I noticed arduino I/O-pin give only 40mA and I want use super bright leds @ 350 mA. So if I put transistor to there, can I still dim it (by PWM)? what kind transistor should I use?
Yea... im beginner....
There is also a limit of 200ma total current for the chip.
Yes, you can dim it using pulse width modulation.
Sorry I can't help with chosing a transistor. This is a fairly common question so you can probably find something useful through Google...
http://www.google.com/search?ie=UTF-8&q=current+led+40+site%3Aarduino.cc
2n2222 should be fine, they have a max current of 500-650ma (depending on who made it)
dont forget to put a 1 or 2k resistor tween the pin and the transistor, or nasties might happen
I'd look for a 1 amp switch -- cheap and common.
It is not that easy. The LED is supposed to receive a constant current of 350mA, so you need a constant current source driven by the uC pin.
You need a constant current LED driver like one of these: BuckPuck DC LED Drivers or something similar. You may also find them cheaper on another site. Shop around.
High power LEDs need careful monitoring of their drive current and heat levels. Make sure you also mount the LED to a good heat sink.
What are you going to use the LED for?
Thanks for links... im making similar mood light than: Arduino mood light
So maybe i can use ULN2003? Can i still dim it?
The schematic:
arduino ULN2003
+-----------+
| |
9 -------+ 2 15 +-R1---LED--+
| | RED |
| | |
10 ------+ 3 14 +-R2--LED---+
| | GREEN |
| | |
11 ------+ 4 13 +-R3--LED---+
| 8 | BLUE |
+------+----+ |
| |
GND +5V
R1 = 8.2 Ohm or 6.8 Ohm (Led=2,6v)
R2 = 6.8 Ohm (Led=3,8v)
R3 = 6.8 Ohm (Led=3,8v)
If im using arduinos normal +5v pin can i get enough Ampeers? (if I example use 12v @ 1A power supply) Should i connect power directly to power supply or can I use andruidos power pin?)
There are two things wrong with this.
First of all you can't switch 3 time 350mA from a ULN2003, although the individual current ratings are not exceeded the power dissipation for the package is.
See:- http://www.thebox.myzen.co.uk/Tutorial/Power_Examples.html
You can't use resistors to limit the current in a high power LED you need a constant current supply.
The web page you referenced is a typical example of crap design. It is not reliable, repeatable or sustainable. That is it will fail after a time and is running components way above their limits and if you make it it might not work even if his did.
You will not get much control over the fade if you use a transistor like this... you will get about 10 levels of brightness (not 255!) and that will not be linear - it will be either dim, bright, very bright or very very bright...
hmm... what do u think if I buy "megabrite" Macetech
The main reason I made this version is that I wanted to fit it in to the base of an Ikea Mylonit lamp. This lamp is made entirely from glass and has a small recess in the base with an entry for the power lead. Since it had nowhere to fit a separate power connector but did have a slot for a power lead to pass in to the base, I needed to make a PCB with a DC power connector in the centre of the board. At the same time I took the opportunity to tweak the circuit slightly and left out things it didn't need. The changes are detailed below. Please note that the original version on the main page is still perfectly okay and you may find the single sided PCB easier to make.
*
This version uses a double sided PCB. Dimensions are approximately the same as the original version.
*
Removed the second push button switch and components for the RS232 serial interface.
*
Added 22K resistor between MCLR input and Vdd. This eliminates the need for a jumper on the ICSP header in normal operation. (If you use a 12F683 you can omit this resistor as well since it has an internal pull-up)
*
Provision for a surface mount DC power socket.
*
Removed the power-on LED and series resistor.
*
Added 47K pull-up resistors to the transistor drive circuit. On the original version, when it first powered on the RGB LEDs would flash at 100% brightness very briefly. This is caused by the PIC I/O lines defaulting to inputs at power-up and until the software starts and sets them as outputs the PWM inputs to the ZXLD1350 are enabled. This modification prevents this from happening. NOTE: it's purely for aesthetics, not a fix for a design fault.
*
In all other respects the circuit and software is identical to the version 2 design on the main page.
Mengembalikan Jati Diri Bangsa
Kenali Dan Kunjungi Objek Wisata Di Pandeglang
Now i have looked different shops and I think that im going to buy BlinkM MaxM - I2C Controlled RGB LED - COM-09000 - SparkFun Electronics (there is controller and 3 bright leds) because it's sounds too difficult to build own. Do you know is that good?
You can't use resistors to limit the current in a high power LED you need a constant current supply.
I don't know whether I'd go that far. Using resistors as the current limiting devices for any LED is subject to limits in accuracy depending on the forward voltage of the LED, and that's more variable with high power LEDs due to heating, but I don't think it's that much more difficult to use a resistor to limit the current of a high power LED to "about 200mA" than it is to limit a normal LED to "about 20mA." It's just "philosophically important" to drive the HP LEDs at more accurate currents because you're probably trying to max out the output, and failures can be rather catastrophic.
Of course, you do need higher power resistors as well. Don't put that 5W LED in series with the usual 1/4W resistor !
It's true, but you need high wattage resistors and you're going to generate a TON of heat doing it that way. If you had a 12V power source and you wanted to drive a 3.6V LED at 280mA (~1W), the resistor will be dissipating 2.3W. However, LED drivers like the BuckPuck are much more efficient and hardly get warm at all even if you drive the same LED at 700mA. And the drivers are easily dimmable with the Arduino's PWM output.
Using a 5 volt supply gives a more reasonable 0.4 watts. The actual wattage is actually lower because the driving transistor will drop around 0.4 volts, so in practice it would be not much more than a quarter watt.
I have had good results driving 300ma per LED RGB devices using two 0.6 watt resistors in parallel for the red element (Vf = 3.0v) and one 0.6 watt resistor for the green and blue (Vf= 3.6) using a 5 volt power supply
I don't know whether I'd go that far.
Well OK if you cut down on the current and treat them like normal LEDs then a resistor is adequate. However there are two major problems. First off the high current requires a lower resistor value. Therefore the actual current through the LED varies much more with the on voltage variations of the individual LED. A resistor will only approximate to a constant current supply when it is large in comparison to the required impedance.
Second the low values also put a greater variation in current with respect to the actual values you can get as opposed to those you can calculate. Because high power LEDs run much hotter than ambient the on voltage will also change as it warms up. Having a smaller resistor then causes again a much wider variation in current than you want.
The net is full of articles explaining this sort of thing. Hears one I just found:-
http://www.emcomp.se/editor/upload/docs/High-Power-LED-Driving.pdf