Preventing LED Issue (Thermal Runaway?)

Hey Guys -

I've got a basic project which lights 32 LEDs which are set up in series of 4 all in parallel as shown below. As each LED is 3V / 20ma, I didn't add in resistors or anything to each series.

Unfortunately, I've twice now run into an issue (thermal runaway?) where the LEDs are only running at about 25% brightness when set at 100%. The first time this occurred, I found with the multimeter that each series is getting 3V instead of 12V. Sometimes a series flickers, too. When I replace the series, everything goes back to normal for a brief period.

How can I remedy this to where it won't be an issue in the future after the next repair?

Ignore typo saying x24 LEDs in diagram - its x32 :slight_smile:


You need series resistors for each string of 4 LEDs. Basically LEDs are shorts so you're shorting your circuit.

Overloading your power supply by drawing 'unlimited' current.

No resistors and no supply.

Supply is available. GND is also available everywhere where necessary, but the current limit for the LEDs is missing, so the power supply switches off as soon as it goes to 100% (maybe even a little earlier). He apparently built a PWM control, which is why it also works in the lower area without the resistors, because that reduces the current flow through the LEDs.

The strips in the schem run sideways !

Drawing error, Fritzing problem, Since according to the description of the TO the circuit already works, at least as long as it keeps the PWM low enough, suply must be present.

Thank goodness I didn't build it to test :stuck_out_tongue_winking_eye:

Appreciate the reply!

So does it matter the resistor size? Would something like a 10 work since voltage is close or better to use something larger?


Nope. 150R. This is how you work it out:

(12V - 3 LEDs * 3V) / 0.020A = 150R

EDIT: forgot to say, group your LEDs together in threes, not fours.

You've got 12V going through 4 leds with 3V forward voltage each....

there's not enough push left to get more current through than makes low brightness.

1 Like

Ahh, I get it - Thanks!

As leds heat up (with more than trickle current) their forward V lowers.

Imagine that your leds are on one side of a seesaw and that resistor is on the other. A little extra R won't hurt.

With high power leds you want constant current circuits while resistors are only current limiting.

@bzowk Your LEDs are not the high power type that @GoForSmoke mentioned. They may be sold as "ultra high brightness" or some such, but they are just normal 5mm, 20mA LEDs. Their temperature won't change enough to affect their forward voltage significantly. @GoForSmoke is talking about the type of LEDs that are 1W or 3W for example and can take 100s of milliamps and require heatsinks to keep them cool.

1 Like

It depends of what resistor you use, you can shorten the led life without getting smoke.
When 1 resistor is limiting 12V for 3 leds, that's 3x the change.

According to the makers, COBB leds run hot without damage.

I would not describe COB LEDs as "high power LEDs', because they are in reality arrays of low power LEDs in a single package. Having many, maybe many 10s, of LEDs spread over the package also spreads out the heat, so it's not so intense in any single spot, and the package includes a metal, usually aluminium, heatsink. So the individual leds don't get as hot as a "true" high power led, and their forward voltage does not change as much.

1 Like

About those, I am only informed by the sales info on the old dealextreme site.
After some reading up, COBs are a nice step up from bulbs.

Good Morning -

Just to note, I rewired everything last week with new LEDs grouping them in series of 3 + resistor as suggested and so far it's lasted. Appreciate the help!