EcoSmart 5csbr650stq1d03 LED light

[RukyCon]

Originally Posted by stj

we all need a vice-of-knowlege!
specially with crap small psu's

I have a cheap power supply powering some LED strips, three days later and the PCB is turning brown.

[jazzie366]

Originally Posted by stj

we all need a vice-of-knowlege!
specially with crap small psu's

Yes, we most certainly do

[momaka]

LEDs will always have shitty life if they are overdriven with current or run too hot. Typically the former causes the latter, but now always! Sometimes an LED can fail simply from being over-driven with current.

I myself was disappointed with LED life way back a few years ago when a friend and I set up lighting for his workshop with a bunch of those flexible 12V white LED strips (the cheap rolls you can find everywhere on eBay and elsewhere online). Driven by a spare PS3 power supply, the strips were getting nice, regulated 12V power (about 12.3V IIRC when I measured it a few times) and so they ran absolutely cool to the touch - even the individual LEDs themselves. Nonetheless, most of the LED strips lost a substantial part of their brightness after just three months of being powered 8-12 hours per day. A lot of them even managed to blow a hole through a few of their LEDs.

So then I took one brand new LED strip home and one of the "bad/weak" ones to compare them. The weak strip was pulling almost 2x the amount of current compared to the new strip. Thus, it looks like the LEDs in the bad strip were starting to get leaky or short out. But why? The new strip was drawing only 10-15 mA of current per LED at 12V. Didn't seem like much.

As I couldn't find the maker of the LEDs, and thus not able to find a datasheet, I just took the new strip and tried running it on 9V. The strip seemed only a little less bright than at 12V, but it was drawing substantially less current - about 1/5 less.

So I think the LEDs in a lot of these strips and light bulbs are just over-driven so that the manufacturer can get away with using less LEDs. Combine that with a hot-running bulb, and it's no surprise that LED bulbs don't last.

So that's why your LED bulbs are bound to get poor life when left to run 24/7. The really expensive bulbs may have better, more efficient LEDs (Cree, for example) and with better heatsinks. But the cheap LED bulbs - forget it! I wouldn't trust the life rating on their box for squat. 10 years? Sure, if you never use the thing. Then again, you can claim that a Pinto is the safest car in your lot - after all if you never drive it, you'll never get in a collision and the statement would be true indeed.

As for what I use (in terms of lighting)?
- I have a decent stash of incandescent and halogen-incandescent bulbs. Should easily last me the next 10 years, if not more. I don't use CFLs at all, as I find the UV-rich light too harsh on my eyes. The only CFLs we have in the house is one set with 2 T8 bulbs in the garage and a desk lamp (with a specialty CLF and linear ballast) that I use for my work area (which I almost never turn ON, as I work during the day most of the time, and I have a big window with plenty of natural light).

[jazzie366]

Originally Posted by momaka

LEDs will always have shitty life if they are overdriven with current or run too hot. Typically the former causes the latter, but now always! Sometimes an LED can fail simply from being over-driven with current.

I myself was disappointed with LED life way back a few years ago when a friend and I set up lighting for his workshop with a bunch of those flexible 12V white LED strips (the cheap rolls you can find everywhere on eBay and elsewhere online). Driven by a spare PS3 power supply, the strips were getting nice, regulated 12V power (about 12.3V IIRC when I measured it a few times) and so they ran absolutely cool to the touch - even the individual LEDs themselves. Nonetheless, most of the LED strips lost a substantial part of their brightness after just three months of being powered 8-12 hours per day. A lot of them even managed to blow a hole through a few of their LEDs.

So then I took one brand new LED strip home and one of the "bad/weak" ones to compare them. The weak strip was pulling almost 2x the amount of current compared to the new strip. Thus, it looks like the LEDs in the bad strip were starting to get leaky or short out. But why? The new strip was drawing only 10-15 mA of current per LED at 12V. Didn't seem like much.

As I couldn't find the maker of the LEDs, and thus not able to find a datasheet, I just took the new strip and tried running it on 9V. The strip seemed only a little less bright than at 12V, but it was drawing substantially less current - about 1/5 less.

So I think the LEDs in a lot of these strips and light bulbs are just over-driven so that the manufacturer can get away with using less LEDs. Combine that with a hot-running bulb, and it's no surprise that LED bulbs don't last.

So that's why your LED bulbs are bound to get poor life when left to run 24/7. The really expensive bulbs may have better, more efficient LEDs (Cree, for example) and with better heatsinks. But the cheap LED bulbs - forget it! I wouldn't trust the life rating on their box for squat. 10 years? Sure, if you never use the thing. Then again, you can claim that a Pinto is the safest car in your lot - after all if you never drive it, you'll never get in a collision and the statement would be true indeed.

As for what I use (in terms of lighting)?
- I have a decent stash of incandescent and halogen-incandescent bulbs. Should easily last me the next 10 years, if not more. I don't use CFLs at all, as I find the UV-rich light too harsh on my eyes. The only CFLs we have in the house is one set with 2 T8 bulbs in the garage and a desk lamp (with a specialty CLF and linear ballast) that I use for my work area (which I almost never turn ON, as I work during the day most of the time, and I have a big window with plenty of natural light).

They claim almost 20 years now on some of these ones

Also, those cheap LED strips are 100% the reason you had problems. I've sourced cheapo led bulbs that I could sell for 50 cents each and make a profit, and they lasted 6 months at best? They all ran not too hot, they just died easily thought because of heat and overcurrent. However, this circuit was running under 50v likely. There were 6 LEDs iirc. I'm guessing no more than 30v, so 6V per chip. Assuming these are multi-chip LEDs (I will microscope them to confirm this) They can be dissipating 3v or less per chip. I've yet to see non multi chip leds that can take over 3V.

If these are 3 chips or more, the dissipation will be very low. That also helps with color temperature, which on these was very good.

However, if that's true then they are going to run only slightly warm each, however combined with the shitty heatsink these had, it'll probably run a little hot. However, they're not being driven hard if that's the case and therefore should have a decent life.

Also keep in mind, it wasn't the LEDs fault for the failure, it was a small SMD resistor that should've been bigger.

[RukyCon]

Most cheapo LED strips have 3 single chip LEDs in series, if we do the math 3.2V x 3 = 9.6V! And they say these are made for 12v.

[stj]

they do have a resistor with those leds though.

[RukyCon]

The resistors on my LED strip are 39Ω each and there's only one resistor per set of three LEDs.

[kc8adu]

or if a bad tube is rectifying.
the ballast WILL overheat!

Originally Posted by stj

the only way a ballast can overheat is if a starter welds closed.
rare - but i'v seen it.

[momaka]

Originally Posted by jazzie366

They claim almost 20 years now on some of these ones

But 20 years -WHERE is the main question here. 20 years in the landfill? - Ya, I can see that!

Originally Posted by jazzie366

Also keep in mind, it wasn't the LEDs fault for the failure, it was a small SMD resistor that should've been bigger.

Well, it could also be that the SMD resistor burned out because the LED started to draw more current as it approached its (short) end of life.

Like I said, a lot of my cheap 12V strips nearly doubled their current use after a few months of 8-12 hours per day operation. I wouldn't be surprised if an LED or two in those chip LEDs shorted out and drew a massive current through the resistor. Though the only way to know for sure is to buy one of these bulbs and measure the current going through the resistor. With:
W = R * I^2, you can easily find out how much power the resistor is dissipating and whether its power rating was chosen properly or not.

Originally Posted by ruky con

Most cheapo LED strips have 3 single chip LEDs in series, if we do the math 3.2V x 3 = 9.6V! And they say these are made for 12v.

Actually, most white and blue LEDs drop a little less than 3V at low power. As you start pushing more current through the LED, then you might see the voltage drop raise to 3.2-3.4V.

That's why I am able to run mine on a 9V power supply (SMPS, regulated - not some linear transformer outputting 16V unloaded ). And at 9V, the LEDs seem to be running quite happily! (Though I will admit there are a few LEDs in the strip that don't quite fully light up until I give the strip a little over 10V - but that's a manufacturing variation to be expected with these cheap LEDs. You sure as heck won't see that from high quality LEDs.)

Originally Posted by ruky con

The resistors on my LED strip are 39Ω each and there's only one resistor per set of three LEDs.

Gee, and I thought the 150 Ohms resistors in my strips were bad enough.

39 Ohms is insane, actually. At 12V operation, there will be approximately 3V for that resistor to drop. So at 39 Ohms, you are looking at 1/13 of an Amp, or 77 mA passing through each LED! That's 230 mW of power dissipation per LED. I doubt those Chinese LEDs can handle that. Heck, mine burned out with the 150 Ohm resistor (that's a little less than 60 mW per LED at 12V operation).

[RukyCon]

Originally Posted by momaka

But 20 years -WHERE is the main question here. 20 years in the landfill? - Ya, I can see that!

The 20 year life is really the shelf life.

Originally Posted by momaka

Actually, most white and blue LEDs drop a little less than 3V at low power. As you start pushing more current through the LED, then you might see the voltage drop raise to 3.2-3.4V.

The reason i used 3.2V when multiplying when figuring out the maximum voltage of the LED strip is because 3.2v IIRC is the max voltage that you can put into an LED without damaging it, but i think you can go as high as 3.4v but that will most likely damage the LED if used on that voltage for too long. You can also go down to 2.8v per LED and still get some light output.

Originally Posted by momaka

Gee, and I thought the 150 Ohms resistors in my strips were bad enough.

Now we need to look for one that does not have a resistor at all.

Originally Posted by momaka

39 Ohms is insane, actually. At 12V operation, there will be approximately 3V for that resistor to drop. So at 39 Ohms, you are looking at 1/13 of an Amp, or 77 mA passing through each LED! That's 230 mW of power dissipation per LED. I doubt those Chinese LEDs can handle that. Heck, mine burned out with the 150 Ohm resistor (that's a little less than 60 mW per LED at 12V operation).

That's a lot of heat those LEDs will be putting off and depending on where they're mounted or what they're mounted to, the LEDs will likely overheat.

[momaka]

Originally Posted by ruky con

The 20 year life is really the shelf life.

Hmmm... I don't know about that. Even Japanese caps will have a hard time "waking up" after 20 years sitting unused.

Originally Posted by ruky con

The reason i used 3.2V when multiplying when figuring out the maximum voltage of the LED strip is because 3.2v IIRC is the max voltage that you can put into an LED without damaging it, but i think you can go as high as 3.4v but that will most likely damage the LED if used on that voltage for too long. You can also go down to 2.8v per LED and still get some light output.

LEDs are not "driven" by voltage. They are driven by current. The current passing through an LED pretty much determines how bright the LED will be.

That said, the more current you try to pass through an LED, the higher its voltage drop will become. It's a non-linear relationship, though, so doubling the current won't really double the voltage drop.

Originally Posted by ruky con

Now we need to look for one that does not have a resistor at all.

Yeah, I've seen LEDs driven by special LED driver ICs. If the LEDs are crappy, though, that still wouldn't change much of anything. A shorted LED will just make the LED driver shut down instead of the resistor burning out. But then who cares? The end result will be the same - no light from the bulb.

Originally Posted by ruky con

That's a lot of heat those LEDs will be putting off and depending on where they're mounted or what they're mounted to, the LEDs will likely overheat.

It's not that much heat, actually. But if the LED is not made to handle it, it will have a rather short life. So with most LEDs, you pretty much have to choose between high brightness with short life or low brightness with longer life. If you want both, then you just have to use more LEDs over a bigger surface area to keep things cool. That's where the third factor - cost - comes in. And then it becomes a triangle trying to choose between any of the three.

Regarding those cheap SMD LED strips: I also read a technical article somewhere sometime ago that these LEDs use have a soft lens/case to cover the die. Because of that, cheap LEDs tend to have poor seals and it said this was one of the main reasons why they failed so quickly. On the other hand, the article said that cheap standard through-hole 3 mm, 5 mm, and 10 mm LEDs with clear epoxy lenses/cases didn't have that problem, as they were generally sealed better. And personally, I think there is some truth to that, because I haven't seen those fail as often as cheap SMD LEDs.

So if buying cheap LED bulbs, you might be better off with ones that have those through hole LEDs. Of course, they won't last long either if their LEDs are over-driven with current and they run hot. But if not, and if they are powered by a good PSU inside, then they should last a more reasonable amount of time.

[jazzie366]

Originally Posted by momaka

Hmmm... I don't know about that. Even Japanese caps will have a hard time "waking up" after 20 years sitting unused.

Haha, I thought the same when I used some 23 year old Nichicon PL's I had. They seemed to work fine though for the month or two that I used them for :P


Anyways, I tested the LEDs on the board of the lamp, they're 6V LEDs! So two chips per I assume? I put 7v across one and it went to full brightness immediately (by this, I mean the chip was absolutely blindingly bright.)

So x6 chips equals about 32V that this would've been running at. None of the LEDs drew excessive current going through my meter, so I'm going to rule that out.

Also, as you said before, the LEDs could've been shorting causing the resistor to go bad, the only problem I have here is the resistor is the size of a quarter watt SMD resistor, and it's on the + side of the rectifier, going straight to the capacitor, and then to the output. I think that's probably wayyyy too small for a job like that. I keep forgetting to upload pics, therefor right after I post this I am going to upload some.

Edit: added pics

[stj]

R1 is burned, interesting the board can take a parallel pair.
i wonder how many franctions of a cent they saved by only fitting one!!!

[jazzie366]

Originally Posted by stj

R1 is burned, interesting the board can take a parallel pair.
i wonder how many franctions of a cent they saved by only fitting one!!!

R1 is what failed. If there were two, they probably wouldn't have had this problem.

[momaka]

Originally Posted by jazzie366

Anyways, I tested the LEDs on the board of the lamp, they're 6V LEDs! So two chips per I assume? I put 7v across one and it went to full brightness immediately (by this, I mean the chip was absolutely blindingly bright.)

I know I've done that mistake too when I first got into electronics, so I'll try not to be too harsh here but........ seriously, you don't connect LEDs to a voltage source with unlimited current capability. Period.

If you don't have a variable current source, use a resistor with a voltage source. It takes all of 20 seconds to run through a simple Ohm's law calculation to determine what resistor you need. Otherwise, you risk burning your LEDs or severely shortening their life. Depending on the current capability of your source, sometimes even a split second is enough to nuke the anode's wire bond on the die.

Originally Posted by jazzie366

Also, as you said before, the LEDs could've been shorting causing the resistor to go bad, the only problem I have here is the resistor is the size of a quarter watt SMD resistor, and it's on the + side of the rectifier, going straight to the capacitor, and then to the output. I think that's probably wayyyy too small for a job like that.

Interesting design. That resistor is NOT what is providing power to the LEDs. Inductor L2 does, which is connected in parallel with R1. Most likely they placed R1 there to limit inductive kickback spikes from L2 as the PSU switches. But that's a very poor design. Should have used a diode or a snubber network for that.

[jazzie366]

Originally Posted by momaka

I know I've done that mistake too when I first got into electronics, so I'll try not to be too harsh here but........ seriously, you don't connect LEDs to a voltage source with unlimited current capability. Period.

If you don't have a variable current source, use a resistor with a voltage source. It takes all of 20 seconds to run through a simple Ohm's law calculation to determine what resistor you need. Otherwise, you risk burning your LEDs or severely shortening their life. Depending on the current capability of your source, sometimes even a split second is enough to nuke the anode's wire bond on the die.

Oh bah, I didn't give a shit about these LEDs. The lamp was dead and I'm never going to use them again, I smashed the lamps case to get it open. I'm going to try some of those Sylvania bulbs as they are made a bit better.

If I was working on sensitive electronics I would've used a proper PSU with current limiting.

[momaka]

Originally Posted by jazzie366

Oh bah, I didn't give a shit about these LEDs. The lamp was dead and I'm never going to use them again, I smashed the lamps case to get it open.

I know what you mean.
But even when I don't care just like you, I still take a random spare resistor from my scrap box and use it. Typically 1 to 10 KOhm will do the trick fine just to test the LED. Thus I always keep my resistor box handy.

Besides, I sometimes find the need to discharge large caps for safety reasons.

[jazzie366]

Originally Posted by momaka

I know what you mean.
But even when I don't care just like you, I still take a random spare resistor from my scrap box and use it. Typically 1 to 10 KOhm will do the trick fine just to test the LED. Thus I always keep my resistor box handy.

Besides, I sometimes find the need to discharge large caps for safety reasons.

Oh yeah, I have two 560Ohm resistors in series for discharging. Now before you say it, yes, it makes large sparks but I don't have the patience to wait 10 seconds for a cap to decharge (either that or I really like the sparks. Sparks are fun. Sparks.)

[stj]

lol
big clive checks for a charge with his fingers!

[momaka]

Originally Posted by jazzie366

Oh yeah, I have two 560Ohm resistors in series for discharging. Now before you say it, yes, it makes large sparks but I don't have the patience to wait 10 seconds for a cap to decharge (either that or I really like the sparks. Sparks are fun. Sparks.)

1 KOhm doesn't make that big of a spark. Try 24-Ohm heating element.
Though I usually use a 4.7 KOhm, 5W resistor. Takes 2-3 seconds tops, usually.

Originally Posted by stj

big clive checks for a charge with his fingers!

I used to do that too. The problem is, I'm not worried so much about my fingers as I am about my multimeters and other tools.