Q: LEDs blowing in Parallel circuit. Any ideas?

[MM3Canuck]

Hi!

Im new to this whole wiring LED's thing. But have some basic knowledge!!!:thinking:

I have 12 LEDs in parallel, here are the numbers:

Power source: 12Volt Motorcycle battery (will be installed on a bike as well!)

LED specs:

Emitted Colour : Yellow
Size (mm) : 3mm
Forward Voltage (V) : 1.8 ~ 2.2
Reverse Current (uA) : <= 30
Luminous Intensity Typ Iv (mcd) : 4000(Typical) ~ 5000(Max)

I used this site to calculate the resistor needed for my 12 LED parallel circuit. http://ledcalc.com/#calc

Values I plugged in:
Supply Voltage
VOLTS
Voltage Drop Across LED
VOLTS
Desired LED Current
MILLIAMPS
How many leds connected


The answer was: 33Ohm - 5Watt resistor.

Everytime I connect the resistor and turn on the power, all LEDs blow instantly. Is the resistor too big? Am I not using the right numbers in the formula??

Any ideas are appreciated.

Dennis​

 

[Turbo Guy]

The LED wizard I use is here:
http://led.linear1.org/led.wiz

It shows a series/parallel arangement, 4 in series and 3 sets sets in parallel. Each series set requires a 180 Ohm 1/4 watt resistor. Not sure way you would want to run them in parallel and have no idea how hard you want to drive them, I used 20 mA.

For a single 12V / 2.1V / 20mA 560 Ohm 1/2 watt.

 

[Mr Happy]

Well, a 33 ohm resistor on a 12 volt supply should allow a current of about 360 mA, which should give you about 30 mA per LED. On the surface that is OK, but maybe that is not the problem.

The problem is that you should not connect LEDs in parallel, it is not reliable. Unlike bulbs, LEDs do not automatically distribute the current between them. You will end up with one LED getting more current than the others and failing. Once it fails the next one will fail, and so on like a chain reaction.

What you should do instead is connect a few LEDs in series and give each series chain its own current limiting resistor.

So for instance if you have LEDs with Vf = 2 volts, then put three in series to get Vf = 6 volts, and now you need to drop 12 - 6 = 6 volts over the resistor. With a current of 20 mA (= 0.02 amps) the resistor would be 6 / 0.02 = 300 ohms.

Repeat four times in parallel, like this:

+12 V ---+--- resistor --- LED --- LED --- LED ---+------ 0 V
         |                                        |
         +--- resistor --- LED --- LED --- LED ---+
         |                                        |
         +--- resistor --- LED --- LED --- LED ---+
         |                                        |
         +--- resistor --- LED --- LED --- LED ---+

 

[asdalton]

Yes, that web page gives the option for designing a purely parallel circuit with 1 resistor, but you never actually want to do this. The part-to-part variability of Vf between LEDs makes this unreliable.

The main page on that site ("The Guru") gives the much more practical series/parallel arrangement that Mr Happy described, with individual limiting resistors for each series chain.

 

[PeAK]

.
.
.LEDs do not automatically distribute the current between them. You will end up with one LED getting more current than the others and failing. Once it fails the next one will fail, and so on like a chain reaction.
.
.
.

Just to inform yourself on the point above, you can do a google search on the terms "current hogging", "parallel", and "thermal runaway".

In terms of an analogy with 5 parallel hoses (i.e. LEDs) transferring water from a common spout (i.e. Battery), the hose with a larger diameter (Lower Vf) would transfer more water (i.e. current) and let's suppose it heats up with this added flow due to friction. Now if this heat causes it to expand further in diameter(i.e. lower the Vf) , then even more flow (and heat) would occur until it failed. In order for this explanation to be consistent with all of the LEDs blowing, the expanding hose would have to suddenly fail by having something clamp it shut to stop(i.e. open circuit) all flow in it. This process would repeat for the next largest hose remaining with the next largest diameter.

PeAK

 

[TranceAddict]

LED doesn't has fixed resistance, take a look at the curve, it is dynamic and different to each other even though it is totally identical (same bin, same flux).

 

[JWP_EE]

Just a thought.

You say they all blow instantly. Even with slightly different Vf I wouldn't expect that. I would expect some to be brighter than others for a short time, burn out and the remaining one getting bright and going out but not all going at the same time. Do you actually see them light? If not maybe you have them wired backwards.

BTW you never stated what the rated forward current is. If it is less that 20ma that could be a problem.

 

[VidPro]

i like Mr Happys plan with multiple series parellel, unfortunatly most of the online calcs dont work that out to well, make sure you include reality into the math.

we use that idea here, no use wasting all the power out something that doesnt output light :-) proper balance of leds to get to a nice voltage, and resistance to insure some flexibilty in voltage.

also to note a 12V battery is not 12V , it rarely is there :-) sometimes higher sometimes lower, but rarely actually at 12. this is not to say that one cannot parrellel and series multiples or resister them. they should just do it with the understanding that the voltage varies , and then run them lots lower and use lots more of them. or better still use higher powered ones that can cope with greater variances. like even in the 5mm led the "pihranna" square flat ones , have a much better ability to remove the heat.

then it always helps to start LOWER at any rate, for testing toss in an (extra even) wariable resister , potentiometer or whatever you want to call it set to its highest resistance level, then bring it on down slowly , untill you see the light, , , , then burn it out after :-)

i have run hundreds of parellel leds off a single tied in resister (instead of the proper one for each), and as long as you stay a bit under, they all light fairly evenly, with a few weird ones acting different, like in a string of 40 , 38 of them would happily act similar , one will decide to croak and one will be brighter than the others. its sorta like batteries, you can potentially assemble items that match eachother pretty well, or you can assume it just cant be done at all. course i only paid 10c each :-) so i didnt have much to lose.

also why burn out a whole array, when you can burn out just the first few series sets instead :-) once you have a single fully operational single series set, or resistered single or whatever, testing that one, is little different than testing entire arrays of that same curcuit repeated. (in the situation where the battery voltage is not dropping under its load, like here).
once you get a single "string" of the leds running, and you test it with the various votlages, then you can make many similar strings that would act the same, then just tack them in.

 

[Turbo Guy]

Just in case there is a problem with the link I provided in post #2 or my explanation was to brief here is a schematic

 

[LukeA]

nm

 

[BoarHunter]

12V gives 36.4mA trhrough a 330 Ohm resistor directly connected to it and 30.9mA if one LED (1.8V drop) is connected in serie. This is not capable to blow the LED ! With 12 in paralell, they should barely tight up and in now way blow up !

So something must be basically wrong with your set up, may be just the resistor that could be for example 3.30 Ohm !!! Did you measure it ?

Anyway, as stated above, this set up is not adapted to LED.

Besides, I am amazed one need a computer program for this kind of basic arithmetic.

One would gain a lot by learning the basis of electricity to have some understanding of it instead of relying on magic recipe that lead to an erratic approach to problem solving.

 

[Benson]

12V gives 36.4mA trhrough a 330 Ohm resistor directly connected to it and 30.9mA if one LED (1.8V drop) is connected in serie. This is not capable to blow the LED ! With 12 in paralell, they should barely tight up and in now way blow up !

Sure, but that doesn't matter, since it's a 33 Ohm resistor, so you'd get 10 times the current.

Besides, I am amazed one need a computer program for this kind of basic arithmetic.

One would gain a lot by learning the basis of electricity to have some understanding of it instead of relying on magic recipe that lead to an erratic approach to problem solving.

Agreed; though if one is doing a number of LED arrays, the program will save time, it should not be necessary.