Multiple and different coloured LED plant light

So I am trying to make a plant light and ideally I would like to be able to include 50-100 LEDs in the light and would run off the mains which is at 12V.

My research tells me the best way to do this is in a series parallel circuit.

From what threads and websites I’ve read I can’t see that this many LEDs would be a problem, what I’m a little bit confused about is whether I can use 4 different coloured LEDs with very different resistor needs?

I have used the LED series/parallel array wizard (http://www.led.linear1.org/led.wiz) but this only does single types of LED. Can I take the schematic drawings of each type of LED and put them together?
Will all these LEDs make the light to hot?

LED info:
LED colour &Wavelength Forward Voltage (V) Forward Current (mA)
Infra-Red, 680nm 1.8 50
Red, 660nm 1.9 30
Blue, 465nm 3.3 30
UV, 400nm 3.5 30

I hopefully would have 20 Infra-red, 30 red, 30 blue and 20 UV

Your design shows proper thinking except that LEDs aren't very dependable about their exact forward voltage. The red setup particularly scares me because you are only allowing .6 volts across the resistor. Six LEDs in series can easily be "off" by a tenth of a volt each and that could cause the results to be either no light or smoke or something in between, depending on which way the Volts Forward went on that batch.

To be more conservative (safe) you should allow more voltage to be wasted in each resistor and measure the results of each string as you build them, just to make sure. (My rule of thumb is to waste 15% to 20% of the voltage in the resistor.)

I have not analyzed the heat aspect.

cheers #12,
I was worrying more about blowing my LEDs up so that is useful to know. I will measure the LEDs voltage and try and match them. I suppose I'll just have to measure heat as I go along and maybe install fans. I can't believe how much difference there is in voltage use in LEDs.
thanks

This whole pile uses about 12 watts. Whether heat becomes a problem depends on how much space you are going to use. 1 CFM would fix this if they are all in a wad. That seems to indicate that a fan is not necessary if these will be spread out over a square foot or so.

I usually avoid the problem of uncertain and variable LED Vf by using a linear regulator in constant current configuration. The efficiency is about the same as resistors.

What's the lowest "waste" voltage that a current regulator would use up at 30 or 50 ma? That answer will tell how many LEDs the series strings will be allowed to use.

I ask because things are constantly changing in that part of electronics.

I usually just use an LM317 and assume that I need to allow for 3V of headroom both to accomodate the junction drops in the regulator and the possibility of several higher than typical LED Vf's in series.

The 317L will work for this. @ 55C/W you can theoretically push it to 1.8 watts and this case only needs about .15 watts.

So if I put a LM317 at the start of my circuit, the voltage after it drops to 9V.I have been reading Bill Marsdens stuff on it and I understand that it regulates input voltage but how does it regulate the voltage drops of the LEDs?

And I have adapted a LM317 IC design and was wandering if you guys think its alright and how do i choose the resistor values of R1 and R2?

That's the wrong circuit topology. Take a look at the LM317 datasheet. There's an application example showing how to use the device as a current regulator. There will be only one resistor. You will need one regulator and resistor for each series string of LEDs. You can't mix 30mA and 50mA LEDs in the same string.

Do you really want to operate the LEDs at their absolute maximum rated power?
Why not derate them so they last longer and run cooler?
22mA in the red LEDs that probably have a maximum of 30mA.

It's going to be back to your original circuits but each resistor will have an LM317L chip as its companion and the LED strings have to be figured to leave at least 3 volts for the regulator and resistor to use. It costs more than twice as much money but it's safe, reliable, consistent, and repairable because it compensates for trying to use parts that don't exactly match in breakover voltage.

My calculator says 27 ohms for the 50 ma strings and 42 ohms for the 30 ma strings, but that 42 ohm number is scary close to no safety margin. Better use a 47 ohm.

Edit: Guru has a valid point. As usual, I just assumed that the given information was correct and didn't think about evaluating the original decision.

You should look at the American datasheet for the LM317L from National Semiconductor who invented it, not the Italian copy that is full of errors.

The Italian datasheet does not say how much heat causes it to quit and shows a wrong graph with an output of 1.2A.
National's datasheet lists its thermal resistances and shows a graph of it limiting the current starting at 180mA.

I got this from National. It says Texas Instruments. I assume that is because (I heard) TI recently bought National or the other way around.

Yes, Texas Instruments bought National Semi a few months ago. TI's datasheet has the date 2004 on it so it is not the newest one.

National's datasheet has (had) graphs on it.

I would like to try and keep the LEDs running for as long as possible, so the LM317 chip sounds good, some of the LEDs were quite hard to get.

I'm not sure about the orientation of the LM317 chip, looking at the specs pin 1 (the adjust) will have the resistor and the LED string, pin 2 (in) will be coming from the power source and pin 3 will be out. I have tried to draw this in the attachment.

On the last string where would I connect the third pin to?

I really appreciate all the help guys, without it I would probably have destroyed all my LEDs

Instead of connecting your circuit completely wrong, why didn't you copy the circuit of the current regulator from the datasheet of the LM317?
Didn't you know that all the current regulators have exactly the same input from VCC?
42 ohms is not a standard value. 82 ohms is standard and two connected in parallel make 41 ohms which is close.

Here is a copy from the datasheet and your circuit with corrections:

EDIT: I made a mistake and copied the errors on the posted schematic where the OUT and ADJ pins were swapped. It is corrected now.

Guru already said it but I'm not going to waste this drawing I did while he was typing.

I goofed and made some mistakes copying the original schematic. The ADJ and OUT pins were swapped. It is corrected now.

thank you so much guys, I apologise for my slowness on the uptake. I guess there's nothing left for me but to make the circuit.

Thanks again