Hello,
I'm making an RGB light for video shooting.
I have this led and a 36V power source.
I want to control it with Arduino Nano and bluetooth board through an app in iPhone. But the problem is that I don't know a max current for each channel? Each of 3 segments consumes different current if I measure it without limiting and of course it grows while it becomes hotter. I can't find any datasheets with this information. Do you have any thoughts?
Tofer:
if I measure it without limiting
What is that supposed to mean?
Do you mean to tell us you connected that LED to a random power supply without the current control circuitry that you know is essential from your reading of the threads here? 
Paul__B:
What is that supposed to mean?Do you mean to tell us you connected that LED to a random power supply without the current control circuitry that you know is essential from your reading of the threads here?
You are saying it in a way like I've connected my dog to a random power supply. ))) It is just a small peace of plastic. I was checking if it works or not. It is a 100W led but it was consuming just 2A maximum. I'm sure it's going to consume more after it get hotter, but the question is different. How can I limit the current if I don't know which current it needs per channel? Does anybody ever has something to do with this kind of led?
I found one topic here and if linked me to this. But I can't find it in my country. I can recreate it myself. But still I don't know the value i should to limit the current to for each channel. I may assume it is just 2.78/3A, but I'm not sure.
Tofer:
You are saying it in a way like I've connected my dog to a random power supply.
Interesting comparison, but - yes!
Tofer:
It is just a small piece of plastic. I was checking if it works or not. It is a 100W led but it was consuming just 2A maximum.
I'm not quite sure what you are saying here.
Is that with all three channels in parallel?
A 100 W RGB LED would obviously use 33 W in each channel - more or less. At about 33 V that would be about 1 Amp each. So it is either massively over-driving one channel, or under-driving three.
But you simply cannot connect all three in parallel because the three colours have significantly different voltage drops so the red would be bright and consuming almost all of the current but the other two colours very dim.
Tofer:
I'm sure it's going to consume more after it get hotter, but the question is different. How can I limit the current if I don't know which current it needs per channel?
If you know the voltage - and it is about 33 V - then you know the current is 1 Amp.
Tofer:
Does anybody ever has something to do with this kind of led?
Quite a few people here I fancy. I have the 20W version (I think) with IR remote control hanging from a shelf here.
Tofer:
I found one topic here and if linked me to this. But I can't find it in my country. I can recreate it myself. But still I don't know the value I should to limit the current to for each channel. I may assume it is just 2.78/3A, but I'm not sure.
That is a nicely designed board. It has three switchmode current regulators with PWM applied by a transistor each from a WS2811 chip. Yes, that is what yours should look like. Replicating it yourself is not advisable given your present state of knowledge (as in - you would not be asking these questions if you had the expertise to do it! 
) It may well not be available in your country - that is why you buy it by mail order from Tindie!)
If you want quality light source for filming, an RGB led is not the ideal choice, you'll get
significant colour distortion. Art-gallery grade white LEDs are more what you want, they
use a range of phosphors to generate a sunlight-like broadband spectrum.
RGB is however better than cheap white LEDs which have a gap in the spectrum.
However RGB will be less power efficient than almost any white LED.
To control the RGB style you'll need to be able to trim the relative currents for the colours
to set neutral white - normally three adjustable constant current sources would be used.
Interesting comparison, but - yes!
I'm not quite sure what you are saying here.Is that with all three channels in parallel?
A 100 W RGB LED would obviously use 33 W in each channel - more or less. At about 33 V that would be about 1 Amp each. So it is either massively over-driving one channel, or under-driving three.
Yes it is in parallel! But the thing is that the Aliexpress description says it is a 32-36V led. I should be about 3A all.
Do you think I need a current limiter at all as far as I have a 35.5V power source and a 36V led?
But you simply cannot connect all three in parallel because the three colours have significantly different voltage drops so the red would be bright and consuming almost all of the current but the other two colours very dim.
It seems like the green consumes more than two others.G - 1A, R - 0.65A, B - 0,55A.
That is a nicely designed board. It has three switchmode current regulators with PWM applied by a transistor each from a WS2811 chip. Yes, that is what yours should look like. Replicating it yourself is not advisable given your present state of knowledge (as in - you would not be asking these questions if you had the expertise to do it!
They don't sheep to my country)) I've done myself a nice PWM 25kHz dimmer base on Attiny13 for a white led panel. I've spent some time, but it works fine.
MarkT:
If you want quality light source for filming, an RGB led is not the ideal choice, you'll get
significant colour distortion. Art-gallery grade white LEDs are more what you want, they
use a range of phosphors to generate a sunlight-like broadband spectrum.RGB is however better than cheap white LEDs which have a gap in the spectrum.
However RGB will be less power efficient than almost any white LED.
To control the RGB style you'll need to be able to trim the relative currents for the colours
to set neutral white - normally three adjustable constant current sources would be used.
I need it to add some color to the scenes. It will work fine for my purposes.
You should use a constant current or current limiting supply for each channel.
I mean, you just-barely-can use a carefully trimmed constant voltage supply (ie, a DC-DC converter driven from your 36v) for each channel (where you've measured the current while carefully adjusting the voltage), if the LED is well heatsunk (since there's positive feadback if it gets hot - I use a monster CPU heatsink (the kind overclockers use) for 100W and 144W LEDs, with the fan always on); with these large LEDs the I/V curve is not as steep as for indicator LEDs; the "dynamic resistance" over the voltage range of interest is enough that CV supply is viable. I have multiple systems, working reliably, where I adjusted the voltage so the current was about 80% of spec'ed current, to give myself some margin, and called it good, and then put nailpolish over the voltage adjustment screw, and a matching streak of nailpolish on the heatsink that I thermal-epoxied the LED to.
This certainly isn't ideal, but I wasn't able to source an acceptable constant current supply that I could PWM for a price I was willing to pay (PWM is problematic with constant current supplies that aren't designed for PWM dimming) - I accepted that there was a risk of burning out the LED and bought extras, even though that hasn't happened (yet).
Regardless of whether you do it right, or halfass it with carefully trimmed constant voltage supplies, you still need a separate supply for each channel, though.
That all said - your claimed current measurements at 36v (a method of measurement that is a terrible idea, because it could easily burn out the LED - you start low, slowly ramp the voltage up while measuring current) raises two issues:
For a given configuration of LED dies (ie, X many strings of Y leds in series, connected in parallel), since red has the lowest Vf, if you apply the same voltage to each channel, the red should be drawing far more current. But it's not in your case, which is very strange.
If at 36v, assuming you haven't already damaged the red and blue channels, you're not getting 1A, the applied voltage needed for 1A is higher than 36v, so a supply based on a buck (step-down) topology won't work unless you start with a higher voltage.
Thank you for taking on the task! 
It was getting a trifle tedious.