Powering led strip with old psu?

[LTSU]

I have 5m rgb led strip and I would like to control it with arduino, but obviously arduino is not cabable of powering it. If I use a old computer psu that has 12v 10A output, will it burn the leds or will the strip draw only the amount of amps that it needs? I'm asking because I once forgot to put resistor to led and it obviosly broke and i really don't want to fry the strip..

[You_are_a_cunt]

The ATX spec for PSUs mentions a maximum of 10A (iirc), with actual power draw being lower (Probably closer to 2A). You can safely power the strip using the +12V off of a Molex connector (Not entirely sure where you connect the GND to, in therms of pins on the arduino)

[mariushm]

The current value (10A) represents the maximum amount of current that the power supply can produce. Think of it like the maximum pressure water can have through a pipe before pipe starts to crack and break.

The voltage is fixed at 12v, think of it like the diameter of the water pipe. Power supply will try to keep the voltage to 12v, or as analogy it will try to keep the pipe constantly full of water and not have air bubbles or gaps in the water flow. ... no matter how much current something consumes, the power supply will put more pressure (more current) trying to keep the pipe filled completely up to a point. If the power supply says it can do maximum 10A, that means if something consumes more than 10A, it will have no choice but to reduce the voltage ( deliver same quantity of water in a unit of time) or just stop. 

 

If the led strip is designed to work from 12v , then almost always it will be segmented in groups of 3 leds and each color in that segment will have is own resistor limiting the current to some reasonable value, something like 10mA.. 20mA per color.

If you want a rough estimate, just assume 10mA per color , so 30mA for each 3 leds. Let's say the 5m strip has 100 segments of 3 leds.. in this case your led strip will consume 30mA x 100 = 3000mA or 3A - IF all the three colors are turned on all the time.

When you'll adjust the "brightness" of one or several of the red,green and blue colors, the total power consumption of the led strip will consume less.

 

You can figure out more accurately how much the led strip will consume by measuring those resistors in a segment and using the formula: 

 

Input Voltage - number of leds x forward voltage of one led = Current x Resistance.

 

For example, if the red channel has a 1 kohm resistor, then using a slightly higher value of 2v than the typical value (1.7v..2v) for red leds because it makes the math easier (2.2..2.4v for green or less depending on chemistry and around 2.8..3.2v for blue) the current for the red part of the 3 led segment will be (12v - 3 x 2v ) / 1000 ohm = 6/1000 = 0.006 A or 6 mA

Red leds are more efficient than the ones made for other colors, so often you get the same brightness/intensity with less current so you may see different value resistor compared to the other colors.

Also, human eyes are more sensitive to green compared to other colors so you may also see a slightly different current for green compared to blue.. but most cheap rgb strips don't care about this and just limit the maximum current to a sane value like 10mA or mA and assume user would pwm the color channel to reduce the brightness

 

So you can just use the formula and add up the currents for each color and you'll have the value for a 3 segment when the segment is 100% white. then multiply with the number of segments.

 

[W-L]

11 hours ago, LTSU said:

I have 5m rgb led strip and I would like to control it with arduino, but obviously arduino is not cabable of powering it. If I use a old computer psu that has 12v 10A output, will it burn the leds or will the strip draw only the amount of amps that it needs? I'm asking because I once forgot to put resistor to led and it obviosly broke and i really don't want to fry the strip..

You will need to use auxiliary power and not have it directly through the ardunio, only for it to provide a PWM control signal and use mosfets to be what connects everything together.

http://www.jerome-bernard.com/blog/2013/01/12/rgb-led-strip-controlled-by-an-arduino/

[You_are_a_cunt]

12 hours ago, W-L said:

You will need to use auxiliary power and not have it directly through the ardunio, only for it to provide a PWM control signal and use mosfets to be what connects everything together.

http://www.jerome-bernard.com/blog/2013/01/12/rgb-led-strip-controlled-by-an-arduino/

OR you could use ULN2003A Darlington Arrays and have it be quad channel and controllable via an OS interface

http://www.instructables.com/id/RGB-led-control-PCArduino/?ALLSTEPS

 

[mariushm]

ULN2003A is a darlington array, it's much slower at turning on and off completely, so you wouldn't be able to turn on and off each individual color as fast as with mosfets. Basically, if you plan to implement some kind of fadeout , face in, decrease brightness slowly, transitions between colors, with uln2003a it would be more sluggish, more chunky, with mosfets you'd have finer control.

 

Also, keep in mind that at high currents, the voltage drop between collector and emitter (ground) pins can be as high as 1.6v (maximum voltage drop at 350mA in the datasheet...  so you have to take that in consideration when you calculate the resistor value for the limiting resistor if you need one (because uln2003a will simply act as an on/off switch) and with such high drop voltage it's really not recommended to use strips working on low voltages like 5v for example.  You'd have a white or blue led with 3v forward voltage, you add 1.6v and you're up to 4.6v and the limiting resistor will raise the minimum voltage requirement to around 4.7-4.8v ... with long strips (let's say over 1 meter) , the resistance of the traces on the strip is enough for the leds to see less than 5v at the far end of the strip. 

 

Of course, ULN2003A is a 7 element array, so for a RGB strip, you could pair two channels for each color and have less current per darlington element (divide by 2), which in turn would lower the maximum Vce voltage drop. that would solve the voltage drop issue, but the slow turn on / turn off time would still be an issue.

 

[You_are_a_cunt]

24 minutes ago, mariushm said:

ULN2003A is a darlington array, it's much slower at turning on and off completely, so you wouldn't be able to turn on and off each individual color as fast as with mosfets. Basically, if you plan to implement some kind of fadeout , face in, decrease brightness slowly, transitions between colors, with uln2003a it would be more sluggish, more chunky, with mosfets you'd have finer control.

 

Also, keep in mind that at high currents, the voltage drop between collector and emitter (ground) pins can be as high as 1.6v (maximum voltage drop at 350mA in the datasheet...  so you have to take that in consideration when you calculate the resistor value for the limiting resistor (because uln2003a will simply act as an on/off switch) and with such high drop voltage it's really not recommended to use strips working on low voltages like 5v for example.  You'd have a white or blue led with 3v forward voltage, you add 1.6v and you're up to 4.6v and the limiting resistor will raise the minimum voltage requirement to around 4.7-4.8v ... with long strips (let's say over 1 meter) , the resistance of the traces on the strip is enough for the leds to see less than 5v at the far end of the strip. 

 

Of course, ULN2003A is a 7 element array, so for a RGB strip, you could pair two channels for each color and have less current per darlington element (divide by 2), which in turn would lower the maximum Vce voltage drop. that would solve the voltage drop issue, but the slow turn on / turn off time would still be an issue.

 

I found the ULN2003A approach a bit more, well, elegant, since what I needed had to have a tiny footprint (using a Nano V3 controlling 2 12V strips and the control unit BARELY fits in the housing I needed to use). To be fair, I'm not using color shifting effects and the main usage is for solid colors, though, in testing I haven't seen any sluggishness when on a color shift mode, likely because of the external supply for the strip and the fact that they're 12V. I'm curious on the "performance" of the darlington array versus mosfet (which I might just test when my component shipment shows up later this month. Been going through Nanos like a fat kid goes through candy lately

[mariushm]

Yeah.. perhaps I didn't give the right example.

 

Imagine an  4x4x4 cube ( 4 layers each made of 4x4 led squares, each layer on top of the other). You use 16 i/o pins of the microcontroller (or 2-3 pins and a  couple of regular 8bit shift registers) to turn on or off the 16 leds that make up a layer. and then you connect an individual layer to ground. 

You multiplex the 4 layers - you turn off the old layer, you set the 16 leds states for the new layer, latch (if needed), turn on the new layer,  repeat .. 4 times for each refresh of the cub. So if you want 60 updates a second, you have to set the led states and turn off/on layers 4 x 60 times a second.

 

Because the darlington elements turn off and on slowly, you may have ghosting and brightness fluctuations (where previous layer is still turning off when the new layer is turned on so basically you'd have two layers on at same time, and half the current through each led) .. so you'd basically have to add some tiny delay before turning off the old layer and turning on the new layer, a dead time between updating layers.

Mosfets will have "sharper", faster switch times.

 

Another example would be if you want to make a 7x128 scrolling text marquee thing or something like that .. you'd naturally multiplex the leds, so with uln2003a the scrolling text could look ugly, you may see smears or double vertical lines from time to time if you're not careful.

With low currents it's not a problem, but if you'd make a signage board for outside where each led could be 50-100mA or even higher, uln2003a is no longer a solution.

 

if anyone's curious, here's how led advertising boards are made, with 10-14 bits of light intensity, multiple methods used together to adjust brightness and to reduce the power consumption of the whole panel and lots of awesome stuff :

 

 

[You_are_a_cunt]

5 minutes ago, mariushm said:

Yeah.. perhaps I didn't give the right example.

 

Imagine an  4x4x4 cube ( 4 layers each made of 4x4 led squares, each layer on top of the other). You use 16 i/o pins of the microcontroller (or 2-3 pins and a  couple of regular 8bit shift registers) to turn on or off the 16 leds that make up a layer. and then you connect an individual layer to ground. 

You multiplex the 4 layers - you turn off the old layer, you set the 16 leds states for the new layer, latch (if needed), turn on the new layer,  repeat .. 4 times for each refresh of the cub. So if you want 60 updates a second, you have to set the led states and turn off/on layers 4 x 60 times a second.

 

Because the darlington elements turn off and on slowly, you may have ghosting and brightness fluctuations (where previous layer is still turning off when the new layer is turned on so basically you'd have two layers on at same time, and half the current through each led) .. so you'd basically have to add some tiny delay before turning off the old layer and turning on the new layer, a dead time between updating layers.

Mosfets will have "sharper", faster switch times.

 

Another example would be if you want to make a 7x128 scrolling text marquee thing or something like that .. you'd naturally multiplex the leds, so with uln2003a the scrolling text could look ugly, you may see smears or double vertical lines from time to time if you're not careful.

With low currents it's not a problem, but if you'd make a signage board for outside where each led could be 50-100mA or even higher, uln2003a is no longer a solution.

 

if anyone's curious, here's how led advertising boards are made, with 10-14 bits of light intensity, multiple methods used together to adjust brightness and to reduce the power consumption of the whole panel and lots of awesome stuff :

 

 

That's actually overcomplicating what is actually a very simple build that doesn't really require as much finesse as, like you've used as an example, scrolling text or a LED advertising board.

For a cheap, fast and dirty DIY project, the darlington array is just fine. If you want something more professional, you'd definitely be looking at something else and most likely it wouldn't involve an Arduino. This is, after all, a Charlie Cheapo LED strip, not a signage board