Yes they are that bad

[JohnR66]

I tested "super bright" white LEDs from various ebay sellers. LOV and JELEDHK's LEDs were the 55,000mcd rated versions and the CECE ones were rated 20,000 mcd. Given the throw (intensity) is about the same, I conclude that the rating of the 55cd ones are way overstated :laughing:

LEDs were tested at only 12.5 ma for 11 days. After 150 hours, the downward trend started. The good news is that after 270 hours they are still well above 70% original light output. I estimate the light output to fall below %70 after 500 hours and perhaps as soon as 400 hours if the curve steepens. I expect the latter to be true.

So there you have it. After less than 3 weeks the LEDs would have faded significantly at only 12.5ma :shakehead

Notes: Y axis is relative light output (not lumens)
Inconsistent test times were due to my availability to test them.
LEDs are known to increase in brightness slightly after a few hours of operation which might explain the slight bump I see in my tests.

 

[Christexan]

Re: es they are that bad

I've used a lot of the JELEDHK ones in bikelight usage that I built a couple of years ago, running around 20mA each, and they've not received extensive usage (maybe 20-30 hours total) and appear to be quite a bit dimmer than they were on day 1.

That's only a "mental" appearance, however I remember how I was pleased with them out of the box (cheap as dirt, so wasn't expecting much), and now I turn them on and am more like "wow, are these dim or are my eyes going?). Apparently you've verified it's the LEDs, not my eyes!
I've used the 5mm and 10mm variety (both in mixed patterns actually), and it seems across the board, they don't last well at all even at rated specs.

 

[JohnR66]

Re: es they are that bad

I've used a lot of the JELEDHK ones in bikelight usage that I built a couple of years ago, running around 20mA each, and they've not received extensive usage (maybe 20-30 hours total) and appear to be quite a bit dimmer than they were on day 1.

That's only a "mental" appearance, however I remember how I was pleased with them out of the box (cheap as dirt, so wasn't expecting much), and now I turn them on and am more like "wow, are these dim or are my eyes going?). Apparently you've verified it's the LEDs, not my eyes!
I've used the 5mm and 10mm variety (both in mixed patterns actually), and it seems across the board, they don't last well at all even at rated specs.

You might be imagining it. At 20ma I doubt they would have changed enough to be noticeably dimmer at anything under 100 hours. These may work out okay in battery driven products where it can take lots of time to accumulate hours.

I call these cheapo LEDs out because they end up in a lot of fixed lighting products that see several hours per day usage and end up fading in a short time. This gives LEDs a bad name to the general public and will slow their acceptance down. Are you listening Lights of America?:poke:

 

[Christexan]

Maybe, but they are also over 3 years old, and who knows the quality of the phosphors which may just naturally fade even without power applied on these cheapos (100 for $12, I think, or something like that when I bought them). I'll have to "A/B" them with some that have never been used (I have plenty left, have only used maybe 20-30), and see if I can tell a difference.

However the assertion that running them around 20mA (really closer to 22mA) won't make much difference, I'm not sure... you ran them at roughly half the max, and lost roughly 80% output at 200 hours. Doubling the power to them at least doubles, and probably triples the degradation rate, so that 80% at 200 hours loss becomes 80% at 50-100 hours loss... so given that, 20-30 hours in, it seems it'd be easy to notice a decrease, probably 20-30% drop by that point already (the graph you provided just plummeted).

But without an objective comparison, no way to tell for sure.

 

[Phaserburn]

Yuck. Anyone ever ID a source to buy Nichia leds?

 

[JohnR66]

Final update to the test. At 520 hours two of the LEDs have dropped below 70% and the LOV LED is at 77%. The fading is more or less on a linear downward slide now. I estimate the LOV LED to drop under 70% at around 600 hours.

Test conclusion: These LEDs couldn't even remain above 70% of their initial light output at only 12.5ma which is well within their rated maximum current. A month is around 700 hours, so they didn't even last a month at low current.

In the other test at 30ma, I only tested the JELED, and it fell to 64% after only 4 days. a Pathetic.:shakehead

 

[jtr1962]

These results are basically telling me that eBay LEDs aren't even worth my time to lumens test them any more. I wonder how the LOV LED might do at 5 mA? I've been saying for a long time that eBay LEDs are fine if you think of them as 5 mA LEDs, not 20 mA. It seems the degradation rate isn't directly proportional to the inverse of the current. The Jeled took 4 days to fall to 64% at 30 mA. If the relationship with current were inversely linear, then at 12.5 mA it would take about 10 days to fade the same amount. However, it took more than twice as long. I might estimate 2500 hours or even longer at 5 mA to reach 70% based on this data.

Another thing which occurred to me is perhaps these LEDs have a critical current density above which they degrade very fast, and below which they last a fairly long time. In any case, these results are very discouraging. It seems eBay LEDs are mostly suited for toys or other uses where they'll rack up at most a few dozen hours.

 

[JohnR66]

These results are basically telling me that eBay LEDs aren't even worth my time to lumens test them any more. I wonder how the LOV LED might do at 5 mA? I've been saying for a long time that eBay LEDs are fine if you think of them as 5 mA LEDs, not 20 mA. It seems the degradation rate isn't directly proportional to the inverse of the current. The Jeled took 4 days to fall to 64% at 30 mA. If the relationship with current were inversely linear, then at 12.5 mA it would take about 10 days to fade the same amount. However, it took more than twice as long. I might estimate 2500 hours or even longer at 5 mA to reach 70% based on this data.

Another thing which occurred to me is perhaps these LEDs have a critical current density above which they degrade very fast, and below which they last a fairly long time. In any case, these results are very discouraging. It seems eBay LEDs are mostly suited for toys or other uses where they'll rack up at most a few dozen hours.

At 30ma, it is over it's rated current, so it might have degraded faster, although what you said is probably correct. I was hoping these would be useful around 12-15ma as light output is usable. At 5ma it seems I have to tie my hands to make a crap product last and light output is weak. I agree that these LEDs are suited for toys and such (battery use).

OTOH, these ebay LEDs are completely pointless now since one can get the white Crees with better tint, brighter and long life for $.45 each shipped. The so called 55,000 mcd ebay LEDs cost in the $.26 to $.30 range, so it make sense to spend a little more. ($ based on 100 qty packs).

JTR, I appreciate your lumen testing on the LEDs. I don't know if you tinker with 5mm LEDs anymore, but if you want a Cree C503 series white, red, amber, blue and green to test, I can send you some. Just PM me.

 

[jtr1962]

JTR, I appreciate your lumen testing on the LEDs. I don't know if you tinker with 5mm LEDs anymore, but if you want a Cree C503 series white, red, amber, blue and green to test, I can send you some. Just PM me.

I already did a test on the Cree C503 white in January. The results averaged 79.3 lm/W, a bit less than the Nichia GS, which was 96.7 lm/W. As for the color LEDs, I'm not set up to test colored LEDs. My light meter response is way off for non-white light, and I have no way of determining the corrections. Any colored LED test results would therefore be meaningless. If you want to send me a sample or two of the colored Crees just for me to play around with, however, that's fine. At least I can subjectively compare them to the other colored LEDs I have.

And although it's not official, I'm probably more or less done with testing 5mm indicator type LEDs. I'm not seeing the point any more given the recent price drops in power LEDs, as well as the general stagnation in efficiency ( maximum efficiency for most 5mm LEDs has remained at ~80 lm/W for the last few years while power LEDs have continued to improve ).

 

[JohnR66]

Hmmm, you have piqued my curiosity. I went back and read your review of the C503B LED and see the full part number is the same that DigiKey has but is not the ones I bought. If you drive the LED at low current and observe the die through a maginfier, you may be able to see the bluish outline through the phosphor. Using a blue filter will help. In mine it is a long rectangle shape. Is it the same with yours? My Part number is LC503FWH1-15P-A0. The mcd is 24,000, the 3C503B-WAN-CCACB231 is listed as 18,000 mcd which seems to be in error in the catalog.

Now here's the wierd thing. Back in March Sgt LED sent me a couple of the LC503FWH1-15P-A0 LEDs which is the same part I just got. It is not as bright and has a narrower beam (the one that just dropped out of my endurance test). See this thread: http://www.candlepowerforums.com/vb/showthread.php?t=215553 Sgt. LED was debating on which of these two LEDs to get and it appears he got the ones I did.

I'll send you the colors and the white one to compare with what you have. PM me and I'll send them off when once my order arrives.

 

[kmossman]

Re: Yes they are that bad >>>>NOT TRUE!!!!!<<<<<

A rose is still a rose... A LED is not a LED

Buying what is really no-name LEDs which are sourced from China is a gamble. THE BUYERS demand/select crappy LEDS to maximize their profit.

The BUYERS (the folks you buy from) are to blame. NOT the factories. The usual factory supply at the target price they are given. If the buyer demands a low price, then they get low quality. It is a simple equation.

THERE ARE high quality LED factories in China. The price is much better than the brand names, though the lumens are not as high. Other quality aspects are as good.

We have run high power Chinese LEDs in harsh conditions for over 10,000 hours with little loss in output.

BE FAIR people!