CREE and LumiLEDs binning

NewBie

*Retired*
Joined
Feb 18, 2004
Messages
4,944
Location
Oregon- United States of America
Some folks have asked about the differences in LumiLEDs and CREE binning. I'm working on a better chart, but here is what I have so far.

creelux2.jpg


Higher resolution:
http://www.molalla.net/~leeper/creelux.jpg


It is useful to note how some of the LumiLEDs luxeon bins are sub-divided by up to three CREE bins, which means much tighter white tint control.
 
Last edited:

balazer

Enlightened
Joined
Aug 22, 2005
Messages
486
Cree binning is kind of moot if you can't get more than one bin from Cree. I've only heard of WD color bins being available.
 

HarryN

Flashlight Enthusiast
Joined
Jan 22, 2004
Messages
3,976
Location
Pleasanton (Bay Area), CA, USA
I just received some Cree WD 3 watts (not tested yet). (WD tint - only available)

Just curious - the 1 watts I purchased earlier in the year had a blue ring near the base, but interestingly, even though they were pretty white looking on the wall, they seemed to be really good at green color rendering compared to my normal TWOK LIII.

Any comments / ideas on these two aspects as they relate to the WD 3 watt Xlamp ?
 

McGizmo

Flashaholic
Joined
May 1, 2002
Messages
17,290
Location
Maui
Newbie,
Thank you for the data and your efforts here! I think uniformity of color is more significant than the integrated x,y measured for a LED especially if the light is collimated with an obvious bi-tint result. Until an apparent unified tint is achieved with the phosphor coating, any binning or CCT assignment is a bit moot, IMHO. The fact that Cree and others are getting better is promissing news to be sure!!
 

NewBie

*Retired*
Joined
Feb 18, 2004
Messages
4,944
Location
Oregon- United States of America
The CREEs have improved significantly, for the color temperature across the part, showing a minor issue just on the edges of the part. LumiLEDs still has them beat for this.

CREE has them beat for efficiency, and should soon release their 70 lm/W parts.

One of the big issues with reflectors, is they mainly focus the edges of the beam, and the center part of the beam does not even hit the reflector. A deeper reflector is important with the CREE, to catch more of the premium part of all LEDs, more of the center of the beam, which is also the highest output area. A bit of surface texture on the reflector will also help mitigate the issues for those trying to use a reflector, along with the deeper reflector.

As HarryN mentioned, most of the available CREE white are bin WD, and it is hard to get any of the other bins. It seems CREE has a much better control on their process, holding the parts much closer to the target. Imagine if 90% of the LumiLEDs parts were WO bin!!! (lol, fat chance)

Anyhow, here is another chart, notice the tighter grouping on the CREE part- where the Luxeons have a much wider spread in colors, and the "sub-binning" as compared to the Luxeon (note-Unfortunately, these full color charts look different on different monitors):

creelux3.jpg
 
Last edited:

HarryN

Flashlight Enthusiast
Joined
Jan 22, 2004
Messages
3,976
Location
Pleasanton (Bay Area), CA, USA
I tested one of the Cree 3 watts on the bench today. The blue ring seen on the 1 watts purchased in Feb is GONE. Color is nice. Throw with my custom optics remains a challenge, but interestingly, they work better than a Lux III for flood applictions.

Vf was a bit higher than the TWOK I have as my "benchmark".

Overall, nice, quality product.
 

NewBie

*Retired*
Joined
Feb 18, 2004
Messages
4,944
Location
Oregon- United States of America
HarryN said:
I tested one of the Cree 3 watts on the bench today. The blue ring seen on the 1 watts purchased in Feb is GONE. Color is nice. Throw with my custom optics remains a challenge, but interestingly, they work better than a Lux III for flood applictions.

Vf was a bit higher than the TWOK I have as my "benchmark".

Overall, nice, quality product.


Yes, they have been working on them...enjoy!!!
 

evan9162

Flashlight Enthusiast
Joined
Apr 18, 2002
Messages
2,639
Location
Boise, ID
The XL3 7090's seem to have a much higher Vf than Luxeons. Now, the Vf at 350mA may be comparable to Lux I/IIIs, the V/I curve is much different.
While the luxeon curve is logarithmic, the cree curve is linear.

Code:
current		XL3-7090	Lux III TY0L (100hrs)
100		2.95		3.05
200		3.14		3.21
300		3.32		3.33
400		3.48		3.41
500		3.66		3.48
600		3.84		3.54
700		4.03		3.59
800		4.23		3.63
900		4.49		3.68
1000		4.79		3.72
 

chimo

Flashlight Enthusiast
Joined
Sep 16, 2004
Messages
1,905
Location
Ottawa, Canada
evan, that Cree curve shape is very odd. Much different shape than a 1W Cree curve I plotted a while back. I wonder what they did differently. I look forward to getting mine.

Paul
 

NewBie

*Retired*
Joined
Feb 18, 2004
Messages
4,944
Location
Oregon- United States of America
evan9162 said:
The XL3 7090's seem to have a much higher Vf than Luxeons. Now, the Vf at 350mA may be comparable to Lux I/IIIs, the V/I curve is much different.
While the luxeon curve is logarithmic, the cree curve is linear.

Code:
current		XL3-7090	Lux III TY0L (100hrs)
100		2.95		3.05
200		3.14		3.21
300		3.32		3.33
400		3.48		3.41
500		3.66		3.48
600		3.84		3.54
700		4.03		3.59
800		4.23		3.63
900		4.49		3.68
1000		4.79		3.72


Since you are testing again, it would be *very useful* to see what you get for the light output curve also, since we are only seeing half the story here.

On the luxeon, it's Vf curve looks awfully similar to what folks have measured for the light output versus current, but offset a bit.

On the CREE, what are you using for a heatsink and coupling to the heatsink?

.
.

It looks like you must have a very bad connection, because I just looked more at your data, and something is very wrong with your measurements. I also just ran the test myself, and don't see anything like what you posted. FYI, at 1 amp, I see 3.7V. Are you measuring the Vf right at the CREE terminals, or through your wiring?

.
.
I'll do up my data in a graph and post it shortly.

.
.
 
Last edited:

NewBie

*Retired*
Joined
Feb 18, 2004
Messages
4,944
Location
Oregon- United States of America
evan9162 said:
The XL3 7090's seem to have a much higher Vf than Luxeons. Now, the Vf at 350mA may be comparable to Lux I/IIIs, the V/I curve is much different.
While the luxeon curve is logarithmic, the cree curve is linear.

Code:
current		XL3-7090	Lux III TY0L (100hrs)
100		2.95		3.05
200		3.14		3.21
300		3.32		3.33
400		3.48		3.41
500		3.66		3.48
600		3.84		3.54
700		4.03		3.59
800		4.23		3.63
900		4.49		3.68
1000		4.79		3.72


Yes, you definitely had some issues, here is what I got for the CREE:
Code:
mA		Vf
10.9		2.645
21.2		2.702
53.1		2.798
95.0		2.879
196.0		3.02
315.3		3.148
414.7		3.237
496.7		3.304
610.0		3.398
694.7		3.464
799.3		3.544
904.7		3.624
1007.3	3.7

creelux4.jpg



I'm pretty sure you had a bad connection and must have not been measuring Vf directly across the LED, on the terminals of the LED itself.

From my data, and looking at yours, I figure you have one ohm of resistance, here is what the corrected data would look like, correcting for the "error"
Code:
Amps 	evan9162 CREE 		Corrected for 1 ohm resistance
0.1	2.95				2.85
0.2	3.14				2.94
0.3	3.32				3.02
0.4	3.48				3.08
0.5	3.66				3.16
0.6	3.84				3.24
0.7	4.03				3.33
0.8	4.23				3.43
0.9	4.49				3.59
1	4.79				3.79

Doing this correction makes your data curve match my data curve. Did you happen to have a 1 ohm sense resistor, and measure on the wrong side for Vf? It is always best to measure Vf right on the terminals of the LED itself.
 
Last edited:

evan9162

Flashlight Enthusiast
Joined
Apr 18, 2002
Messages
2,639
Location
Boise, ID
I'm using one set of wires for power, and another set of wires to measure voltage (no sense resistor). Both are soldered directly to the same point at the top contact. The second set of wires act as kelvin contacts, negating any effects of resistance in the power leads.

I wired up another one tonight, and saw very similar results (maybe 0.05V difference on the second one). I just went out and tested the resistance from the end of the power lead to the bottom contact of the cree (so testing the solder joint) - there was less than 0.1 ohms of total resistance, and about 0.01 ohms of resistance between the wire right before the solder joint, and the bottom contact on the cree (again, the wire is soldered to the top of the contact). This held true for both power connections. I don't think any solder joint is responsible for this. I tested again bypassing the wire, and using the probes of my DMM to touch the power contacts directly - exact same results.

here's my heatsinking setup: The power wires are soldered to the top contacts. I have a copper strip of foil the width of the heat sinking contact on the bottom. This props the cree up so the bottom power contacts don't short out, and provides the thermal path to the heat sink. The cree is then clamped to the heat sink (large, finned variety with fan) with a aluminium strip that has a hole for the dome. I used thermal goo on both interfaces. The thermal interface is good, as the thermal goo squeezes out, and after removing the clamp (which is left on during testing), the cree and copper strip are both stuck very well.
 

NewBie

*Retired*
Joined
Feb 18, 2004
Messages
4,944
Location
Oregon- United States of America
Well, something odd is happening on your end, not sure what it is yet.

I've measured this stuff many times on quite a few of the CREE XL7090, and they pretty much match the curve I posted from the tests I just did tonight.

Interesting.
 
Top