Electronics Questions & Answers ← browse Q&As

I would like to pick the brains of anyone here who has worked with LED's. For a single LED it is simple to define a suitable ballast resistor using the VI curve. I have an application requiring 40 to 50 RGB LED's, so 120 to 150 units total. A possibility is to connect all of one colour in parallel, add up the currents on the VI curves and define a single ballast resistor for the parallel group. Sounds ok but raises the question of current sharing if individual VI curves are slightly different. Does anyone have any insight on this? Another possibility is to dispense with the ballast resistor and use a current source power supply but the current sharing question remains. Another possibility is to connect all of one colour in series, add up the individual voltages to define a composite VI curve and define a ballast resistor based on this (and the higher voltage) or to use a current source power supply of appropriate voltage and dispense with the ballast resistor.

I would be interested in the thoughts of any one on this subject.

After looking over the data sheets for these things it almost looks like they were designed to be impossible to connect them in series so I'll have to give up that idea. So back to the drawing board. Foxx

Responding to where you say "use a current source power supply but the current sharing question remains."

Constant current is an interesting way to go. Remember that current is the same at every point around a circuit. The power supply's voltage may get high to do it, but each LED will see exactly the current you specify. If your constant current power supply is sourcing for example 20 mA, every point in the circuit will see 20 mA. Every wire, every LED's PN junction. And you won't need a single resistor. You will need LEDs with six terminals, rather than RGB LEDs with only four pins (ie not common anode nor common cathode).

I can imagine three constant-current power supplies and 50, six-pin RGB LEDs. You may or may not like the fact that the CC power supplies are putting out over 100 V DC to pull it off, but each LED would get the specified current, without you having to use any current-limiting resistors.

Here is an example with one color and three LEDs, trying to illustrate that the voltages add up, and the currents are the same in a circuit.

In your case you'd need 47 more LEDs, and then two more instances of the circuit for all three colors.

Another thing to not like is that if a single LED fails open, all the LEDs in that circuit (presumably that color) will go out. Interestingly, if an LED were to fail shorted, only that one LED would go out, the rest would remain lit, and would continue to get the specified current. The CC power supply would continue to source the configured current, and a side-effect would be that its voltage would go down.

You can consider three Mean Well LCM-40 CC supplies and a number of six-pin RGB LEDs which draw 350 mA per die. Assuming the highest die voltage is 3.8, from a cursory look you are limited to 26 LEDs in your circuit. Not bad for a $35 power supply which is in stock at Digikey.

https://www.digikey.com/en/products/detail/mean-well-usa-inc/LCM-40/7704706

https://www.meanwellusa.com/upload/pdf/LCM-40(EO)/LCM-40-spec.pdf

I don't know your application, but here is one example of a six-terminal, 350 mA RGB LED:

https://www.digikey.com/en/products/detail/seoul-semiconductor-inc/F50360/2181973

https://www.mouser.com/datasheet/2/363/seul_s_a0001754609_1-2281995.pdf

Please watch out for the high voltage which CC power supplies can put out open-circuit (or in your case, when operating at a high output voltage due to the LED string configuration.)

By definition of a potential, the voltage between two points, A and B, has to be the same (statically) whatever the path is. So, with two LED in parallel, a red one and a green one as example, only the red one will turn on, say, at 1.6 Volt while that voltage won't be enough for the green one. Even with two red ones, if one is at 1.5V while the other is at 1.6V, the lowest one could burn off long before the one at 1.6V gets reasonably well lit. That is a reason why you need a resistor FOR EACH LED in parallel.

In series, the total current is acceptable. Using a pulsing very high voltage can be a solution. Take a look at ElectroBoom - Driving a LED

Thanks for your thoughts to Marshal and Vandergast. I think I've found out what I was looking for---Your thoughts are about the same as mine and there seems to be no magic solution; just the brute force type.