Reliability of LEDs, and the curious case of a burnt out resistor

[Dr.Glum]

Hello your electronics experts,

I've been using LEDs for about 18 months, and in that time have only had one failure. That was my fault. I forgot to include the resistor in the circuit when I was testing a newly-bought one, so I applied 12volts across it! A brief flare of light and then it's dead.

 

Up till now the LEDs I've used have been from Maplins, or model suppliers such as Modelex. This week I have been deploying ones from a cut up Christmas set, which have the advantage of 1) being cheap and 2) securely prefitted with wires soldered to their legs. I read elsewhere in RMweb that such sets are often made up of cheap LEDs that fail to meet qualtiy control. I don't know if that has a bearing on what I've just experienced.

 

I wanted to test one such (call it item B, for b****r) before soldering a resistor to it. My little test rig of wires and chocolate blocks had just been used to test an identical LED (call it A) with a 3.9Kohm resistor. So I swapped B for A, and also changed the resistor for a 10Kohm one, in order to see how dim that might be. I reached across to plug in my power source and looked back to see what the light was like. No light, but curls of smelly smoke. "Oh blimey", says I. Actually that's not what I said, but nevermind. Hurried unplugging.

I removed the cooked resistor, checked my wiring to see if somehow I'd connected things up wrong. Nothing obvious, so swap in a new 10Kohm resistor. "Oh lawkes a mercy, you'll burn the house down!" That one did the same: one grey extremely smelly slug. Luckily I had no more 10K resistors! As the LED has shone no light in any of this, I conclude it has failed.

 

So here is the question: was the initial failure in the resistor or the LED? Have you had any LEDs fail in service, and if so, what sypmtoms (apart from the bl****** obvious)?

 

If anyone can explain why the resistor fried, I'd feel happier. My supply was notional 12volt going through notional 10Kohm gives 0.0012amp, watts = 0.0144, the resistors are rated at 0.6 watt continuous, so that's a factor of 41.6 safety!

I can no longer read the colour code on fried resistors, but my notes say they were marked brown/black/black/orange/brown. Now that I look again at my resistor code chart, I think that works out at 100 times 10 to the 3, which is 100Kohms. OK, they maybe gave me the wrong ones, but that would make the current tiny (10 times LESS). (I've bought some more from Maplin (and checked them on a meter) and they are marked brown/black/black/red/brown.)

I checked my supply: measured open circuit 11.6v and I calculate that to at least equal 0.6 watts (watts=IV or I=watts/V) the current would be 0.05amp and you'd get that on 12v with a resistance of value 240ohms (V=IR). (And that's assuming there is now zero resistance in the LED.) Makes no sense to me.

Maybe that LED (item B) is still OK - who knows?

[bertiedog]

You will not burn out the limiting resistor on any normal LED, so it must be a short somewhere that drew current, faulty LED, wiring etc, even applying the voltage across the resistor (of the value quoted), should not cause it to fry, unless the resistor value is all wrong and much much lower, drawing the current.

 

This sounds the most likely explanation, the resistance value is low and current high, and the LED blew instantly, and the current burst fried the resistor. 560 ohms is enough to work, but any up to 3K would allow the LED to work at lower output.

 

560 ohms would apply to 2 volt led and 20ma, so 1000 ohms would be completely safe. Different colours require more voltage, and may draw more current, but this is typical case for red. 1000 ohms is a safe test value for all leds on 12volts.

 

The Xmas LED lights, if the bulb is working, is a good source of LED's, they are second grade, which means a bit unpredictable but working, they just do not meet full specification. A lot of Ebay ones are seconds, the seller usually specifies if they are top grades. Seconds are no more likely to fail really, just give a dimmer output, or bad colour.

 

Hope this helps.

 

Stephen.

[bertiedog]

I can no longer read the colour code on fried resistors, but my notes say they were marked brown/black/black/orange/brown. Now that I look again at my resistor code chart, I think that works out at 100 times 10 to the 3, which is 100Kohms. OK, they maybe gave me the wrong ones, but that would make the current tiny (10 times LESS). (I've bought some more from Maplin (and checked them on a meter) and they are marked brown/black/black/red/brown.)

 

The colour code should be a three band plus tolerance, and brown, black, black, would be 10 ohms and explain the problem. The fourth band should be silver or gold. There are four band resistors made as well, but are unusual from Maplins.

 

12 volts applied to 10 ohms will cause 1.2 amps to flow, with the led in series it will not flow, but the LED failed in the burst, and probably burnt out leaving contact, which fried the resistor.

 

Stephen

[Dagworth]

The colour code should be a three band plus tolerance, and brown, black, black, would be 10 ohms and explain the problem. The fourth band should be silver or gold. There are four band resistors made as well, but are unusual from Maplins.

 

12 volts applied to 10 ohms will cause 1.2 amps to flow, with the led in series it will not flow, but the LED failed in the burst, and probably burnt out leaving contact, which fried the resistor.

 

Stephen

Most of the Maplins resistors are 5 band marked and have been now for several years.

From experience ALWAYS verify that you've got what you asked for when buying from Maplins.

 

Andi

[bertiedog]

Most of the Maplins resistors are 5 band marked and have been now for several years.

From experience ALWAYS verify that you've got what you asked for when buying from Maplins.

 

Andi

I don't buy many from Maplins*, surprised they do 5 band, still I think that the one that failed was 10 ohms.........

 

*the last value bag Maplin ones I had were three plus tolerance....I tend to buy bulk from Ebay these days.

 

Stephen

[meple]

The colour code should be a three band plus tolerance, and brown, black, black, would be 10 ohms and explain the problem. The fourth band should be silver or gold. There are four band resistors made as well, but are unusual from Maplins.

 

12 volts applied to 10 ohms will cause 1.2 amps to flow, with the led in series it will not flow, but the LED failed in the burst, and probably burnt out leaving contact, which fried the resistor.

 

Stephen

 

 

brown/black/black/red/brown = 1+0+0*100= 10k brown = 1% tolerance.

[bertiedog]

brown/black/black/red/brown = 1+0+0*100= 10k brown = 1% tolerance.

 

I am not querying your reading of the values, only querying whether it was 10 ohms or 10K

But.... if it was the full 4 plus tolerance code and 10K it simply would not have burnt out, more likely Maplins supplied the wrong value. (10K on 12v would only pass .0012 amps).

 

Stephen

[Gordon H]

Perhaps they might have been inductors and not resistors (maybe 10mH?).

Difficult to know now that both examples have been destroyed.

[bertiedog]

Perhaps they might have been inductors and not resistors (maybe 10mH?).

Difficult to know now that both examples have been destroyed.

 

Could well be, but codes differ, some with printed on code numbers rather than bands,

Stephen.

[Dr.Glum]

Thank you, guys.

I attach a photo of the burnt resistors, plus the replacement type that is supposed to be 10kohm and measured at that on the meter.

These are standard blue body resistors with five colour bands, of the sort that Maplins have been selling to me for about 18 months.

I'm attracted to Stephen's proposal that what I actually had was 10ohm resistors, especially as my original typed up notes described the markings as: "brown/black/black/gold=orange/brown" meaning that I originally thought the fourth band was gold, but looked at my colour code charts and found that that didn't make sense.

However, it does make sense if the bands were: number number power-of-ten tolerance quality

I have been given wrong ones in the past (56ohm for 560ohm), so it's possible. I have two observations about my local Maplins: 1) the staff are incredibly helpful and eager and 2) the stock control for small components is a worthy and whimisical contender for the Booker Prize for Fiction.

So, assuming a different theory isn't put up here in the next few days, I'll assume it was low value resistor trouble, which I can avoid in future with my recently bought meter. I won't be going back to demand my 34p back!

Cheers, Tony

 

PS the reason I went to the trouble to write the original post was a worry about whether I was unknowingly building time-bombs into my scenery - imagine, one dark night I briefly leave the loft to have a wee, and a resistor burns out and the card scenery next to it starts to smoulder, etc. etc.

Doesn't bear thinking about, or rather, it makes you think . . . .

[bertiedog]

Those are really over heated, the LEDs must have internally shorted and the power flowed for several seconds, but another check, are you really sure the voltage was only 12 VDC? I have just tested a similar 10 ohms resistor on 12 volts, and it takes about 5 seconds to overheat, and 15 to burn out, and this is without any load from the Duff LED added into the circuit. Taking it to 25VDC gives an instant blow, with the burnt finish.... They go "pop" on 150volts....(don't try this at home!).

 

Rule number one with electronics, never trust any thing, check and test the obvious, however much your sure..........A digital meter is vital to check resistors anyway!!

 

Once set up, and checked, there is little chance a 560/1000 ohms etc., would blow with LED's in the future, but you could limit the current supplied with a simple current regulator,( chip or transistor), in the 12V feed for the lighting, stopping all potential troubles, just set to match approx number of bulbs consumption.

 

For "buildings" and "gas lamp model"s, LED's can be under run a lot, most glow with 3000 ohms on 12 volt, cutting current to lower values as well. You may find the Xmas type LED's are very variable with the cut back low current, it is one of the reasons they are seconds.

 

Stephen

[Dr.Glum]

Hello Stephen,

I'm pretty sure it was only 12volt (nominal). After it happened I put my new meter on it and read 11.6volts open circuit, and the same again just now. I've just mimed the 'plug it, look at smoke, hurried unplug' and would say it was applied for about 5 seconds.

 

I wouldn't recommend it: it took a while to air the smell out of the room!,

Cheers, Tony

[bertiedog]

Hello Stephen,

I'm pretty sure it was only 12volt (nominal). After it happened I put my new meter on it and read 11.6volts open circuit, and the same again just now. I've just mimed the 'plug it, look at smoke, hurried unplug' and would say it was applied for about 5 seconds.

 

I wouldn't recommend it: it took a while to air the smell out of the room!,

Cheers, Tony

Then it's got to be a 10 ohm resistor..... those in the shot are not inductors....

Stephen.

[meple]

Then it's got to be a 10 ohm resistor..... those in the shot are not inductors....

Stephen.

 

 

of course if it's brown/red/black/black/brown that gives 120 ohm 1%

[hollywoodfoundry]

of course if it's brown/red/black/black/brown that gives 120 ohm 1%

 

Sorry, but brown, red, black (forget about the rest) is 1, 2 and No Zeros, so it is 12 ohms.

[Shadow]

Not correct for 5 banded resistors, (it is 120 ohm at 1%)

 

see link to chart/diagram, Resistor Codes

[bertiedog]

I did think about reverse, and 120 ohms would heat up at 1.2 watts, I just tried it and it takes about 30/40 secs to burn on 12 volts.

Anyway the value is behind it..... blame the idiots that designed a colour code with red, greens and browns, when such a large percentage of men are colour blind.

Stephen.

[96701]

I'm afraid that I have no idea what the colour bands mean. No1, I've hardly ever been able to make them out (and I pass all the colour blind tests for working trackside) and No2, I can never remember what the values are for the different colours, but I do have a belting digital multimeter!

[Dagworth]

I'm afraid that I have no idea what the colour bands mean. No1, I've hardly ever been able to make them out (and I pass all the colour blind tests for working trackside) and No2, I can never remember what the values are for the different colours, but I do have a belting digital multimeter!

 

The colour code is quite simple and is largely based around the spectrum.

0 Black

1 brown

2 red

3 orange

4 yellow

5 green

6 blue

7 purple

8 grey

9 white

 

It appears in a lot more than just electronic compoments and is pretty much universal.

 

Andi

[96701]

The colour code is quite simple and is largely based around the spectrum.

0 Black

1 brown

2 red

3 orange

4 yellow

5 green

6 blue

7 purple

8 grey

9 white

 

It appears in a lot more than just electronic compoments and is pretty much universal.

 

Andi

 

I'm sure I could commit the above to memory, but there seems to be a discussion above about 5 colour bands equalling either 12 or 120 Ohms, so I'll stick to using my meter.

[Dagworth]

On a four band resistor you have two significant numbers, then a multiplier (number of zeros after the significant numbers), then a tolerance code, five band resistors are the same but have three significant numbers rather than two.

The multipliers have a couple of oddities in gold (0.1) and silver (0.01)

The tolerances are brown (1%), red (2%), gold (5%,)and silver (10%)

 

This may be useful http://diyaudioprojects.com/Technical/Electronics/

 

Andi

 

 

[Dr.Glum]

Hi guys, again,

 

I earlier wrote: "my original typed up notes described the markings as: "brown/black/black/gold=orange/brown" meaning that I originally thought the fourth band was gold, but looked at my colour code charts and found that that didn't make sense."

 

Now that I've looked at the diagram on the site that Dagworth posted ( http://diyaudioprojects.com/Technical/Electronics/ ) I can see that this shows that the 4th band can be gold or silver which I hadn't noticed on the charts I was using. Therefore I read the fried resistors as:

one, zero, zero, times 0.1, 1% = 10ohms (as Stephen said)

If I'd had that chart or my new meter when I first brought the resistors home, I would not have been "frying tonight!"

 

Cheers, Tony