Avago 1W Warm White LED

AntronX

http://mouser.com/Search/ProductDetail.aspx?qs=sGAEpiMZZMvbrKVFTEWTlqgHDE6JOMztGMzMxlfclmI%3d

Just received 10 of these, ended up costing me $3 each after shipping. Hooked up 8 of them in series to run directly from 24V battery bank. They are rated at 3.5V and 0.35 A continuous. At 25.2V each LED is seeing 3.15V x 0.22 A = 0.7W, so eight of them total 5.5W. I chose to run eight instead of seven because my system goes into absorption mode at 29V which is 3.625V per LED.

I placed them behind my computer monitor where before I had 9W CFL. Brightness is the same, or at least very close. I cannot tell the difference. Total load on the battery with 9W CFL and inverter was 14W. Now its 5.5W. This is quite an improvement in efficiency.

Mouser sells a reel of 1000 of them for $1500, $1.5 per LED. I would like to get 100 more for myself, anyone want's to join a group buy?

Pictures are on the way...

niel

Re: Avago 1W Warm White LED

i wouldn't mind a couple of them, but i can't go in for that many to warrant a large buy.
if you have a dmm i would like you to observe the current when the led is operating to see if it slowly rises. if it does it could reach critical after quite a few hours and burn out.

BB.

Re: Avago 1W Warm White LED

Be careful with your LEDs running on "24 volts" and no ballast resistor or circuit... At the very least, make sure you put a 1 amp fuse or something in series.

The issue is that LEDs typically have a drop in junction voltage as their temperature rises.

So, if you take the 3.625 volt junction at 25C on your "29 volt" battery bank... If they get hot (over 100C), your current will possibly increase to over 500 mAmps (which decreases the Vf even more and causes the LEDs to run hotter--see data sheet here PDF download).

Normally, even if it "wastes" a bit of power--you should probably replace at least one of the LEDs with a resistor:

R=V/I=3.6v/0.350a=10.3 Ohms
Wres=V^2/R=3.6v^2 / 10.3 Ohms = 1.26 watts nominal Resistor power dissipation...

So, you should have one ~10 Ohm 5 watt resistor in the string to at least provide a linear element to reduce the effects of thermal run-a-way in the string.

Power wise with LED lighting--typically you can barely see a 2:1 ratio difference in power output (1 watt vs 0.5 watts of "light" is very difficult to see the difference). So if you are trying to save power and keep things a bit cooler (less issues with heat sinks and longer life)--you can experiment with a larger resistor to pull the power back down a bit (try ~20-30 Ohms total resistance).

Of course, you can build/buy current regulators that would regulate the LED current (and remove the voltage dependency).

-Bill

AntronX

Re: Avago 1W Warm White LED

Thank you Bill. I did not know that about LEDs. I am letting the heatsink cool right now, to take more accurate current reading with another DMM. I will take a look at current regulator circuits.

niel

Re: Avago 1W Warm White LED

btw, your cfl was probably more like 20w if not more due to power factor.

AntronX

Re: Avago 1W Warm White LED

Today, my voltage went up to 25.4V in daytime. I hooked up LEDs, but current was 0.4A right away. They were also brighter, too bright for my application. I hooked up 25 ohm 10W resistor in series. Now its 0.12A at 25.4V. Light output has dropped noticeably, but so did the heat. After 30 minutes their heatsink is still cold. Are they getting higher Lumen/Watt ratio at lower power?

BB.

Re: Avago 1W Warm White LED

If you look at figure 11 on the PDF spec. download (earlier post), you can calculate the normalized lumen output to current ratio:

"1" @ 350 mAmps => 1/0.35 = 2.9:1 ratio
"0.18" @ 50 mAmps => 0.18/0.06 = 3.6:1 ratio

So, the LED's are somewhat more efficient at lower drive currents.

If you look at lumen output vs junction temperature (fig. 15), you will find that white LED's loose another 25% of their output at they approach 130C (max temperature). So, keeping them cool also helps efficiency.

You can even work a little harder to Find Vf at the two currents, use P=V*I, then figure out lumen:watt output too.

Roughly, a 2:1 change in current would result in a barely perceptible change in light output, a 10:1 current change would be a "night and day" change in output (10x would swamp 1x output). Our eyes look at things on a logarithmic scale. So, the ratios are what matter.

-Bill

AntronX

Re: Avago 1W Warm White LED

Ok, so lets figure out how much more efficient they are at lower power.

0.39 A * 25.26 V = 9.8514 W, 60 lm * 8 * 1.1 = 528 lm
0.12 A * 22.48 V = 2.6976 W, 60 lm * 8 * 0.39 = 187.2 lm

528 lm / 9.8514 W = 53.6 lm/W
187.2 lm / 2.6976 W = 69.4 lm/W

69.4 / 53.6 = 1.295 = 29.5% increase in efficiency.

Let's include lumen loss due to higher operating temperature at higher power:

9.8514 W at 50 C = 94%
2.6976 W at 35 C = 98%

9.8514 W = 53.6 lm/W * 0.94 = 50.384 lm/W
2.6976 W = 69.4 lm/W * 0.98 = 68.012 lm/W

68.012 / 50.384 = 1.34987 = 35% increase in efficiency

The above does not include efficiency loss due to series resistor:

25.26 V - 22.48 V = 2.78 V * 0.12 A = 0.3336 W

or:

2.6976 W + 0.3336 W = 3.0312 W
187.2 lm / 3.0312 W * 0.98 = 60.52 lm/W (as seen by my battery)

BB.

Re: Avago 1W Warm White LED

I got a little lost in the Higher Temperature equals Higher Efficiency--should be lower efficiency (lumen output drops for White LEDs as temperature goes up).

If you are taking into account that the LED Vf drops as they get hot--that will increase their electrical efficiency (at the cost of phosphor efficiency--perhaps a wash?).

-Bill

AntronX

Re: Avago 1W Warm White LED

I meant it more like a derate factor. But the calculation shows that lumen output does decrease the hotter LEDs get. At 9W temperature is 50 degrees C and the graph in data sheet shows derate of 0.94 compared to 25 C. At 3W temperature is 35C, and derate is 0.98. That's why when temperatute variation is included efficiency difference jumps to 35%, up from 29.5% without temperature factored in.

mike95490

Re: Avago 1W Warm White LED

Heat & LED's
White LED's are usually Blue or UV leds with a dab of white color mix phosphor, like lines a fluorescent bulb interior. This phosphor degrades with heat, and the LED diode junction also changes characteristics with heat (in a bad way) too. So usually, diode junction is spec to 150F max. Junction is usually quite a bit ( 20F) hotter than the heat sink.