cree vs luxeon

[Axkiker]

So just today i recieved my first set of blue cree emitters.... I immediatly opened the package and pulled one of these little suckers out and hooked it up to my powersupply

My first reaction was "this sucks". I really dont see them being any brighter than the luxeon emitters I was buying. These crees are supposed to put out at least 30lm at 350 ma. The luxeon were supposed to put out 30 lm at 700ma. So I naturally thought at 700ma the cree would blow the luxeon away. not so much. I havent had a chance to set them up side by side. I plan on doing that tonight. I dont expect much of a difference though.

My next thoughts are a little better. The cree didnt seem to get nearly as hot as the luxeon. This is a great advantage to me.

have you all had the same experience with the crees not getting as hot. If for no other reason the heat issue may be a reason to keep using the crees. However if not then back to the luxeon I go. The luxeon are cheaper and easier to get.

any thoughts.

 

[Gryloc]

Well, first of all, the Cree is operating at a lower power (more than half as much), so that is why they are running cooler. Secondly, are you attaching the emitter or star to the proper heatsink? What are you using to power up the emitters? How can you be sure they are getting the exact current? If you have a constant current power supply, you could always hook them in series and power them at both 350mA then at 700mA to see how they compare side by side.

Where did you purchase your blue Cree? How can you be sure if it is the 40lm binned blue? Few places offer the higher efficiency blue emitters. I think DX sells the X-Lamp blue emitter, but it uses older technology (which is equivalent to Lumiled's luxeons in therms of efficiency). Cutter may sell the 40lm blue emitter, as well as LED Lighting Supply. I am interested in hearing more about these...

-Tony

 

[HarryN]

I think your initial observation is closer than you think. Cree and Lumileds are at technologically similar points in time - there is not a 2 X difference between their products in real life / use.

It is important of course to mount them all for proper heat sinking, even those tiny little rebels with their anti - droop technology. (reduced temperature sensitivity.)

BTW - blue is the most dim of all visually, and you really need a large change to notice the difference.

 

[Axkiker]

Well, first of all, the Cree is operating at a lower power (more than half as much), so that is why they are running cooler. Secondly, are you attaching the emitter or star to the proper heatsink? What are you using to power up the emitters? How can you be sure they are getting the exact current? If you have a constant current power supply, you could always hook them in series and power them at both 350mA then at 700mA to see how they compare side by side.

Where did you purchase your blue Cree? How can you be sure if it is the 40lm binned blue? Few places offer the higher efficiency blue emitters. I think DX sells the X-Lamp blue emitter, but it uses older technology (which is equivalent to Lumiled's luxeons in therms of efficiency). Cutter may sell the 40lm blue emitter, as well as LED Lighting Supply. I am interested in hearing more about these...

-Tony

Well I was running the cree at 700ma... I have been using a large piece of aluminum with paste as a heat sink. So im not sure what the difference in heat is yet. Im going to do exactly what you said and hook them up in series side by side to see what difference there is if any...

I purcahsed my cree's from led supply. I was very specific when I ordered them. I got on the cree website and looked up the tech bulletins and made sure I ordered the 30lm at 350ma emitters. Now who knows if thats what I got but its what i ordered....

I havent had a great deal of time to screw with them so once I do ill report back and let you know what I find

 

[Fallingwater]

BTW - blue is the most dim of all visually

Nope - red is.

 

[Gunner12]

... even those tiny little rebels with their anti - droop technology. (reduced temperature sensitivity.)...

I don't think that those are used with the rebels or any production LED, but I could be wrong.

Yes, red should be the least visible per watt of light.

IIRC it takes 4x the light for a source to seem 2x brighter, not sure if those numbers are correct(or work for blue light) but it does take a lot more light for one source to seem 2x as bright as another. So you might not see a huge change in output when there actually is a decently large gape between the outputs.

 

[HarryN]

Ok, I guess I answered a bit fast in my attempt to help.

When I think red (650ish nm ) I tend to actually think of red / orange ( 617ish nm). These R/O LEDs are quite efficient per watt and easy to see (used as red lights)

Blue emitters (460 ish nm) on a per watt basis have fairly low visibility by comparison, at least to my eyes.

The main point is that you will likely not be satisfied if you expect that the LEDs from one market leader will be dramatically better than similar LEDs from another market leader. This would be similar to expecting a Ford pickup truck to be 2 x as fuel efficient and 2 x as fast as a similar Chevy.

I suppose next, someone will say that Cree is 10 years ahead of everyone else and that their ____ doesn't stink either. :shrug:

 

[Oznog]

Well it depends on whether you're talking about photopic (daytime) or scotopic (low-light adapted) vision.

http://en.wikipedia.org/wiki/Photopic_vision
http://en.wikipedia.org/wiki/Scotopic_vision

AFAIK that is scaled as visible response per mW of light energy. Note this is not per lumen, that would be recursive nonsense. Lumens are a unit of visibility, already adjusted as per the photopic scale.

All said, 50 lumens of blue should be perceived as "similar" to 50 lumens of green or 50 lumens red in photopic conditions. During scotopic conditions, the reaction to red drops way down though and 50 lumens of red may appear "weaker" than 50 lumens of green.

What the photopic chart tells us is that far more mW of light energy is needed to make 50 lumens of blue than 50 lumens of green. When an LED mfg tells us it's 80 lumens/W for the greens and 40 lumens/W for the reds, that's not necessarily saying the greens are more efficient in terms of mW of light power out since the mW out have been highly biased for the green ones.

Unless you have equipment to measure or place 2 LEDs side-by-side for comparison, brightness estimates are next to useless. For one, the human eye response is inverse-exponential. If you show someone light source A and ask him to pick what seems "twice as bright" from another light source by his judgement, he'll pick one with 4x the lumen output not 2x. In fact some of us have a hard time noticing 350mA vs 700mA outputs even though the light meter reports a 75% power difference.

And that's totally blown away by the human eye adaptation to ambient conditions. What seems like retina-burning in the dark may seem "normal" in indoor lighting and not even visible in direct sunlight. With judgement this inconsistent you just can't quantify the source this way, not within like +/- 50% unless you're an especially well trained eye.

 

[Axkiker]

Well here is my first impression of my new blue cree leds compaired to my blue luxeons

Well the first test was actually soldering them to the thin flexible pcbs I make.

The cree is by far harder to reflow. The cree has those little pads on each side of the bottom which you need to solder to the board. Well I was able to do it however it just seems like a PITA. You have to position them just so so in order to make sure you dont short the little bugger out with the thermal pad on the bottem. The luxeons have those little leads sticking out each side which you can easily apply paste to and them reflow. Now this was my first shot at reflow soldering so maybe im doing something wrong. I just find trying to set them exactly in place so only the very ends get solder is a pain.

The next test was burning them side by side in series. I supplied 700ma with my power supply and I can see a difference with the cree. It is fairly brighter. I cant say its a huge difference but with blue every bit helps.

The biggest thing that I have found is that the cree runs much cooler. I cant give you an actual temp as I am only checking by feel. I however am positive that the cree is running cooler.

Im gonna keep experimenting with them just to see if either or has any major advaltage or disadvantage. As of now the only thing I dont like about the cree is trying to reflow them.

 

[spencer]

The Cree is more efficient. That is why it runs cooler. It uses less power for the same amount of output. That is a huge advantage for it.

 

[Oznog]

What Cree and what Luxeon are we comparing?

Efficiency makes very little difference in heat generation. IIRC these can be something on the order of 20% efficient. So a vastly more efficient emitter might generate like 700mW of heat instead of 800mW.

Heat at the front has far more to do with the thermal conductivity of the board (and the LED power level) than the LED efficiency.

Reflow is cake. Solder paste, tweezers, heat gun with a temp-control dial standing on end.
Slowly lower the device closer to the heat gun, with the air heating the back. You want to first observe the flux boiling out. Once you see this start, back off a bit and wait like 10 sec for it to finish. Then finish lowering it until you see the solder melt, hold for a few sec, then pull it out of the stream. You really can't screw this up, the solder self-positions the device(s). Just don't overheat them and burn up the MCPCB insulator.

Or just use a hot plate, whatever.

 

[asdalton]

The Cree is more efficient. That is why it runs cooler. It uses less power for the same amount of output. That is a huge advantage for it.

But he said that the Cree was running cooler at the same drive current, not at the same light output.

If so, one reason is likely that the Cree LED is putting more of its energy into the radiated light rather than into heat that has to be conducted away.

It is also possible that the Cree has a lower Vf, which would lead to less power consumption at the same current. This would be easy to check with a voltage meter while the LEDs are running.

 

[HarryN]

It would be helpful to know if this is a Lumileds K2 or Lux V or rebel - I assume a K2 ? Which Cree model ?

Perhaps for all of us curious people, you can show a picture of the setup ? :thinking:

 

[2xTrinity]

BTW - blue is the most dim of all visually

Nope - red is.

You're both wrong -- IR, and UV are tied for the most dim visually/ watt output power. Sensitivity "peaks" in the green, and falls off in BOTH directions.

What the photopic chart tells us is that far more mW of light energy is needed to make 50 lumens of blue than 50 lumens of green. When an LED mfg tells us it's 80 lumens/W for the greens and 40 lumens/W for the reds, that's not necessarily saying the greens are more efficient in terms of mW of light power out since the mW out have been highly biased for the green ones.

As far as light power out / electricity in is concerned, green emitters are the least efficient of the visible-color LEDs.

Green at ~530nm has a theoretical ceiling of 600 lumens/watt. Real devices are cranking out 60-80, corresponding to about 10% power efficiency. Blue and red LEDs conversely have been made to approach 40% power efficiency. That is why a blue LED with a lossy phosphor coating on it (eg a white LED) can still cranking out more lumens than a green LED at the same power.

White phosphor LEDs are able to top 100lm/watt because of the insanely high efficiency of the blue dice compared to the other colors. RGB on the other hand will top out at more like 50-60 (if you consider the "best" blues, greens, and reds get 30/80/40 lm/W respectively)

My first reaction was "this sucks". I really dont see them being any brighter than the luxeon emitters I was buying. These crees are supposed to put out at least 30lm at 350 ma. The luxeon were supposed to put out 30 lm at 700ma. So I naturally thought at 700ma the cree would blow the luxeon away. not so much. I havent had a chance to set them up side by side. I plan on doing that tonight. I dont expect much of a difference though.

In order to judge the relative brightnesses, you should run the LEDs at close to the same power as each other, and do A-B testing with ceiling bounce. Your eyes will be able to tell the difference between light that is uniformly spread out, than light in a bright "point" (as a point source of just about any brightness will cause pupil constriction and saturation of the retina)

 

[lyyyghtmaster]

Yes, ceiling bounce will definitely prove more accurate. Also the lottery applies here. Blues are of all different wavelengths typically, sometimes even in the same batch, and shorter wavelength blues will appear slightly dimmer at the same mW output. Try adjusting distance from a white target to each LED so their brightness is equivalent and see if one appears to be deeper blue than the other. It might help to include various light-colored fluorescent and non-fluorescent items in the "target space".

BTW I have noticed that Luxeon whites use a slightly deeper to equivalent blue die to that typically used in Crees, but Seoul SSC often seems to use a somewhat shallower blue, thereby possibly explaining the slight edge in luminous efficacy they (used to?) have.

 

[yellow]

so, hooked in series, running on the same current, the Cree is brighter and runs cooler than the Luxeon...
... what is the problem then? Its totally correct.


(but I thought - with colored led - there is not much difference and the old Luxes still are good) :thinking:

 

[Fallingwater]

You're both wrong -- IR, and UV are tied for the most dim visually/ watt output power. Sensitivity "peaks" in the green, and falls off in BOTH directions.

Technically yes, but we're talking about LEDs useful for intense visible light here.

 

[ledstein]

I was lucky enough to play with all power leds from all the "big names": " Cree, Lumileds, Seoul, Osram".
My important advice for you is simple: forget the numbers, just test the leds. Lumens are for leds as mhz where for cpu´s: pure marketing with the big difference that you actually can´t test the numbers yourself or be sure from what bin the leds are.

More to the point: Cree leds should be the best right now because of one simple reason, they are designed to run at lower current (350mA). At 700mA and beyond heat becomes a big issue and few producers are willing to implement oversized heatsinks in their products. No one will use high currents in real life nomatter what the producers say about the performance at those levels.
For DIY my advice is to use only star versions and ALWAYS use a constant current powersupply.


ps: If you care about the numbers i found this website where you can compare all the brands/types of power leds :
http://www.ledrise.com/index.php?cat=c2_High-Power-LEDs.html&language=en

 

[HarryN]

I was lucky enough to play with all power leds from all the "big names": " Cree, Lumileds, Seoul, Osram".
My important advice for you is simple: forget the numbers, just test the leds. Lumens are for leds as mhz where for cpu´s: pure marketing with the big difference that you actually can´t test the numbers yourself or be sure from what bin the leds are.

More to the point: Cree leds should be the best right now because of one simple reason, they are designed to run at lower current (350mA).

ps: If you care about the numbers i found this website where you can compare all the brands/types of power leds :
http://www.ledrise.com/index.php?cat=c2_High-Power-LEDs.html&language=en

I am having a hard time agreeing with this information.

a) Designed for 350 vs 700
- Normally, an LED designed for 700ma operation will be more efficient at 350m.
- I may be entirely wrong, and one die may be different than the other technically, but this is not the reason IMHO.

b) Part Type
- So far, we still have no idea which LED package(s) the poster has used or how well the heat removal really is.
- For us to speculate on the results without this info is not useful.
- It is difficult to take the results seriously - we don't even know if it is a single die vs multi die package being compared.

c) Binning vs Marketing
- While your general concept might have some merit, that binning is a factor in marketing, it is not "just marketing"
- While it is quite difficult to stand around and shine an LED on the wall and exactly tell the bin, I know plenty of people that can easily identify which "similar" LED is brighter, has a nicer beam, and has a more pleasant color "for their eyes".
- I suggest you take the time to build or buy a few flashlights and go for a walk through the woods, or try cooking a steak by the light of an LED flashlight before you decide that color temperature and binning are not useful and significant.

d) The web link
- Surely you do not believe the numbers from an on-line retailer as your main source of information about LEDs ?

e) Lumens per die
- The "typical" results today from the major producers are all in the range of 100 Lumens per 1 x 1 mm die for similar packages and white color temps.
- After that, it is a matter of binning, packaging, and marketing.
- I will admit that some suppliers are more active in promoting their products on internet forums than others through stealth memberships.

 

[Oznog]

I am having a hard time agreeing with this information.

a) Designed for 350 vs 700
- Normally, an LED designed for 700ma operation will be more efficient at 350m.
- I may be entirely wrong, and one die may be different than the other technically, but this is not the reason IMHO.

Generally you'll find these devices' power ratings has a linear relationship with the package's thermal resistance. The thermal resistance of the package often presents more of a limit on the maximum power than any other factor.