Continue to Site

Welcome to our site!

Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

  • Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

white leds problem(flickering/die:S)

Status
Not open for further replies.

TnF

New Member
Hello electronics experts. I am a uni student studying MEng in UK. Anyway i'm from Cyprus and last year i made some custom led lightning for some bumper lights of my car. I connected many leds in parallel, and each row had 4 leds in series. The leds are ebay quality;p and the listing gave the following specs:

Technical Summary
RoHS Yes
Package F 3mm
Packaging bulk
Dimensions F3mm
Lead Spacing 2.54mm
Color White
Brightness 11000 ~ 13000mcd
Wavelength 8000k-9000k
Lens Waterclear, White
Forward Voltage 3.5V
Current Rating 20mA
Reverse Voltage (V) 5
Continuous forward current(mA) 20
Peak forward Current(mA) 40
Power Dissipation (mW) 100
Operation Temperature Range() -20~80
Storage Temperature Range() -20~80

Since the voltage supply was standard 12V-14V(since the alternator of the car produces about 14V) i assumed power supply to be 14v. So with a given forward voltage 3.5V i connected 4 leds in series at each row as 14V/3.5V=4 After using them couple of days(not even 500hours of use), some started to die, others started to flickering, others sometimes from dead came alive, and finally they all died-note: I remember that they died rows by rows-not a single led died alone in a row) Then i replaced them all again, same problem, and finally last christmas when i went back to Cyprus, i replaced them again putting a resistor and a diode in the power supply...The intensity dropped but at least for the time i was there, not a single one died. What i am assuming is that the forward voltage is 3.5V and working voltage must be lower(since resistance may change with temperature increase since they were hours opened each time)..Also another thing is that despite sealing them well, when i wash the car, if water goes into the contacts and affects it somehow...

Any help appreciated(what is really weird, is the flickering and the reborn of the leds..)

Cheers, Ken
 
Last edited:
No resistors in series with the LED groups?
 
Last edited:
No resistors in series with the groups?

no...at first i haven't used any resistor at all...but even so the circuit is between the specs...why i had those problems? The power supply is less than 14V for sure. Then(last christmas) i added a resistor between the leds and the power supply. I recently learned that resistors must be connected one for each series, and avoid parallel configurations...so why i have this problem, and what is the correct configuration i must use?
 
I suggest 3 LEDs in series, and 180 Ω in series, or 2 in series and 360 Ω. Have as many strings in parallel as you want, but each string has its own resistor and no LEDs should ever be directly in parallel with another.

The forward voltage for LED is not the voltage that you should run them at. The current is what you should run them at, and the forward voltage is what you should allow for.

If you alter the supply voltage a little bit, the current changes a lot, so the variations of voltage between the engine running and not are enough to go from the LEDs being off to being burned out, with 4 LEDs and no resistor. With 3 LEDs in series you will get quite a bit of variation, and with 2 LEDs the current won't change significantly.

I've recently converted some car interior lights to white LEDs. They have the additional complication of being dimmed rather than switching off. For the first, I used two LEDs in each string and a simple resistor, but of course the LEDs turn off abruptly as the voltage drops below 6 - 7 V.

On the second, I used a boost converter to increase the supply voltage and I ran 5 LEDs in series. 5 is enough so that they won't turn on from the battery voltage alone. The boost converter is arranged to control the current rather than the voltage. I also made that current proportional to the supply voltage, so that the light dims as the voltage is reduced by the other electronics in the car.

The other advantages of a boost converter are that the current can be accurately controlled, so that it doesn't change with supply voltage, and the efficiency is better as there is little or not voltage drop in resistors. The circuits are more complicated.
 
I suggest 3 LEDs in series, and 180 Ω in series, or 2 in series and 360 Ω. Have as many strings in parallel as you want, but each string has its own resistor and no LEDs should ever be directly in parallel with another.

The forward voltage for LED is not the voltage that you should run them at. The current is what you should run them at, and the forward voltage is what you should allow for.

If you alter the supply voltage a little bit, the current changes a lot, so the variations of voltage between the engine running and not are enough to go from the LEDs being off to being burned out, with 4 LEDs and no resistor. With 3 LEDs in series you will get quite a bit of variation, and with 2 LEDs the current won't change significantly.

I've recently converted some car interior lights to white LEDs. They have the additional complication of being dimmed rather than switching off. For the first, I used two LEDs in each string and a simple resistor, but of course the LEDs turn off abruptly as the voltage drops below 6 - 7 V.

On the second, I used a boost converter to increase the supply voltage and I ran 5 LEDs in series. 5 is enough so that they won't turn on from the battery voltage alone. The boost converter is arranged to control the current rather than the voltage. I also made that current proportional to the supply voltage, so that the light dims as the voltage is reduced by the other electronics in the car.

The other advantages of a boost converter are that the current can be accurately controlled, so that it doesn't change with supply voltage, and the efficiency is better as there is little or not voltage drop in resistors. The circuits are more complicated.

Thanks very much. I don't want to add extra weight in the car so i won't be running a controller. So the best thing to do is to take measurements of current along the leds when the engine is running, so i can find a resistor value which is stable at all times. Also is there a specific current that gives maximum luminosity(and probably lower lifetime) other than the recommend 20mA? Also can you explain me the effect of the led flickering?
 
Measure your cheap LEDs and you will see that only a few might have a forward voltage of 3.5V. Datasheets for LEDs always have a range of voltages because they cannot make them the same.
The ones that burnt out had less forward voltage than 3.5V.

If you want maximum brightness then you apply too much current but then they won't last very long (1minute? 5 minutes?).
 
Measure your cheap LEDs and you will see that only a few might have a forward voltage of 3.5V. Datasheets for LEDs always have a range of voltages because they cannot make them the same.
The ones that burnt out had less forward voltage than 3.5V.

If you want maximum brightness then you apply too much current but then they won't last very long (1minute? 5 minutes?).

before i fitted them in the housings i tested 3 in a series over a 12 voltage for a couple of minutes and they still worked...though in the car i had those problems...what i noticed also before the fail, it that despite the higher brightness, the brightness dropped significantly before they failed...
 
Automotive electrical systems have inductively-induced glitches on the electrical system. Every time a fuel-injector, motor, coil, solenoid, turns off, there is likely a spike on the 12V bus. When you connect LEDs directly across the bus with no current limiting resistor in series, it is the glitches that are likely damaging the LEDs. That is why the failures seem so random.

Running LEDs with no resistor (or other current limiter) is a cardinal sin...
 
Thanks very much. I don't want to add extra weight in the car so i won't be running a controller. So the best thing to do is to take measurements of current along the leds when the engine is running, so i can find a resistor value which is stable at all times. Also is there a specific current that gives maximum luminosity(and probably lower lifetime) other than the recommend 20mA? Also can you explain me the effect of the led flickering?

The boost controller for a few LEDs at 20 mA only weighs a few grams. You can also have a constant current circuit which would weigh even less.

The resistance of a resistor is generally very stable. Don't run resistors beyond half their maximum power as they get very hot at maximum power. The voltage will vary when the engine starts and that will vary the current a lot more.

For example, if you go with 3 LEDs and 180 Ω, when the engine stops and the voltage is lower, say 12 V, the LED voltage will be much the same, at about 3 * 3.5 V = 10.5 V. That only leaves 1.5 V across the resistor so the current will only be 8.33 mA, while it is 19.44 mA. There is less current variation with 2 LEDs, and virtually none with a constant current circuit or a boost controller.

The light from LEDs is virtually linearly related to the current. So at 10 mA you get just about half the light that you get at 20 mA. You probably get slightly more than half the light at 10 mA. You can run them at any current you want. It is a trade off between brightness and life. If you never go above 15 mA they will probably outlast the car. If you go to 25 mA they may not last for hours.

As Audioguru points out, there is a range of voltages, and if they are low and you use a resistor there will be more current, which is why a larger value resistor would lead to more reliability. It is another reason to use constant current circuits.

I wouldn't worry about the flickering or other funny effects. You put the LEDs in a circuit where there was too much current some of the time, and they died. The fact that they didn't die suddenly isn't relevant. If you treat them correctly they will work for years.
 
The boost controller for a few LEDs at 20 mA only weighs a few grams. You can also have a constant current circuit which would weigh even less.

The resistance of a resistor is generally very stable. Don't run resistors beyond half their maximum power as they get very hot at maximum power. The voltage will vary when the engine starts and that will vary the current a lot more.

For example, if you go with 3 LEDs and 180 Ω, when the engine stops and the voltage is lower, say 12 V, the LED voltage will be much the same, at about 3 * 3.5 V = 10.5 V. That only leaves 1.5 V across the resistor so the current will only be 8.33 mA, while it is 19.44 mA. There is less current variation with 2 LEDs, and virtually none with a constant current circuit or a boost controller.

The light from LEDs is virtually linearly related to the current. So at 10 mA you get just about half the light that you get at 20 mA. You probably get slightly more than half the light at 10 mA. You can run them at any current you want. It is a trade off between brightness and life. If you never go above 15 mA they will probably outlast the car. If you go to 25 mA they may not last for hours.

As Audioguru points out, there is a range of voltages, and if they are low and you use a resistor there will be more current, which is why a larger value resistor would lead to more reliability. It is another reason to use constant current circuits.

I wouldn't worry about the flickering or other funny effects. You put the LEDs in a circuit where there was too much current some of the time, and they died. The fact that they didn't die suddenly isn't relevant. If you treat them correctly they will work for years.

thanks...i may run smd resistors in each row since wattage of the resistors would be only 0.07w per resistor at a 3leds row running at 14V. Or i can use this regulator your using...this **broken link removed** will work? (or is there anything more compact(since i'll have to put it behind the light housing and must be protected from water and so on..?

edit: found also this one which is more compact: **broken link removed**

also which is the most effective led's by surface area?(most luminosity by area?) smd's or t1 type?(or bigger ones?)
 
Last edited:
You must put current control in each series string of LEDs. It makes no difference where in the string you put the current control. If you want to use SM resistors, that's fine, and one might take the power, but you should go to 1206 size or larger.

If you are using smaller ones it would be better to use a few to spread out the power. The resistors can be in series or parallel, but you need to adjust the resistance either way. You can have the resistors anywhere in the string.

Those two regulators are buck regulators, and they are both designed for higher current systems than you are using. The first is adjustable between 40 mA and 600 mA, and the second is fixed at 1A
 
You must put current control in each series string of LEDs. It makes no difference where in the string you put the current control. If you want to use SM resistors, that's fine, and one might take the power, but you should go to 1206 size or larger.

If you are using smaller ones it would be better to use a few to spread out the power. The resistors can be in series or parallel, but you need to adjust the resistance either way. You can have the resistors anywhere in the string.

Those two regulators are buck regulators, and they are both designed for higher current systems than you are using. The first is adjustable between 40 mA and 600 mA, and the second is fixed at 1A

Thanks...I'll go with smd resistors per row or driver module...I remember i was using 4*12rows= 48leds and i want more lightning than this..I was using t1 leds, if i go with smds, lets say 96smd leds will i gain a significant amount or light? or let's say which is brighter? a t1 or a smd?(btw i am using white/clear leds)
 
Most cheap very bright LEDs have a focussed case that produces a narrow beam that cannot be seen if it doesn't shine at you.
Every SMD LED that I have seen are bright and have a very wide beam angle.
 
Hello there,

Some LEDs that blink on and off and then go out completely do that because of over current. Running them at too high a current causes them to overheat and then strange effects take place like that.

There were some commercial white LED flashlights made with dropping resistors who's values were too low and what would happen was after running the flashlight for a few minutes or so one of the LEDs would start blinking on and off, then eventually go out completely. If the light was kept on then another LED would start blinking, then go out. This would happen until all the LEDs were dead completely. Increasing the series resistor for each LED fixed the problem.
If the light was turned off after the blinking started but before the LED died completely and the light was turned back on after a cool down period the LED would come back on fully lit and not blinking, but after a short time would start blinking again and die if left on too long.
 
Last edited:
All traffic lights in my city were replaced with LED lights to "save money" on replacing light bulbs that burn out.
The new LED lights and the labour to replace them all cost a fortune.
It has been a few years and yesterday I noticed that every traffic light I saw had some LEDs burnt out. I saw one LED blinking. I guess it is safer to have a traffic light with a few black spots instead of an old burnt out traffic light bulb that is completely black.
 
Thanks all for the responses! I'll go ask also to the Electrical department for more clarification. The light housing i'm using is a reflective one, located at the front bumper, so i don't need wide angle of beam(it's like daytime running lights, mostly not to open the pop-up ones;p i have a 200sx(240sx in us)) I'll stick with the 3mm leds i have, add more, and see what happens;p Smd will be difficult now to refit..Also i think the driver module would still cause me problems, since i will run many led rows in parallel...so let's say i want 20mA per row, and i have 5 rows, i adjust the driver to give 100mA...though if resistance of one row is less than the others, then it would allow more current to pass from there and still cause me problems...and if it burns that raw then the 100mA will pass through the other 4 rows an burn me all the leds....i'm i correct in this or no?
 
My rule for how many LEDs in series is such that the current limiting resistor is 100 ohms or higher. I'll go as low as 47, but I Don't Like it. But NEVER under 47, just too little margin for voltage absorption to limit net current.

You'll use more resistors & take longer to build... but you'll do it Once. G.H. <<<)))
 
Yeah resistors are better... Can anyone explain if the driver module can still have the same problem?(see my previous post)
 
Hello again,


The resistor selection in theory is always:
R=(Vs-vLED)/I
where
R is the series resistance,
Vs is the source voltage,
vLED is the total voltage of all LEDs in the string,
I is the required forward current.

That's the theory, but because of the way many voltage sources work (they vary a little bit) the way we have to apply this is to follow up the calculation with a quick sensitivity analysis. The idea is to figure out how sensitive the current and the light intensity are to a change in source voltage.
The critical point for the current is when Vs goes to its highest level, and the critical intensity point is when Vs goes to its lowest level. If you can stand the decrease in light when Vs goes to its lowest level then you are ok with the intensity, and if your LEDs can handle the current when Vs goes to its highest level then your LEDs will be ok with the highest expected current level.
If you want to do a thermal sensitivity check too then all the better, but combining this with the voltage sensitivity works well. Figure the LEDs forward voltage each changing by -2mv/degreeC, so worst case is with Vs high and temperature T high also. If they survive that then they last for quite a while. Worst case light output is with temperature low and Vs low also...if you can stand the decrease in light output then this design is a go.

The problem with some LED products including traffic lights is the right information was not conveyed soon enough for some designs. The outcome of this was usually early LED failures. The manufacturers seemed to act like the LEDs could never fail and designers did not pick up on this until it was too late.
 
Last edited:
Hello again,


The resistor selection in theory is always:
R=(Vs-vLED)/I
where
R is the series resistance,
Vs is the source voltage,
vLED is the total voltage of all LEDs in the string,
I is the required forward current.

That's the theory, but because of the way many voltage sources work (they vary a little bit) the way we have to apply this is to follow up the calculation with a quick sensitivity analysis. The idea is to figure out how sensitive the current and the light intensity are to a change in source voltage.
The critical point for the current is when Vs goes to its highest level, and the critical intensity point is when Vs goes to its lowest level. If you can stand the decrease in light when Vs goes to its lowest level then you are ok with the intensity, and if your LEDs can handle the current when Vs goes to its highest level then your LEDs will be ok with the highest expected current level.
If you want to do a thermal sensitivity check too then all the better, but combining this with the voltage sensitivity works well. Figure the LEDs forward voltage each changing by -2mv/degreeC, so worst case is with Vs high and temperature T high also. If they survive that then they last for quite a while. Worst case light output is with temperature low and Vs low also...if you can stand the decrease in light output then this design is a go.

The problem with some LED products including traffic lights is the right information was not conveyed soon enough for some designs. The outcome of this was usually early LED failures. The manufacturers seemed to act like the LEDs could never fail and designers did not pick up on this until it was too late.

I can understand this. Your talking about voltage spikes since voltage of the car is not constant. If i use a led driver module, that would stabilize the voltage for me. Though what i'm saying is that if i try to run exactly 20mA for every row(leds in series) and have connected many in parallel, then difference in resistance between each row, will allow current more than 20mA to pass through the row with the lowest resistance, and that will result in a failure of the specific row. But then if lets say i have 3 rows running at 20mA each one, i would have the driver module giving output of 60mA. Then if one row gets burned, then the 60mA would go through the remaining 2 rows, so each of the remaining rows would run at 30mA which will result in instant failure of all the rows. How can i protect the leds from this scenario? A solution is to run the driver at a lowest output so i will be everytime under 20mA, but this will affect the total light output.
 
Status
Not open for further replies.

Latest threads

New Articles From Microcontroller Tips

Back
Top