LED T8 Tube light Many failures, WHAT????

SLK001

Joined Nov 29, 2011
1,549
All of the computer monitors that I have had fail to light had CCFL backlights. Probably TVs are a different quality.
Ditto with me. On a laptop, a CCFL is shoe-horned into a tight space, making them susceptible to breaking. TV's have the space a CCFL needs.
 

Thread Starter

MisterBill2

Joined Jan 23, 2018
18,176
So finally I totally disassembled two lights and spliced the working sections to make one good light, so I thought. So then I installed it and a few hours later it is dead. So much for that method.

But when I slid the LED strip out I find writing on it, as follows:
DE RA 2835 96LED 24C4B 1165=10
(the = sign is actually a black blob that may be a different character) And now I see that actually they are 4 strings of 24 LEDs and at 2.85 volts or so they are bright at 20 Ma, and at 3.0 volts they are very bright at 30Ma. This shows how very non-linear LEDs are, since with about 0.15 volt increase the current jumps up by 50%. But we all know that diodes are not linear devices. That is rather intrinsic.
If I had a source of that LED adhesive backed strip I could build my own 12 volt lights using these tubes, reflectors, and ends. But producing 50 of the tubes would be a big task, and then I would need to come up with a large constant duty UL approved power supply. I could do that but it would not be a simple task.
AND I would still be stuck with a bunch of the new bulbs already bought and installed, and so far those are working OK. So far.

But how do you control the current in each of 3 series strings of 24 each, all 3 in parallel?
 

Thread Starter

MisterBill2

Joined Jan 23, 2018
18,176
Apparently, the manufacturer did not know the answer. Perhaps this is why they are failing.

The answer is: with 3 controllers.

Bob
I have considered the possibility of adding three series resistors, but that would be some quite delicate surgery indeed, because of the way the things are assembled.
 

Thread Starter

MisterBill2

Joined Jan 23, 2018
18,176
I tried re-soldering a section that was only just barely glowing. I discovered that these surface mount LEDs are evidently glued onto the board before being soldered. Now that section does not light at all. So much for that direction. I will try to unsolder and remove some of them and see what I can learn. I am not optomistic. I did learn that they seem to have a small bit above 3 volts at the terminals. No wonder they are so bright. My hope to be able to replace failed LEDs is diminishing rapidly.
 

One1

Joined May 2, 2019
10
A thread on multiple led failure and not one person said the words duty cycle. Resistors cannot help you, voltage cannot help you, current cannot help you, when the issue is time. These poorly designed control boxes that power these are not ramping up and down the frequency of on time for these leds or they wouldn’t be blowing. Control the frequency and you’ll quit blowing leds.
 

Tonyr1084

Joined Sep 24, 2015
7,853
A thread on multiple led failure and not one person said the words duty cycle. Resistors cannot help you, voltage cannot help you, current cannot help you, when the issue is time. These poorly designed control boxes that power these are not ramping up and down the frequency of on time for these leds or they wouldn’t be blowing. Control the frequency and you’ll quit blowing leds.
Early on @MisterBill2 stated that the failures are occurring on the outdoor lighting but not on the same type of lights used indoors.

To be honest, I'm not the expert on LED's. I know how to use them and to light them properly but I am not sure I understand, or agree with where you're going with your comment about voltage, current and resistors. To me an LED needs a set current to control brightness. That current is derived by subtracting the forward voltage of the LED from the power source, then divided by the desired current to determine the proper resistance to achieve that current at that voltage.

If you're talking about PWM - that doesn't vary the current it varies the average voltage. That along with the fixed resistance will vary the amount of current going through the LED. Unless I have more to learn, I'm not sure I follow or agree with your comment. Nevertheless, I'm certainly not the author of all knowledge. Educate me (us) if you can.
 

Sensacell

Joined Jun 19, 2012
3,432
A thread on multiple led failure and not one person said the words duty cycle. Resistors cannot help you, voltage cannot help you, current cannot help you, when the issue is time. These poorly designed control boxes that power these are not ramping up and down the frequency of on time for these leds or they wouldn’t be blowing. Control the frequency and you’ll quit blowing leds.
Nope. Nothing to do with frequency.

It's poor quality LEDs.
 

Thread Starter

MisterBill2

Joined Jan 23, 2018
18,176
Nope. Nothing to do with frequency.

It's poor quality LEDs.
I am now suspecting that indeed it is poor quality of both the LEDs and the process of attaching them to the PCB, and that unfortunately the solution is going to be abandoning this particular style of tube and using some other kind. The problem there is that these look so very good with the individual spots of light so very visible, while the other versions that I have seen are all trying to copy the very even light from a florescent tube. reworking the existing tubes to use the 12 volt adhesive backed LED tapes could be done, and then rewiring the fixtures to supply 12 volts DC could be done, but neither would be easy or cheap. And unfortunately the system we have is already paid for.
 

Sensacell

Joined Jun 19, 2012
3,432
If the LEDs are crappy or damaged, its a waste of time trying to fix it.

As soon as it's working, the next crappy LED will fail and you are back to zero.
 

shortbus

Joined Sep 30, 2009
10,045
These have an aluminum extrusion and a clear colorless plastic cover. The LEDs are on a narrow strip of PC board that slides into the heat sink, the power supply is in the hollow back section of the heat sink
How is that heat sink anything more than a "artistic" design element if it is hollow and not in contact with the PCB?
 

One1

Joined May 2, 2019
10
Nope. Nothing to do with frequency.

It's poor quality LEDs.

Shakes head..... lol. There’s no need to further that, moving on......

Until the leds are powered correctly there will be no solution. I remember back in school 20 years ago one kid kept melting leds off his board and turned the voltage down so low his lumens were barely visible. He managed to run the board for a whole day before the led died. We fired ours correctly like an led is supposed to be ran with a steep ramp to max voltage and a 40% on time then a steep ramp to 60% off time and it ran the entire year on the test bench brighter than advertised. The value of an education.

These cheap chinese led power supplies have no design whatsoever and simply pulse 100% on time and 100% off time.
 

One1

Joined May 2, 2019
10
As I don’t have the power supply obviously I cannot say what they are using (digital or analog) to control it however i can go to Vegas and bet with 100% odds they are using 100% on time in their pulses if digital. That’s because the cheap chinese companies don’t care if it blows up in ten mins. Nobody is shipping a $15 led strip back to china and they know it. Why engineer it?

LEDs can be ran analog, but no led.....NO LED.... is designed to be ran at 100% on time whether it can be or not. That’s not how it works. People that need to brush up on led power requirements can reference this quick link..... first one on google.

https://www.eetimes.com/document.asp?doc_id=1281013

Now, let’s say they use digital pulse to control it, they still aren’t using ramping and splitting the on time.....because china. At the very best, it’s a 100% pulse on and no ramping. No time control. The wave is a simple square. It should look more like this if they want to control the ramp.
 

Attachments

Sensacell

Joined Jun 19, 2012
3,432
Shakes head..... lol. There’s no need to further that, moving on......

Until the leds are powered correctly there will be no solution. I remember back in school 20 years ago one kid kept melting leds off his board and turned the voltage down so low his lumens were barely visible. He managed to run the board for a whole day before the led died. We fired ours correctly like an led is supposed to be ran with a steep ramp to max voltage and a 40% on time then a steep ramp to 60% off time and it ran the entire year on the test bench brighter than advertised. The value of an education.

These cheap chinese led power supplies have no design whatsoever and simply pulse 100% on time and 100% off time.
Sorry this makes no sense.

LEDs are current driven devices- as long as the current controlled within specs, and the device is mounted on an adequate heat sink, all is good.

PWM refers to a method of reducing the apparent brightness in an efficient manner, nothing to do with longevity other than reducing it's total ON time.
 

Thread Starter

MisterBill2

Joined Jan 23, 2018
18,176
Heat is definitely an enemy of LEDs, the LED makers will tell you that, and testing proves it every time. So pushing them too hard will cause problems for sure.
I just got a brilliant suggestion yesterday evening, from a brilliant rabbi, a radio amateur who I occasionally contact, and who has a great deal of insight into things. Now I need to try and see how his idea works. The idea is that for the failed sections, jumper the failed LED with a resistor. CHEAP AND EASY! I do need to determine the best current to run them at, which will be less than they presently run, and determine the voltage to drop. The best part is that it will be simple to solder a leaded resistor across the glued-in-place LED. And once the current is limited to a lower level they should last a lot longer.
 

shortbus

Joined Sep 30, 2009
10,045
and the device is mounted on an adequate heat sink, all is good.
Which is what I was saying in post #33. The heat sink is hollow and where the power supply is, and the led boards are in front of the heat sink. So the heat sink isn't doing anything for the leds at all. Sounds like the heat sink is shaped like the letter 'C' with fins on the outside and the leds across the open side..
 

Thread Starter

MisterBill2

Joined Jan 23, 2018
18,176
Which is what I was saying in post #33. The heat sink is hollow and where the power supply is, and the led boards are in front of the heat sink. So the heat sink isn't doing anything for the leds at all. Sounds like the heat sink is shaped like the letter 'C' with fins on the outside and the leds across the open side..
actually there is heat sink material immediately below the PCB with the surface mount LEDs ob=n it, and the edges of the circuit board are wrapped around by heat sink material. So heat may be conducted off the edges of he circuit board, but how much is not obvious.
 
Top