The Tech Blog

Last time we saw the pros and cons of driving LEDs in a series drive. An alternate approach is a parallel drive. As the name suggests, parallel drive involves sets of LEDs in series, interconnected across each other. Since you can have many parallel arms, the overall voltage required across a string is lower, hence even with a constant current or a buck regulator, you can drive a lot of LEDs. That’s the biggest advantage of a parallel drive. The source will ideally split currents across the different arms. Practically though, because of variations in Vf’s, each arm will have a slightly different forward voltage, which will cause a slight mismatch in currents through the arms which may cause some strings to glow brighter.

To fix this, you would need a current mirror circuit across each of the arms. For 2 parallel arm circuits a simple NPN current mirror will ensure that the current in one arm mirrors the other. If there are more than 2 arms, there are certain LED driver ICs that can control multiple arms and ensure that the current is the same in each of them. But they are expensive.

In parallel LED setups, if one arm’s LED fails as an open circuit, the whole arm goes out, and the constant current source diverts excess current to the remaining arms. If you’ve fine-tuned the LED drives to operate at peak current in each arm, a failure in one arm can lead to an increase in current through the others. This, in turn, heats up the LEDs, causing a drop in Vf and even more current flow. It sets off a chain reaction, ultimately leading to the failure of all arms over time.

So when driving in parallel mode always make sure the current in one arm is not driven at LED max current. Additionally, whenever possible, incorporate more parallel arms (as long as the input current source can handle it) for increased longevity. With more arms, even in the event of a failure in one arm, the rise in current in the others will be a relatively small factor, as there are more arms to share the current load.

We will cover one more variation of LED drives next time.

We started off with what it takes to drive a single LED. Now the easiest way to drive multiple LEDs is to connect all the LEDs in Series and limit the current in through them via a normal resistor. If you drive it with a large enough voltage. It works, but is inefficient. Once you find the resistor value, the key thing will be to figure out its size for power dissipation calculation. For high current LEDs and with nominal resistors, Power-dropped is related to the square of the current and power rating needed increases drastically. Well this power is dissipated as heat on the resistor and is not something we want if we need in energy efficient systems.

So to fix this we use a constant current source which drives the LEDs with a constant current. Current is usually fixed by setting a resistor which will be referenced to a smaller voltage so no need to worry about the sizing of that one. This is probably the most common mode of driving LEDs out there. But there is a problem here, the drive voltage output of the string of LEDs should always be higher than the sum of Vf’s as the most common topology will be a buck mode. ie) If your input power supply is say a USB of 5V, then you can probably drive only a couple of LEDs in series(that too if it’s not white as they have a Vf of 3V+). if you have a lot of LEDs to be lit up, then this is not a right topology for you. For that you now need boost mode topology or parallel drive, which is a whole topic in itself. Boost topology will increase the output voltage relative to the input but only to a certain point. One point of consideration is frequency of switching of your buck/boost regulator. We will get into the why of that later on in this series. Also, the major issue with series drives is if one of the LEDs blows up, the entire string stops working.