Picture of High Power LED Driver Circuits
High-power LED's: the future of lighting!

but... how do you use them? where do you get them?

1-watt and 3-watt Power LED's are now widely available in the $3 to $5 range, so i've been working on a bunch of projects lately that use them. in the process it was bugging me that the only options anyone talks about for driving the LED's are: (1) a resistor, or (2) a really expensive electronic gizmo. now that the LED's cost $3, it feels wrong to be paying $20 for the device to drive them!

So I went back to my "Analog Circuits 101" book, and figured out a couple of simple circuits for driving power LED's that only cost $1 or $2.

This instructable will give you a blow-by-blow of all the different types of circuits for powering Big LED's, everything from resistors to switching supplies, with some tips on all of them, and of course will give much detail on my new simple Power LED driver circuits and when/how to use them (and i've got 3 other instructables so far that use these circuits). Some of this information ends up being pretty useful for small LED's too

here's my other power-LED instructables, check those out for other notes & ideas

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

ssharmagdr2 days ago
dear sir I have a super bright led. 10 watt, 12 volts and 1 ampere. pls help me how can it be flashing from bike.
Jonners1 month ago


Wow, great article!

Can someone please tell me if the circuit in step 10 - the analog adjustable driver - is a constant current driver, and is stable across ambient temperature changes?

Many thanks,


Matto707991 month ago
I have a 96W COB chip that has unusually high voltage requirements. It needs 70-90VDC to light at 1.5A. I want to build a driver for it. Can you recommend parts I would need to use 110VAC as my supply.

another thought

could I just use an extra 3W 1A bead to forward another 3 volts to the 10w chip

I believe the 10w LED draws 10.5V and if I used another 3.4V LED that would put me at 13.9 V my 100 AMP PSU can be overvolted to 13.65V

would this keep the LED's in a range where they will not try and draw more current?

also if I used 2x 500ma PTC parallel would that then handle 1amp with 2 amp cut off ?

I was thinking instead of using diodes to drop voltage why not just use an LED and an 3w LED bead should have the same current rating as the 10w

I bought 10w LED
I believe they have 3x1w beads in series and 3x parallel runs ( 9 beads )

when combining LED's in series does their running current also increase ?

ie) if they where 1w 3.5v 350ma beads would they cut out at 1050ma ?
or would they still cut-out at 350 current and use 1050ma power ?

I had ordered 500ma PTC but now thinking I really needed something a lot lower

RamanaR12 months ago
I have 12v, 10w bright led, can u help me to build drive circuit from 220v
RobertK132 months ago

I can confirm that this circuit works, but a couple things to mention. You definitely want to do your math right or your MOSFET will get pretty hot. It wont take much to get this baby hot. Second, if you have a stable power source from your power supply, there really isn't a need to control your current in such a manner. Simply using a resistor in series with your LEDs will work. The MOSFET can and should be used to soft start your LEDS extending there overall lifespan. Also when using a MOSFET, adding some TVS diodes to protect it is a good idea specifically gate to source.

KiSwiSje3 months ago

I really don't think the pptc is regulating the current. It is just protecting the battery from excessive current draw and posibly explosion.

As the pptc needs time to warm up under overcurrent conditions, the led will probably die before it trips. And if it does not trip, the resistance over the pptc is fairly low. Transition from not-tripped to tripped is fairly fast.

This sounds more like the 'throwie' principle where the battery's internal resistance will limit the current, combined with the small margin the driving voltage is over the needed forward voltage of the led.

I wish I had the necessary components to test this out. My 250mA pptc did not play nice with my 20mA leds :(

Clark574 months ago


I have a Constant Current LED driver module that I removed
from a LED grow light, rated 80-100V @ 600ma. It currently runs a 45 LED
string. It says that it’s a HP060 but it doesn’t look like anything I’ve seen
that matches that on the web. I can’t really see any of the part numbers, but
there’s a transformer surrounded by smaller components, including the only
adjustable thing, a little blue potentiometer. (I have no schematic, but I
assume this might be a “typical’ circuit.)

What I want it to do is to run a 23 string @ the same
rating. The PROBLEM is that when I do, the supply kicks the total voltage down
to equal the same voltage across each LED. Basically, I am trying to double the
voltage at the same current, but it won't let me. (The LEDs are rated at over
double their current power so I am not worried about blowing them.)

Real world:

I measure across the LED while 45 of them are on and the
voltage is 2.225VDC @ 600ma. I then short the string to 23 and measure and it
is still 2.225VDC@600ma, but I want 4.45VDC@600ma across each LED. (2.225@1200ma would do but I don’t think this
thing will do it.)

Like I said, there's a micro blue potentiometer on this
module and I don't know what it does (nor do I have a schematic) but I need to
know what to do to adjust this up to 4.45VDC.

I’m assuming there’s some kind of resistance sensing circuit going on, but I’m
looking at it and I see the transformer, a couple of what look to be SCRs on
heatsinks, a choke, three electrolytic caps (and some smaller caps, too), and
the potentiometer. I’m not comfortable randomly adjusting the pot when I don’t
know what it does.

(As well, a miniature supply that could do what I want would be considered,

I need this, like, 8 years ago so please help!


conquer4 months ago


I wish you a nice day!

We are looking for LED drivers and dimmers for project. We
have finished with wiring.

Lighting system in that project have about 450 meters of LED
strips inside aluminum profiles recessed in ceiling, separated in parts
with lenghts 0.75 m, 1 meter, 1.5 meter and 2 meters.. Also, there are parts
with individual lenghts of 5.4 m, 3.2 m.... Type of LED strip is 3014 SMD
14,4 W/m, input voltage DC 12V, 120pcs/m, pure white color. So, that is
450(m)x14,4(W/m)=6480W, around 6,5kW power of whole system at full load. Lighting system
must be dimmable and wireless controlled. I need help with choosing right LED drivers and controllers for system. You can see project in attachment.

Thank you in advance.

For LED drivers.jpg
node-xD4 months ago

Trying to wrap my head around this and I've got a few questions that hopefully somebody can clear up.

1. " LED current is approximately equal to: 0.5 / R3."

Is he using the value of 0.5v because that is the "Vce Saturation (Max)" value for Q1, the 2N5088BU transistor? If not, where does it come from?

2. "The input voltage can be as little as 0.6 volts higher than the output voltage."

Is this also determined by the "Vce Saturation (Max)" value for Q1? If not, then where does it come from?

3. "The power dissipated by R3 is approximately: 0.25 / R3."

Where does the 0.25 come from? Is it calculated as 1/2 of whatever the 0.5 volts is from my first two questions?

4. "Q2 can only handle 2/3 watt before you need some kind of heatsink. with a large heatsink, this circuit can handle a LOT of power & current - probably 50 watts and 20 amps with this exact transistor."

Is he talking about the MOSFET burning up 50 watts on it's own, or 50 watts for the entire circuit, LEDs included?

5. Considering that the "dropout" remains constant at higher voltages, and assuming that the input voltage stays tuned to this level; would it be true to say that this circuit becomes more efficient with higher power LEDs?

boredom.is.me4 months ago

I've been studying and designing some modules based off of your schematics.

I am interested in an *even simpler* setup to power two sets of 3W LEDs. They will both share a 24V power supply. The first set uses six 3W LEDs (3.2-3.8V). This should be straight forward. My concern pertains to the second set. The second set only has two 3W LEDs (3.2-3.8V). Will the circuit drop the voltage to what the LEDs only need?

I received my parts and can confirm that the "even simpler" takes care of the voltage difference too. :)

ozdenakca5 months ago


for step 10, how can i calculate the voltages and currents? I want to modify this curcuit for different types of power leds.

thank you and have a nice day! :)

Krolo made it!6 months ago

This is the first pcb I designed that uses this instructable to power 16 1w LEDs. I later made a 48 output design. The most I have daisy chained is 109 for a light wall in a reastruant.

2015-02-08 10.54.11.jpg2015-02-06 04.05.36.jpg
xenonion10 months ago


I've had great success with Schematic #4. I'm dimming 4 - 3 watt LED's. I'm powering them with a 20 volt supply. My question is, since this is a constant current device, can I power a single 3 watt led (forward voltage of 4 volts) with the same 20 volt supply? Is this a good idea?

Thank You

Hi there, this is a very good instructable with nice presentation.

I am planning to build constant current source #1.

Led : 10 W

Forward Voltage : 9-11 V.

Forward Current : 1050 mA.

Input Voltage : 12 V 5 A.

I assume R3 is 0.476 ~ 0.5 Own. Is it right?

I can't find the parts you suggested in local electronic stores. I could find only 2N2222, BC547 (Q1) and IRF640N (Q2).

If I use those transistors instead of the ones in the guide, what changes do I have to make?

AjayV110 months ago

your circuit is good cost effective need thermal compassion .

dapackers made it!10 months ago

Thanks for your Instructable! Using your even simpler current source and a
10W RGB LED my pumpkin really glows!!!!!! I used the cheaper
LM1084IT-ADJ regulators and some IRLB8721PbF N-Channel Mosfets to
turn the circuit on and off using an Arduino UNO. The video can be
seen on YouTube http://youtu.be/ry2G-STAhl8

library for the Arduino UNO was another instructable here:

to great contributors like you, this week I made a pumpkin that can
be seen from blocks away and I found out that the Hedgehog averages
about 6 miles a night on his wheel! I'm off to start my next
Instructables project!

xenonion10 months ago


This is a great tutorial, I have all of the parts and built the #4 circuit. Now I want to use it to power 4X - 4.1 Volt 700 mA LED's. I tried to test the current output using 18 regular green LED's and the circuit will only pull ~ 40 mA's according to my DMM at full on no PWM signal. I am powering with a 40 volt 2 amp powersupply. I have the components listed and am using a 0.18 ohm resistor (R3) with theoretical current available up to 2.5 amps. R1 = 100k and the zener is 5.1 volts. What am I missing?

Question: What can I use for a dummy load to test my circuit? I have seen tutorials that use regular diodes (1N4005's) in series and then a series resistor to give the proper current draw? I've got a bunch of 1N4007's around, so i'm tempted to make several strings and test with that?

Thank You!

Hi Dilshan, I am building a battery powered LED torch with 4X1.2V 1200mA batteries, and 3X3W LEDs I am trying hard to find a driver which could do the job of connecting them together

the LEDs I have are,
LED, HIGH POWER, 5000K, 70CRI, 275LM
LED Colour: White
Luminous Flux @ Test: 369lm
Forward Current @ Test: 1A
Forward Current If Max: 1.2A
Forward Voltage @ Test: 2.86V.

could you please help me what should be the specifications of the driver??

.your voltage is very near so no need of regulators why not a resistor actually this regulator is needed if supply is fluctuating its output.

according to your calc.=0.64ohm resistance will work

so use 2=1ohm and 2 ohm 5w resistances in parallel to get .66 ohm

1.94/0.66=2.93A for 3 led and 0.97A for 1 led

remember NI-MH cells have voltage somewhat1.35-1.15 volts in my different calculations but doesn't makes difference too much i.e. doesn't kills led. Especially if led are connected to heat sink

I would avoid cranking them to their maximum current (If Max of 1.2A) because that will shorten their life span (and also increase heat output).

4x 1.2V = 4.8V (when fully charged, of course). Batteries are very likely 1200mAh (the little h is important: milli Amp HOURS)

3 of your LEDs in series = 3x2.86V = 8.58V, so nearly 2x your max battery voltage: you'd need a boost circuit to achieve that (i.e you won't get far with this circuit driving them because your input voltage isn't high enough). In series, you'd pull the 1A (at test), which is near the total Ah capacity of your batteries (but this would first require your batteries to provide 8.58V+, which they do not). Assuming you added 3 more 1.2V batteries (7x1.2 = 8.4V), you wouldn't have especially bright LEDs, and your battery voltage would still plummet like a rock, and eventually not actually drive the LEDs.

If you instead ran the LEDs in parallel (requiring lower voltage but higher current), you could drive them at a little more than 1/2 of your battery voltage (a switching regulator would be a decent approach). In such a configuration however, you would need to provide _3_AMPS_, which would cause your batteries to heat up, and their voltage would drop even faster.

In general, lower current, higher voltage (LEDs in series, not parallel) is preferable. Ohms Law will kick your butt: high current causes small resistances (including what occurs within the batteries as they heat under load) to result in larger voltage drops. This is the basic premise behind high voltage power transmission lines.

LEDs like this are intended to run off of something more than a few rechargeable AA batteries. An 18650 Li-Ion gives you 3.7V rated at say 2.4Ah (2400mAh, but that's at 3.7V not 1.2V). One such battery would be circa 1 hour of runtime for your lighting setup, and as the power supply for Dan's circuit, would not require a significant voltage drop (and thus dissipation).

I've had decent results from using a switching "buck" converter driving a 3W LED using a partially depleted 9V battery. At 8V or so, the 9V has outlived it's useful life as a 9V in whatever plug-in application it had, but clear on down to as low as 5V or so, it's still a good power source for a buck converter - you can drive a 3W LED quite nicely. Not for hours, but the battery was essentially trash to start with.

fche4141 year ago


Could l know how do you come up with the equation I = 0.5/R3 please?

splud fche4141 year ago

An easier to understand and use explanation might be, given a known LED current rating of say 20mA (0.020 A), to determine the necessary resistance of R3:

R3_in_ohms = 0.5 / (desired_current_expressed_in_amps)

0.5 / 0.020 = 25 (ohms)

Does that help?

I have also a problem with this equation, were do you get the 0.5 from?


(I apologize for my crappy english)...

This is the way I see it. As you can see on the diagram, the NPN transistor emitter is wired to ground and the base is wired to the power resistor. When the transistor is well... "polarized" (don't know how else to say it), a fixed voltage should appear between the base and the emitter. In the case of a 2N3904, the base to emitter voltage is approximately 0.58V, which is the same voltage drop on the power resistor... So, if you use this transistor, the current flowing would be I = 0.58/R3 (this is Ohm law V = I*R).

I hope you can understand my english...

"biased" is likely the term you're looking for.

The FET is effectively a voltage controlled variable resistor. It (and well, the diode series) forms one half of the voltage divider circuit (coming from Vcc reference), which R3 is the other half of (going to ground reference). The lower the R3 resistance, the "closer to ground" the base of Q2 will be, and therefore the higher current (in the form of a lower FET D-S resistance) can flow through Q1. Raise R3, and the base of Q2 will be more positive, thus conducting the Q1 gate to ground, causing the Q1 D-S to have a higher resistance, and reducing the flow through the diodes.

The 0.5 figure seems like a reasonable ballpark figure for the voltage at which the Q2 base would be within the "region of operation", while not being saturated. The value would vary with the transistor being used - chiefly with the part type (refer to the data sheet), but also individual characteristics.. If one had a good transistor tester, you could get the Vbe(sat) of an individual candidate transistor and tune the resistor value to that.

Note that use of a Q2 with a higher Vbe(sat) voltage would allow for use of a nominally higher resistance value for R3, which may be easier to acquire. You don't want to use too low a resistance value, or the BJT will burn itself up.

Vbe(sat) for the PN2222A transistor is 0.6-1.2V for 15mA base current and 150mA of collector current. This represents _saturation_ (full-on), which is not where the transistor will (or should) be operating while it is regulating the current

Remember, the total resistance between the resistors and the ground is R3 + whatever the FET has been regulated to. The FET is the primary current regulator here.

FTR, your English is entirely understandable, and better than many who presumably communicate in it as a first language...

The Freak1 year ago
Link to PWM 555 thingy is dead (end of step 8).
ChronoX5 made it!1 year ago

Thanks for the shematic! The LM317 driver is too wasteful for one LED but this is just perfect. I used a 2N2222 NPN Transistor which has a Base Emitter Saturation Voltage of 0.6V. I didn't have any 1 Ohm resistors so I had to take about ten 10 Ohm ones ;)

2014-08-31 19.08.46.jpg
synthdood made it!1 year ago

I have built a bunch of these little circuits, and they work really well. Thank you for posting this! Here I built four driving a color LED with an micro controller and a pot. I encased it with liquid plastic for protection. You can adjust the LED cycle speed with the pot. I use a single lipo to power everything. Fun!

pratham1231 year ago

I need to drive 5 3W RGB LEDs in series using an arduino. The maximum current required is 350mA each for R G and B dyes.
and minimum current requirement is 2.4V, 2.4V and 3.5V respectively.

What should I use to interface the LEDs with my arduino as the output current of each arduino pin is 50mA maximum

You need MOSFETs my friend (though a darlington may well suffice), i would recommend you read up on the topic of signal amplification and driving power using transistors :)

Hi Pratham, were you able to solve your query since I am also looking for the response to a similar query

yuenong1 year ago

Hi, I was into final exam on making LED lamp.

May i know which circuit the best for driving 48 LEDs simultaneously?

the LED forward voltage will be 2.8-3.2v and max forward current is 20mA.


sbadgujar1 year ago

How to Connect 10*1w led to this circuit. what changes I have to make in circuit

Would a TIP41C be a good substitute for Q1 (the 2N5088)? I'd like to make sure I have enough headroom heat-wise, as I plan on putting the circuit in an enclosed area with no ventilation. Are there any other possible modifications I might have to make to the circuit (adjusting the value of R3, etc) to accomodate this? Thanks!

Upon looking at the TIP41C's datasheet more closely, it seems the forward gain is significantly smaller than the 2N5088. Would perhaps a TIP122 be a better replacement instead?

ovenam made it!1 year ago

Muy bueno!!!! Ya lleva poco más de dos años funcionando!!!


Ref to http://www.instructables.com/id/Yet-Another-Daft-P... any idea how I can get y query solved.

I want the same matrix display on the higher powered LEDs ( 3 W would be good enough).

Thanks in advance

remi.serriere made it!1 year ago

Thank you for these schematics. I used them on my sequential taillights for my car : http://www.dodgeintrepid.net/showpost.php?p=269747...

I shared the link to your article in this forum, I hope you don't mind. There's also a video of the final result.

The picture included is just a "fly bug" design ^^ IRF640 and BC547B, I used it for testing ;)

cfsterpka1 year ago

Awesome tutorial on LEDs and drivers thereof, thank you!

charika1 year ago

thank you fr this excellent tutorial, I need help for a project, I need to use the high power LEDs listed bellow, 2 of them in series, can you help me to drive them please?


Mr--P1 year ago

Hi all... Can anybody tell me if it would be possible to adapt circuit #5 to allow for either direct voltage in or PWM. In other words, can I add to the circuit and have it so that it work via PWM but if there is no PWM signal it will still work without it. Something like this...

pholme1 year ago
Hi dan good work buddy just wanted to as how many leds can you run on circuit 3# the one you say is the best
Regards phil


I want to power my bike headlight for 1watt X 2 leds and my alternator may be powering some where to 14.5v -16.5 v at 4-5 amps

battery capacity at 12v 7 amps

Will your basic circuit work without failure

svdstoep1 year ago
I've made this diagram with an IRL540 as Q2 and for Q1 I used a BC549.
The calculation for R3 was slightly off. In the example is given 0.5/Milliamps=R3. I've tried this for 700mA 0.5/0.7=7,142857142857143 (0.75 Ohm resistor). But my current was way to high with this resistor. Then I slightly increased the resistor to 1 Ohm. at this moment the current is 640 mA. I believe this is due to the fact the collector saturation voltage of the BC549 is somewhat higher 600mV.

All seems to be working now. There's only one thing I wonder. What about the power consumption. Is there a large difference between the fancy LED drivers you can buy in the shop for a lot of money?
You2132 years ago
I want to create a 8 led light bar( each led requires 3.2 volts and 350ma ) Is this the circuit i need because the #'s aren't adding up. if I don't use one of these circuits i am just going to use resistors alone I just wanted something to condition the power more because I will be using a 7.2 volt 1600 mah battery and as it drains I don't want the leds to dim
JTHinton2 years ago
I am getting things dialed in now. Built my first prototype. Realized I made a mistake with my math on calculating the R3. Not a problem. I chose too much resistance. Even with the lesser current going through the LED's, I am seeing that the parallel setup I made for the MOSFETs is not drawing power equally. One is getting hot, the other is barely warm. How do I keep a steady current through both in parallel? Because they are not at 100%, they are fighting each other? Will running a separate resistor into 'G' for each transistor do the trick? Thanks again for all the help.
IansMusical2 years ago
Hello, Can anyone tell me if the Rds(on) value is important and what range it should be within if selecting a new mosfet? Also are there any none logic level mosfets that can be used, even if driving/using PWM from a 5v PIC micro? Thank you in advance, Ian
I think Rds won't matter in this case. I have used and tested IRF540 and IRF840. You can use them as they are cheap and can handle very heavy current.
krytcz2 years ago
Hi Dan,
I have built the circuit #5. I am using a microcontroller to control the switching on and off of the LED. However, i have a problem. The LED does not turn off immediately when the signal at the Gate terminal is terminated. It takes quite sometime for it to turn off...With this, i am not able to dim the LED..Could you help me?
rikil krytcz2 years ago
MOSFET gate has large capacitance which has to be grounded heavily to turn it completely off. Better use a NPN signal BJT as a switch which can act as a strong pull-down resistor. I use MCT2E which gives me isolation as well as pull-down.
JTHinton2 years ago
Great instructable here. I am currently attempting to power 6 Cree XM-L2's. They run around 3.35v 3000mA. I am building an off-road lamp for a vehicle. Power source will be vehicle voltage. Between 12.6v-14v DC. I am trying to adapt your Constant Current #1 to accommodate the current. First question. Do I need to run the 6 LED's in a series parallel? Seems like in series the voltage drop would exceed my input voltage. If I do, I am having trouble finding a resistor with such a low resistance, and a high wattage rating. Any help would be great. I can't seem to find DC drivers to power this build and maintain a small size and price. I am no electrical engineer that's for sure.
rikil JTHinton2 years ago
You can use multiple resistors in series and parallel to create the desired value. Which will reduce the wattage rating of your resistor, making this circuit more reliable.
Derek Lewis2 years ago
Hi Dan,

Loving this design! It's working really well for me so far, but I'm curious about component selection.

I'm using a (purchased) high powered buck regulator to get the voltage down to as low as I can, and still have this circuit operate. (~0.6V above the LED Vfwd) I'm building 36 of these circuits, to individually power and PWM twelve RGB LEDs (with 3 of the buck regulators, one for each colour), at 700ma per channel, so these serve as excellent low part-count CCRs. However, I'm not sure how various characteristics of Q1 and Q2 relate to the efficiency of the design. At 700ma x 36 any efficiency gains I can get by choosing the correct components will help reduce the amount of heat I have to deal with.

What kind of things should one look for in Q1 and Q2 to help with efficiency, either to allow input and output voltages to be closer, or anything else?
Dan, This post was a great starting off point and took care of a few questions I had been trying to work out. So thanks for the inspiration and great info. I just finished my Bike light and will be working on getting a full instructable up shortly. Here is my blog post, http://crumpspot.blogspot.com/2013/04/power-led-bike-tail-light-with-arduino.html And this is the schematic: https://www.circuitlab.com/circuit/b6r5h8/rear-bike-light/
hello Dan,

I'm looking at a similar project to drive 3 MCE cree LED's via a LM317 and 555 giving the PWM option for a bike light
can you advise what sort of circuit schematic and values I'd be looking at? I remember seeing a PWM circuit diagram you had on the instructible, but couldnt find it
janly2 years ago
Hi Dan,
I want to build this project http://blog.makezine.com/2012/01/02/luminch-one-offers-lumens-by-the-inch/ using RGB high power LEDS, forward current of 350mA. I will be using the Arduino Uno and I want to supply a constant current to the LEDS. Can this be done using LM317, if so how can I do it?
nathan.0192 years ago
Hi! I need to design a LED driver which can power 70 typical (20mA) LEDs. i need to design it in a way that it will light up when a signal sends to turn it on. is it possible?? please help me..
jonesakysiu2 years ago
I'm planning to blink IR LED at 36kHz with microcontroller, and I'm thinking about using Your circuit from step 8 (http://cdn.instructables.com/FHY/YUTR/RCQEWP86JFY/FHYYUTRRCQEWP86JFY.LARGE.jpg).
But as far as I understood from the text, it limits current by "turning on and off" Q2 and LEDs. So the question is will it interfere with my PWMing rate? I need to have exactly 36kHz and if transistor will also turn LED back and forth it may cause problems...?
Davidko2 years ago
Love the instructable but would love to ask anyone out there if they can help me out. I made curcuit #5 work perfectly and i am using an Arduino uno to work the pwm. my problem is when i try controlling more then one driver it doesnt "drain" both drivers. (one dims and the other blinks high/low like it is unable to pull the transistor down) how do i build a curcuit that willl be able to control 30 drivers with only one pin on the uno?
1 pin si gonan be a bit difficult because one pin just can only send one piece of info at the same time, so you would need a seriaol protocol, but then also you need the AC circuit to understand a serial protocol. I dont think it is really possible in practice
Just in case I was dimwitted. Do you actually mean you wanted to drive 30 with one pin in total or 30 with each their own pin? Both impossible on the uno. You could add one or more multiplexers that you drive with a minimum amount of pins
i would like to drive 30 drivers with one pin on the uno. i was thinking (if it is possible), using one pin on the uno to drive a transistor or MOSFET to turn ON an external 5-9v that would in turn, drive all 30 drivers in the circuit. That is possible, BUT when the main circuit is OFF how do i "Drain" all of the circuits so that the power drivers turn off? and how who the main circuit be hooked up?

i know a little about transistors, maybe could use a NPN for turning on the power and PNP for turning off and draining? i really dont know..
My knowledge is limited, but i wish to learn.
ok so you want to switch on 30KED's at the same time with one pin right?, that in fact is not so difficult. If you are using 9 volts you do not need 30 drivers as you can drive several LED's in series albeit not 30 on that Voltage. If you use a mosfet to turn on an external 5-9 Volts, That same mosfet can turn them off, no need for extra circuitry there.
How much current do your LED's draw. Are they just regular LED's? or power LED's
i sent you a private msg a little while back stating the conditions and what i was thinking of doing. but here it is again
i would like power around 50-100 high powered LED's...
i have and would like to use:
12vdc 10A power supply (maybe 24vdc and just hookup more LED's in series to makeup the difference and use less drivers in the future)
IRF540N 33A/ 100V/ 0.040 Ohm/ N-Channel Power MOSFET with heatsinks
2N3904 transistor
0R75 ohm 2w resistor
100K ohm resistor 1/4W (i could use different values if specified)
3x 3w epistar white 3.2-3.8v 700mA LED (per built driver) and i can adjust the voltage on dc power supply(11vdc/3 = 3.66vdc each LED)
a fan for the breadboards
and an Arduino uno to control the PWM with one or two pins

now i would love to build this in a way that can PWM control these drivers with one or two pins on the Arduino. i starting building the driver in circuit #5 (and 3x 3w LED's) with great success but when i hooked up TWO drivers to one pin that is when driver #1 LED's would flicker and driver #2 LED's fade out and work like it was suposed to...(programmed the arduino to fade in and out slowly for testing purposes)
so that is when i needed help. i got to thinking.... i suppied circuit #5 with 12vdc to the LED's but i controlled the N-Channel Power MOSFET with my ARDUINO SO what if i suppied the MOSFET or transistor with 5-9vdc or 12vdc and made it supply ALL the drivers somehow. but if i do that, when the power is ON everything will work(i hope) and when the power is OFF, everything might flicker because there is no "DRAIN" hooked up to that circuit.(just like my experiment showed hooking up the drivers directly to the Arduino UNO) i dont know circuitry too well to accomplish this task so i needed some help.

any feedback would be greatly appreciated.
Thank you.
I replied to that message. didn't you get that reply?
Well there might be various reasons for the flickering
I asked about the value of R1 (the resistor connected to the Uno pin). It seems most likely that the two drivers are influencing eachother. diodes maight be the answer but I need to know the value of your resistor R1 is that the 100k Ohm you refer to?
Just a bit more of a reply
Although theer might be different solutions to your problem (e.g. separating the circuits with diodes or transistors, your thinking of switching all the drivers on and off with another FET is good.

I do not really foresee the problems you mention (and are guessing about).

One of the problems you may have is that you are using an IRF540N FET. As I understand that is designed to work with about 10 Volts as a trigger and though it is odd there would be a difference in how the two circuits you built react, it is well possible that one of the circuit sjust does not get enough trigger voltage from the Arduino.

Nevertheless, driving all the drivers as a load by another MOSFET should be possible (then ofcourse you would need to change yr circuit accoring to say #2 or 3#)

Nevertheless, you may try something else:
Change your circuit into circuit #2 or #3 (does not matter which one) and put a transistor in parallel with R2 or Z1 and add a 10 k in its baseline. Drive that 10 k with your uno. That should work. (although it will invert the signal: high = off, Low=on)

As I am not building this myself it is a bit of trial and error, but subsequently you could go two or 3 ways. The easiest would be to give each of your drivers their own transistor and connect these all to one Arduino pin (make sure you do not draw more than I think 40 mA) or you could hook up all the circuits to one transistor (but I'd suggest you would use separation diodes then).

I suggest you adapt the two circuits you built in the way i describe and see if you can drive both of them with one pin
I might have found something for you, akin to what i describe
diy_bloke, you are the BEST! i do not know how to thank you enough. with the circuit that you have provided me, i have made 2 drivers to work from ONE Arduino uno pin using PWM EXACTLY the way i wanted. I really cant thank you enough good sir. now i am on my way to constructing about a dozen of these circuits. Thank you!
I am glad I could have been of help. It took a while before I exactly understood what you wanted, but I am happy you got what you wanted. I hope it all will still be fine with a dozen of them, you may need to trim the value of R3 perhaps.
Go dluck and if you need more help don't hesitate to ask
That looks amazing. You are the best. I thank you for helping me and I will definitely try making that circuit tomorrow as it is way too late today to even start. (when I get started I can't stop till its done) I really appreciate your patience and knowledgeable responses. I will write back as soon as I build the circuit you described. Thank you again and I will keep you posted.
My pleasure. We all have to learn
David, I recently replied already in a private message
I do not really understand your worries why wld you need to drain anything
Currently i cant see thereply i thought I send to you, so i just try again You do not need 30 drivers as you can feed a couple of LED's in series.
Wether you can connect the amount of drivers to 1 pin depends on a few things:
Please tell me the following:
what is your value of R1?
What current do your LED's draw?
Do you just want to switch them on and off or do you actually send a PWM signal to the LED's?
If your circuit is now working perfectly, it would seem the easiest to either increase the voltage so you can feed more LED's in series or drive 2 led's in series, so you need 15 drivers and then drive those via 15 diodes connected to the UNO
However, there are two concerns: will the voltage drop over the diode affect opening the FET (but you can try this with one diode), and will the 15 drivers not draw too much current from yr UNO (that is why I was asking about R1)
snotty2 years ago
Yeah! I made a couple of helmets like this. One with a 555 running bright red flashers. And one with a thin film solar charger built in.

Solar powered version:

dreadnaught2 years ago
Hey can we run this in series with other circuts?
josephx863 years ago
I want to build a constant current source which has input voltage between 3-311v.
how can i make it with low cost and transistor based circuit?
Thank you very much.
Oh, I forgot write i want to drive 1 up to 80 3.2v / 20ma white LED with 311vdc.
that is kind off a stretch. Do i understand correctly that you want 1 circuit that can drive 1 LED with 311 Volt as well as 80 LED's with 311 Volt?
if the LED is 3,2 Volt you would need a transistor that can take at least 310 Volt as Vce or as much as (311-80x3.2)=55 Volt when you use 80 LED's.
Going by the circuit I just uploaded before (but you are going to need a different transistor) you would need an emmitter resistor of about 30 Ohms (1/4 Watt more than enough). But I would not know a transistor that can take 310 Volts just from the top of my head. Even the BU111 can only take 300 Volt, but even then, that would be overdoing it as it can take 6 amps and u only need 20MA.
The MJ16110 can take 400 Volts, but at 15 amps and tht is looking launching a scud missile to kill an ant. I am really not aware of a 350 Volt 50-100mA low power transistor.

If beforehand you would already know that you will drive 80 LED's then it becomes a whole different ballparc because then a BC546 would be enough.

I am not saying you should do things different from what you plan but as far as I understand your plans you are setting yourself up for trouble and high cost, whereas it could be simple and cheap.

Also, are you really bound to 311 Volt? because even with 80 LED's you still will be turning 55 Volt*20mA=1 Watt into heat.

Perhaps your project will still be possible if you rethink a bit what you are doing
you could use a 3055 or comparable powertransistor, as in attached picture. the LED curent is roughly 0.6/R2. 3 Volts input voltage might be pushing it a bit but should be doable
earned guy I was in need of such a scheme
happy i could be of help :-)
diy_bloke2 years ago
Though circuits 1-2 will most likely work with almost any power NFET, for the circuits 3, 4 and 5 it is pretty important that indeed a 'Logic' type NFET (usually with an 'L' behind the typenumber) is used because 5 Volts may not be sufficient to completely open up non-logic level NFET completely. Now ofcourse this won't alwaysbe necessary, but if drawing higher currents it might well be.

One may want to consider taking circuit 4 and replacing the zener with a transistor that has its base pulled low via 1 M resistor and that receives its PWM signal via a (say 4k7) base resistor
bloftin2 years ago
Thanks for a great instructable. Since high power LEDs are not that cheap, some good advice from Winder's book "Power Suppliews for LED Driving" is to use a zener diode (1N5334B) in place of the LEDs until you have tested your circuit. Example is to have one like this for each 1w LED in circuit.

You have helped me more than you will ever know. This circuit is amazing if your Vf on your led's is close to your input voltage. It then becomes super efficient.

I bought a bunch of MC34167's hoping to run some led's with it, only to find that it doesn't have current sensing. So you saved my reer.

To make sure of the Forward voltage drop being close to your input voltage though, the MC34167 is a tremendous choice. It is a single chip step up/down converter with a minimum of external components available that will get you to the correct voltage for super cheap and not have to worry about heat as it is crazy efficient.. I now use this along with most of your circuit for current control for 4ampere's worth of 100lumen/watt led's in an automotive environment....... CRAZY bright. And no heat given off anywhere except for the LED's themself. Best of all, no need for an O-scope and you can get this much efficiency. Lets face it, if you have a $4000 scope you can make your own PWM circuits and not have to worry about any of this. But if you don't it is good to know there is still ways to do things.
electricmic2 years ago
Great tutorial!
When calculating the wattage of resistors in version 1, it would be of great educational value to demonstrate the calculation using the Power=I*I*R formula
Davidko2 years ago
Excellent Instructable, i would just like to add and you might want to inform people that if this is done on a solderless breadboard with inproperly connected wires the curcuit will not work properly. i was getting strange reads of 400-500mA. when i soldered the parts together, perfect 690mA. As well the draw may vary depending on the voltage. Very imformative instructable.
Great job.
vallka3 years ago
Hi, I would like to run 3 Cree Cree XLamp XR-E Q5 off of a 14.8-18.5 li-ion battery, will this driver work to run them all or one driver for each?
This will be for a mtn bike light so the heat sink will get plenty of air flow.
I only have a p-channel FET (U9024N). How do I modify this circuit to work with it?
Here's an attempt I did with Circuit Simulator:


Unfortunately, the current doesn't seem too stable. :(
chikoroll5 years ago
hey mate, i want to hook up a 25watt LED to my car

the voltage is 13-14v
my battery runs 12v (car off) and 13.6v (alternator charging)

would this be acceptable to run direct from battery, or should i run a dc-dc converter

the link is what i'm looking at

im planning to make a lighting array with similar to this


with similar 10w LEDs


i chose 10w ones because they are more cost effective as it seem that the higher power in watts for each unit they get less efficient, also easier to find lenses that i can use.

im also struggling to find out the right way to power them, as i believe the power in a car battery can range from 11-16v (because of cranking, charging...etc) so im not sure how to go about making a power supply i would also like to make one or 2 rather than 20 separate ones for each LED. im actually starting to lean towards using 3w LEDs like the xmitter, as its easier to get the right lenses.

just some things to think about, let me know how you go.

To avoid the varying voltage issue, I think you could add a 12V voltage reg before the current reg, granted you will need a meaty heatsink on it if your voltages are varying that much. Can someone check me on this
A regulator will probably have too much voltage drop in this situation.
Loafman3 years ago
Hi, first thing I have to say is I love this circuit. I have built one and have been doing some testing with it. The main difference with mine is that for Q2 I am using an IRF540 (I had about load of them already), but it functions basically the same as it's still a power N-FET. The thing I have found though is that when calculating the value of the current set resistor (R3) I kept coming up with worryingly high readings (I was aiming for 300ma), and found that R3 = 0.5 / 0.3 = 1.67 ohm to be giving about 360ma (according to both my multimeters). The solution (at least for me) was to change it to R3 = 0.6 / 0.3, which gave a resistance of 2 ohms and a near perfect 300ma. I'm not sure if anyone else has had the same issue (I can't find any issue with my equipment), so thought I would post up here in case it helps.
Anyhow, thanks for a great post!
Loafman Loafman3 years ago
Hi, just a quick follow on from my last post. In the last post I was using a 12v DC power supply, and found that the formula to work out the resistance of R3 (at least for my circuit) was 0.6 / 0.3a (300ma was my required current).

I have since chosen a 36v DC supply ( so that I can connect more LED's in each series and maximise efficiency) and found that it once again alters the formula slightly. With the 12v supply I had a steady 302ma current, but with the same circuit and 36v input I am showing a steady 326ma. I am currently using a breadboard, so I'm not sure if the higher voltage is proving more efficient at driving the current through the connecting points... maybe, but I'm rusty as hell with electronics!

The good thing is that the Vgs on the MOSFET is sitting around 3.6v, so no danger there!

Anyway, thanks again for this fab cicuit, and I hope my findings help some others with their experiments!
nnadine19733 years ago
Small add: if instead of r2 we connect there a fotocell, we have a great system that is swiching off during the day, and light up automaticly, at constant current, all night.
nnadine19733 years ago
It is working incredible considering the parts involved! I become member here just to say thank you to you.
So, Thank You!!
bobking1443 years ago
just a fast know it seems to help out with a 1000uf+ cap on the outgoing wire to the led's, I am using a 3000uf 16v cap.
bobking1443 years ago
ok I am kind of rusty for the Constant Current driver do I need to use the same voltage as my le need? led need 15volt so in put needs to be like 17volt is that right?
anyone know.
hello, could you explain to me how i would dim with an arduino? i'm a little fuzzy on this with circuit #5, i'm a lot more comfortable programming wise.
i'm using 5 xre q5s in series so around 18 volts with a 22volt power supply. do i need a common ground to control the circuit with a micro controller? how would i do that with an arduino?
bobking1443 years ago
this is a great drive I love it, what voltage did you build yours with? cues I see that when input voltage is lower output amps goes down and when input voltage is higher amps out goes up.
TheGreatS3 years ago
One of the best instructables I have read. My favorite method would have to be the analog adjustable driver.
diy_bloke3 years ago
I understand wanting to protect the gate agains high voltage, but why would you want to do this if you are feeding with 10 Volts as in your description?

And if you would 'use this with a microcontroller' (though I do not understand how a microcontroller would deliver such high currents) the Voltage would most likely be 5 V to begin with. Or am I missing something?
OK I understand the microcontroller now. I thought you wanted to feed Vcc from a microcontroller output pin
sunil85r3 years ago

I am building a constant power source driver for UV LEDs (4 nos) (High power 11W UV LEDs). The data sheets are as below.

Could you please let me know if I can use your circuit for this.

Thanks in advance
ggodinho3 years ago
Hello, Sir Dan. I've liked a lot your circuit and I've been making some updatings in order to fit with my project. So, I'd like to share my experiences with all and I'd like your technical remarks and sujestions.

First of all, I couldn't understand why did you insert the R2... Since I needed more input power, I just verified my NFET datasheet (IRF260) and Q1 (BC548) components and I noticed Q2 supports until 200V. So, I connected all the circuit at 180V, (an local 110V plus half-wave retific and filtered voltage). This power in drives a string of 40 5mm white leds (40x3.1=124V).

I know I could put more leds until reach near 180V, but I preferred put 3 paralel strings of 40 LEDs (a resistor was added in series on each LED string in order to fit all the strings to the same current). Of course, NFET will waste the extra voltage, but it doesn't matter to me, the important is to sustain the same 60mA constant current (3 strings of 20mA LEDS) and for this, the circuit is perfect.

So far, so good but, I have questions to ask for the group: what would happen with my 120 LEDs if Q1 openned its base or other else problems with this component ? There is a way to introduce a protection for them?

and more: How this circuit would behave if I strobe the LEDs feed with any flip-flop? it seems that the voltage delay a bit to fall when I just insert the leds... so, in a pulsed condition I belive the leds power will rise a bit.

leven ggodinho3 years ago
The purpose of R2 is to create a voltage divider and keep the gate voltage of the FET below the maximum rating, which is usually lower than the drain-source voltage.
Spuzzum3 years ago
This works great! I just finished my first box, consisting of 7 drivers, and will be making 2 more boxes, for 21 drivers in total.

I found that anything above 500mA and I needed a heatsink on the fets, so I bulked them up. I then found that the current actually drops after 900mA.. wasn't the driver, it was my tester diodes needing a fan on the tester heatsink. :p Now it runs steady at 1000mA, and can be dimmed to as low as 100mA. I modified with a 10ohm 1/2w trimpot, and a 10ohm 1/2w resistor paralleled with the trimpot.. making a 5ohm 1w variable resistor. Results.. an "adjustable" constant current driver. :)

Works out to about $8 a driver.

I also found that splurging for 24v sources is actually cheaper than using 12v from an atx psu... less drivers needed in the long run.

Stenvar Spuzzum3 years ago
Hi Spuzzum,
I have been watching the progress of your project with interest, I am constructing something very similar myself and you are a couple of steps ahead of me which has been quite useful. I'm interested in how you have incorporated the PWM signal into the constant current driver. How are you generating the PWM signal? I have made some analogue, adjustable PWM generators using 555 chips, I read in an earlier post that you were outputting the PWM signal through 2N2222's, so it sounds like a similar output to mine.
Any advice appreciated, Thanks.
Spuzzum Stenvar3 years ago
Also again.. :p

I went back to using 3 10ohm 1/2w resistors in parallel with the 10ohm 1/2w trimpot. This gives a 2.5ohm 2w variable resistor. And with a 24v power supply.. it starts at 250mA and goes to 1200mA or so. I've only gone as high as 1100mA.. then dropped right back down to 1000mA.
Stenvar Spuzzum3 years ago
Hi Spuzzum,
Thanks for the info, I will look into it.

In the batch of pictures you recently posted, there was a circuit layout diagram. Could you explain the route the pwm signal takes please. It looks as though you have it coming into the base of a transistor. Sorry to be a pain, just need this small point clearing up.

Many thanks :)
Spuzzum Stenvar3 years ago
No problem... :)

It's taken from a schematic that used to be posted here.. I just swapped the 2m2222 transistor with another FQP50N06 in it's place.

All parts, except the boards and standoffs, were ordered from Jameco

- Trans MOSFET N Channel 60 Volt 50A 3-Pin (3+Tab) TO-220 Rail - FAIRCHILD FQP50N06
- RES, CF, 10 OHM, 1/2 WATT, 5%, (100 BAG)
- POT,10 OHM,3299Y-100,25-TURN,.5 WATT,CERMET
- HEASTSINK, 531202B02500, TO-220 POWER W/PINS BK - 7.5°C/W
- Connector Terminal Blocks 3 Position 5mm Solder Straight Thru-Hole 16A
- Connector Terminal Blocks 2 Position 5mm Solder Straight Thru-Hole 16A

The 100K power, and 1K PWM resistors I already had on hand.

I also just did up an Eagle CAD layout.. I may just have to try etching instead of actual wiring.. it looks soooo much easier :p.

You should be able to see what's going on with the PWM signal now.. I labeled the pins. In "my" circuit, the PWM is selectable by the jumper pin just above the input terminal block. Positions 1/2 = normal mode. Positions 2/3 = PWM mode.

Just remember to connect the PWM's ground to the power source's ground.. I did this by twisting the ground wire of the Arduino with the ground wire of the 24v source, then inserted into the terminal block. I plan on having a junction box, with all power supplies and PWM signals, where they get rerouted to their various destinations. This will result in the Arduino being grounded within the junction box, and the 6 PWM signals can then be shared by as many drivers as you want.. within reason. I have 21 drivers in total.. 6 wavelengths, with 1 wavelength spread over 2 drivers per box. So.. 6 wavelengths that can be PWM'd. And seeing as I have 21 drivers, or 18 PWM instances.. I can share each PWM signal with 1 driver/wavelength per box.
Stenvar Spuzzum3 years ago
Hey Spuzzum,

Thanks, that's exactly what I was after. Nice diagrams btw ;)

One other small thing, I assume you are running the arduino pwm generator at a fairly low voltage, say 5V or whatever, and the led driver has a 24V supply. So all I need to do is power them from their respective voltages and tie the grounds together like you say?

Many thanks for your help so far, I'm really close to getting this going now.

Spuzzum Stenvar3 years ago
Yep.. Arduino gets fed 5v, and the PWM signal is 5v. The 2 circuits just need to share a common ground. :)
Stenvar Spuzzum3 years ago
Oh and I like the sound of your pwm'ing, very nice. ;)
Spuzzum Stenvar3 years ago
also.. Google Arduino.. all the board files and schematics are open source.. ie: free :D.
Stenvar Spuzzum3 years ago
Also where are you buying your 10 ohm trimmers from?
Spuzzum Stenvar3 years ago
Hi Stenvar :)

I made an Arduino for the PWM. It's an Atmel based microcontroller board. The actual script I'm using comes from the coral reef community.. fades up, fades down.. sunrise, sunset. The chip I'm using is the atmega328.. 6 PWM pins. There's bigger atmegas that have something like 14 PWM pins, but that's surface mount.. haven't done that yet :p. You can buy premade.. but where's the fun? :D

Google "Who wants a cheap, simple, Arduino-based LED controller?" at reefcentral.. has the script there, as well as a pcb project for the board. My board was a different design, but they added more things.
Spuzzum Spuzzum3 years ago
OK.. found a bug in "my" circuit. I used a 2n2222 transistor for the PWM transistor, but it died at 450mA's. Friend suggested using another FQP50N06 in it's place.. "Drain connect as collector. Gate connect as base. Source connect as emitter.". It now works fine at 1000mA's in PWM mode, withe the fet being cool to the touch. :)
peterlonz4 years ago
This seems to me to be a very imprecise way of driving an LED.
Any key component with a set point that may vary by 100% is surely utterly useless.
You will not know untill you receive the PTC exactly what its trip point is.
The example shown raises more Q's than A's:
You may end up driving at the desired 500mA or the unwanted value of 250mA.
The circuit copied by the poster possibly employed this method but I'll bet the PTC 's were carefully sped'd & tested to be sure they were within an acceptable range, not something the hobbyist can easilty do!
dan (author)  peterlonz3 years ago
yes, it is imprecise but very cheap and simple. for some applications that can be useful.
innowati dan3 years ago
Hi,I like your ideas.
In the battery pack of Laptop are Ptc used? You can take them?
what do you suggest to drive 1 watt luxeon leds from an lm3915 output? tnx
profpat3 years ago
great tutorial, will try it out got lots of LED BULBS for projects
sprocketme23 years ago
I understand that R3 is calculated as R=(Saturation voltage of Q1)/(desired current through LEDs) If I am wrong please correct me. Now for circuit 5 specifically, is R1 still functioning as a pull up (100kohm) like the other circuits, or is it functioning as gate resistance for Q2? In which case it would be in the order of 25-100ohm.
i need to kwon how calculate the set current, because i have a 18v and 3,5 A power source and i want conect 4 red led 3w and a blue one 3w at 700mA each one
johnjuan3 years ago
Not sure if I understand right - does the Zener in fig. 3 make so the circuit will output more than 20v, or just intake more than 20v? I want to connect to 24v supply. Will Zener be enough, or do I need a better NFET? If I need better NFET, can you suggest one capable of 24v or more?
toyds3 years ago
i want to make a simple high power led driver for this one

Input supply : 12V-13V motorcycle battery
# of LED : 2 LED in series at around 350mA

does this driver be possible to use?please reply
thank you very much
rguzmán3 years ago
Hi, i just tested it with a IRF520...

works fine, too cool really

the IRF520 (at 800mA calculated R3) was "cool", no heat at all

so I tested the consumption of the three 3W leds I use... only 540mA with R3= 0,66 ohms

so I changed R3 with a New Voltage Drop value... to get close to 800mA. R3= 0,45 ohms. But Consumtion was only 570mA.

could it be that with IRF520 or any mosfet I would need to use a really low R3?

or should I test another NPN (I'm using MPSA18 which has much more gain than 2N3904)
the Voltage at GATE from IRF520 was 10,2 , but at SOURCE onle 0.29V...

maybe here is the problem,.... as I read that on BASE from the NPN there should be more than 0.5V for it to work right.
kmantesla4 years ago
Great instructable, I just threw this together to get the hands on experience of a current source. I designed single led desk lamp driven at just under it's max rating 1.0A. I am very pleased with the results, and I wanted to chime in on everyone's rants on efficiency, because most people don't understand this circuit and real world efficiency. Granted, it's not the most efficient circuit in the world, but it does EXACTLY what it's designed to do, change your input to an exact output, keeping the leds happy and healthy.. but at a cost.. all the "changing" is converted directly to heat, not an ounce is saved.. My circuit uses a switching wallwart; 5.2v @ 1.0A.. we can safely assume this type of adapter has an efficiency of 85%, if not more. My circuit "inputs" 5.3v @ .90A and my load, the leds get 3.0v @ .88A. Yay, this is exactly what it's designed to do, my led is nice and BRIGHT and happy, but all that stepping down and regulation doesn't magically convert, not with this circuit, it's converted to HEAT. So, my "regulator" circuit has a weak efficiency of about 56%, coupled with the wall adapter efficiency, the total circuit is around 50% efficient. Yikes, thats a scary number, right? But this is why we use leds, because an amateur can still be "green" even with a lossy, dirty regulator like this.. My lamp is SUPER bright, at less than 6W. I guarantee you 10 of these would EASILY rival a singular 60W bulb, in brightness and cool factor no doubt. Plus, I built mine out of spare parts, you could easily increase efficiency ALOT by designing a circuit with specific values on the adapter and fet.. and my led was an old rebel, newer ones are even more efficient.

Great instructable! It reignited my desire to custom build drivers over store bought ones, so I can start to really push leds, like the newer XML's to 3amps..
Following this build, I immediately bought some high quality, low dropout, high EFF adjustable voltage regulators with built in buck/boost options all in a T0220 package, got them for around 2$ apiece. I currently use luxdrive's drivers for all my projects, but now I'm going to start playing around again and see if I can come up with something just as rock steady and functional as my beloved buckpuck.

Dear Sir.Your right and your wrong also. The right part you have figured out and said yourself.The wrong part is that when your using 5 v @1a input to suit ur 1 led need that entirely is your choice,but you shouldn't calculate the heat loss or efficiency in this instance as you have not yet used the fet to its minimum potential also.
I say this because when u run 3 or 4 led of 700mA each on a 13 v dc 3 amp supply.Dan's simple but polite circuit doesn't even heat up and it remains cool and efficient for even more that 36 hours at a stretch (my bench test) . It only heat up to waste voltage and current when you use 1 or 2 leds or 300mA or 700mA.
Dan's circuit is superb in almost all ways.Sometimes the most simplest things are the best things in life.
Well yes, but in the above post I was mentioning the "excess voltage"... 3 or 4 leds is going to be pretty much RIGHT at 13v so you're not going to have to drop much if any, hence your driver is probably doing just about nothing and you'd be better with one series resistor. In my case, if I hooked up just the led to my 5.3v source the led would be instantly fried, and this circuit is used to drop the excess, which gets converted to HEAT :)
-I wanted to add; people trying to calculate efficiency using any AC settings on an ordinary multimeter (less than 100$) are probably not getting accurate results... at all... Use DC for all measurements, and estimate the mains to circuit supply adapter efficiency.. Depending on the type of adapter and load, assume it can vary as much as 65-95%.. Good AC meters and true rms meters are hundreds of dollars. If you understand them, generic meters are basically worthless for AC measurements outside of simple residential mains checking.
Without going through the laborious process of reading all 256 replies dating back several years apparently...

Would a recycled computer power supply converted to a benchtop power supply do the job?

You have a complete set of tightly controlled AC and DC voltages that could be used.

I raided the dead computer repair shop and collected quite a few and tested them. Most still performed flawlessly.
This is exactly what I have just done.. a 350w antec, and will be spreading my led's over the 5v, 12v1, and 12v2 rails. If just using palin resistors, and splitting up over 3.3v, 5v, 12v1, and 12v2.. I calculate 92% efficiency. That's for 127 Cree's at 350mA. But I want PWM capabilities.. so went with this driver. Can't run it on the 3.3v rail, but 5v and the 2 12v rails have plenty of amperage to offer. Just don't go over the maximum 10 amps the psu's 18 awg wire's good for. But at 4 or 5 lines per rail, that's still quite a lot of power to be tapped. :D

I mentioned in my last post that it wasn't working on my 5v line.. it is.. must've been a ghost. And it's working with just the 100K R1 resistor as well, didn't need the 22K.. nor did it need the zener. At least not with the 5v and 12v sources the psu gives.

Just need to look at the specs on the psu (usually on the side, or on the site). Make sure you don't go over the maximum amps per voltage rail, and note that 3.3v and 5v rails are "combined wattage" maximums. And some have "combined maximums" on the 12v1 and 12v2 rails as well.

As for the 10ohm 10watt "dummy resistor" needed for the bench supply mod.. I found using an led and 470ohm resistor on both the 5v and 5vsb lines for the "power on" and "standby" led's. Uses about 1/2w. If you look on your motherboard, you'll see (or should see) a green led.. that's your pc's "dummy load". That's also connected to the 5vsb (5v standby).

For 127 led's at 700mA, I calculate about 15 drivers over the 5v and 2 12v lines.
3 leds work best on the 12v lines.(3 volts or less go for the consumption of the cc regulator) So dont bother with the 5 volt or lesser voltage lines. depending on the ampere required play around with R3.
Cree XP-E Reds use 2.3V a piece.. 2 LEDs per 5V line wastes less than 5 LEDs on the 12V line. The Cree XP-E Orange-Red uses 2.1V. Osram Golden Dragon Plus "Hyper Red" uses 2.15V (@ 400mA).

So if you "only" use the 12V line, then you have wasted voltage. You'd still have waste using the 5V line, but less. And although it's only 2 LEDs in series on the 5V line.. you can wire up more in a series/parallel set-up. There's quite a bit of Amperage available, so even though each series is only using 4.2-4.6V per 5V line, you can still fit up to 9Amps worth. Each voltage line is good to 10Amps, but use less just to be safe.

Solution 0: 2 x 3 array uses 6 LEDs exactly
+----|>|----|>|---/\/\/----+ R = 1 ohms
+----|>|----|>|---/\/\/----+ R = 1 ohms
+----|>|----|>|---/\/\/----+ R = 1 ohms

Also.. don't forget.. you need a "minimum load" on every rail.. 3.3V, 5V, and the 12V. It'll work without it, but it's safer, and better for the psu, if you ensure the minimum load for each rail.
I think you just raised my IQ by at least 2 points with that response.

THANKS My Friend. Good information to keep handy!
You're very welcome. :)
rguzmán3 years ago

i'm trying to calculate the dropout voltaje of a JFET or MOSFET...

any help to build this?
csepzol3 years ago
Can anyone help me to calculate the R1, R2, R3, R4, Q1, C1 with an lm317?
nrao23 years ago
I need to rig up aLED white cold light source of approx 100w, what best circuit you would recommend & the White LED part.
bau053 years ago
Hi Dan

I use a group of 32 Cree MC-E leds(4 in series x8 in parallel), mounting in a XLarge heatsink with fans.
An IRF3710(Vd=100V,Rds=23mΩ, Ιd=57Α) drives these leds.
All work fine, except that leds open/close "unevenly" during first/last steps(Sunrise/sunset mode)
(it isn't a PIC problem, cause i tested with four 5mm simple leds that worked "smoothly")

Q: Can i replace the two resistors(100R, 10K) with your circuit and drive safely the leds?


*See IRF.sch below.
It sounds to me like you just completely paralleled the led's.
Here's the problem:
a constant current source works by changing voltage to stay at set current.
because of this, you have to make sure each led shares current evenly, as they each have different turn on voltages. you can do this by putting a small resistor in series with each one. (10 ohms) The ideal solution would be a separate driver for each, but that might be expensive.....
A friend gave me yesterday his ebay cheap Led Wireless Dimmer(with remote ctrl). It uses an smd ATMEL microchip inside.
After testing this i noticed that ATMEL doesn't minimizes the light at all, plus remote ctrl step to step cliks are different.
So with a digital voltmeter, i deside to measure the PWM output steps.
1)My PIC diy dimmer steps: 0.00>0.01>0.03>0.05>0.07>.....>5.02 Volts
2)ebay Atmel dimmer steps: 0.02>0.04>0.08>........>4.98 Volts

That's the answer:
the different voltage PWM steps in each microcontroller's software.
A smooth opening light mode should have 0.00>0.005>0.01>0.015>0.02>0.25>....voltage steps(maybe this will be done with an analog potentiometer).
I guess that there is no solution, except someone re-programs the .hex file. Cause i dont "speak" the PIC-language, my circuit will stay as is.
Anyway thanks for answers.

*To help others:
A 10x10cm IRF heatsink gone useless. IRF+heatsink were TOO HOT!
I cooled the burning IRF, mounting it to a cheap 5x5cm PC VGA heatsink fan!
IRF now has the room temperature!
polilies4 years ago
Hi dan!
firstly thanks for your project!it's great, i have try it few days ago. and it works with one led.But the NFET heating too much . Do you know something about it? i've try it with 5v zener. and when i tried the circuit with three power led it wasn't heating but the leds were not bright enough.actually too dead.. ?
whitehawk824 years ago
I ran across this post recently as I kept buring high power LED's or had very hot resistors. So, I tried this method, I already had a fet connected in the manor shown in Circuit 5. So I disconnected the Source on my fet and put R3 (3x1ohm 1/2 watt resistors) and the Q1 parts in. For the FET @ Q2 I was already using a NTD5806N and had a 3904 transistor laying around to use for Q1. R1 from my ATTINY85V was 500ohm as anything more and the FET doesnt fully turn on. So, I rigged this up and tried it to limit the current to 1.6Amps on 13.8vdc,. I fried yet another LED. When I put my meter in the line I was hitting close to 8 amps of current! I know I did the calculations right.. Any thoughts and wisdom?
suguimoto4 years ago
Hi Dan! Thank you very much for your article!
I´m building my own RGB LAMP, I bought a 9W (350mA per color) RGB 6 Pins (Chinese-Star) Led and i´m using an Arduino to control it.
My main problem was that I couldn´t find many of the IC described above, so I bought similar IC´s that are working quite fine:
Q1: BC 547
R1: 100k 1/4W
R3: 1ohm 1W

I really dont know how efficient it is, but it´s working fine and the components didn´t get so hot.

I hope this can help some other people that are willing to work with high power leds.
Can u pl's share the circuit diagram of what you have done?

can it be used with 1w LED?
adi19774 years ago
My dear friend

I dont have enough knowledge in electronics so i'm asking you for help. I want to make a LED Tube on a PVC batton with 1watt or more wattage of an led when given a power supply of direct 220 - 240 Volt so that a single tube lightens up the entire room.

My dear friend can you please design a circuit for me and mail me along with all the parts to be bought and their specification . (i.e. resistance of how much wattage, capacitor, diodes etc.)

A little help from you will be highly appretiated.

Thanking You

Adit Pal
erikp4 years ago
Can someone help me out regarding the proper heatsink to use for Q2? My LED (part number 475-2581-1-ND on digikey) draws 1.4A with a forward voltage of 2.5V. I am using a wall-wart that gives me about 10V. My LED runs smoothly, but the two heatsinks I've tried using with Q2 get *very* hot to the touch. I bolted both of them on using thermal compound. Both are "U" shaped. One is about 2cm tall by 1cm wide (about the height and width of Q2). The other is about 3cm tall by 2.5cm wide.
burton26634 years ago
i have a vu meter using 3915 with 10 small LEDs, im trying to find a way how i can convert those little LEDs to 3w LEDs using simple transistor drivers. any diagram is appreciated thank you very much
drbogger4 years ago
Will a 1N4733A 5.1 volt zener diode with a Voltage Forward (Vf) of 1.2V @ 200mA work?

There is also one that doesn't specify a Vf... Which one should I get?
bwrussell4 years ago
Maybe I'm just missing something but how do I determine the value of R2 if I'm not concerned about interfacing with a microcontroller but need an input voltage greater than 20v?
letniq5 years ago
Hi, Great article but I didn't understand how to calculate the how many ohm should be R3. For example I'm having 36 3W power leds. 12x3 leds Source: 12V PC PSU LED Forward: 3.2V LED Current: 750mA I will have 9A through Q2 (IRF3205) Q1: 2N3904 R1: 110K It will be great if someone can help me. Thanks.
wolf1728 letniq4 years ago
I see my website has been referenced. A month ago, the domain name 1728.com was stolen from me and so that link should no go to parallel resistor calculator
pmj_pedro4 years ago
If i use one of this circuits like the 5, and d the supply font is 18V, and the 4 high power leds only need 13.6V, should i use a resistor or something to not burn the leds?
No, the point of these circuits is that you can use most any power supply that meets the LEDs min total forward voltage without changing the circuit. Q2 acts as a variable resistor and steps down the voltage. Read through page 6 completely, it's explained more completely there.
karthik_dm4 years ago
Thanks a lot!! your circuit works well..
Hello Dan...
Thank you for the nice and simple circuit which works very well!
Dan, I have a question, I'm using circuit #5 with my microcontroller, but, even when output is 100% low (off) or even hard tied to GND the LEDs still a very little lit, touching the underside of the driver board with bare fingers makes the LED to light strong.
I'm using IRLZ14, BC547B and 0.68 resistor as current set..
uC is AT89S52, 4k7 pull-up, LEDs are red 750mA 2v (series of 3) supply is 12v.
Thank you.
(PS: changing IRLZ14 to IRF640 it works well, but IRF640 is not logic level :( I worry to burn my uC )
ArtemKuchin4 years ago
I am thinking about using this schematic to power just 100mA RGB LED (pulses 1/12 of time). But the power is only 5V and Vf for blue can go up to 4V (for 20mA). So, the question is: do i have enough voltage here? As i understand the minimum voltage drop will be
V across R3 (Vbe which can be 0.7V) + drop across Q2ds which probably can be ignored because it will open fully when needed and Ron is 0.021 Ohm.
But the led Vf can be higher than 4V for 100mA and who know how much higher. But it seems like i have only 5-0.7-4=0.3V headroom.
Do you think this is not enough and i should switch to higher voltage power supply? For my project this means 12V. If so, i would probably go for switching regulator because of power dissipation on the FET.
I not even sure that with 100mA the Vf will be within 5V :( No graph for such case in the datasheet, but 100mA is allowed with 1/10 PWM. Any idea what could it be?
oranges0da4 years ago
I'm looking to power 1 or 2 1W LEDs with a 3.7V LI-ion battery... will this circuit work with it?
Would this circuit need to be modified to run multiple parallel high power LED's? if so, could you help this beginner out?
cegu4 years ago
Great instructable, but doesn't work for me. I only have 3V (2 batteries), well 2.3 when they get empty.

Would you suggest to boost the voltage or just make them paralel?
Spuzzum4 years ago
Fixed the rising current issue.. needed to tie the free leg of the trimpot to the middle leg. Everything's stable now.. and adjustable from almost nothing, to well over 1500mA. No "set resistor" required :D.

Still won't go very high on 3.3v and 5v sources though, and I even removed the zener diode.. reading back, I noticed that was for 20v sources and higher.. obviously not needed for 3.3v and 5v.

Oh well.. a 12v driver's still good enough I guess. Thanks man.........
Spuzzum4 years ago
Well, built the circuit.. with a few mods. Added both the 100K and 22K source voltage resistors, selectable by a DIP switch. This way, the board's good for 10v or less, as well as 10v or more. I also used a 100K trimpot in place of a set resistor for RSense.. I want it adjustable. Problem is, anything above 300mA, and the thing just keeps on climbing.. on it's own! That's on the 12v rail of my mo'd psu. The 3.3v line won't go above 500mA for some reason.

Could it be the pot's just not rated for such a current? I used the exact n-fet and npn transistors.. local shop sells Fairchild.

Where'd I go wrong?

LED Driver3.jpg
Spuzzum4 years ago
The LM317K is rated at a "minimum" 1.5A, "typical" 2.2A, and a "maximum" 3.4A.
Spuzzum4 years ago
 "perhaps you'd like to put a variable-resistor dial for R3? unfortunately, they don't make them in such a low resistance value, so we need something a bit more complicated to do that."

I've been looking at modifying to be adjustable.. I'm thinking you can use a 10ohm 25-turn trimpot:


Then use the pot in "parallel" with the other values needed to create 1000mA, then turn the pot "up" to increase the ohms, therefore lowering the current. (4x) 2ohm redistors, in parallel.. gives 0.5ohms. As long as the trimpot goes down to 2ohm.. then you just need (3x) 2ohm resistors, as well as the trimpot. 


Sounds good anyways................ :)
do you think you could drive an sst 90 off of 1 18650 using this?
phidelt854 years ago
Great read and easy to build.  However, I'm having some trouble with my build.  I used Circuit #3.  I'm powering 6- 3W Luxeon K2 Vf=3.65V.  I'm using an R3 value of .56/2W to drive the 6 LEDs at 1A.  The problem is that when I take my amp reading, I'm only getting a reading of 166mA.  I'm showing 3.3V on the G leg of the FET.  I'm not sure what I'm doing wrong.  I get 166mA whether it's one LED or all six in series.  Any help would be appreciated.
Any help would be appreciated... Another question I have regarding the supply voltage to the FET which is currently at 3.3V, do I have to reduce the R1 value in order to increase the voltage to get the FET to turn on fully?
SpiffyChee4 years ago
So I'm confused about which diagram I should use for PWM with my arduino uno microcontroller. Should I use #4 or #5.
Same goes for if I use a 555 chip, would i use #4 or #5?
could someone please enlighten me?
I'm no engineer, but in my opinion it doesn't matter too much with an arduino. Their pins can both source and sink regulated current. If you're using it set up as 5v then #5 would be easier to hook up and program. If you're forced to use the 3.3v setting or other analog stuff like a 555, then #4 is going to better for a full range of control, however I suck at coding and for some reason I'm pretty sure it would be a nightmare getting what I want with #4 because you're going to have to use pins as inputs and outputs whereas I think #5 is a simple 5v on off situation, floating pins and imperfect code would not be (as) big of a deal.
"if your micro-controller is 3.3V or less, you need to use circuit #4"

Arduino Uno's digital out pins are 5V so you will want to use circuit #5. Same goes for the 555 as long as you are running it at 5V.
I understand these circuits when using the same LED's but I have 3 series of 3 LED's wired in series. 2 of those series chains have a different voltage drop across them then the 3rd. I am assuming I need to add resistors to the front of those chains to equal out the voltage?
acidjc4 years ago
I have 10 x 10W LED, Vf = 9V and current is 600mA.

I plan to put them 2 series and 5 parallel, using a input of 24V DC,

ie. 2 x 10W in series = 18V, 600mA
5 row in parallel = 18V , 3A drawn from the source

Plan to use the above "even simpler current source" on each row to limite the current flow to each row. Any problem?? do I need to change the Votlage regulator to other part?

can anyone pls kindly advise. Tks
larams4 years ago
Does the power dissipation of R3 depend on how much of a voltage drop there is? He says .25/R3 will tell us what wattage we need but doesn't really say where the .25 comes from. Does that formula change depending on the supply voltage?

dan (author)  larams4 years ago
P = V^2 / R. V across R3 is constant at about 0.5.
larams4 years ago
Does the value for R1 matter in circuit #5?
dan (author)  larams4 years ago
not much, 10k should work.
Where can i buy a current set resistor (labeled as R3)?
popeyescu6 years ago
Hi Dan,

it seems that the link to "555 circuit" doesn't work. Do you have a schema/link for this circuit?
I found another schema for 555: http://www.dprg.org/tutorials/2005-11a/index.html
Is it good?

You can find the 555pwm ckt here:
Trimaran4 years ago
5V zener was not enough in my case. I got 0.17 ampers while i should have get .32 according to my configuration and with one 5V zener, so I thought maybe mosfet was not working normal with 5V gate voltage and put 2x5V zeners in series to make it 10V and it worked. Btw, I put them in series just because there was not 10V rated single zener in my stock.
conntaxman4 years ago
Question about a 20 watt led.I just got a 20watt led ,it's the type that has about a 1 inch square back plate with the +and- on each side.It's13-15 volts draws 14 to 1700 ma.or 1800. I bought a wall wart that puts out 14.5vdc. I have ran this on 12 volts ,it ran good. I hooked it up to the wall wart, with a 1 ohm res. in line. It was nice and bright [white in color] had it on for about 30 seconds and them moved it alittle and then the led went Dim a little bit.Say it started out at 1500LM and then went to about 1200lm,This was all in about 45 seconds. The next day I turned it back on and all the Small leds[20 of them] were lite but still dim.Would anyone know what happened to the led?
My last test showed it still was drawing 1.7 am;and 14.5 volts dc. It still lights.but dimmer ,kind of like if you hooked it to 12 vdc.
I know when you blow them out they Don't work. lol I have done that.
kjjohn5 years ago
In the schematic, the LEDs appear to be wired in series. I have multiple 3v 700mA LEDs, but a 3.7v power source, so I would like to wire the LEDs in parallel. So do I calculate the R3 value for 700mA * the number of LEDs?
switch62 kjjohn4 years ago

You should not put the LEDs in parallel.  It sounds good in theory, but there is no control as to how much current each LED will draw.  You will only control the total current.  That means that if one of the LEDs starts to draw a little more than the others it could overheat and then blow.  If that happens the current in the other LEDs will rise, the circuit keeps pushing the same current, and they will all eventually die.

In your case I would recommend a separate driver for each LED.  Circuit #1 would work but may not give 100% brightness.  If you had 4.5-6V you could guarantee 100% brightness.  At a few dollars each it would be the cheapest and easiest solution.

The other option is to use a boost driver to raise the 3.7V to something higher.  The LEDs would be in series.  The required voltage would be at least Number of LEDs x LED Voltage.  So 4 LEDS at 3V = 12V  the currrent setting would be 700mA.  Unless you can find a suitable design to build you'll need to buy one.

damidb5 years ago
hello, I have a problem using dan's interesting LED circuit for a high power led. (I want to use it as a better lighting source then a bulb for my microscope, using a 3W LED 3,2V 1000mA). I included the circuit as png file so you know what i'm talking about. I first simulated the circuit with LTspice IV simulation software (see other picture). The problem: Dan mentioned on instructables: "The combination of R1, R2 and R4 set the output voltage (= LED brightness) of the regulator. the output voltage is approximately equal to: 1.25 * (1 + ((R2+R4) / R1))." So I calculate with my chosen values R2, R1 and R4 (this is a potentiomter from 0 to 2.2 KOhm) as output voltage for the voltage regulator: 1,25 * (1 + ((3188 + 2200)/3580)) = 3,13 V But in LT spice simulation (see .png picture) the output voltage from the voltage regulator is quasi constant and independent from R4 (R4= the potentiometer) (see picture please) and about 1,48V. instead of 3,13 V. I checked everything. What did i do wrong ?? Or is the simulation software LTspice wrong ? thanks in advance for any answer.
Thanks Dan, just whipped up #1 and it works great. One note: The FQP50N06L FET was out of stock at both Mouser and Digikey, so I subbed in the FQP30N06L. I think the only difference is that I can't handle as much continuous current. I never plan on drawing over 10A so it is not a problem.
Here is one in action. You guys may recognize the LED array
Cairie5 years ago
I only want to have ONE Luxeon 1W, how should I change this schematic?
grimgroper5 years ago
any ideas to run 20 10w 12v LEDs off in a car i want to run them around 700-900mA?
mertg5 years ago
 Ok. I have a question about this topic. Some people posted comments about this circuit, being not so efficient. I want to ask them, how to measure the efficiency. I built this circuit with an IRF630MFP  and a C3198 from my old CRT monitor chasis.  I put 2 ammeters. 1 for LEDs, 1 for the whole circuit. If I apply 12 VDC and drive 3 PowerLEDs with 349mA, I read 351mA from the ammeter connected to the main supply. The FET is not connected to a heatsink and no heat is produced on the FET. It supplies constant current for leds. Thus can I calculate the efficiency like 349/351 ???
mertg mertg5 years ago
I think I've missed some points. if i calcutale P_in and P_led (P_led / P_in) it becomes 3,35Watt / 4,2 Watt. Thus the circuit is 79% efficient. Is that correct?

Thank you.
godofal5 years ago
wow, great instructable!
this is really something il keep faved for future use, its really neat, handy, and all in 1 place! (instead of being scattered in 20 instructables, by 15 ppl, with all theyr own tricks and tips :D )

thnx for this, its really helpfull!
ac-dc7 years ago
This is another very lossy circuit, not what one would want to use to power an LED except as a good learning exercise about linear regulator control. As xsmuft mentioned other regulators could be substituted but I would suggest using an LDO, Low DropOut type as you then don't need as high a supply input voltage. However there is a problem with this circuit. Suppose 3 x white LED at 3.6V forward and 1A current. That's 10.8V total and at up to 28V input to the regulator we have 17.2W of heat!! That is a very unrealisticly sized heatsink and in fact any use with linear regulators driving 1A LEDs would require a larger assembly and more costly heatsinked cooling strategy. In other words, it's just too lossy. It might be a better way to hack together a circuit for driving a few encapsulated 100mW (~ 20mA) LEDs instead of high powered 1-3W LEDs.
dan (author)  ac-dc7 years ago
if you want to maximize efficiency you need to use the correct number of LED's based on desired input voltage. for example if you have 24V input, then use 6 LED's in series to minimize loss.
ac-dc dan7 years ago
Agreed, but in most situations this is the reverse of the way things work out, that one has X # of LEDs they need, and Y PSU voltage to work with to drive them.
 ac-dc, you regularly chime in about loss and efficiency.  What would you recommend as an alternative?  I was looking into the STCS1 from SMT Electronics used in another instructable.  That is until I realized the IC itself was 2 x 3mm.  I can't work with components this small.  All the other IC drivers I found suffer the same flaw.  I need something I can build with my hands and a soldering tool, not a commercially fabricated PCB, a microscope, and re-flow oven.  Seriously, I am killing myself to find a do-able driver.  

I am trying to run 12 1 watt LEDs with 3.7Vf off of a 12v motorcycle @ 900 mA. I realize I will probably have to have 4 drivers each running 3 LED's.  Please, any suggestions for an electronic noob?
Yakumo5 years ago
I don't know how to say it but the Instructable was a little bit confusing even if the schematics are easy to understand. Please don't get me wrong but the "parts used" and "setting the current!"-part are the most important and its kind of hidden in this much extra information. ordering the textparts a little bit different will make it easier for the reader ... I think^^
Thanks for the Instructable made me read a lot about stuff I didn't understood before oh and my LED's are actually emitting light now ^^

@Everyone who cares
I replaced Q1 with an IRF1010N (because of the lack of Fairchild FQP50N06L in the German market) and Q2 with a 2N3904 (hat too much lying around) and it works nicely
zspzs5 years ago

Dear Dan,
Thanks for the article!
I would like to use your current regulator to provide 5A. (For a DC motor.)
Based on the above calcularions in my circuit R3 should be 0,1 Ohm / 2,5W.
Should I modify anything else or that's all?
Can I use 2N5089 instead of 2N5088? The supply voltage is 18V.
Thank you in advance for the answer.

Orikson5 years ago
Hey there,

First of all, I guess the symbol for Q2 is wrong. You write, that you are using a N-channel FET, but in the scematics there are p-channel FET symbols!

Also I'm trying to build this circuit, but it isn't working as I expected. I'm tried a 100k and 1M for R1, 0.85 Ohm (5 watt) for R3 (should be around 600mA), white power LED (3,4 - 3,6V @ 700mA), BD548C for Q1 and IRL540N for Q2 because I already have them. The power supply was a regulated powersupply. I set it to 5 Volts and set the current limiter to 100 mA which I increased slowly up to 600mA.

The problem is, that the voltage wasn't regulated by this circuit! In the end the LED got 4 V at 550mA!

So where is the problem? My Q1 and Q2 seems to have nearly the same technical data, don't they?
Q1 & R3 set up a clasic constant current supply (by def voltage will vary and is not regulated).  The base emitter junction drop is approximately 0.7 volts, as the author noted, this voltage will vary as temp varies (temp variation of diode junstion - exp of absolute kelvin temp).  The current is set by 0.7/R3.  He set the circuit to effect the gate voltage and allow the FET to regulate the power (heat), the R3 power drop is constant i^2 * r, but the FET power will depend on the current, led voltage drop and supply voltage P(fet) = (V+ - V(led) - 0.7) * (0.7 / R3).  As you noted the drop of the LED is not as listed, welcome to the real world.  This is usually a result of heat build up within the LED, but could be normal for your LED. 
While this circuit is not the most efficient from a power view, it is easily one of the best to understand and sets the current with a minimum of components.  This is a good circuit and should perform well with a wide power supply tolerance.  Just monitor the heat dissapated by the FET.
xsmurf7 years ago
In the FET circuit, can I dim the leds by changing R1 to a potentiometer or multi-position switch?
dan (author)  xsmurf7 years ago
no. see the dimming options in the following steps
gopskochi dan5 years ago
cant i adjust the brightness by connecting a pot instead of R1, pls can u brief the theory behind how R3 adjust the current in the led.
bobbyg dan7 years ago
Dear dan, I am designing a light for argicultural applications I want to use 36 k2s red 700 ma 2 watt at 75 lm and 10 royal blue at 620mw 1500ma appx 7 watts I am disturbed because they are charging me almost $300 for the materials and only $100 of it is actually leds the drivers are so much I want to do a 24 x 24" box w/6" overhang where do I start?
gopskochi5 years ago
can i adjust led brightness by making R1 a variable resistor, also i am using a lead acid battery, which will not be connected to any charging circuits when lighting leds. will the brightness remain constant for a few hours as the battery discharges.
lamikam5 years ago
Hi.  I want to use circuit #5 example, but I bought some RGB leds from Hong Kong (cheap) but they are all common anode instead of common cathode.
How would I change circuit #5 to handle this?

osmana5 years ago
hi all,
i want to drive 12x3w power LED's. LED's forward voltages are 3.2V. Total i need 38.4V. How can i drive it.
conntaxman5 years ago
I want to 3watt led, forward voltage 3.5/4 at 700ma draw.Will a 7806 IC work.I want to put the led in series.guess the cir is just the basic.Also what is the most of the same leds that you could run in series. It will be of a 12 dc source.Tks
12V to 6V converter by 7806.JPG
andybuda5 years ago
i want to make a few leds that get charged up in the day and switch on at night ... i need a solar cell photo cell, relay switch and rechargable battries any ideas how to connect the thing up its to light up a plant at night and charge by day...
doctek7 years ago
How did you make these measurements? One way I can think of is to put appropriate current limiting resistors in series with a led and measure the voltage drop across the led. Is that indeed what you did or is there a much cleverer method I'm missing?
Nyxius doctek5 years ago
LED are incapable of limiting current on it's own.  Therefore, voltage drop across the LED is dependent on total potential and total load in series with the LED.  Too much potential and not enough load means that the voltage drop across the LED will be high enough to "fry" the LED.  This is because LED have very small impedance compared to regular light bulbs. The voltages listed on the spec sheet are recommended drops that will ensure long lasting lifetime.
kreston5 years ago
 Hi there. I have a 70watt single led. In the item description it says 24-28V 2800mA

I'm planning to use LD1585CV with 30v input (220v to 30v transformer with bridge rectifier)
like in this webpage : www.electronics-lab.com/projects/power/016/index.html 

do you think its ok?

conntaxman5 years ago
Hello Dan and everyone else.Im trying to find a cir, to run 2 of these 3 watt leds, and I looked all over.I will be useing a 12v car battary for the source power.I know that the voltage can go up to 14vdc.here is the spec.for the 3 watt led.voltage in 3.0v-3.27v Current 700ma - brightness 80 Lumes - color temp 6700k- .This is going to be used when the power goes of in the house.I want to put one of these in each room for light.I was thinking of just putting 5 in seires and that would be 15 volts.I know about the brightness will varry,but I dont think that it would be that much.Also I know I would need a heat sink on these.So I guess I would like to use 3-4-5 of these in one light.
Thanks for the help and I did read many of the cir. here.This is not my trade,And that is why I ask for help.LOL.
I did build a wind-mill from this site also.This is a good site,with many smart people.and Friendly people.
Thank you all and have a nice day [or] night.

mrmarshall15 years ago
does anyone know where I can get a driver for this


led? i want it to run off 120 volts if possible, but 12 to 24 volts is fine to. if you could do this it would be greatly appreatiated for my first Instructable.
Re-design5 years ago
Great info. I've got it bookmarked so I can really study it later.
Mezbah5 years ago
Hi Dan Thanks a lot for the instructrable. Your circuits are very nice. I would like to know if I can drive, say about 200 white high bright LED directly from 220 AC by any of your design of this instructrable? What modification, if any I need for that. The idea is to drive as much LED as possible and practicable for my new LED tube light project. Thanks alot in advance and have a good day
DarkRage36 years ago
While I'm not one to normally take the quick and easy way out, however I was hoping to get a suggestion on a SIMPLE LEDsetup for a light-box I'm constructing for my son. Space is very limited and there are currently multiple variables that are up in the air because of my lack of knowledge in the subject. The box would have at least two settings. I'm completely new to LEDs and have only rudimentary, and mostly theoretical knowledge of circuits. Normally I'd do extensive research for a project like this, but time is not on my side right now. ANY help would be appreciated.

Possible parts I've been looking into include:

-random DC power adapters ranging from about 5V - 19V

-a wide range of LEDs (because of my lack of knowledge I can't choose)


--5mm InGaN White Slow Burning* wavelength(nm): 6000~12000K
Vf(V): Typ.3.2 Max.3.6 mcd:400~1000
half intensity angle(deg):85°~100°

--5mm White Slow Burning* wavelength(nm): 6000~12000K
Vf(V): Typ.3.2 Max.3.6 mcd:1000~2000
half intensity angle(deg):60°~70°

The exact number of LEDs to be used is also up in the air because I'm uncertain of the brightness and uniform distribution of light. Best guess, (based on the number of LEDs on a prefab. LED strip I found) 66 - 108 LEDs

I'm attempting to use the two diffuser screens of an old projection TV (it seems like it should work well with the lights shining through the sides, toward the center) and an evenly white frosted glass from an old cabinet (looks just like the real light-box diffuser screens) so the number could be considerably less, however I have no single LEDs of comparable power to test it.

Again I'm unaccustomed to asking for help, but ANY at all would be greatly appreciated, Thank you.
These circuits are great. If you want to interface it to a microcontroller, you should really use a logic level FET. Otherwise make sure that the turn-on voltage is low enough. A lot of commonly available MOSFETs don't fully turn on until like 8V. That's why the IRF540 mentioned by gerojalo won't work. It will work, just not with a microcontroller or a lower input voltage.
starwarts6 years ago
Hi Dan, Can a TIP122 (BJT NPN with VBE of 2.5V) be used inplace of the NFET? What needs to be modified? Thanks
Peregrine76 years ago
Hi Dan,
Thank you very much for so much good info here. However, I'm still at a loss on how to proceed with my project. I was wondering if you could help me fill in the gaps. I think the analog adjustable driver would be the best option in my case.

I'm doing a circuit that will use LEDs in car taillights. So the input power will be a standard automotive 12V battery, which usually has a voltage that varies between 11V -14V. The LEDs will have 2 stages: full brightness when in stop light mode, limited brightness when in parklight mode. There'll be 4 identical lights, so let's concentrate on only 1:

The driver for each light will accommodate 2 series (in parallel) of 4 LEDs in each serie. So 8 LEDs total. The following are specs and parts I already have on order:

Regulator LM317
Red 1W LED, 2.1v-2.3v, 300mA

4 of the above LEDs in serie is (2.2 x 4) + 3 = 11.8v necessary to support the LM317 regulator.
Since there will be 2 of those series in parallel, it will require 1.25v / (0.300A + 0.300A) = 2.08ohms R1 resistor (I believe closest is 2.1).

1) Can you help me fill in the gaps for the rest of the parts in your circuit (R2, R3, R4, Q1, C1) ?

2) Do you think this would be the best solution based on my above specs? The 2 stage brightness is what complicates my circuit. The R1 would be enough if I only wanted to run them at full brightness. But since I also need them to switch to limited brightness, I think that's what your circuit can accomplish. Initially, I thought I was gonna just do 2 inputs on the regulator separated by diodes, and one input would be with some resistor that would cause the LEDs to be dimmer, but I'm not sure if that would work because I don't know how the regulator would affect that.

Thank you in advance!
rhosman6 years ago
if I changed Q2 to something like http://www.fairchildsemi.com/ds/FD%2FFDB52N20.pdf and added a Bridge Rectifier that was 120V 20A rated I should be able to hook this circuit up directly to a US power main? Maybe through a light socket?
TS846 years ago
Dan can u help me,.... if i'm using IRF530 and 2N3904 for this circuit.is it works fine?can you give me advise because your listed parts are hard to find......
jasmclbs6 years ago
Hi Dan, Can you kindly help me with the components' selection? Do you mind sharing with me the part no. of Q1 and Q2 that you have used? As I am adopting circuit #5, what will be the preferred R1 and R3 values for an 5V output from my MCU I/O pin. The voltage supply to the LEDs will be +5V too. For my LEDs, I am using OSRAM LUW W5AM which has a IF of 350mA and VF of 3.2V. (White) Please kindly advise. Thank you in advance.
soovui6 years ago
Dear friend,

I like to make use of this circuit for my LEDs about 18 LEDs in a row. It groups into 3 LEDs in serial and 4 parallel (3 LEDs in serial) in a row of 18 LEDs. Each LED is 1 Watt x 18 = 18watts in a row and also LED is 3 Watts x 18watts = 54 watts in a row.

Anyone can help me on what is the right resistors value for LED circuit 18watts and 54watts?

Your advise and help is very much appreciated.
Thanks and best regards,

ac-dc7 years ago
Efficiency of this circuit is usually as bad as with a resistor. The one large benefit is the variability to different supply voltage like when batteries are used. It does not work as well as a switching current regulator, it does waste as much power as a resistor in typical uses since even one "poorly configured" is an equivalent state to just picking a different resistor. So it is indeed better than a resistor to control current as supply voltage drops but I feel we can ignore this method altogether as there's little point. Switching controllers are not expensive anymore, and smaller, and require fewer supportive parts. Switching controllers are better and cheaper, and faster to implement. Even so, it is good to be thorough showing this method, it just isn't ever the better option _today_ to power LEDs. See my prior comment for sources of the current regulating controllers for about $2 each and no time prototyping to use them.
ewitte ac-dc7 years ago
Please show me a *cheap* switching controller that will drive 56 3w ;) Multiples is ok but when it costs $20 to drive only a small handfull the cost adds up.
alah ewitte6 years ago
Look at Alah´s comment above. Thic IC can drive 6 X 3,6v LEDs at up to 1A from a 24v 10A PS. So you will need ten @ $1.56 = $16.00.
ac-dc ewitte7 years ago
Why would a 168W specialty controller be cheap? It won't be. Your best bet is to get a bag of low ohm resistors from some electronics surplus 'site (to lower cost) and then use a typical ATX PC PSU. Because these PSU are so common the commodity pricing makes them most affordable but since they are voltage regulating you'll need the series resistors to limit current. Any online LED resistor calculator will show the right value to use, then if you select a PSU with a potentiometer adjustment inside you can fine tune the resultant current. Anyway, that's about as cheap as it gets, you could pick up a PSU with suitable current on it's 5V rail for about $20 and the bag of resistors is $2 or so if you find the right 'site with a deal on the values you need.
ewitte ac-dc7 years ago
My plan is to run this circuit at 20v or so with 5 on the blue and 6 on reds. There shouldn't be that much loss. I can actually get a 210w 24v power controller for about the same as a 450W computer PSU. Or I might just use several old laptop ac adapters.
ac-dc ewitte6 years ago
Then consider these values:

6 x blue LED (@ 3.6V fwd drop) = 21.6V
(24V - 21.6V) * 1A = 2.4W

2.4W dissipation is going to require a larger chassis vent system and significant (for what it is) heatsink, both of which add to project cost and overall size. Granted, the transistor could be heatsunk to the same 'sink used by the LEDs, but either way this does still increase that heatsink size, overall size, and project cost by more than a typical switching regulator.

A 210W PSU seems overkill when your load is under 50W, but I could see the practical side of using what you have or are familiar with - in that situation the feedback circuit of the supply could be tweaked to drop the voltage down closer to that of the sum forward drop of the LEDs in series, OR use of a voltage divider in the feedback loop could cause the target current to result in the voltage drop the PSU's comparator circuit uses to regulate (similar to how it's already done in a typical LED switching regulator) and arrive at the target current with very little addt'l power loss and heat.
ac-dc ewitte7 years ago
Some of the kaidomain and dealextreme drivers have selectable steps to change brightness. Others do not.
smartroad ac-dc6 years ago
Valid points, although I feel you are being a bit harsh saying this can be "ignored". I am designing a circuit which doesn't need to worry too much about efficiency as it is still more efficient then the normal bulb it replaces. It has to fit into a very tight space on the circuit board and this circuit does it nicely. Most switching circuits I have seen need far more then 4 components, and also need coils and capacitors which again take up more space. So yes, if efficiency is the aim of the game and you have the space for the circuitry then don't use this circuit, find a good switcher. If you have something in a tight space with limited board area and you don't need anything more efficient then a resistor, but have an environment where the voltage can change, then this could be the kiddy. Its cheap and effective.
ac-dc smartroad6 years ago
If the space on a circuit board is tight, it's unlikely you will have enough space to heatsink the transistors. An entire switching regulation circuit can fit within 1 sq. in, less area than is needed to heatsink the LEDs so that swamps the area difference from the regulation circuits. While you are right that it becomes more useful when the input voltage changes, generally this is a backwards approach to engineering. Instead of taking a limited purpose design and envisioning where it might be used, generally one looks at the functional requirements and THEN designs a circuit to meet those. Plus, you still had to have an input voltage higher than the sum forward drop of all LEDs in a series so once again a switching regulator makes more sense. The voltage needed to be close to the forward drop of these LEDs so there wasn't excessive heat buildup, and yet in this case the scenario of having fluctuating voltage is necessarily going to be minimized. With switching controllers costing $5 or less, or to run several high powered LEDs you have a bit higher cost for the controller but still a small cost relative to the cost of all those LEDs, heatsink, enclosure, etc. There's very little reason to use such a lossy design unless you were in a hurry and happened to have the parts in your spare parts bin. Otherwise postage will cost more than the parts.
smartroad ac-dc6 years ago
I for one never said I was use a power LED, in fact I am using 5mm leds and not that many of them. I have changed the FET to better meet the power and thermal requirements to not need a heatsink. Yes switching CONTROLLERS are getting cheaper, however they still need a number of support components that can push this price up fast. I didn't take this design and "envision where it might be needed", I had already decided what functionality I needed then started the hunt for a simple design that would achieve this. This one happened to fit all the needs I had plus I am able to get all the parts from my local electronics supplier on my way home from work. I appreciate that you have a passion for SMPS and high efficiency systems, however although this is a less efficient design it is simple and effective, it does what it says it does without fuss or a high component count.
ac-dc smartroad6 years ago
In that case, yes it's a good way to cut cost with low powered LEDs, as you can't even get a typical controller with low enough current to run low powered LEDs unless they're in a parallel topology with a resistive current limiter (or separate IC for this). I was not necessarily claiming the circuit would not work acceptably with any specific purpose, rather than it is not assumed universally useful without considering the project requirements first. The component count need not be very high on a switcher though, and when I spoke of switching controllers I was speaking of the entire circuit like used in flashlight drivers, not just the IC which needs support parts, BUT these were mentioned in the context of using higher powered LEDs instead of the 100mW 2.5-10mm encapsulated type.
larams ac-dc7 years ago
Except those current regulating drivers your talking about from places like kaidomain can't be PWM to control brightness like you can with dans circuit can they?
ac-dc larams7 years ago
Some of the kaidomain and dealextreme drivers have selectable steps to change brightness. Others do not.
alah6 years ago
It´s good to use your brains doing something like this, but if you prefer just using your hands you can buy a simple, up to 1A controller, ( CAT4101 )for just $1.56 at futureelectronics. You will only need one resistor. Please read and UNDERSTAND the datasheet.
bliz236 years ago
Dan, I've looked at everything I can think of but something is still wrong. Used your pwm circuit #5 with r3 set at 0.6 ohms (read with multimeter). The led runs with about 505 ma of current. This is kinda strange, but i'm cool with it. So I want to increase the led current, so i decrease the resistance to 0.45 (the multimeter toggles between .4 and .5) but the current only increases to about 580ma. Theoretically it should increase to ~670 ma. The microcontroller produces 5v output. Any ideas on why or how to correct this? Also the leds voltage drop is about 3.2v at 580 ma and the power supply gives out about 5.5v at the load i'm working with. Any help would be nice, thanks. R1: 100k R3: 0.6 or 0.45 Q1: ? Q2: IRF510
bliz23 bliz236 years ago
I've been trying to play with the led's more: I have the microcontroller part separate from the rest of the circuit board. I connect the two parts together with two separate wires, one for ground and one to 'turn' on Q2 (the +5 pin). When i disconnect the wires on the microcontroller side, and touch just the plus wire, the led will light up and when i let go the led stays lit up. But when i touch both wires, the led will stop lighting up. Now i know there is some capacitance in Q2, which makes sense as to why it stays lit up when i stop touching the wires, but why does it light up in the first place? Also i tried to see if i had a natural voltage drop between me and the ground, which could cause this, and when using the multimeter, i got like 50mV drop and the led turned off. Some very strange things are happening, if anyone has any ideas, please let me know, this is starting to drive me kinda crazy. But i figure if nothing works out, i could just make the wire into some touch pad and see if people can light it up, almost like a magic trick, but it would be cooler to use it like i wanted to
bliz23 bliz236 years ago
I'm beginning to think that my power supply is not constant enough. I may need more capacitance in my smoothing section of my power supply because I can notice my leds flickering, which would cause the non ideal ratio, i'm hoping. Anyways, I'm going to try to find a better capacitor and hope that works, i'll let you know how it works, if anyone is reading this
Hi, I am new to the world of electronics and you seem to be a person who knows a lot more than what I do, So I have a quick question for you... I have a toy RC and I want to replace it's boring and dull headlights with bright white LEDs. I want to use typical 8mm White LEDs, I think they fall in the category as below: Typical values for regular "small" LED's with 20mA are: green/cyan/blue/purple/white: 3.5V drop The car runs on brand new 4 x 1.2V rechargeable cells 2800 mAh (a custom upgrade from 600 mAh) The car has 2 motors. I have already soldered the new LEDs to the circuit, but they somehow don't light up, whereas if i attach the old lights, they work just fine. What do I do? Also, while the car is turning, it tends to get jerky and it feels like the motors are not getting enough voltage because the car seems to be running at full load at that time... I think 3V for the drive motor and 1.5V for the direction motor and 7 V for the LEDs. Later, I plan to use LEDs for the taillights as well which will work only when the car is reversing (2 x 2V 5 mm Red LEDs) Load on Battery: Forward: Drive Motor + 2 White LEDs (3V + 7V i.e. 3.5V+3.5V) Reverse: Drive Motor + 2 Red LEDs (3V + 4V i.e. 2V+2V) Forward while changing direction: Drive Motor + Direction Motor + 2 White LEDs (3V + 1.5 V + 7V) Reverse while changing direction: Drive Motor + Direction Motor + 2 Red LEDs (3V + 1.5 V + 4V) So is there a way i could have regulated voltage to all the electric components so that the car doesn't slow down or gets jerky (as if it's switching on/off) while taking turns etc? or at least tell me what do I need to have the LEDs work with having to upgrade the battery : (
Hi Dan. I want to drive 3W RGB led (1W per color) 350mA per color 2.2~2,8v red 3,0~3,8v green and blue. Can you teach me how to calculate R3 for Circuit 5 and can you give me some easy to find substitutes for Q2? Can I use somethin like TIP31? Thanks a lot.
Wesley6666 years ago
how much power will these work with before parts start burning out? It would be nice if this worked with 120v or 240v.
nyeskus6 years ago
Thank you Dan, it works very well
torptorp6 years ago
Dan can you help me understand your math for step 6
"setting the current!

the value of R3 determines the set current.

- LED current is approximately equal to: 0.5 / R3
- R3 power: the power dissipated by the resistor is approximately: 0.25 / R3. choose a resistor value at least 2x the power calculated so the resistor does not get burning hot.

so for 700mA LED current:
R3 = 0.5 / 0.7 = 0.71 ohms. closest standard resistor is 0.75 ohms.
R3 power = 0.25 / 0.71 = 0.35 watts. we'll need at least a 1/2 watt rated resistor.
How are you coming up with the R3 number? If I was trying to set my current at 800mA then R3=0.5/0.8=0.625ohms? correct?
R3 Power=0.25/.625=0.4 Watts So go Minimum 1/2 Watt but I would go safe with a 1Watt?

BTW my wife doesn't think my new hobby is the greatest. But I really thank you for your great instructables! I have built 2 ultimate headlamps and just ordered a load of parts for "other experiments" I really appreciate the depth or research on the control circuits. I hate having to pay almost $20 for the bucks. Now I won't have too. =)
The most fun I'm having is getting creative with the heatsinks and enclosure methods for a finished product. I have a number of Ideas currently in the works.
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!!!
tiledz6 years ago
Sorry for my question im a noobs in led. i want to use 12 x 3.2v 300ma led on a single 12v power supply. Im thinking about 4 group of 3 led in serie with a 8.2 ohm resistor. then the 4 group will be wired in parallel to the 12vdc power supply. My cricuit total is 1170.7ma. I already have a 12v 2amp power supply, will this fit or i need a smaller one? i can eassyly find a 12v 1.2 amp here. Thanks!
gerojalo6 years ago
all i could manage to get hold of was an IRF540A for Q2 ... and a BC549C for Q1 ..... would these be ok ?
nukte8 years ago
Can't find the specified voltage regulator anywhere. Do you have the specs for it or do you have a substitute number? Thanks
xsmurf nukte7 years ago
I believe you can use any voltage regulator. An LM317 for example will handle up to 1.5A. Just find one that matches your specs. Linear and TI both make some.
Jugfet xsmurf6 years ago
If using a an LM317 be sure to get the right one ie LM317T as there are 'L' and 'M' versions that have significantly lower ratings.
geeklord6 years ago
Okay, so I'm tryin to figure out how this works, because I always do that. I can't decide if that's a curse or a blessing :D. So Q1's base is "monitoring" the current going through the LED's, the FET, and R3. If the current through that part of the circuit is too high, it pulls that gate of the FET low, lowering the current through the LED's. Right?
wow, I didn't think I would receive all of that, thanks! I learned some new stuff today.
How It Works

The gate of Q2 is held near its threshold voltage (around 1 - 2 V) by the potential divider formed by R1 and C-E of Q1. But as the voltage on the gate rises, the resistance of D-S of Q2 falls (perhaps as low as 0.02 ohm). As that happens, the voltage on the base of Q1 rises, which turns on Q1, pulling the voltage on the gate of Q2 down. Which feeds back by increasing the D-S resistance, which lowers the voltage on the base of Q1. It reaches equilibrium quickly.

In practice, both transistors will be near their turn-on points.

I built this with the same FET (FQP 50N06L), a similar NPN (the very common 2n3904), and a single 3 W LED from Avago (ASMT-MYE0).

The FET has gate threshold of 1.0-2.5 V, the 2n3904 has base saturation voltage of 0.65-0.95 V, and the LED has a forward drop of 3.6-4.3 V, current of 700 mA. (Values from datasheets.)

If the supply voltage is enough (above the forward voltage of the LEDs + the turn-on voltage of Q1), then the base of Q1 will basically be at its turn-on voltage (measured at about 550 mV in my circuit). It depends a little on the supply voltage (see later), but mostly on the properties of Q1. It doesn't depend on R1 at all (tried a few values 100K, 150K etc).

This is the crux of the design: the voltage on the base of Q1 is held constant, R3 is constant, so the current is constant. As all the current (except for a negligible amount) is going through the LEDs, the D-S of the FET, and R3, we can control the current through the LEDs by the value of R3.

So now we know that R3 will have all the LED current, and drop 550 mV. Per Dan's calculation, a 1R resistor will give 550 mV = 550 mA * 1R, thus 0.55 V * 0.55 A = 0.3 W.

It should be pointed out that the power used by the resistor is dependent on the turn-on voltage of Q1 and the required current. The power dissipated by the FET is dependent on how much we have to drop the supply voltage to the forward drop voltage of the LED chain.

Attached below are the scope traces for 100R, 10R and 1R setting resistors. The top trace is the voltage at the gate of the FET, the bottom trace is the base of the NPN. We see some start-up bounce, then the gate falls while the base rises until they are in equilibrium. It's quicker if R3 is lower.

The start-spike for 10R is shown expanded in the fourth trace. It shows a gate voltage 'shoulder', followed by a fast rise, and damped oscillation to the supply voltage. At this point Q1 is off with Vbase at zero. With 10R, it takes a further 75 usec unti Vbase starts rising. Sorry but I don't know why that is! (Both Q1 and Q2 have switching times in nanoseconds, so my guess is that capacitive effects of the LED or the circuit in general are charging up.)

At 1R, the current should be 550 mA (confirmed on power supply) and the 1 Watt resister got warm. The FET got very hot, and the LED got finger-burning hot. (It should have had a heatsink like they told us as it has an external operating temperature of only 95 C. It lasted about 5 minutes.)

Lastly is graph of measured Vbase against Vsupply.

Hope that's of some use.

Kind regards.
dan (author)  jonathanjo6 years ago
good summary and scope charts!

minor point: the circuit depends *a little* on R1. If you use 10k, Vbe on Q1 must be a little higher to turn off Q2. maybe 0.1V higher, so you do get a 10% or 20% reduction in R3 power losses is you use R1 of 100k or 200k. That's also why i spec the 2N5088BU - it has a slightly lower turn-on point than 2N3904 so again reduces R3 losses by 10%.
jonathanjo dan6 years ago
Thanks Dan.

You're right of course: I've just tried it now with R1 = 12K, and it drives up the Q1 base voltage to 631mV. When R1 was 100K or 150K I got the same 550mV value.
gigi_boeru6 years ago
how can I use the 555 circuit? i don't understand from that link... can you post the complete circuit (555 & this led driver)? thank you
ddickenson6 years ago
Does anyone smarter than me know how to adjust the calculations to use 350mA LED's? I am using "Luxeon I Star (hex) High Power LED Red Lambertian 350 mA" I forget the part number but I need to calculate this for either 4 or 8 of these led's in series.
gnodev6 years ago
Do you know about the ZXLD1350? This simple integrated circuit costs about $1.50 (small quantity) and has a complete LED driver in it for 350mA LEDs. It uses switching technology, so it is efficient. It can be regulated with a variable resistor, or with PWM. And you only need about 4 other components (0.3 Ohm resistor, protection zener, small coil 50 to 200 microH, capacitor) to use it.
Yes but where are you going to find a 0.3 ohm resistor because that might be costly
A wirewound 1W resistor is $0.25 and the thick film equivalent is $0.55 (at a well known internet electronics giant you can easily find on the web). The biggest problem would be soldering those very small components. It takes some skill and a lot of patience.
gnodev gnodev6 years ago
Now that I think of it, you can also use 3 1 Ohm resistors in parallel, which might be something you have lying around some where already. The little chip resistors are easily stacked to make a 3 resistors in parallel combination.
smartroad6 years ago
In the calculations where does the 0.5 and 0.25 come from, or are they just contants?
Actually scrub that, the 0.5 is the voltage drop across the FET I belive from reading your LED Instructable. I wanted to know as I have just tried building this for a different project at much lower power levels and changed the FET. Somehow I really messed up the calculations for the resistors LOL back on course now. I have used a BS170 N-FET which isn't logic level but works just as well from a 12v source :)
nevery6 years ago
Why was my comment removed?? I'm looking for a circut that will drive a high powered Luxeon off a variable power supply. from 3v or so to 19v, up to 2amps. can someone help me figure this out? Thanks
kdris6 years ago
would circuit #5 work for an led array with a source voltage of 36V or would the zener diode still have to be used since it is over 20v. Thanks in advance!
Quazar6 years ago
Dan, Great project! I'm looking for a design for a universal input module that can take anywhere from 3VAC/DC up to about 24VAC/DC. I thought I could use this circuit to drive an opto-isolator and add a full-wave-bridge on the input side. Does this seem like a reasonable approach, or is this circuit overkill for my application? Thanks in advance, - Dean
dan (author)  Quazar6 years ago
i dont think this is what you want. maybe you just want a 2-FET input buffer, it depends on what this optoisolator wants.
Quazar dan6 years ago
Dan, thanks for the speedy reply! I haven't picked a particular optoisolator yet; I figured I needed to get an idea of the driving circuit before I settled on one. Originally, I was thinking that a 5W zener with a suitable series resistor could get me in the ballpark to drive the opto, but I don't have much experience with constant current sources, particularly when used with wide voltage ranges.
geeklord6 years ago
couldn't you use a pair of AA's in series and then connect them in parallel with another set of 2 AA's?? How man mA does 1 AA typicaly put out??
Just seen your post,and without looking further,I know of some (rechargeable) AA cells rated @2500Ma,in fact they're all I'll use in my High power LED headlamp (by DAN last year).
You sure thats not milliamp hours?
Well yeah,but 2500Ma is 2.4 amps??
2.5amps, which is surprising to me...... What device runs on 1.5v and 2 amps?
Does'nt matter "what device runs on 1.5v and 2amps," that is simply what the cell is capable of supplying,so if your device uses L-E-S-S,then your power supply will last longer,..............................according to Ohms law at least.
I know that... Just thinking out loud.
ooooooooh,not the thing to do on a site like this ,methinks??
aburton6 years ago
I'm looking to power 5 arrays of 41 parallel LEDS. I'm a mechanical engineer by education, so I'm not entirely familiar with the world of microcontrollers. If I wanted to be able to dim them, how would you recommend I go about doing this?
lux_manish6 years ago
can you please tell me how I can drive aa 1watt LED with LM1117 or other low drop out voltage IC as my input voltage is 5-6.8volt
brianpxbd6 years ago
If you are into model control, you probably use LIPO batteries, if you have such a battery with a duff cell, you can use one of the other cells as a battery for a 3W lED. The voltage of approx 3.9 volts is about perfect to drive a white 3W LED. No series resistor should be needed, if you have a worry about the battery being short circuited then place a fuse or fusible resistor in series. Naturally one must be take special care with LIPO cells as they can burn or explode if abused by short circuit or overcharging. If you make a LED torch using such a LIPO cell then be sure it is safely protected when in use and only recharge with a correct LIPO type charger. Lipo cell are usually spoilt if the cell voltage drops below 2Volts, but this shouldn't be a problem with Your LED torch, as the power led will go out as soon as the voltage drops to about 2 volts. Take great care when stripping the old battery assembly that you do not accidentally short any of the cells, and do it outside away from any objects that could burn. I have been using such a torch now for about 8 months and find it super light and efficient, only requiring charging after many hours of use. Once again *Take care with LIPO cells*"
samphantom6 years ago
Hello Dan I was looking in the FEt transistor, is this a N-channel or P-channel, I'm implementing this same circuit connected to AC power line ovbiously increasing the capacity of this driver, using FQPF10N50CF as Fet and C3743 as the transistor, but still I'm confusing with the FET 'cuz in your drawing appears as P-channel and the number in your post is N-channel. Thank you for your time
dan (author)  samphantom6 years ago
it is an NFET
South_B7 years ago
what is the maximum number of LEDs that these circuits will power. Would it power enough LEDs to work for grow lighting Say 20 3W LEDs? I am thinking about aquaponics and I was considering LEDs as an altenative to fluorescent lighting PS. where did you learn your electronics
sjmusic27 years ago
Dan, If I wanted to use an LED array of 6 strings containing 3 SSC p7's (3.6Vf; 2.8A each) how would your 555 timer circuit look ? Secondly, would it be better (more consistent lighting) to provide individual Q2's & R's for each string - how would this circuit look and would it change the value of R1 or R3 ? PSU will be 12v, 20A (cctv psu) TIA (from a noob)
batchku7 years ago
Hello, I'm working with a mictrocontroller whose output PWMs go from 0 to ~ 3.2 V. In the instructions, it is indicated: "if your micro-controller is 3.3V or less, you need to use circuit #4, and set your micro-controller's output pin to be "open collector" - which allows the micro to pull down the pin, but lets the R1 resistor pull it up to 5V which is needed to fully turn on Q2." what exactly does that mean? does it mean change the value of R1 according to the voltage range of your microcontroller? or do i simply need a different Q1 or Q2....? thank you in advance. batchku
msmall7 years ago
Dan, I am quite new to the LED scene so please bear with me as I will probably ask some pretty moronic questions. But I have been thinking about using LED's for the lighting of chicken houses. Yes Chicken Houses...BIG ONES! 24,000 chicks per house! I have always noticed how the chickens will concentrate around a spot of sunlight that reaches the floor. The present method is the use of 60 watt incandescent bulbs (about 80 per house)mounted on the ceiling. I would like to try to illuminate their feed pans only with the use of LED's. I have 4 augers with 100 Feed Pans that I would require 4 LED's per pan or 400 LED's per run with a total of 1600 LED's per house. It would be helpful if the intensity could be reduced during their growth cycle. My questions are: Can a constant current power supply be made that can handle this number of lights? Can a multi-position selector switch with different resistors be used to allow for the control of intensity? Another concern is to create a system that is energy efficient. Do you have any suggestions? And remember...Eat more chicken. LOL
wow! That is one big setup. I would assume that you are going to build the 4 LED cluster packs your self? If you are doing something like this, I can recommend doing it very cheaply or very complicated. Cheap - Simply wire enough LEDs in series to run directly from 120VAC and wire them through a dimmer. At 20 ma / LED, you would be looking at 32000 mA. This works out to 32 amps. This is quite a draw. (nothing compared to the incandescents. At 60 watts, the bulbs are drawing 500 ma each!) They would be wired similar to christmas lights. One thing I must note is that the LEDs will only run at half duty because they only function in one direction. So there would be a slight flicker in your light, at 60 Hz. I would recommend staying away from this method because eventually, you would be running a high amount of current through the first LED. This method is very dangerous. Expensive - the recommended method above with a hefty DC power supply. Something like a computer power supply (maybe multiple). These are just suggestions, don't take them seriously, they are just thoughts on your idea. If you hurt yourself, I am not responsible.
If, as you suggest, you wire enough LEDs in series to run directly from 120V, then the current through the whole string will only be 20mA, not 32A as you suggest. If they were in parallel, you would add the currents, but in series, you add the voltages. I still suspect you would need some kind of controller circuit, as techchese has done above with his square array.
Derin7 years ago
jsbarrie7 years ago
Dan- This looks great! I am about to work on an LED lighting system in Latin America. We are using 3.6v 1W white LEDs (Lumileds) and locally they like to use 12v photovoltaic systems (probably because so many things work off of auto current) Anyway, how would you maximize the efficiency of your design considering these two parameters? Thanks, jsbarrie
dagorald7 years ago
hi Dan, this circuit looks very useful, but I have some questions about it, how many LEDs of 1W can you use in each circuit?, what's the minimal and maximum Volts you can use on this circuit? the number of leds depends the volts you give the circuit? Thanks. And Congratulations, you are great.
stib7 years ago
Thanks for all your instruction. But one criticism: LED does not require an apostrophe to make it plural, LEDs is fine.
Kevlarster7 years ago
You realize the an LED setup can be operated from a car, motorcylce, or lawnmower BATTERY. What an ideal solution for someone growing something in a remote location...
tecchese8 years ago
Hi Dan....I built your circuit #1 and it works fine for low voltage input, but for very high voltage doesn't... LEDs are the light of the future, so I built a #81 (9x9) 10mm (100mW) White LED array (see picture #1) driven by 220V AC, a 4x1N4007 bridge rectifier, a 100microFarad/400V electrolitic capacitor and your circuit #1, but I get only a Low brightness from the leds. Putting an amperometer in series to the leds I measured only 2.5 mA (I calculated R3 for 30mA based on your formula, and it is well calculated because with few leds and 12V input voltage I get 30mA effectly). From the output of the bridge there are 284 Vcc, between collector and emitter of Q1 I measure 2VCC, and between all the led array about 215VCC. Not even Your suggestion to put a resistor R2 to limit the voltage of the gate of Q2 to 10/20 VCC works, so the same for a 5V zener diode...no result I get ... I used for Q1->BC237B , for Q2->IRF740, for R3->18 ohm, for R1-> 100kohm (I tested other values for this last with no effect). There is a way for making a good work for your circuit with 220VCA ? Any idea? I'm trying to make a test using a cheap LM317 but it is a risk because its Input-Voltage differential limited to 40V, but it could be the right direction to light the LEDs correctly. Thanks
dan (author)  tecchese8 years ago
the circuit should work fine, but components must be chosen carefully.

284V / 81 led's = 3.5 volts per led. depending on your leds this may not be quite enough for full current. try tapping the led tree at around 70 led's and see if current goes up. remember the 284V measured by a multimeter is not very accurate because it is bumpy even with the cap - average could be a bit lower, and the peak will be higher (peak is 1.42 * 220 = 312V)

you need a 400V rated Q2. you must build circuit number 2 or 3 to protect the gate on Q2.

if you've blown Q2 by exceeding Vgs even once, it needs to be replaced.

the LM317 probably won't work (see circuit in step 11) - the problem is that you are likely to exceed the 40V drop rating at your peak V input of 310V+

let us know if you solve the problem!
tecchese dan8 years ago
Hi Dan....thanks for your reply. I found the problem...Q2 was destroyed! I replaced with a new one and now it works! Zener Diode *MUST* be present in any case for voltage input more than 20 V and Q2 must be a 400V certified Mosfet like IRF730 or IRF740 (cooled by a thermal dissipator). I inserted a diode 1N4007 in series to the #81 led array to prevent the inversion of the polarity. About 274VCC is present from the output of the bridge, while between the led array there are 252vCC (every white led gives 3.1-3.2 V each one, so we have 81x3.15V=255v).Maybe minor leds give better performance. A 100microF electrolitic capacitor in parallel to diode zener should turn on the leds softly (I haven't tested). If needed , I'll post the schematic and PCBs.
ajreef tecchese7 years ago
tecchese, (or anyone else for that matter) can someone elaborate on how possible the 'soft start' -talking about the 100uF in parallel with the diode) that you mention is possible? -Can push this soft start to lets say 10-15 minutes (from o% to 100% brightness)? -Will the LEDs also turn off gradually in 10-15 minutes with this method? I am looking for a 'dimmer' solution for some arrays, hence the Q.
ajReef, A soft-start is a circuit used to limit the inrush current to electrical devices. Its used primarily on electric motors starting under load. These devices are quite complicated & expensive. To answer your question...No, the time period involved for this capacitive-type 'soft-start' is on the order of milliseconds (1/1000s of Second) and just pulls some of the current away from the LEDs to charge the Cap, thus reducing the amount of initial current flowing through the LEDs.
Apolo8 tecchese7 years ago
[Quote}If needed , I'll post the schematic and PCBs. Thanks [Quote\] I am really interested on it
kmantesla7 years ago
Cool this looks really easy and cheap but I'm pretty new to the whole electronics scene... I'm used to the ol resistor and batt deal... So i have a few questions; when i'm calculating my total current, say 3 leds at 700ma each does that stack? Do I build the circuit for 700ma or 2100ma? They would be in series.... and lastly, in the basic first circuit you provided you're talking about Q2 acting as a variable resitor and only being able to handle 2/3 watt... and the formula for that was (voltage drop x current) Can u provide an example? For example if i'm using 12v and 3 leds thats a aproximately 3volt drop i would need, and running them at 700ma is (3x.7)=2.1 2.1 watts? so is it already too much for the transistor? I would really appreciate anyone who could help thanks! Cool instructables dan!
Da_Fudge7 years ago
Thanks! these work really well! also, great macro photos. I think that really makes the difference between a great 'able and a crap one. +1.
I have this circuit driving an array of 15 5mm UV LEDs. Q1 and R1 get burning hot when it's left on for long amounts of time. Is this normal?
Correction: R1 and Q1 get very hot, very quickly.
dan (author)  travail.studios7 years ago
not normal. i would guess you have something mis-wired. i assume you are talking about the circuit in step 6.
Could it be possible that I'm using the wrong value for R3?
up the wattage on the resistor. Maybe use a 5 watt with an attached heat sink.
kuan.fl7 years ago
Thanks for all the great information. I want to build a couple of the options given. Re: sourcing of LEDs: it's a year after you wrote up the article, and the prices that I can find for the Luxeon I, Luxeon III are $6-$8. I expected prices to down, not up; is it due to volatility like computer memory prices? > 1-watt and 3-watt Power LED's are now widely available in the $3 to $5 range
dan (author)  kuan.fl7 years ago
yes, it does seem that prices have gone up over the last year.
ac-dc7 years ago
While it may seem pedantic, this description of buck or boost regulators is wrong. They don't constantly monitor the LED current, they don't monitor current at all. Buck and boost mean voltage change - nothing more. Current regulators use a different kind of feedback than boost or buck, generally they sense a change in voltage through a very low ohm feedback resistor.

As for the cons, today they are much less than $20 for the current regulating type, see websites like http://www.kaidomain.com or http://www.dealextreme.com
Frankly at their low prices (typically between $1.50 to $5 per board delivered) it makes little sense to do anything else but buy the regulator boards from them. Only major issue is the most economical regulator boards are designed for flashlights so if the application needs board isolation you have to devise a method (like epoxy to attach to a mounting frame, and to solder on the supply leads where the battery and flashlight body contacts are). I have used several of these and they work very well for the price. I see no reason to spend $20 on a buckpuck, it's highway robbery these days opposed to the past when shoppers were more limited in what was available.
doctek ac-dc7 years ago
Looking at the web sites you mention, I don't see any LED regulators except some for LED flashlights. These seem to be only for CREE LEDs and then only 1 to 3 of them. Can you explain 1) Am I just not finding more "general purpose" LED regulators? 2) Is it possible to use the regulators on the sites in a more general fashion - that is to drive (for example) 30+ 100mA leds? 3) If that's possible, perhaps you could explain. Thanks.
ac-dc doctek7 years ago
1) Yes, dirt cheap flashlight regulators. In various voltages to suit the need. Surely if you can implement the rest you can solder a positive and negative lead to the bottom of the circuit board, keeping in mind that based on how a battery is designed, the center is positive. I built my bike headlights using a couple and they work great, you can affix them using epoxy or I just slipped some heatshrink tubing over them.

They aren't just for Cree LEDs, they are current sensing based on their individual specs so they regulate to that spec'd current level. Why would you use 30+ less efficient LEDs? Are you sure they are 100mA? The small 5mm encapsulated type are generally 100mW, which is 20mA conservatively or about 30mA with good heatsinking of the leads to the PCB copper.

You have to do some math to determine the logical arrangement as well as consider the input voltage. You give none of these so it would be excessively long postings to cover every scenario. I will give basics instead.

1) Determine your input voltage, keeping in mind the max the regulator can handle (same with any regulator actually).
2) Determine the voltage drop of the regulator. For example if your input is 8V and the regulator drops 1V, you have 7V out maxium.
3) Determine forward voltage of the LEDs. Suppose it's 3.4V @ 300mA. You can do the math to see 7V=n*3.4V, n=2.06
4) Rounded down 2.06 is two, you could put two of the example LEDs in series. That means you would have 15 parallel series of 2 LED to arrive at 30 total LEDs, except you claim 30 LEDs at 100mA each which is 3 amps. This is one area where I wonder if you have 20-30mA LED, or if buying some new why you would buy less efficient 100mA LED instead of 350, 700mA or more LEDs. Since the regulators linked were about 1A, with 100mA LEDs you'd need three boards and splitting up the LEDs, 15/3 = 5 parallel series of 2 LED per board.
5) LEDs don't have exactly the same forward voltage as spec'd, it varies a little so with your 2 x 3.4V series you have 6.8V forward from 7V and 100mA drive current so (7-6.8)/R = 100mA.
6) R = 0.2V/0.1A, you'd put a 2 Ohm resistor in series with each series of 2 LEDs, so you'd need 15, 2 Ohm resistors to make it work in this example.

In summary, yes there are many ways to configure LED arrays you just have to do the math to see how you need to go to get from where you are starting to where you need to be. With some combinations and input voltage options the flashlight regulators may not be ideal, if your Resistor value goes too high it becomes more lossy, but on the other hand a cheap way to lose some current through a handful of resistors may not be a problem if the power source is not a low capacity battery.

Doctek I think where's your stuck is on the basics of combining LEDs in series and parallel combinations and calculating out the parameters. Beyond examples you need to be more clear about defining the project requirements whether they be max brightness, minimal size, max efficiency, minimal cost, max resistance to extremes of temperature, max lifespan, max battery life or some combination of these. Once you have clearly defined your goal then choosing parts and adding complimentary circuits to reach the goal will become a more clear process.

I suggest looking at some of the LED calculators found via Google search and doing some more basic circuits before trying to tackle something more advanced, costly and time consuming.

doctek ac-dc7 years ago
Thanks for the thorough, thoughtful, and helpful response, ac-dc. It is greatly appreciated. Here's some more info on what I'm doing and the LEDs I'm working with. What I'm trying to make is an "IR Spotlight" for night time videography. As suggested in this forum, I found BG Micro to be a good source and bought a kit having 36 LEDs and a round circuit board. The LEDs are BG Micro part number LED 1101. They're rated 1.5V, 100mA, 850nm, 20 degree. I've built the little circuit and it works pretty well - maybe not quite as powerful as I'd hoped, but I'm still experimenting. So far I've powered it with a nominal 16V, 700mA wall wart. The LEDs are wired in 4 rows of 9 LEDs plus a 100 ohm resistor in each row. My goal is to have this totally battery powered, with eventual solar charging of the batteries. The camera I'm using has an RF link. I'll use a motion detector and a AT-Tiny to only turn on the camera when there's motion, and only turn on the IR LEDs if it's dark. So having an efficient power system is important. Keeping it cheap just adds to the challenge. After doing a bit of Googling, I'm checking into the LM3410; but if there's something already out there that gets the job done, I'm open. Thus my interest in the chargers you mentioned. I just want to know there limits. The two web sites you mention are a bit thin on specs. Thanks again for your help.
ac-dc doctek7 years ago
Presuming you want to use that circuit board the flashlight requlators aren't suitable since a series of nine of those (assuming BGMicro spec is correct) is minimum 13.5V out. The LM3410 does look like an option, though with 1500mA out limit you'd use more than one, the math works out nicely that you can use one LM3410 driver circuit per each series/row of 9, and not need that 100 Ohm resistor since you can set the LM3410 to the output current you want (ie 100mA). Since LM3410 has max input of 5.5V, getting good runtime from batteries would require using fairly high capacity cells, it's unfortunate it wasn't a higher input so a series of 2 Li-Ion could be used. Therefore I would think about using 4 x C cell (roughly 4000mAH) NiMH cells as a pack, and it's more forgiving of unregulated charging than Li-Ion would be w/o adding separate charging circuit. You can do the math on that to determine what runtime you will have, or if less runtime is needed then perhaps AA NiMH to save size and weight but think about how much power the circuit uses, it's going to require a fairly large, expensive solar panel and a bright day to charge a pack in a reasonable amount of time. I always find solar panel specs are crap, you never get even close to the peak rating in normal use so to avoid frustration (or maybe I just don't know enough about solar panel calculations) it's better to assume you won't get even 1/3rd rated power and overbuild the panel array. Frankly I think you should avoid solar charging unless you really, really need it, and grab some random wall wart limited to a low current (C/15 or so) trickle charge if that is fast enough for your needs, or of course you can built yet another more elaborate charging circuit for that too, but the time, cost and size keeps going up.
dan (author)  ac-dc7 years ago
A buck/boost LED driver regulates current. we are not talking about voltage-regulating buck/boost device since that requires using additional circuitry such as one of my circuits shown in this project. In the year since I wrote this article buck/boost LED drivers have come down in price.
xsmurf7 years ago
I'm not very familiar with Zeners, what voltage/rating should be used for a 3.3v µC? Thanks
dan (author)  xsmurf7 years ago
i'm suggesting to use a 4.7v zener, but you need to check your micro to see if it is OK with pin voltage being higher than its power voltage. if not, you either need to find a FET that turns on at 3.3v, or you need to make a fet-switch between micro and the gate (using one added FET).
xsmurf dan7 years ago
Sorry to bother you again, after moving on to Cree XR-E, with the higher drop the efficiency went to ~90% so I'm exploring the options again... As I understand Q2 G can be anywhere from ~4.7v to 20v, if I had a 7v zener laying around, that would do?
Aquatope7 years ago
I have noticed in the previous comments that a few were going to use your circuit for horticulture. I am looking to do the same for Aquaculture. I know that it would just be a matter of just selecting the appropriate LED spectrums ( for which I am using cool white and 455nm Lambertian with Vf=3.9V and 1000mA) I am new to this in a way so I apologize if i butcher any terms. I was planning to use 10 of each on different drivers. Because of the nature of these animals that i am growing I need for the leds to slowly dim on over the course of an hour so as not to shock them ( preferable the blue led dimming slowly on and then the white) Any Ideas or means of me doing this with out to much heat output...
heikoh7 years ago
If I use two LEDs in series with different Vf values (Vf=2.9v and Vf=3.15v) and build circuit #3 to supply 700mA, what will then happen?
dan (author)  heikoh7 years ago
works fine.
ac-dc7 years ago
This extensions of the basic circuit are still pretty lossy. It is a good way to learn some basic electronics but I feel it defeats some of the purpose of LEDs to achieve utmost efficiency.
SteveDavis7 years ago
These circuits would not work well for stepping down a very high voltage to the voltage needed to drive one LED. I would probably recommend against wiring to the phone line, since there are some sharp pulses when the phone rings. That said...

The reason that these circuits don't work for high voltages is that all the current going through the LEDs needs to drop the entire supply voltage. Using a buck regulator, which steps down the voltage using an inductor, allows you to have less current drawn from the supply than to the LEDs, and lower voltage across the load than the supply.

Here's a wikipedia description of a buck regulator: http://en.wikipedia.org/wiki/Buck_converter

And here's a listing from the company I work for (Arrow electronics) of driver chips that can be used to create a regulator circuit (look for buck in the type column): http://www.arrownac.com/mktg/nac/microsites/lighting/selector_guide/fall_2007/drivers_guide.pdf
whiteflour7 years ago
I'm building your circuit circa 2007-1-7, and I was wondering if I needed to heatsink Q2, I am putting three Cree power LED's in series in the circuit, they draw 700mA's and total 10.5V and my transformer puts in 700mA and 12V, so I'm thinking Q2 power=1.5V times .7A=just over 1Watt, did I calculate this properly? So, how would I heatsink, component:glue:mylar:glue:heatsink? And what is the size of the heatsink I would need?
dan (author)  whiteflour7 years ago
that is all correct. 5cm x 5cm heatsink should be plenty.
whiteflour dan7 years ago
Awesome, I built it today and tested the low setting with one of them hooked up to a 3.7V/350mA wall-wart, I'm really happy with the result these things put out some major light, I can't wait to try to get it on a three way switch, a 12V/700mA transformer, and add two more led's so I can really see the power of these things. Thanks for sharing.
leotavali7 years ago
Hi Dan, I've been thinking about creating a multi-LED 'emergency' light to be powered by my phone line so when the power goes off I still have lights. Since the power source is substantially more than just a few volts and not limited like it would be in a battery powered light, which circuit would you suggest? I'm thinking you might suggest the last circuit in this Instructable since it is probably the simplest and can handle larger voltages. However, I have two concern that maybe you can clarify for me ... 1- I've been told the phone line provides 70 volts continuous and 100+ when ringing. 2- I want to make the light work even if it is plugged in with the polarity reversed. Would your last circuit deal with both these issues without modification, or would it be necessary to make changes in the components -- or the circuit itself -- to eliminate these potential problems?
Power LEDs draw a lot of current. It's probably not a good idea. Although they do provide useful voltages, you wouldn't want to draw more than a few miliamps. At best you could trickle charge a small cell. Look into using a lead acid battery.
andyy327 years ago
I've been using circuit # 1 which works great but I would like to use some Li-ion cells. From what I understand a low voltage cut-out is necessary. Could you point me in the right direction for adding such a thing to the circuit or would it be beyond a beginner at circuits?
leidulv7 years ago
okay. Another question is why you didn`t use 3w leds in your headlamp?
leidulv7 years ago
Hi Dan I am a beginner. In this circuit you introduce R4. Does R4 have the same value as R3?
dan (author)  leidulv7 years ago
it can, up to you. with the switch closed it is like having a smaller value R3 because R4 is connected also. (search web for 'parallel resistor calculator'). with the switch open only R3 is in use.
g7mjv7 years ago
Very useful circuit. l'd like to be able to switch the load with something like a 2n2222, could you make a suggestion please. l already have a low power small amber light bar that i run via a 4060 circuit and 2n2222 but the load from this will be a lot higher with my group of prolights.
ajreef7 years ago
Dan, in step 6 (and then onwards) you refer to this:

Q1: large (TO-220 package) N-channel logic-level FET (such as: Fairchild FQP50N06L)
Q2: small (TO-92 package) NPN transistor (such as: Fairchild 2N5088BU)

Isn't this backwards? Q1 is small NPN and Q2 large N-channel FET?

Also, what would I need to substitute to have Vin=48VDC and Vout as close to 48VDC as possible, at 1000mA ( I understood how to control the 1000mA part, but the voltage is a problem for me sourcing a proper FET)?

Should I look for Vds of 48V+ or Vgs as you mention. I ask since I cannot source any FET with more than 20V Vgs!? Is this right?
dan (author)  ajreef7 years ago
thanks i fixed the text regarding Q1/Q2. for high voltage see step 7, and there are comments below about that. you can't get FET with higher than 20V Vgs, hence the need for a circuit change.
g7mjv7 years ago
Very useful, many thanks... My prolights will twinkle nicely :)
wellsw7 years ago
Hey Dan Great info I just ordered 5 "5027-PW14 Luxeon K2 Star - White, 130 Lumens @ 1500mA" I would like to power them using an old PC power supply. am I correct is thinking I should be able to power them directly if I get the right combination. this is what i am thinking. they have a 3.85 foreward voltage. the power supply has a 3.3 v rail. would that drive them fairly well?? or It has a 12V rail. 3 in series is 11.65V. would that overdrive them and could I use just a resistor in series since the power supply is regulated? Thanks
ajreef wellsw7 years ago
Short answer yes, both scenarios are doable on paper. Caution though, since you do not want to operate the Luxeons at 3.85V all day, since it is their MAX voltage. Also, the 12V rail with 3 LEDS in series will give them 4V each. You need to fine tune things a bit with resistors (mind the 700mA as well).
dan (author)  wellsw7 years ago
i think reading this instructable should answer your questions!
askjacob7 years ago
Very nice collection of LED powering options here! Something I would add though, is that LEDs are not consistant with their colour purity if you dim them the good old 'analog' way (i.e. by reducing the current). I have seen Green LEDs go through some weird colours (orange and yellow-y brown) when trying to dim them - whites also slide into a blue-y colour... that plus efficiency is why manufacturers recommend PWM for dimming.... Yet more often than not I still just 'use a bigger resistor' for desired effect :)
McJosh137 years ago
I created constant current source circuit #1 and etched it onto a copper clad pcb. I will try to post it here for all you pcb etchers out there.
Jamshed7 years ago
Hi Dan, What happens if an LED is open circuited. How the current will complete its path. My thinking is current will flow through R1, Q1. Am I right. Waiting for your reply Thanks
pjsquared7 years ago
Hi Dan et al Great site and led project stuff. I am working on an LED grow light and will likely use the pwr supply specs here for two channels with PWM (first from a 555 then an ethernet device from PC). There will likely be 30-40 Luxeon Rebels per square foot running at 350-500 ma, two colors red and blue. Will need to optimize the supply voltage and also deal with heat sinking them. Any advice would be very welcom vis the two power supplies and the heat sink arrangement. Many thanks. Peter
kellykungfu7 years ago
Hi Dan and Everyone, I just got some cheap Luxeon K2 LXK2-PW14-U00. They are rated at 130 lumeons at 1500mA. For the life of me, I'm not too sure how to calculate the resistors for the full 1500mA. I would be upgrading a bunch of my flashlights that would have 3 volts (small flashlight), 6 volts (flashlight with two 123 batteries), 4.5 volts (three D cell) and lastly a 18 volt flashight that came along with my cordless drill. So the big question, what resistor size/values/capacity do I need? Can I get it from Radio Shack and/or Fry's Electronics? Thanks a million and God bless. -kelly
pappyjoe7 years ago
dan, if i wanted to add more LED's to the circuit, what changes/adjustment do i have to do?thanks!
rmccurra7 years ago
Dan, Me again. I tried to build circuit 1 and something is not working. I have a 1W star (3.8V, 350ma). I calculate R3 as 1.42 ohm and P as .176W so using .5W, 1.5 ohm resistor. When I hook up a 9V battery, I am getting 508ma through the circuit. I'm pretty sure my breadboard is set up ok. Any ideas what I may be doing wrong? Thanks
Dan, If I understand what you are saying, you mean try increasing values of R3 to bring current down. At 2.3 ohms, I was at 380mA At 2.5 ohms, I was at 350mA At 2.9 ohms, I was at 300mA Do these numbers seem to indicate the circuit is set up properly and I'm just getting normal component variation? I got my Q1 and Q2 from Digi-Key and I believe they are the Fairchild components you specified. My room temperature is about 70F. Would it be better for accurate results to be using 6V vs 9V since the LED only wants 3.8? I'm mainly just trying to demonstrate/verify that I can set up the circuit at this point. Thanks.
dan (author)  rmccurra7 years ago
yes, test with a different voltage. doesn't matter if it is less or more, just something different. if it is working, you will see the same numbers you list above with the different voltage. 6v or 12v would be good to test.
dan (author)  rmccurra7 years ago
try other R3 value. 9V battery may be misleading since it probably would do only 500mA under a direct short-out, so you don't really know if you circuit is limiting the current at all. i would expect up to 20% inaccuracy between calculated current and actual, depending on the NPN you have and your room temperature.
rmccurra8 years ago
Dan, I have a source that goes up to 18V. I plan to use 5, white luxeon 1s (300mA, 3.3V drop). Which circuit do you recommend as the most efficient choice?
StCanna8 years ago
I'm planning on making a LED array for horticultural purposes which means i would need the brightest, constant-on driver. I will use a different driver for each color and am asking which of these circuits would provide the most efficient energy conversion? I am also considering making a circuit that can be plugged into 120v ac mains or 12v dc. Efficiency is my main concern... What do you suggest?
dan (author)  StCanna8 years ago
oh neat, i looked at several of the commercial 'horticultural / grow lights' and they seem to have the usual property of being overpriced and easy to build yourself. all they are doing is using red/orange and blue LED's, because the plants do not respond to yellow/green light. according to the patents on some of the lights they use specific ratios of red and blue which they think are optimal. you can find the specific recipes for red/orange and blue led's in the patents, so you can make your light exactly light the commercial ones. my circuit #1 is good for either 12V or 120V use, for 120V see the comments below from 'tecchese' and myself.
squalho8 years ago
Hi dan, these circuits are great, and they works fine, but they are based on the assumption that a DC power supply is available (you mentioned a battery for example). I wanted to use these circuits to build some lamps to substitute some of the old incandescent light bulbs i have at home, but the problem is to step down the wall voltage (220V in europe) to like 12 or 5 DC volts. How can i do this? I could use a transformer but i should find a small transformer to put in every light bulb (doesn't sound as a good idea, right?). A possible alternative is to use some "transformerless" circuit. I looked for some of them on the internet but they can just supply low current (up to 100mA). Switching power supplies are expensive.... so, what can i use?
dan (author)  squalho8 years ago
use a 'wall wart' type transformer, they are cheap and common, and you should have no trouble finding old ones to recycle. you can usually power several led's from a single transformer. for example, see my 'power led's - simplest light' project.
xsmurf8 years ago
Would I need a battery cut off if I ran this circuit on an NiMh battery pack? If so, does anyone know of a simple one?
dan (author)  xsmurf8 years ago
no, you only need cutoff for lead-acid and lithium batteries.
charlie_r8 years ago
Just breadboarded this (circuit # 1) for a few low power LEDs, to see if I understood the math that goes with this. THANK YOU for clear and understandable instructables! I calculated R3 to be 27 Ω for 20mA and actual reading on meter was 18.6mA........I think my meter (el cheapo) is cutting the current a slight bit. When I decrease the R to 23 Ω, I read 20.9mA. Does this sound right? Also, I used a Toshiba 2SK2545 for the MOSFET.....a lot lower power than what you used for the high power driver, but it seemed to work ok in this app. I paralleled 3 more sets of LEDs into the circuit, and still no change to the meter reading. Great work! Again, THANK YOU for the work, and for sharing it with everyone.
RSABear8 years ago
Hi Dan, I can only echo the words in many of the comments. Thank you for the work and effort which you have put into this fine instructable. I am still battling with the theory and operation of Circuit No 4. ? During the PWM pulse the micro is going to turn Q2 on, current through R3 is going to get Q1 to its bias voltage and current through Q1 b-e is going to turn Q1 on and hence the current limiting effect by turning Q2 off again. However, this might happen n times during the PWM high pulse. Will the micro not blow its output stage as Q1 is basically getting shorted it out? I hope this is not a dumb question. Any comment will be great.
dan (author)  RSABear8 years ago
see the text for that step - the microcontroller output pin must be "open-collector" mode or "high-impedence pull up" mode. that means the micro can pull down the node but only has a weak or no pullup. R3 pulls up the node. that way the micro can't short out. most micros give you this option of "normal" or "high-impedence" pullup, but if your micro (or other) chip can only do pullup/pulldown mode, then put an extra resistor in series with the pin like in circuit number 5. good question though, you cant tell from the drawing only the text.
RSABear dan8 years ago
Hi Dan, thank you for the comment. I have constructed the circuit x 4 on a bread board using IRF530s and BC458Cs for testing and it works like a charm. I used all the small 12V globes I could find in my components box and connected them in parallel during the tests. With my test circuit I got up to 370mA and then no more – wonderful! I have not connected the microcontroller yet, however I am very comfortable with the operation circuit #4 now. The purpose of all my work is to drive the LT-001-XXX modules from Kingbright. They have a forward voltage between 2.5V and 3.5V for the various colours. In order to reduce heat from power dissipation I am going to drop the supply input voltage to 5V. In the next few weeks I should be able to complete the circuit and battle to code into place.
dan (author)  RSABear8 years ago
sounds great!
ionic8 years ago
Dan, I have chosen the analog adjustable driver in step ten. I wil be using 3 Cyan Luxeon LHHL-LE3C rated at 1000mA. I've calculated the value of r3 to be .5 ohms at 2W. What I can not figure out is the values for R2,R4, and C1 Can you give me a hav\nd with this! Thakns. Brian
ChrisTexan8 years ago
Some comments about the comments and testing results....

I used the first circuit (great circuit, thank you, JUST what I've been looking for), I've previously used LM317 type circuits....

If you want a linear regulator, use an LM108x (LM1085 or LM1086 for example, adjustable versions), which is a low-dropout (quasi) version of the LM317-adj. They cost $0.98 US at mouser.com, compared to $0.32, but you get about 0.5-1 volt "free" (depending on amperage used in the circuit) in your circuit for a tiny price increase. You are going to lose 1.0-2.5V in your circuit with them however.
Efficiency - using an LM type circuit, with a single LED at 3.6V at 1A for example, you will need at least 5V from your source... if it's battery powered (NIMH for instance), the next closest you'll get is 6V (5*1.2V batteries). So you'll have a 6V source driving a 3.6V sink, meaning 2.2V are "wasted" (this is based on actual testing of Cree XR-E LEDs). At 1A, that's 2.2 watts of power going up as heat, and you'll need to heatsink your regulator as well at that point. I've been doing this (mounting the LM1085 to the aluminum chassis with a mica insulator), but it seemed like overkill. Using 4 batteries (4.8V), the best it can sustain is around 300mA of current (way low). It won't even light at 3.6V (3 batteries).

Thanks to Dan's circuit, I can now run the same LED from a 4.8V source at full current (4*1.2V NIMH batteries)... I've also tested it with 3 batteries (3.6V nominal) to simulate a low battery condition, and it was still holding around 500mA of current (basically LED limited by the voltage at this point)... in testing, the circuit is dropping around 100mV (0.1V). This is tremendously efficient for battery operation, and removes the need for a regulator heatsink by closely matching the input and output power (although I still chassis-mount the tab just to give flexibility).
This is why commercial regulator ICs like the 3904, etc, are not useful in battery/bikelight (my application) usage, because to operate the circuit requires a much much higher source voltage than output voltage just to power the circuit.
Dan's simple (#1) circuit is also nice because I was looking at a 12V source to barely operate 3 LEDs (10.8Vf) at full current, or a 13.2 or 14.4 source for safety overhead. Now I can operate off 12V with plenty of overhead, or even consider a LiIon 11.1V source. And you can run a single LED with a single LiIon at 3.7Vf and expect to get full, or nearly full output for at least half the batteyr life.
Oh, yes, I also tested on **2** batteries, (2.4Vf), and still managed to get some light out of the LEDs (only around 7mA though, it was enough to maybe light a doorlock in the dark to put a key in, but pretty dang dim)....

The point being, for parts count 9and cost), and simplicity, and efficiency, this simple circuit is hard to beat.
If you need multiple light levels, add a simple physical switch DPDT for instance, and wire in different resistances (Off/ 1ohm/0.5ohms for instance). This gives you 0/50%/100% power. No blinking or anything (use a circuit as appropriate for that obviously), but again, it's simplicity. That is what I've done with the linear circuits, and now works just as well with the Mosfet circuit.
Thanks Dan!!!
canaldoc8 years ago
Great site! I am trying to construct covert K2 LED internal windshield light-bar with Luxeon K2s Cyan colour (colour - I'm from UK!). Your circuits are very helpful - but how do I generate the various emergency flash patterns to strobe the LEDs? I can't find any commercially available units for DIY The likes of Whelen and Sho-Me are very costly - and anyway they say they are unable to supply CYAN colour - which I need as a UK Police Medical Examiner. Hope you can help!
dan (author)  canaldoc8 years ago
you can do simple flashing like in my ultimate bike light project, more complex blinking patterns require a microcontroller. circuits in this project show you how to connect a microcontroller to the led's, there are plenty of websites devoted to learning to use microcontrollers.
hey Dan I could set up your basic circuit to control 4 of these http://www.futurecb.com/Store/ProductDetails.aspx?ProductIdentifier=LXHLPH09LUMILEDSLIGHTING4502159
run off a 12v sealed lead acid, right? I know I would need to heat sink everything because of the high current but I think it would bake a pretty cool emergency light, to mount a small container with the LEDs and controller on the top of the battery
dan (author)  John Culbertson8 years ago
joedm8 years ago
Dan, Do you think it would be easy to add a flashing capability to one of your led circuits? Thanks
dan (author)  joedm8 years ago
see my new ulmitame bike light instructable, which has a rear blinker
It's a great instructable. Way too hot!!!(or Cool. should I say) Either way, i'm going to try and build my own bike lights from Ur design.
they have an instrucable for this mounted in an altoids can, you should check it out.
Can I run 3 white LEDs in series using a constant current source from 12 volts DC? If so, how can I determine when the drop out will affect the LED drive current? I have been reading up on this situation at http://www.lunaraccents.com/technology-high-power-LED.html
Hey Dan, cool instructable, just what i was looking for. I was wondering, in the circuit in step 9, you use two resistors to change the brightness. Would it be possible to use a pot instead? If so then how many watts and what resistance would the pot need to be? Also how efficient is the circuit in step 9. I want to run two white one watt luxons off 4 ni-mhs (4.8v) to make a head lamp/torch thingo. I need somthing with reasonable efficiency. Third question. Would it be better to run the two luxons in series or parallel? or does it not matter?.
Dan, first of all I would like to say thanks this is a great instructable. I am about to order the parts, and am not very familiar w/ Mofset thing and transistors. Could you check these?
and for higher than 20v would a transistor like this work
thanks alot

dan (author)  John Culbertson8 years ago
the first two look fine, although the mosfet you found probably is a 'normal' mosfet which means it needs 6V minimum to turn on. so if your power source is more than 6V no problem, otherwise you need a 'logic level' mosfet. you do not want a darlington transistor. see my circuits 2 & 3 in step 7 about how to go above 20V power source. you should have no trouble finding NPN transistors that can do 100V or more, the 2N4400 looks like a 40V part. one of the problems with buying from places like allelectronics.com as compared to digikey.com or mouser.com is that you don't really know what you are getting - no datasheet or manufacturer links are provided. sometimes you dont care, other times you do.
Hi Dan, I've become obsesses with this instructable, but because I'ma beginner, I'm having issues deciding which MOSFET to use, the choices are: 60V 30A 60V 55A 100V 40A I'm running 3x3W LEDs, so I'm assuming thatI'll take either of the top two, should it matter which one I use (I'm having trouble sourcing the 30A model)?
dan (author)  pierrelenfestey8 years ago
all of those are fine. the point of this circuit is you can use nearly any MOSFET and it will work.
swizzle8 years ago
Ok I just found the site and I want to say this is awesome. Second I love LED lights. I think they are amazing and hold a ton of potential. I'm wondering how long it is until someone creates a home made project that will put out the same amount of light as a coleman double burner lantern in a more compact size. Preferably something with a handcrank and a battery capable of maintaining a bright bulb for hours per charge. I'm not very electronic savy and all my electronic idea's have been done by my friend. I just can't make heads or tails of schematics at all. Any help would be great. Swizzle
BeDub8 years ago
Hi Dan, thanks for this instructable. I don't understand circuits 2 and 3, however. If the zener diode (circuit 3) sets the Q2 gate voltage at 5 volts, how can Q2 ever be turned off? (Also, can you give me the name of your "Analog Circuits 101" book so I don't have to ask you any more dumb questions??)
dan (author)  BeDub8 years ago
Q2 is turned off when Q1 turns on and provides a low resistance path to ground, lowering the voltage at Q2 gate.

this is the benchmark:
although it is starting to get a bit dated.
Transience8 years ago
I was thinking of using this circuit for PWM control. Would the extra 5% on the top and bottom matter, and if so, how would i go about using it?
dan (author)  Transience8 years ago
dont understand the question. for lighting applcation you probably want a pwm that goes all the way to down to 0% duty cycle, the high duty cycle does not need to go to 100% though.
Transience dan8 years ago
ok, That PWM chip sounds interesting, but the link doesn't work. Do you have the part number handy?
dan (author)  Transience8 years ago
thats weird, link works for me. anyway: Microchip TC642
dan (author)  Transience8 years ago
also, i've been thinking about trying one of these:

which is a super-simple 8-pin PWM chip needing only 2 resistors to set the output.
Vortexe8 years ago
Hi Dan, excellent projects :-)
Just wanna know, for the cheapo version with Q1 and Q2. On the MOSFET, what kind of minimum drop can you have over Vds? Eg, is 12V enough to run 3 Luxeonns in series, assuming 3.5V drop over each? By my calcs, this leaves just 0.8V to drop over the FET. (I'm trying to design max efficiency here)
Running in the ohmic region, min Rdson = 0.0025ohm, so it should be OK?

dan (author)  Vortexe8 years ago
power mosfets have very low on-resistance, as you say perhaps a few milli-ohms, so min vdrop is likewise a few milli-volts for the fet, plus 0.5 volt for the sensing-resistor.
rfarkash8 years ago
Quick question on caculating efficiency when driving a single LED. vF = 3.75, i = .7a. Assuming using R3 = .75ohm and battery pack at 7.2 volts, how can I determine efficiency? In particular, I'm looking at the step which indicates efficiency is related to R3 and voltage drop.
dan (author)  rfarkash8 years ago
efficiency: how much of the battery power is actually being used to power the led:

efficiency = (led power) / (power out of battery)

power out of battery = (battery voltage) * (current)

led power = (led forward voltage) * (current)

in your example, (3.75 * 0.7) / (7.2 * 0.7) = 52%

to get high efficiency you must put led's in series so their forward drop is nearly the battery voltage, or else you must use an expensive buck/boost regulator.
rfarkash dan8 years ago
Thanks for the prompt reply. Sounds like I need to use a 4.8v battery for a single led to get decent efficiency. I already have a buck regulator, but can't afford another $25 for each light I want to build. Hopefully the prices for good current regulating circuits will come down in the near future. Thanks again Ron
dan (author)  rfarkash8 years ago
4.8v: yes. or else put 2 led's in series. dont expect buck regulators to get cheaper very fast, the technology has been around for a long time.
danny3698 years ago
I run LEDs in a series string, about 110 to 120 to a series. Power supply is simple 4xin4007 bridge configuration and resistor is 10W added in series to the LEds. Can someone suggest a suitable circuit giving constant current adjustment and voltage setting? Mains power is 220 V AC.
joedm8 years ago
Dan, Do you think it would be easy to add a flashing capability to one of your led circuits? Thanks
larams8 years ago
Hello I have a question about the R3 sense resistor. According to the formula .5/.7 the resistor value is .71 is that the correct value. Just asking because that seems small and hard to find so far. I want to make sure the decimal isn't in the wrong place or something like that thanks....
LED Boatguy8 years ago
Hey there: I noticed you have upsized the wattage of R3 over the formulae you gave for the two circuits. Curious as to why. Of course it could just be my crappy math. BTW: THANK YOU for what you're doing.
dan (author)  LED Boatguy8 years ago
you want a resistor rated for at least 2x the actual power you think it will dissipate, so that it is not burning hot. at rated power a resistor will be over 100degC - ouch!
LED Boatguy dan8 years ago
OK. Thanks for clearing that up.

I make LED area lighting for boats. On a typical boat's electrical system are Radar, GPS, Chart Plotter, VHF radio, etc. Any suggestions on whether a cap or two would be warranted to keep all the electronics happy?

The Digikey shipment arrives Monday. I've been using constant voltage (LM2940 w/correct ESR caps) driving between 3 and 66 5mm LEDs, three in series, then parallel with each series getting its own resistor, but the constant current method seems to be the way to go and will drive the parts count way down.

Here's the latest creation: 66 of the brightest white LEDs I could find behind Fresnel lenses in a Bud potting box. After everything gets burned in, the box gets filled with epoxy to waterproof it. The sucker lights my living room!


Luxeons are next, but I'm having trouble finding a way to mount them in a box that's waterproof and presentable, i.e., sellable.

dan (author)  LED Boatguy8 years ago
wow, nice light. looks like a lot of work, if you switch to luxeons you will probably only need 5 or 10 of them. but yeah, mounting seems tricky. perhaps mount them on a metal plate, then just put a clear-ish plastic cover over the the whole front of the plate.
rfarkash8 years ago
Wow, great instructable. I'm a complete noob to electronics and had a basic question ...... 1- I'm looking to power a single Cree at 350ma and 700ma. Which of the above drivers would be the best (i.e efficiency, ease of building) running off of a 7.2 or 9.6 V nimh battery pack? 2 - I presume the DC-DC converters such as a buckpuck or bFlex have some additional capabilities or advantages to command their $20 price? Thanks! Ron
dan (author)  rfarkash8 years ago
i think all your questions are covered by this instructable already
hi Dan, I wanted to know how you calculate values 0,5 and 0,25 in the R3 calculations thanks
dan (author)  alexhumbertoa8 years ago
those are based on the threshold voltage of Q1: about 0.5 volts.
how can be connected a luxeon 800mA LED to 110v directly?
dan (author)  alexhumbertoa8 years ago
please see my 'power led's simplest light' instructable.
camp0s8 years ago
Hi nice circuit ! i have 4 (short) doubt: 1) what parameters do you look for finding alternative MosFet ? 2) from the datasheet, where do you see the stability of a transistor to the temperature? 3) how do you measure the efficency of the circuit? 4) and last, i added a small cap in parallel to R3 to smooth voltage oscillation, is it good? Thank you !
dan (author)  camp0s8 years ago
1) for a particular application, this depends on the LED current, the input voltage, and what you are using to control the gate. as a replacement for the suggested TO-220 general-purpose large mosfet, look for:
package: TO-220
Rds < 100 milli-ohm
Id > 5 amp
Vdss > supply voltage
Vgs max = 16V to 20V
Vgs threshold: look for 'logic level' (3V or lower) if you want to run it from a 5V gate control.

2) there is a feedback loop, this compensates for the temperature effect on Q2. the only temperature effect is to reduce the current limit when Q1 heats up, and if you keep Q1 away from Q2 it will not heat up.

3) LED power / power input.

4) i don't think it is necessary, have you noticed any problem? anyway, i think it will work better if you put it from Q2 gate to ground since Q2 has much higher gain than Q1.
Goose8 years ago
Ahoy there. Im new on here so forgive is if the Q has been answered before. I ride motorcycles and as a group we do alot of riding at night in the canyons. I been looking at installing 3w or 5w leds on the bike to light the corners when you lean into them. What would be the a good place to buy everything i would need??? Im new into electronics so what would be a good way to supply power to about 4 LEDs on each side??? any help would be great. My email is sashaa1 @ hotmail.com thank you all
Aeshir8 years ago
I did a quick search for "circuit maker" on google (always try putting the word maker into a search for an app like that) and found, well, Circuit Maker. Took me a bit to learn it but I'm a very quick learner. So I dunno. It gets the job done though.
dvawolk8 years ago
Hi, Dan!
You really have a way of explaining things usefully. But i had some problem.
That Fairchild 2N5088BU is making me troubles. I was trying to locate it here in my country (Slovenia, EU), but couldn't find any. Anyway i am pretty sure that there are some here under different names. It would be great if you could tell me what speciffications to look for. I will find the transistor myself but if you have some simmilar resistor name at hand, do not hesitate to write it...

dan (author)  dvawolk8 years ago
hi dvawolk, none of the parts is very critical, as it says 'any small NPN transistor' is fine. 2N2222 i am sure you could find.
dvawolk dan8 years ago
Thank you, that one i found.
But the problem apeared somewhere else:
The don't have FQP50N06 in stock... Is that one replaceable, too? What should i look for?

Thank you, again, 2w
larams8 years ago
How can a noob like me build any of these circuits if there is no part values? Am I just blind and not seeing it? I read thru every page and never saw any parts list.... That aside the author is very smart and creative. You have alot more brains in your noggin than I do thats for sure. :)
dan (author)  larams8 years ago
there is a parts list on step 5 about 2/3 of the way down the text. i was a bit negligent on some of the later circuits. i added part numbers in step 6 and 10. for step 9 as noted the parts list is in my other instructable which is linked there. all the part numbers are available at www.digikey.com and you can just enter them into the 'search' feature there. let me know if you have any further trouble!
dan (author)  dan8 years ago
also i updated step 1 in regards to your question also.
sysadmn8 years ago
I haven't worked with high power leds before. Sorry if this is a dumb question - what's wrong with using a dedicated led driver made for this application? The National LM3402 runs $2 quantity 1 from digikey. Looks like it requires lots of DC (application notes show circuits using 24V and 60V), but furnishes constant current with undervoltage and thermal protection and has PWM dimming. It it just that it's overkill for experimenters, or that designing your own buck is not worth the effort?
dan (author)  sysadmn8 years ago
the chip is very small SMT part, and thus not very suitable for beginners. if that doesn't stop you, then it looks like a great part to use.
5Volt8 years ago
This one is a PWM driver I simulated with Orcad PSpice demo version; the components are the one I could find in it. The PNP transistor is a medium power one. R8 is the load and must be replaced by the LED and current limiting resistor series. A real resistor was placed there just for the simulation. The lm555 is an oscillator (1kHz here) and at the capacitor C1 is a sawtooth waveform swinging from 1/3 Vcc to 2/3 Vcc. This is compared against the value set by the potentiometer whose wiper swings the positive input of the comparator from 1/3 Vcc to 2/3 Vcc as well. When the former is greater than the latter input, the transistor is in full conduction, otherwise it is off, turning the LED on or off accordingly. The PWM duty cycle yelds different LED intensity. The potentioneter and the two accompanying resistors can be replaced by a variable voltage thus making a voltage controlled PWM. The resistor in series with the LED must be calculated for the maximum current we want to flow through the LED under any circumstances (i.e. continuos on). Multiple LEDs can be controlled independently from one another using the same sawtooth generator and a separate comparator and driver for each LED. 5V.
dan (author)  5Volt8 years ago
nice, but still need 2 chips and 8 discretes. i've been trying to figure out what is the easiest way to do a pwm with a single 8-pin DIP, i think i've found a pwm fan-motor driver chip that can do it with only like 3-4 discretes. i'll post after i've tested it.
Fake_Name8 years ago
Can I reccomend ExpressSCH for doing schematic stuff? (http://www.expresspcb.com/ExpressPCBHtm/Free_schematic_software.htm)

It's part of expressPCBs CAD suite, and it's free, because if you want to have PCboards made from it, you have to go through them. (Or make them yourself)

However, for simple Schematic stuff, it's by far the easiest application to use that I've found. The learing curve, for me, was literally ~5 minutes.

Now, compare that to something like EAGLEcad...
dan (author)  Fake_Name8 years ago
i know eagle quite well, i just prefer paper sometimes.