.4v)
or
3 x AAA (Normal 4.5v, rechageble 3.6)
or
4 x AAA (Normal 6v, rechargeble 4.8)

Driver / Step up circuit simple plan
Simple and can get in any electronic shop. Voltage step up and current constant / boost or
anyway that can drive the Leds efficiently.

Please email me. Thanks a lot

[e-mail]

[richie1479]

Is a good design

[akimirza]

it is best...
thanks

[budabob07]

Very clever, certainly easier than using a microcontroller to monitor the current and adjust the PWM accordingly.

[snowluck2345]

is there an easy way to adjust brightness in this circuit? I was thinking of a potentiometer as a replacement for one of the resistors. Or could you use a pwn circuit?

[Spuzzum]

I used a 100K 25-turn pot for R3, and needed to tie the free leg to the center leg.. otherwise the current climbed on it's own. This way, it's a "true" variable resistor.

Works great. From 100mA or so, to well over 1500mA in my tests. A 10ohm 25-turn might be better though.. tighter calibration.

And if you want to tie in a PWM signal, follow the author's other post...
http://www.instructables.com/id/Circuits-for-using-High-Power-LED-s/

[snowluck2345]

could youe use a 555 timer to switch a transistor to switch this circuit on and off, with a potentiometer on the 555 timer to control the speed of switching?

[scottinnh]

Probably. Don't ask me for details though.

Here is a constant current 555 driven led driver (see part, Inari):
http://web.jfet.org/ignignokt/

[Spuzzum]

This is great bro! I've already modified an atx psu to supply a panel I'm making, and was looking at the ncp3066 circuit for a constant current source. But unless I use 24v with it, it's only 75-79% efficient. Powering off the psu with just resistors is already 87% efficiency. So I scrapped that idea, but still need a constant current regulator. I then looked at the lm317, but that wastes 1.25v just to make it run. This circuit uses what.. .5v? Even better! Now if it can handle 1.4 amps I'm set. Only need 1000mA, but want some wiggle room.

Anybody know the maximum current this can handle? At least 1.5 amps?

[lasermaster3531]

dude read the instructable. all the info you need is right there.same for Jo
-burg.

[lasermaster3531]

sorry I meant jo_burg.

[jo_burg]

Hi Dan!

Awesome! Only one slight tiny problem, I'm a idiot!

I don't understand a thing about circuits. Circuits for Dummies is way past my pay-grade. But can monkey-see-monkey-do better than most engineers.

Tell me to get lost but I'll try anyway.
What parts do I modify to run a 1W LED (up to 3 LED's) with a forward current 350mA & 3.2Volts?
...and the same values, but for a 4W LED?

Yeah, sorry.

[MexicoDoug]

Can R3 be replaced with a 2.2K ohm resistor for a very, very low current application to give 0.227 mA constant current or is that not within the operating parameters of the transistors? If it is ok, except for the power disipated in the LED and resistor, will the rest of the circuit contribute to significant inefficiency (Is there a way to estimate the mA drain of the circuit at this low current not including the LED and R3)

I want to use it to make a constant current 3V battery tester and count the hours it takes to discharge the batteries.

Thanks!

[dan]

i think that will work. just check it with a multimeter because the accuracy might be +/- 10%.

[ovven]

what value transistors did you use as i am building a very similar set up but cant find a decent place to get the same model transistor.

Cheers

[unaffiliatedperson]

im tryin to run two leds in series they work at 3.7 V and i need 1A current. could this circuit work? i was using two 18650 batteries in series but they exploded on me over the weekend.

[ihwild]

Future Electronics changed their web address. It's now http://www.futureelectronics.com/

[jufreese]

I have done the BLOB method on a circuit before....turned out the regulator got a little too hot and melted all the hot glue off of it. Keep that in mind.

[frollard]

hence the use of silicone glue, not 'hot' glue :D That could go poorly - as it did when I "repaired" my landlord's laptop power supply. It remelted and stuck to the floor. Doesn't normally get hot, but if you leave it on carpet with a blanket over it, it does get hot :S

[seadweller43]

I was hoping to get a little help learning how to utilize this instructable to connect two separate series of LED's to an Arduino. The first set of 1w LED's will be 8 white and the second set will be 4 blue. I'd like them to be connected to the Arduino for a automated dimming feature. I want to create a sunrise/ sunset effect for an aquarium.

Thanks in advance.

[technodude92]

almost forgot the circuit, sorry for the double post

[technodude92]

I'm assuming you will be using PWM to dim the LED's here but this circuit while pertinent to your project is not exactly what you need for dimming. For that you will need to switch power to this circuit on and off quickly. The arduino cannot source the ammount of current necessary to run these high power LED's and so an extra transistor is required. This transistor should be chosen so that it can handle switching the current of whatever power LED's you are using. R2, is a current limiting resistor, chose it's value so as to keep the current draw from the arduino pin within reasonable limits. using Ohms law if your arduino can source 50ma per pin, then at 5 volts you would need a 100ohm resistor (5V/0.05A=100Ohms).
Hope this helped

[frollard]

I've been searching for this solution for a long time! Kudos!

Is there any way to introduce q3 before q2 in the circuit to make q2 act as both current limit AND pwm digital switch?

[seadweller43]

thanks for the reply. I wish I had just half of your knowledge of electronics. My damn bachelor's degree in business management won't help me with for stuff like this. Do you have any good book suggestions or website that will help me learn this stuff easier?

[scottinnh]

Get the Make:Electronics book...

[EmmettO]

 Can this circuit be used to drive multiple LEDs? If so, would they be wired in parallel or series?

[vanmankline]

I'm not an expert on this topic, but I'm learning.... From what I know:

You would need to switch out the resistors and ensure the other components can handle the current you require. The way these work you will get better results if you wire the LEDs in series.

[EmmettO]

Thanks, I eventually figured it out. http://www.instructables.com/id/USB-Mega-Flashlight/

[LEDismylife]

Is it possible to make a series 0f 5x 3 watt LED with just one LED driver? If it is, may i know what type of resistors and regulators i should need...

Thanks.

[conntaxman]

I know that i made one using three 3 watt leds in series, and used the cir from the instructables.I think I changed the chip whith a higher amp, rate,But same styl chip ,and they worked fine. Im now buying some 20 watt leds, have one and it sure is bright.looks like an 85 watt inc.bulb
John

[conntaxman]

For making a power supply is their any way besides using a stepdown transformer will say using a 120vac down to 18vac? Is it possible to use like 3 MC7815CTG [output 15vdc at 2.2Amp] or something different like a LM338K transistor. Im trying to make a power supply for a new 20 watt led with 1000LM at 13-15 vdc with 1400 to 1800 ma.
I ran it so far with a 12 dc wall wart at 1000ma. and it was very bright. And that was down 3 volts. I want to try to make a power supply that will fit in a recessed light housing.
tks
John.

[kingofrandom92]

i am so happy there are people like you on instructables that use the normal symbols for building a circuit draw up like this.

[.Unknown.]

Normal?

[rogueleadr]

 agreed

[MinerJay]

Hi all, I am wanting to power a waterproof LED rope for illuminating a sign. The rope is 6 meters long and has a current draw of 1800mA. The supply will be a 12 V rechargeable battery so assuming that the input voltage is 12 V and the current needed by the load is 1800mA then if I read correctly your selection for R3 would be V divided by I = R therefore 12V/1.8A=6.66 ohms assuming the resistance needs to be 6.6 Ohms and the current rating of the resistor would be 2 amps (24W) this sounds like a very expensive and costly resistor, surely I am wrong please someone help me I am at a loss.

[cdousley]

if im right you just need the right wattage resistor but im not sure.

[cdousley]

cool

[indralho]

Can anybody teach me ? I just made this circuit with 7.2 volt input voltage but it was so dim. I am using the same component.

[Subterranean]

If I was wanting to just draw a constant current from the power supply (batteries in this case) could I replace the LEDs with a dummy load, say a power resistor, and have it still work fine?

Is there a better, but still simple circuit for doing this?

Thanks for any help.

[TS84]

Can any one tell me to replace , Q1: small NPN transistor (such as: Fairchild 2N5088BU) and Q2: large N-channel FET (such as: Fairchild FQP50N06L) its really hard to find both of them in my country.and i want to drive luxeon 3W white LED.please help me.....

[EmmettO]

 This is a very late reply but I've just been working on this circuit. I don't think it matters the exact parts. All you need is a NPN transistor for Q1 and a FET for Q2. I pulled mine out of an old TV circuit board and everything appears to work correctly. I'm not even using a 100 K ohm resistor, I could only find a 56 K ohm in my stash.

Now I'm just trying to figure out how to modify the circuit. If the voltage is constant at 6v then you multiply by .5 of R3 for your current but I don't know what to do if using a different voltage.

[crazyromanian]

have u tried mouser.com, they're based in the US, I think they ship to most any country and they usually have better prices than say Digikey for the same product (the only down side is mouser does not have as many "obsolete" items).

[doctaq]

hey there
so instead of the transistor you reccomended i used
http://www.mouser.com/Search/ProductDetail.aspx?R=BC184L_D27Zvirtualkey51210000virtualkey512-BC184L_D27Z 
i put it together like in your instructions but nothing happens when i hook it up, does the collector current matter? im thinking that maybe somehow i might need a different R1? i used the same fet and used a .75 ohm current set and 100k ohm R1, 12v power supply and 1 cree xr-e, i also tried it with 3 batteries for 4.5ish volts, neither did anything. 
thanks

[1loloo]

i have eight 10W, 7-8V leds. What power supply i need and what change in this circuit i have to make in order to leds give light safely?

[KDS4444]

Okay, I am sorta new to electronics and schematic terminology, but I got me a couple of Luxeon LEDs and I want to build a driver. I understand the function of a resistor, transistor, etc., and I understand concepts like Ohms, Amps, Watts, etc. at a layman's level, BUT I have been reading and re-reading (and RE-reading) over this list of components and I get stumped at your description of "Selecting R3". Though I claim some rudimentary electronics understanding, the sentence, "R3 power is 0.1 watt, so a standard 1/4 watt resistor is fine", as simple and clear as it must seem to most, just ends up becoming jello in my head! Aren't resistors rated in Ohms, not watts? When you say "R3 power" what do you "mean"? (Volts? amps???) And what is a "standard" ¼ watt resistor?? Can you imagine I am a 5th grader and walk me through not just the numbers (0.1 watts, etc.) but through the range of "acceptable" numbers/ ratings (i.e., can you gimme the "too much", "too little", "what happens if you don't do it this way" stuff!)?. I suspect that terms are being used somewhat casually in this step and as a novice the VERY SPECIFIC meanings of terms to me are VERY important until I can begin to ignore them! THANK YOU!!!

[Sharanga]

             Dan, thanks for the circuit idea. I needed something like this.

 

KDS4444: regarding your question on rating resistors in ohms and watts, though this is a bit late, I thought I'd still respond, since the question is still relevant and worth knowing the answer to. Will try to keep it simple (which naturally will make it a bit longer :-)

 

Resistors are indeed measured in ohms, as you have noted. But the “Watts” (the wattage) of the resistor is also important, because this tells you how much power (heating) the resistor can handle without getting destroyed. Ideally you should know both these things when selecting components for your circuit.

 

When you apply a voltage to the resistor, a current passes through it, and it gets heated up to some extent. This heating is just wasted power, which we measure in Watts by multiplying the voltage and the current.

 

The same resistor can be available in the market with different “wattages”; for example, your R3 = 2.2 ohms could be available as 0.25 Watts, 0.5 Watts, 1 Watt, and so on. Resistors are manufactured in these standard wattages; you cannot find some odd figure like 0.1 Watt, 0.4 Watt, 0.85 Watt, and so on in the market.

The higher the Watts (wattage), the bigger the physical size of the resistor (it also gets more expensive). It therefore makes sense to select the resistor with the lowest wattage that can do the job safely, based on your calculations of currents and voltages.  (I wanted to upload an image, but unfortunately the page does not seem to display uploaded images though it offers the option.)  Your circuit will also be compact in size without getting unnecessarily bulky. For example, you wouldn’t need a 1 Watt resistor for a penlight LED circuit, though you could certainly use it.

 

So, Dan worked out that R3 can be 2.2 ohms. Then, he figured out that it does not need to handle more than 0.1 Watt of heating. That is what he calls R3 power. Now, you are not going to get a resistor in the market with 0.1 Watt capacity. The closest is 0.25 Watts. So, this is what you select. In any case, it is usually better to select a resistor with wattage on the higher side, just to keep things cool and safe.
 

This is what “Selecting R3” means.We have selected a resistor of 2.2 ohms, 0.25 Watts

 

To understand things like whether R3 = 2.2 ohms is too much, too little, and so on, you will really need to know a little more about how components like transistors work. Hope this is helpful.

 

(By the way, Dan, is that R3 = 2.2 ohms correct? The picture is a little unclear, but I get the feeling that I can see 3 red bands on R3 there.)

 

 

[andrewortman]

Hi there, maybe I can help a bit.

If you apply a voltage across the resistor, there will be current flowing through that resistor. That current can be modeled by i = v/r (from ohm's law)

The power rating on resistors is the maximum allowed power that can go through that resistor. Power is the product of voltage AND current (p = v*i)

A standard "1/4" watt resistor can take up to 0.25 watts, which is really the most common type of resistor you can buy at hobby shops like radioshack. You can calculate power through a resistor using a handful of equations, all going back to p=vi and i=v/r
if you replace i in p=vi, you get p=v*v/r = (v^{2)/r - where v = voltage ACROSS the resistor and R is the resistance in ohms.}

Hope this helps!

[andrewortman]

eh, the formatting genies got me there at the end.

it should say power = voltage square divided by resistance

[KDS4444]

Now THAT is helpful! Thank you!

[conntaxman]

Dose anyone have a cir. that would run 7  3watt leds together? Forward voltage of each is 3.5 to 4vdc and the current is 700ma each. There are many cir to run 1 3watt led at a time useing LM317T.
tks to all
Johnny

[KT Gadget]

 Depending what power supply you are going to use, your best bet would be to give each LED its own driver, or if saying the power supply is 12V, make 2 and put a high watt resistor on the second because it will only have 3 LEDs in your case on that line and 4 LEDs on the other without a resistor.

[cryptopsii]

Won't work because a darlington is a sort of dual
bipolar transistor and you need a FET (Field effect transistor)
A bipolar transistor drive current if current flow throught
it's base. A FET drive current if Voltage is applied to it's
gate  (it remplace base). And  a Mosfet is different that a fet
and won't work either.

[Zando]

Umm, a MOSFET is a FET, FET stands for Field-effect transitor, MOSFET stands for metal–oxide–semiconductor field-effect transistor, its formed with a thin layer of a glass-like substance between the gate and the drain/source, as apposed to a JFET, which stands for Junction field-effect transistor, which is formed with a revesed biased p-n junction between the gate and drain/source.  JFETs tend to have higher leakage currents, around 10^ -9 amps, while MOSFETs leak only at around 10 ^ -14 amps, MOSFETs tend to stand much higher currents and be cheaper and smaller to make, while JFETS can work at much higher voltages and switch a bit faster.  Your processor is made out of millions of MOSFETs.  Both can work the same if there specks cover it, but JFETs can rarely handle much current

[om-3alawi]

Hello there,

I'm new here and don't know much but project is a line of site of optical link in a short range, and I'm using white Lxeon 3 star. For the 1st part, is the curcuit parts going to change if i used LHXL-LW3C instead of  Luxeon 1-watt white star LXHL-MWEC??????

i hope someone can help
thanks

[conntaxman]

Could I use this cir.to run three 3watt leds ,forward voltage 3.5 to 4vdc and 700ma draw.And what would be the best battary voltage to use.6 or 12 vdc.the link is above.
http://www.instructables.com/id/Power-LED_s---simplest-light-with-constant-current/?&sort=NEWEST&limit=50&offset=100
Also what is the most of these leds that I could run in seires?
tks
John

[Highjump44]

 my friend had conductive silicone and did that  pretty funny

[skantu]

Hi dude,

im using a dynamo 6v 3w for just 1 high power led 1W, what resistenvce do you recomend? im reall stupid in numbers.

thanks.

[jason 77]

 Hey Dan.... First I want to say thanks for all the great stuff you have on here!

I have used this circuit 2 times already for relatively low power leds "under 400mA"  I now built this circuit to out 1A, and then decided to use a PWM to control the brightness of the led. I built the constant current circuit on a proto board and it seemed to work fine with the led... however when I take out the 100k ohm resistor "R1" the led dims but does not go out?  Do you have any Idea what is wrong with my circuit?  I am using 2 1 ohm resistors in parallel to get .5 ohms which according to my calculations should give me about 1A of power.  

[Meigus]

 Thank you so much for the circuit!

I made the circuit using a potentiometer (variable resistor)+constant resistor for r3, and using 8.5V (crappy 9volt battery). I tested the circuit with LEDs, but I plan to use it for other stuff.

You should make a note that if you change the voltage or components, you have to change your equation. You put in step one that the equation is .5 /R3 = amperage. I had to change the equation to 6/R3 = A. I had different components besides the battery. (tip29A NPN transistor and BTS117 N-mode/channel MOSFET. I got both from ebay cheap for 5 packs!)

I got really confused at first, cause my 843 ohm pot. was putting through about 7 mA. Just doing backtracking i realized the equation was different for me.

Hope this helps anyone who tries to make this awesome, cheap, effective circuit.

P.S. Because my values were different, I had to add the normal resistor. With the 2200 ohm resistor and the 843 pot, I get a sweet spot of 1.99 mA averaging measurements from 4 different LEDs tested individually.

I also suggest anyone who makes this circuit to make an average amperage if you use it for unknown LEDs. The variability for my four LEDs was 1.94mA to 2.06, so I guess a nice +/- 3% change. I used GOLD stripe resistors, but the potentiometer was 1 kilohm but actually topped out at something like 843 ohms, so that probably would be the least accurate component in my circuit.

[luvotter]

I recommend you do this before BLOB-IT in step 8. Thanks for this cool compact circuit.

[killjade]

Hey I need some help understanding how this works. Being an Auto Tech apprentice, I have a very good grasp on how electricity works. All the laws etc... In fact I get most of the electrical car problems that come into the shop. But I can't figure out how this circuit works because it's conflicting with my knowledge. I just got into electronics hobbying because I want to build some cool things for my car. Here are my issues. Lets say the the supply voltage is12v. We will be using a LED that runs in 250amps and has a forward voltage of 3.5 for example. That means, using this circuit, we'd have to drop 8.5v before the led. So why is R3 ahead of it? Shouldn't it be before the led? I'm still trying to understand the operation of Q2 and Q1 so maybe that's my issue. This setup makes it seem like 12v will go through the led first when it should only be 3.5. Help :(

[samurai1200]

You have to think of voltage and current in complete loops. It does not matter whether the resistor is before or after the LEDs, the effect is the same. When Q2 is "on" it is as if there is another resistor connecting the two ends of the transistor... this is what is called a mosfet operating in its "linear region" -- it does just what a resistor does: limits current and throws the rest of the voltage away as heat.

[crazyromanian]

yo dan, does it really matter if I use the 2n5088 and FQP50N06L? I have a bunch of 2n3904 and IRF640PFB. Is the main idea here just changing the values of the resistors so they trigger the transistors correctly?

[crazyromanian]

k actually nevermind, i finally realized how this ALL works and it is amazing once you understand it. I made my driver with a D965 and an IRF640 and it works great.

[jmengel]

There is a discrepancy between the parts list and the schematic and rest of instructable. In the parts list, the NPN is listed as Q2, while in the schematic and text, the FET is shown as Q2. Correct labeling is that Q1 is the small NPN for regulation and Q2 the power n-channel FET. Might trip up some with less electronics background. Otherwise, great work.

[dan]

oops i fixed it.

[gchagui]

did it, it burned my 1 watt led with 12 volts......... and with 5 volts, 20 secs, in it burns...... and, at 5 volts it draws 1.8 amps!!!!

[shawn-schro]

Is there a way to make this circuit dimmable? Could R3 be replaced with a variable resistor to accomplish the task, or is it more complex than that?

[lolzertank]

A variable resistor would work in theory, but very low resistance and high power variable resistors are extremely expensive. You could add a PNP transistor between R1 and Ground and a gate resistor which would give you a PWM input. Then you could use a 555 or two and a variable resistor to make a PWM waveform generator.

[rdunung]

Would you be kind enough to draw a quick schematic of what you recommend for dimming. Also you mentioned that you could use one or two 555 timers. Is there an advantage of using two over one. Thank you

[shawn-schro]

lots of schematics available online for PWM circuits using a single 555. It seems that you only need 2 x 555's if you are adjusting frequency AND duty cycle.
Try this one...
http://www.reuk.co.uk/LED-Dimmer-Circuit.htm
(This circuit does not provide constant current)
The only question I have about this is the spec on the PNP transistor and gate resistor, and how one would connect a PWM circuit such as the one I've mentioned to the constant current source in this thread

[starwarts]

Hi Dan. Do you think I can use TIP122 NPN darlington pair instead of the MOSFET? I dont have them available locally while TIPs are lying handy. What should I be concerned about if I use TIP122?

[sign-up]

While I agree that your circuit is more accurate than using a resistor - IF the input voltage is varying greatly, that is. But its efficiency is EXACTLY the same as using a resistor.

Whether it's a MOS or a LM317 or a 25cent resistor, they are there to drop the extra voltage as heat - so the power wasted is exactly the same. Your way costs more, is all.

It would be MUCH more eco-friendly for you to build a constant-current buck circuit that does not lose any more than 10% of the power as heat.

[jmengel]

Um, actually no. If you read the above and understand the circuit, the power dissipated in the set resistor R3 is only 0.11W at 225mA. With power LEDs dropping 3.3V Vf at 225mA the 3 LEDs drawn in series would be dissipating 0.75W each for a total of 2.25W. 0.11W out of 2.25W is about 5% heat loss in the resistor. The dissipation in the MOSFET will be neglible if it is a high power low R-on type with a on resistance <100mOhms. The NPN meters the gate voltage on the MOSFET and will not dissipate much of anything with R1 at 100k. For comparison, if you took the half assed resistor to set current approach and you had a 12V source, with 3.3Vf then you would have 9.9V across the LEDs and 2.1V across the resistor. To get 225mA you would set the resistor to 9.3 Ohms. With 225mA going through the 9.3 Ohms, you have a power dissipation of 0.47W which is 21% of the input power dissipated as heat. Not to mention the other problems with a fixed resistor to set current. So which is more efficient?

[lolzertank]

Um, actually yes. Like the author said... "Q2 acts as a variable resistor." The MOSFET is in its LINEAR region. The whole purpose of this circuit is to adjust the MOSFET's RDS(on) so that exactly the right amount of current is flowing.

[jmengel]

Hmmm, you are correct, thanks for setting me straight. Looking back at my post I can't figure out what my thinking was. I blame it on the morning after the Fourth.

[dagenius]

lets see you make one...

[vargose]

I'll be running between 1A and 2A. Is there another MOSFET I can use that won't get as hot, wouldn't need a heat sink?

[lolzertank]

No matter what, the MOSFET will get to the same temperature since even if it has a lower on resistance, the whole purpose of this circuit is to use the MOSFET as a variable resistor, negating any benefits of a MOSFET with a lower on resistance.

[jack.h]

I want to use this circuit to power leds running off a 12v battery charged by a solar pannel. Would this curcuit reduce the draw from the battery or would it just change all the extra current into heat?

[kawfeegod]

Dan, Very cool. Just what i needed. Could you help with creating a 16 LED bulb? I am using 1watt LEDS and figure that I would need four of the regulators that you build here, but would need to know what to set the current set resistor at. I will be using a 12v power source and mounting the LEDs and others to a heat sink(s). Any help would be great.

[ReCreate]

AllThat To Power an LED?
A Simple 500Ohm Resistor could have replaced The transistor,regulator thingy,and the Useless resistor in there

[dagenius]

But then it would not have been a "constant-current" circuit.

[ReCreate]

Constant current?

[dagenius]

um, look in the title. using a resistor would defeat the whole point of making a "constant current" led circuit.

[ReCreate]

Ok...

[lightime]

I believe this is more efficient than using an lm317 which loses about 3volts I think? tHow much voltage does this setup lose? Thanks

[lightime]

What is the highest amp you can run through this? Thanks!

[ttr232]

i was curious if u could tell me how i could hook up 8 10w leds? im new to electronics lol sorry

[dagenius]

wow. 10 watt leds would be like...... crazy bright!

[fokusco]

... I have a single 12 watt led... :) 900 lumen output... :) hot... very hot...

[dagenius]

:O wwwwhhhhooooaaaahhhh!!!!

[fokusco]

.. well.... i had one... burnt it up a couple days ago... It was a P7 look them up... pretty sweet

[d-lite]

.... and crazy HOT! 80w is enough to bring a quart of water to the boil!

[smelnyk]

I want to build an LED circuit (using blue LED's) to add some accent lighting to my deck that I am building. I would like to run the LED's off a photo cell power supply to some rechargeable batteries. The hope is that the photo cell will charge the batteries in the sun all day, then a photo switch placed somewhere in the circuit will turn on the LED's when it gets dark. Can this be done or am I nuts?

[arduinoe]

if your using a constant current source anyway (like a lab power supply) can you just put a reistor in series and run them of the power supply?

[mellink]

the point is that this varies with whatever led you put on right?

[Wesley666]

I not sure if it was you but someone already did this.

[jimbowen]

I am currently Building A light with 4 x 20w leds and 4 x 10w ,.. Those 20w’s are ace !!!! The easiest way to run them that I have found is to use Buck–boost converter

[148wmcquiston]

On a related note, does anyone know how to make a current regulator circuit using Pulse Width Modulation?

[johnpombrio]

Yep, NO REASON AT ALL to make this anymore. a Constant current source for use in flashlight making are dirt cheap and work great. I have 20 of them that have 5 functions, work off of 2-3 batteries and cost all of Three dollars a piece:
http://www.kaidomain.com/WEBUI/ProductDetail.aspx?TranID=2958

[mathman47]

Your link doesn't work. They must have changed part numbers. Delete all the way to .com and then search on the LEDs.

[johnpombrio]

http://www.kaidomain.com/

right you are. Kaidomain has revamped their whole site. Pull down the Sports/Hobbies and find Flashlights DIY for their listings of flashlight parts. Or just buy some of their great flashlights premade.

[johnpombrio]

Sorry, link no longer works. I have been using
http://www.dealextreme.com/
They have several very nice things for LEDS. One is a 6 Watt 120 volt constant current power supply. Can run 2 3W LEDs off of it:
http://www.dealextreme.com/details.dx/sku.10852
Check in the flashlight DIY for all of their LED goodies.
NOte that it takes a MONTH to arrive. Sometimes less but do not plan on it.

[gardnerpomper]

This link doesn't work and I have not been able to locate the parts you reference. Can you update the link?

[johnpombrio]

Done. Thanks for the heads up. I have been using the stuff from DealExtreme now for a year or so and they a great source of LED goodies as well as stuff for the computer and kids.

[dark sponge]

I have a 3 watt power LED that when run at its reccomended voltage (3.5 v), it draws a little less than the maximum current of 800ma. Do I need any sort of current limiting at all?

[samphantom]

Hello every one. Hello Dan, I like your circuit. I have a Led @1W from digikey part # 475-2587-1-ND. I implemented a driver with an LM317 an couple resistors, the LM317 has heat sink, the problem is the LED needs 1.4A to give the tolerance brightness with no problem @3.5V. This LED needs also a heat sink to avoid overburn, I put a 0.27R resistor @1/2W with no problem but the leds still need a little fan to avoid over heating. Could you help me to make a circuit for this specific LED avoiding heatsinks? I armed the circuit because I saw it on internet but yours seems so simple. Thank you beforehand.

[mattthegamer463]

Voltage x Current = Power
3.5V x 1.4A = 4.9W

Are you actually putting 1.4A through your LED? That can't be a 1W LED, it must be a 5W. Reading the Digi-key page I think it is a 5W LED.

The LM317 puts out 5V? If so, you should need a 1.2 ohm 3W capable resistor, according to http://led.linear1.org/1led.wiz

[stormer1809]

I think you may be making a mistake reading the LED specifications. LED voltage x LED current = 3.5 V x 1.4 A = 4.9 W. The current drawn would make it a 5 W LED.

Your 1 W LEDs should only require about 0.29 A or 290 mA @ 3.5 V

[kim gregersen]

Hello! First off all, great circuit, thanks for showing. I have a question. I want to build a string of 5 3w led. Could i use a sistem like yours? If sow, what power suply would i need?? and what components should i buy? The leds run at 700ma and have a forward Voltage of 3.5V~4.0V Sorry about all the questions but i'm not all that good in electronics! Thanks!!!

[mattthegamer463]

The wizard at linear1.org ( http://led.linear1.org/led.wiz ) will help you, but I'll explain it for you as well.

If you want 5 LEDs at 3.5V:

5 x 3.5 = 17.5V

If you wire them in series, you'll need a 17.5V power supply. Current on that would be around 700mA. That shouldn't be too hard to find. If you wire them in parallel you need a 3.5V power supply, but you'll need 700mA x 5 = 3.5A, a power supply with that current capabilities will be difficult to find.

I would wire in series. Using a 18V power supply with a 1 ohm 1W resistor will work good for you.

Use that website to calculate the array that best fits your needs.

[sharmil]

sir, is there any driver with pwm which can drive five 3w high power LEDs(connected in series) . the leds technical specification includes forward current -350mA, Voltage- 2.9-3.7v, typical intensity-65lm.

[ginho]

sorry for the stupid question. If I can't find Fairchild components what can I use in replacement? What are the specs, or better, what are the models for other brands? thank you

[Derin]

I think you mixed the two last steps up.

[haroun]

I'm thinking of converting a DeWalt 18v flashlight to LED, could I supply this circuit right off the 18v battery pack w/o cooking the circuit?

[Artificial Intelligence]

Very cool, but not the simplest. Watch this http://www.instructables.com/id/Super-simple-high-power-LED-driver/

[dan]

thanks, i covered your circuit here:
http://www.instructables.com/id/EVKHFD64UZEWP86K7M/
the problem with it is that using an LM317 it has a high dropout (3.5v) so your efficiency is poor unless you are using many leds in series. you can do a little better with a different regulator model.

[ewitte]

Would this be good to run with 6 at 12v with a .33ohm resistor? That gives 1.515A with it dropping to .7575 running series two sets of 3. Some are 3W and some are 5W but I decided to run them all between .7 and .8ma because I have 50+ to power in all ;)

[ewitte]

I've got 16 on this now working my way up to the full 56. A tip I have for keeping the efficiency up. Usually what I do is use 3.6v for blue and 2.4-2.5 for red and calculate without the 0.5v of the circuit. When I'm finished its usually barely warm if even that. I'll throw an extra 1v on to see if it brings the light up, if it does I stay if not go back. Usually I'm working at 15-18v.

[Artificial Intelligence]

But I'm from Denmark and the LD1585CV or LM1084IT-ADJ is not available here.

[Gopher]

OK, I need a circuit to power a 3W luxeon LED that is powered by 4AA batteries, which is easy enough in itself, however I would like to be able to have the LED at quite a high frequency (5 - 10pulses per second) and preferably with a constant current (thought the last bit isnt essential) I cant give specifics of the Luxeons as I intend to use it on a range of models/makes etc so something to cover most bases will be great does anyone have a diagram of such a circuit or an idea where to finf one etc

[1up]

This looks nice! I might use this sometime. Could this circuit also be used as a driver for a laser diode?

[raygromer]

Great one Dan. I've used it successfully on a couple projects already. But it won't work on the project I'm working on now because my LED cathode legs must connect directly to ground (battery negative). Is it possible to change the polarity of your circuit so the LED anode side is "regulated" and the LED cathodes connect directly to ground? Thanks!

[krunal_299]

Hi!
My requirement is for 350ma LED and Vf=3.5v.
So does it require Heat sink in this ckt.?
Thanks & regards!

[bilti]

Congratulation Dan for you contribution! I am very interested in your work and I`m wondering how can this idea can be adapted to 220V (Europe). I want to connect few power LEDs. directly to the 220v power . Any suggestion?

[Coffeebot]

I'm looking at building a lamp for my living room using this technique (possibly the RGB adjustable one, if my wife lets me ^_).

The lumileds are rated at about 80 lumens (for the whites), and the RGBs are 30-40 (give or take). An average incandescent bulb is about 1000 lumens.

My question is whether or not the intensity is additive. That is, would I honestly need more than 10 luxeon 3s to come close to a standard 75 watt bulb?

Surely I'm missing or misunderstanding something here.

[computerwiz_222]

lumens are lumens. the lumen is a measure of light output. There is no different scale for LEDs and light bulbs. So yes, you would need over 10 to reach the output of a light bulb. Just think though, the, the luxeons should never burn out.

[Coffeebot]

Well, I guess it's not so much the fact that I need 10 LEDs, as it is the amount of heatsink they would need in order to not burn out. I suppose it could be done with a torch lamp, given the size of the average "bowl" on top. But, sadly, we threw all of ours out after the last move.

[Damian7]

I'm working on a project that will used LEDs and will be powered by a 120v AC input 12v DC output adapter. You can't powered directly LEDs to 120v AC, since they work with DC. As soon as I finished my project I will post it. My idea is to replace regular bulbs where they can fit with the adapter.

[Dipankar]

Check out this Instructable "LED CHANDELIER" it uses 220 v in 12 v out.
And also "LED TUBE LIGHT" this runs on 220v AC only.
Who told you that LED's cannot run on AC only.

[computerwiz_222]

they can be run off of AC. The LED will only run at a 50 percent duty cycle. The LED will have a 60 hz flicker. Luckily, the power companies designed electrical systems to alternate faster than our eyes can see, so you should have no issue. Do some research, it is never safe to connect directly into 120VAC.

[Coffeebot]

I'm curious to see what you come up with. I've been thinking of doing something like that myself. They already have some LED bulbs like this (thinkgeek.com had them...they don't anymore), but I'm skeptical as to how bright they are (note Dan's comment below, and my earlier thoughts on the subject above). I would ultimately like to have them lighting up my office in "futuristic" light bars across the top of the walls...but that would be pricey, and probably kick out a fair amount of heat.

[dan]

no, you are not misunderstanding. that's why you are not seeing LED lightbulbs at home depot yet.

[johnpombrio]

Closer than you think, now (1 1/2 years later)! Home Depot has LED stickies (3 LEDS using 3 AAA batteries), 7 LED light Pucks on a transformer (similar to the Halogen pucks). IKEA DOES have LED small reading lamps which I use at night when others are asleep.

[Richard21]

I would really like to make this project, as the LM317 has too much overheads. Problem is searching for Fairchild 2N5088BU returns no results in Google.co.uk Buying from Digikey makes the whole project too expensive with European postages :-( Does andy one know where to get these parts or equivalent in the UK Thanks... and hoping. Richard

[callmeshane]

Uhhhh N-Fet's, F-Net's, Fet-Net's and Nut-Fat's - does anyone ever use F-it's? I am trying to be funny, and overcome my default setting at retarded, mentally defective and lazy..... Actually this device is pretty much what I want to make up, to run a similar system. Only I like to tell me self, (whining tone of voice, tears in my eyes, professional victim to the max) "Oh it's too hard, my brain hurts, I am too busy... etc.." When all I really need to do is to ge a pencil and possibly a really big sheet of paper and do a few diagrams, a few sums and sort of hope to figure it out, and when I get REALLY stuck... to then look for more clarity and or ask some people some questions... However being a sniveling victim is easier, because I can blame them for my not making it work. Anway... I shall start machining the aluminium billet for my head lamp case 2 knight.. Cheers Shane.

[surftom]

Just a comment/Question your Text says an N-FET but your schematic is a P-FET. I assume you really want an N-FET as the DigiKey part number is in fact an N-FET Tom

[dan]

yes, NFET

[rmccurra]

Dan,
Is there a way to use the simple circuit with a number of LEDs in parallel? I want to keep my voltage down so I can power with a homemade alternator where the input RPM could fluctuate (and therefore voltage). Voltage would most likely run up from zero and hover around target of 6V but could spike if RPM momentarily increased. I would like to use 6 bright white leds (3.4V and 6*.02A=.12A). Is my only option to put a resistor next to each one and make the circuit less efficient? Is there a better way than the simple circuit for this application?

Thanks

[dan]

you can do this if: (1) all your parallel leds are the same (2) they are all mounted near each other, so they will have the same temperature when operating.

[rmccurra]

Just to make sure I understand, you are saying no resistor is ok with the conditions you mention? Would it add any safety factor to put like a 1ohm resistor in series with each led? Thanks!!!

[dan]

correct. if you follow the above notes, i would not bother with resistors unless this is a commercial product you are making. yes - adding a resistor will improve the safety factor, target maybe a 0.2v drop on the resistor, so that's more like 10ohm in this case.

[maroneyi]

I built this but realized it is not supplying enough current (even removed both resistors) I have it hooked up to two of those high power LEDs. Putting 4.5 V at 700 mA through it. What can I do to increase brightness?

[mori]

Hi.
I'm planning to wire up this circuit as an addon for this product; to produce an uber alarm clock.
http://www.soleilsunalarm.com/SA-300W%20Controller.htm
With a 460nm star LED; which according to this:
http://www.jneurosci.org/cgi/reprint/21/16/6405.pdf
is the wavelength of light responsible for stopping production of melatonin.
However... as this "simulates" dawn by increasing and decreasing the voltage to the plugged in devices, thereby increasing and decreasing the luminosity of the light. I'm not quite sure if it does this by varing the current or the voltage of the devices, but from what you're saying this device is either ON or OFF? re: "it keeps the LED brightness constant no matter what power supply you use or surrounding environmental conditions you subject the LED's to".

So is there anyway of modifying the luminosity of the LED via the supply voltage? or is it one brightness solves all?

[mori]

ooo.. Already answered i think by a different thread.
http://www.instructables.com/id/Blue-LED-light-box/

[dan]

see my 'power led circuits' instructable, it has a couple methods of converting this circuit for dimming.

[midlander]

Hi, I wanted to use this circuit to drive 1W LED (3,35 V/350 mA) from 4 AA NiMH cells. You declared 0,5 V dropout voltage, however, from 4,9 V on the battery (circuit switched on) I can get only 3,3 V (330 mA) on the diode. With 6 cells it's fine, so there must be higher dropout for some reason. As I am from Czech Rep. and could not get the transistors used by you I bought BC548C (common NPN) and IRFZ46N (as MOSFET) which should have similar parameters. But I'm not an expert, so maybe I missed something important in the specification. I use 100k and ~1,47ohm resistors. With 6 cells I got 354 mA and 3,34 V on the diode, with 4 cells I got only 3,29 and 325 mA even if I used 1,22 ohm instead of 1,47. Do you have some idea what's the reason for it? Thanks a lot.

[midlander]

Shame on me! It was the bad probe wire! Its resistance was causing additional 1.2 V dropout while measuring the current with the multimeter. :o))) But I have another question: the current is still dependent on the source voltage. Well, not much, but it is. E.g. if I increase the source voltage from 5.2 to 7.8 the flowing current rises from 351 to 370 mA. Is there some kind of formula that can express this dependence?

[Jamshed]

Hi Dan, In step 2, you mentioned turning off Q2 reduces the current through the LED's and R3. Assuming Q2, is turned off, where you will trigger the base of Q1 from. And you have mentioned Q2 is acting as a variable resistor, so if 1 of our led gets short, then the LED Voltage drop will be across Q2?

[dan]

1) as Q2 turns off, voltage at base of Q1 goes up, turning it on. this is a feedback system. 2) yes, if a led shorts out then the voltage drop will be across Q2.

[Jamshed]

Hello Dan, I mean to ask, if an led gets open then there will be no voltage drop across the NFET Q2. So, this means Q2 is off. Now with Q2 off(because of open led), how the base of Q1 is going to be triggered?

[trons]

Could I hook up 2 Luxeon I's in parallel to this circuit? Or can you only safely go parallel with strings of minimum 2 series LEDs? I was thinking of using an mATX PSU's 5V line to drive the Luxeons in pairs, each sharing a driver with 700 mA output. (I have 12 LEDs, 8 are from the same batch 0473468, the other 4 are batch 0473467) Thanks in advance.

[Tomteslakt]

Cool! If this circuit is used in a flashlight with a Luxeon K2 LXK2-PW14-V00 LED running at 1000-1500 mA powered by batteries (3 C-cells?), are there any modifications needed? This is a bit more amps than the 200mA mentioned in the instructable... And what is most important to keep the voltage drop of the circuit as low as possible?

[Tjp]

Cool stuff. I am looking for a way to have a 12 Volt + signal turn off the LED supplied by the 12 V+ supply (nominal automotive supply actually around 13.8 but can be 18VDC in normal operation and as high as 80V in extremes.) Basic concept is to turn one constant current LED source on with +12 VDC and then turn on an alternate constant current source for a different color LED array while turning off the first one... It gets a bit more complex in that the first one needs to have a 30% brightness and 100% brightness but that's another issue maybe. Maybe something that shunts the FET gate to ground when positive bias is applied?

[BlessedWrath]

Can this circuit be used as a driver circuit for laser diodes? I assume so, but wanted to be sure. Is there any reference material I can download to begin to understand why these components are necessary, and what is happening with the physics of the electricity?

[technomorph]

Hey Dan, Love it, How would you add more LEDs to the circuit. In series? In Parrallel? Is there anything else that needs to be modified? thanks

[evox]

Hi Dan i realize this is a stupid question and i am very new to this stuff. but when i build your circuit i get the same voltage on the led end of it (basically what ever the input voltage is the output voltage is the same) i was just wondering if this is what is supposed to happen. i am trying to power up 1w leds. if anyone knows what i am doing wrong they can email me at [e-mail] thanks for the help guys.

[dan]

how are you measuring? with no leds connected, you will see approximately the full input voltage at the output. you must connect leds to see the reduced voltage. also, measure the output voltage across the leds only.

[johnfischer06]

Two questions: 1) Why is R1 100k ohms? What would happen id you were to change it? 2) When deciding the value of R3, why did you divide .5/R3? What is the .5 in this equation?

[dan]

1) as noted it is 'approximately 100k ohms' - anything from about 20k to 500k will work fine. what it does is turn on Q2 in a 'soft' way so that Q1 can over-ride it to turn off Q2. 2) 0.5 is the turn-on voltage for Q1. in reality it is somewhere between 0.4 and 0.6 depending on the Q1 type used, and the air temperature. the specified Q1 model was chosen because it is closer to 0.5 over a wide range of temperatures than most other models.

[KT Gadget]

hey dan. im wondering if a 9V battery will work. will the resistor be able to take 9V and lower the voltage to the appropriate voltage?

[dan]

yes, that's the point of this circuit

[drips]

So easy a primate can do it! Thank you dan for all of your LED Instructables. I built my first one already and it was a-ok. I was checking out some other LED Instructables and found this which uses a large capacitor before the LEDs to get a fade-out effect. Would this be possible with your circuit and if so any recommendations for the caps? I would love to have that nice blue light do a soothing fade as it drifts off into sleep...

[altaria1993]

all the parts just should be a bit bigger to hold :P

[pclark]

Cool project. Since the author stated "Power LED's are now around $3, so this is a very inexpensive project with many uses" I was wondering if anyone could direct me to a good place to purchase leds like the one used here or similar ones. Thanks a lot.

[johnpombrio]

I bought 20 3 Watt Lumi Luxeon 3 LEDs on E-Bay for $60 +$3 shipping. They were JUST the emitter without the star heat sink.

[curlyfry562]

http://www.futurecb.com/Store/InCategoryFilteringSearch.aspx?PartNumber=&SearchType=Category&CategoryName=41301000000

I just ordered mine from this site.

[erckgillis]

The efficiency is very low and heat output high with transistors as drivers. Specialized MOSFET's would be needed to drive 5000mA.

Yup, buck and boost regs are expensive when driivng over 1000mW.

Try a MAX regulated LED Driver on a chip. These are available as samples and online (See my instructable http://www.instructables.com/id/EAZDF6J2S5EPA8L4K4/)
...many models are available for one, three or several LED's in series and/or driven in parallel. Voltage regulated, boost and buck configs avail and with the circuit features:
ô€‚¾ 10V – 65V Input Power
􀂾 Drives One or Multiple HBLED in Series. Drives Three HBLED for 12V input.
􀂾 Constant Current Limit with Integrated Power MOSFET
o MAX5917A: 510mA – 623mA
o MAX5917B: 370mA – 470mA
o MAX5910: 250mA – 310mA
􀂾 High Power Packages:
o MAX5917A/B: 1039mW SO-16
o MAX5910: 470mW SO-8
􀂾 Higher Current Option with Parallel IC
􀂾 Dimmable Through Enable Input (J2)
􀂾 Over Temperature Protection
􀂾 Open LED Alert Signal

[dchall8]

Great Dan. I'm new here and have bookmarked all your Instructables as well as added Instructable and Luxeon to all my spell checkers. Very inspiring. At the risk of me looking extremely dense, your schematic shows three LEDs. Does that mean you could change the wires out and put three Luxeons in series? How about four? Is there a maximum number? If you move to the 3-watt or 5-watt LEDs and add a heat sink, wouldn't you have to expose the Q2 heat sink to the air? I suppose you could still pot it with the silicone and leave the heat sink sticking out of the blob. Or instead of the blob, you could glue Q2 to a heat sink and use the heat sink as the bottom of a mold to make the blob less blobby. Picture a matchbox as the mold. Place the heat sink on the bottom of the matchbox with the rest of the circuit on top. Then fill the matchbox with silicone. Once it hardens you peel the matchbox off and you have a nice looking blob with the heat sink forming one side. Since your circuit is so small, a matchbox would be way too large, but you get the idea. You could still use a matchbox with your small circuit and after it's hardened, use an X-acto knife to cut the blob down to size. A film container would still be too big, but it might give someone ideas for other Instructables. Is there any reason why you could not put spade lugs or some other kind of handy plug out to the lamps instead of hard wiring it? And if you're putting plugs on wires, why not a plug out to a battery holder instead of hard wiring it? Would this circuit blow under the load of a car or boat battery?

[slsmet]

Dan, Really nice circuit. Can you tell me how to calculate total power consumption? The LED consumes, Q2 heats up, 1/10 watt at R3; but how much total power consumption? Does it very with supply voltage? I'm interested in the most energy efficient way to drive a power LED. Thanks for your fine Instructables; they're inspirational. -Steve

[dan]

hi steve,
power = voltage * current.

current through all the devices is the same. to find power for given device, multiply its voltage drop by the current. total power is battery voltage * current.

[maiki]

Hey Dan, great instructable! I've got 2 questions: 1. I read that that resistance of parallel resistors is (1/R1) + (1/R2) + [...]. Is it possible then to replace R3 with two 1.4Ohm resitors in parallel, while one of them is behind a simple on/off switch to switch between 350mA when the switch is off and 700mA when it is on? Or even better: using a 2.5 and a 1.25Ohm resistor, parallel again, both behind switches. Wouldn't that make me able turn the light on and off and to switch between 200 (the 2.5Ohm turned on), 400 (the other one turned on) and 600mA (both turned on)? 2. Just to make sure: A bike light with two Seoul P4 leds (3.2V at 350mA, 100lm) in a row, your circuit and 6 1.2V rechargeables would work and be relatively efficient? And with 2700mAh batteries it would work about 6h? Thanks in advance Michael

[dan]

1) we show circuit for exactly this here: http://www.instructables.com/id/ERVLPDU5XVEWOF31U3/

2) yes

[jeanthibca]

Awesome. Would it works with a 1.4 A K2 Luxeon? If I use your formula R3 would be 0.5/1.4=0.357 Ohms, which is a very small resistor (about the resistance of the wire itself). Am I right? Would it works?
thanks

[dan]

that is correct. wire is a lot less than .357 ohms.

[curlyfry562]

Ya, the resistance of the wire is more like 0.0002 ohms. The formula consists of the constant for the metal x the length of the wire divided by the area of the wire.

[KJM]

Thanks Dan! I got mine working. I got all of the parts at Radio Shack, except for the LED and 2.2 ohm resistors, which I had to order. The output is around 3 volts 220 milliamps. The output voltage is right around 3 volts with a 4.5 or a 6 volt input voltage.

[curlyfry562]

Just found this site, these LEDs put luxeons to shame, someone has to do something with these.
http://www.laminaceramics.com/products/

over 1000 lumens per LED!

[lomeranger]

Dude, I'm assuming your refering to the NT-52D0-0429 Daylight White (7 led array) that is rated at 1100 lumens 24V 1.4A $74 at Digikey and is already lensed so you are limited to around a 30 deg light. You also get more bang per LED from the NT-42D0-0426... Only four left though ;) Happy hunting.......

[lomeranger]

Simple and elegant circuit. I powered 2 K2's in seriese @ 1.35A using about .4 Ohms. Those K2's generate heat, as well as the FET! Using a thin copper plate and a good AL heatsink, must be 130F. Just ordered 3 Lamina's rated at 1A 250 Lumens 8V to make an underwater torch. Should I create 3 seperate circuits, one for each bulb? What about boost circuits for the 24+V 1A? Are they efficient enough? Metal housing for heat. I have considered 20 AA NiMH 2.5Ah batteries, but there must be a better way.

[curlyfry562]

you could try to use lead-acid batteries instead of 20 AA, where are you getting 250 Lumen LEDs?

[lomeranger]

SLA (Sealed Lead Acid) batteries are cheaper, but bulkier & heavier than NiMH or Li-Ion. The major downside is that current draw affects performance much more in a SLA than the latter. Digikey for the NT-42D0-0426. They are the best I have found for raw lumens but pricey considering the link you just posted. They are comprised of four elements instead of one, two pairs in series, in parallel on the die. That's important as you can see the pattern of the four dies in certain lens. In making my underwater torch, I mounted the three on a thin plate of copper. Within 30 seconds, it was too hot to touch and I don't doubt that within 60 sec I would receive 1st degree burns. Also, the lights come mounted on stars. I've already posted a TY on the link, but I want to mention it again. Those prices are the cheapest by a factor of 3. (mouser, digikey, arrow, ledsupply, etc)

[lomeranger]

First - curlyfry562- TY! for the post of the futurecb. Best prices I've seen anywhere. Finished my underwater light using this circuit and a heavy heat sink using 3 Lamina NT-42D0-0426 Daylight White I purchased at Digikey (8V 8W 1A 250lm $15per) and a cast Aluminum housing plus 20 AA (2500mAH) NiMH. My resistor value is a 1 Ohm (10W) and 6 10 Ohm (1/2W) resistors, all in parallel. Thank you DAN for posting this circuit and thanks to my friend for giving me a huge solid copper CPU heat sink.

[KJM]

I'm still waiting for my LED and 2.2 ohm resistors to arrive, but I've got the circuit in my experimenter socket with a 10 ohm resistor instead of the 2.2 ohm one. When I test the voltage between the positive and the drain on the MOSFET, it's exactly the same as the input voltage. If I change the input voltage, the output voltage is the same as the input voltage. 3 in = 3 out, 6 in = 6 out, etc. Is this right? Maybe I'm not testing it correctly.

[jeanthibca]

Awesome circuit. I just saved 20$ from purchasing a buckpuck. Just a simple question for you Dan: what would be the value of R3 using a Luxeon k2 at full current value. According to your formula, it would be: 0.5/1.4=0.350 Ohms??

(I plan to use a Luxeon K2 with 4 NiMH AAs.)

[him27hk]

Can use mosfet instead of Q2?? Should i run 5 watt LED from this board?

[dan]

Q2 is a mosfet already, why did you think it wasn't? Q1 is not a mosfet, and you don't want it to be. you can use this with any led, but you will need heatsinks on the LED and Q2 for more than 1/2W

[homer]

Interesting enough I have the same question. What should I change if I want to use a Luxeon K2 Star OR an array of Luxeon K2 Stars?

[dan]

this circuit will work with any led. you just set the current as appropriate for your led or led array.

[machineage]

Hey Dan - cool circuit!

I wonder if you / anyone could advise me as a newbie (blush)!

I have just bought a self contained RGB controller with PWM'd outputs:

http://www.dotlight.de/shop/product_info.php/cPath/263_266/products_id/828?osCsid=c99bcde36c901d8ae197dcf038b315d3

What I want is to use this for the lighting in a vehicle (12 - 14.6V). The LEDs I have looked at and bought one for testing are:

http://cgi.ebay.com/High-brightness-1-pcs-3W-RGB-full-color-LED_W0QQitemZ230093694925QQihZ013QQcategoryZ66954QQrdZ1QQcmdZViewItem

Now - the vehicle has a removable rear cabin leaving only the front cabin LEDs connected at times. I will need say 6 - 7 for the rear cabin and 2 - 4 for the front cabin depending on eventual light output.

Would this circuit work with the PWM output of the controller?

I guess I would need several circuits to power the array but am struggling on how many and how best to connect them?!?

Thanks in advance!

Chris.

[dan]

hi machineage, you should only need 1 copy of this circuit for each of the 3 colors. make a series-parallel setup for each of the 3 colors as needed, depending on how many total LED's you have. i think that controller is PWM'ing the positive main power? if so then this circuit will work great. if it is PWM'ing only a low-power control signal, take a look at my 'circuits for using power led's' project for how to connect it.

[LED Boatguy]

Success! Got it running an 18 LED array. Need to add TVSS before the epoxy goes in.

http://www.allcoastsportfishing.com/photos/data/500/100091cc10_run.jpg

Dan: Since these are usually 10' from the switch, would a capacitor help? What kind and where?

Thanks man,

Kurt

[BradA]

Hi Dan, great circuit. Thanks!. I have used this with a lamina Bl4000 RGB led array ( Lamina led bl4 )(took a while to save my pennies for this!) . the result is amazing. I have been asked to make an LED pool flood light and have chosen another lamina led array ( lamina bl3 ) . (wow wish i could afford one of these for myself). It's current needs are 1.17 amps per colour. Would this circuit be suitable to power these? Thanks again :)

[LasVegas]

Great Instructable. Kinda wierd that Step 9 repeats twice.

[Ian01]

Strange, I see two of step 10, and no step 9. However no step appears to be missing from the sequence of operations.

[LockOnToIt]

<-- Wicked new guy. I have plans to incorporate this into some under-cabinet lighting using the "warm white" luxeons... what modifications do you think I would need to make to run this out with 4-5 1-watt LED's in series? I'm relatively inexperienced with circuit design and construction. Great project, btw.

[alexhumbertoa]

hi dan, can i directly connect the circuit to the 110vAC mains?

[Baralheia]

This is a very nice trick, and I'm considering using it to make a USB Bawls light powered by a 1W Luxeon. That'd be fun. It strikes me that this circuit could be incredibly scalable. Would it be possible to use a circuit such as this for regulating 14.4 or 14.8v down to 13.2v at a max of 3A? My calculations for R3 would say to use a 0.17 ohm at 1.48W, and the closest one I can find is a .18 ohm, 2W resistor. Assuming the FET was heatsinked properly, would this be feasible with this circuit, or would I be pushing too much current?

[backcountry]

I think the circuit is working fine. With the alkaline batteries, my vom shows 0.5A (I don't have a very precise vom), but with the rechargables it shows 0.2 at turnon and then drops off to 0.04 after a minute. I'm wondering if these rechargables either aren't charging properly or whether they are trashed? As I said, I recharge them by just connecting the wall wart in parallel, does that seem OK? I'm using exactly the part number shown in your tutorial for Q2. For R1, 100K. For R3, 1.25 ohms (1 ohm and 0.25 ohms in series). Certainly will appreciate any help you can offer.

[dan]

connecting the wall wart in parallel to charge the batteries - that might work well, or it might not work at all - part of the problem with that method is the uncertainty involved if you dont have a way to test them. if you have a way to connect 5 or 6 rechargeables to your light, you could test that way. the reason i ask about Q2 is that i changed the recommended Q2 a week after i posted the project :) did you use the (currently recommended) FQP50N06L ? if not, then the problem is that the input voltage is slightly too low to turn on the "old" Q2, and you can either switch to the "new" recommended Q2, or else just use 5 or 6 rechargeables.

[backcountry]

My Q2 is FQP50N06L-ND. Adding another battery or two in series at this point was harder than just adding another 4 rechargables in parallel, so I did that. It brightened everything up so I do think it is a battery problem, not the voltage difference between alkaline and NiMH. Because by putting another 4 NiMH in parallel, I dropped the voltage to 2.4, right? I think my next step is to get some better rechargables (than Radio Shack) and a real recharger.

[dan]

ok, i would suspect the batteries then. let me know if you have any problems after using an appropriate charger.

[backcountry]

With new batteries and a proper charger, I get 4 hours before the bulb starts to dim. That's plenty. Now what I'd really like is a battery holder that doubles as a charger. Plug in a DC dongle to charge the batteries and/or run the light. I'll have to work on that. Charging NiMHs properly isn't simple. But is it enough light? Classical musician opinions were mixed (which means I'm getting close). For those with young eyes it was fine. Older folks weren't quite as satisfied. Actually, I think there's enough light being generated, I just need to figure out a way to focus it into a single plane on a sheet of paper. I tried a lens over the Luxeon but it didn't help.

[backcountry]

Dan - My Lux's finally arrived and I built this circuit into one of my music stand lights. A round Altoids tin contains the Luxeon (star/batwing) glued to the head of a bolt that secures this tin to an arm. The circuit itself sits with the batteries in a rectangular altoids tin, along with 2 1/8 in jacks (one for the bulb, one for DC in) and an an/off switch. This second tin is part of the stand clip. I'm going to try this out with some chamber musicians this week and I'll report back. They poo-poohed my previous effort - with 3 10mm super bright leds - as not enough light for reading dense music. The Luxeon and this circuit is much brighter and I think they'll be pleased. But there is a hitch. While I get kick butt brightness running on a wall wart at 4.5v, or on 4 alkaline AAs, my rechargables can't seem to give it enough juice. Any ideas? Should rechargables be able to deliver like alkaline? I've just been using the wall wart to recharge the rechargables thinking that 4.5v across them shouldn't be a problem, but was plenty.

[dan]

hmm, 4.8V from the rechargeables *should* be plenty, so i'm wondering if you built the circuit correctly. do you have a multimeter you can check it with? if so, check the current through the LED with the rechargeables and with the alkalines and compare to what it was supposed to be. what Q2 part number did you use?

oh, is the metal tab on Q2 bolted / electrically connected to any wires, like the circuit ground? that would be a problem.

rechargeables have considerably more total power in them than alkalines, but the per-cell voltage is lower (1.2 vs. 1.5), so you need 5 rechargeables to get the voltage of 4 alkalines - but the 5 rechargeables will last perhaps 2x longer than the 4 alkalines. so if you are convinced the circuit is correct, then just use 5 or 6 rechargeables.

the wall-wart: the labels on these are confusing - an older "big and heavy" 4.5V wall-wart will put out considerably more than 4.5V at less than its max-rated current, that's why you are seeing the effect you describe. the newer "small and light" wall-warts don't have this annoying behavior.

[alexhumbertoa]

hi dan, can i directly connect the circuit to the 110vAC mains?

[pfwagner]

Simple and sweet. What would the changes be if I wanted to put 12-15 LEDs in series on a 48VDC supply?

[dan]

you just add a small resistor or zener diode to limit the gate voltage of Q2. for schematic see my other instructable here: http://www.instructables.com/id/EZD7IWSZDVEWOF32KY

[hubs99]

Anyone know where to get a 1 to 3watt Infrared high power led??? I've found a 3watt version (lumileds version) on ebay but can't seem to find any other places that sell such a beast. But if its on Ebay I know there must be a place to put in a bulk order. Any ideas/help would be highly appreciated.

[Gthing 2.0]

try radioshack or a hardware store

[Deane142]

Try http://www.roithner-laser.com/ They have single chip units up to 5 watts i think and multi chip units that emit crazy amounts of light. A lot of unique wavelengths too.

[pclark]

Could you please tell me the source you used for the Power LED's. I have seen some RGB ones available but only through eBay. Any 1 Watt or 3 Watt ones would do fine. Thanks

[Gthing 2.0]

3 watt i think

[Gthing 2.0]

wow i mean it dosen't get better then that!keep posting

[backcountry]

Can you help me apply this instructable to a project I am working on? I am trying to build a portable music stand light; bright light, no wires, no brick or dongle, rechargable batteries, and 3-4 hours light on a single charge. My current design uses both a round and square Altoids tin. Power is provided by 4 NiMH AAs, housed in the square tin and attached to the stand clip. The round tin contains 3 "superbright" 10mm LEDs with current-limiting resistors. It works well for jazz gigs (music being just lead sheets and not very dense), but is not bright enough for classical work (dense pages of small notes). I'd like to go brighter without going bulkier. I think my next step up in brightness would be a 1W Luxeon with a lens. Would you agree? Can I get enough juice out of 4 AAs to power a Luxeon for 3-4 hours? If not, is there a good alternative? Do I need a heat sink for the Luxeon? Looking at your various instructables, it seems like the constant current circuit in this tutorial would be fine as the input voltage is limited. Can you help me figure the proper values for R1 and R3? Thanks much in advance for any help you can give!

[dan]

this is a perfect application.

- can you get enough juice from the AA's? 2000mAh / 5 hours = 400mA, which is "full power" for a 1 watt LED. no problem.

- the constant current source in this tutorial is a perfect circuit for this. R1 = 100k-ohm or so (it doesn't matter much). R3 = 0.5 / 400mA = 0.5 / 0.4A = 1.25 ohm, so use either a 1.5 ohm or 1.2 ohm resistor for R3. resistor power = 0.25 / 1.2 = 0.2 watts. you'll want to get a 1/2-watt rated or larger resistor for R3 to keep it cool, the 2-watt parts listed in the project are cheap.

- you need a small heatsink on the luxeon. based on your application, it would be ideal to just glue it to the flat outside of the altoids tin. you'll also want to heatsink the Q2 transistor the same way, you can glue it to the flat inside of the tin.

- not sure if you'll need a lens, it depends how close this is to the music. get a "batwing" luxeon, those are 110-degree angle, and get a 25-degree lens for it, and see if you like it with the lens or without.

- what's a big benefit now that you have this circuit? if you forget to recharge your batteries, you can just use regular AA's instead.

[GalileoPilot]

Dan,

Excellent circuit, I was working on a similar circuit using a LM317T, over the weekend, for a project I'm doing for Burning Man 2007 and didn't want to use a expensive LED constant-current driver module/board.

Thanks man.

GalileoPilot
http://www.trexxis.com/shuttle/shuttle.html

[dan]

thanks, nice website & band!

i just added the 'lm317' type current source here, if you switch to an LD1585CV you can improve efficiency and also get more power since it is a 5-amp part.

[royalestel]

Yep, when I finish making my USB powered flashlight, I'm making this sucka. This is a great way to learn about electronics!

[battletux]

Ok, For some reason I'm seeing step 10 twice and no step 9 on the 'View All steps' option.......

[chardman]

That is one of the clearest instructables I've seen yet. I'm not the sharpest when it comes to electronics, but even I can't go wrong with that. Thanks!

[russ_hensel]

I sort of miss step 8. What value should r3 be? My byte of theory says ( .6 v / ( desired current ) ). Is this what you used?

[dan]

R3 is calculated in step 1. it's actually more like (0.5 / (desired current)), because we used a 100k-ohm pull-up resistor. with a 5k pull-up, yeah it is closer to 0.6. so by using the 100k pull-up, we reduced the power waste of the sense resistor by 20%.

[Tool Using Animal]

Yeah, i get two step 9's, no step eight.

[dan]

fixed.

[crapflinger]

man...i REALLY hope you keep posting instructables...you've got a way to take some higher-level stuff and make it almost make sense...i've enjoyed reading what you've posted so far...and i can only assume that anything that follows will be just as good

[trebuchet03]

That's why he's a part of squid labs :P But yes, if anyone wants to see how to write a good technical instructable -- this is it :D

http://www.instructables.com/about/

[thadrien]

It is awesome! Continue to post!