Luxeon III vs 1W---Graphical comparison

[CM]

OK, more comparisons between 1W and 3W parts. The graph below uses the 1W part as the reference part. However, it also works by using a 3W part and derating to 1W levels.

1W overdriven to 3W levels:
Q flux ==> S flux
R flux ==> T flux

Another way is to derate the 3W parts down to 1W:
S flux ==> Q flux
T flux ==> R flux

The graph above shows the lumen output at several bias points. Summarizing, if you want an S flux 3W, just overdrive a Q flux 1W and use good heatsinking. Same goes with a T flux--overdrive an R and you'll get a T. Lumen maintenance will not be as good as the Lux III's but if it lasts 20,000 hours, who cares? S and T flux will probably be considered "dim" by the standards of the future.

CM

 

[yclo]

I'd suspected that, thanks for the information! /ubbthreads/images/graemlins/thumbsup.gif

 

[MR Bulk]

CM your work here benefits all of us greeatly. Thank You.

 

[CM]

You're all welcome. Just hope the information was presented coherently and not in my typical rambling fashion.

CM

 

[LightScene]

Do you know of any similar graph for the 5W?

 

[McGizmo]

CM,

Thanks for the info. Your mapping makes more sense but I was getting the impression that a R 1W mapped to a S III. Of course I gather this from comparing a small sample of LED's and an inexpensive Lux meter. I was actually hoping that the T rank III's was bring us into some new ground. I don't know how this example fits your data points but I have a 1W S rank with DB700 and it compares in Lux in about a dead heat with a S rank III with a DB1000.

- Don

 

[CM]

Going from 700mA to 1A should give an increase of about 30% in light output for the 43% increase in current. Wonder if the difference is due more to the intrinsic differences between LED's than the difference between the 1W and 3W. I worked out the numbers for a T rank which is middle of the road (76 lumens). From 700mA to 350mA from the graph above we get

76lumens/1.67=46 lumens which is smack in the middle of R flux territory. Same thing for the an S rank:

59/1.67=36 lumens, again within the Q rank rating.

It's somewhat disappointing to see a 1W@700mA compare to a 3W@1A in brightness. 1W or 3W, we're still subject to the same variations in output level. I'll get excited about the 3W when they hit the low end 5W flux levels such as U. I'll go break the bank when I see a V flux 3W. Now how about that same die multiplied four times in a 5W part /ubbthreads/images/graemlins/grin.gif

CM

 

[LEDmodMan]

CM,
Nice graph! I like what you've done there! /ubbthreads/images/graemlins/thumbsup.gif

 

[Reima]

McGizmo, you have S rank 1 watters?
RC

 

[Doug S]

[ QUOTE ]
CM said:
Same thing for the an S rank:

59/1.67=36 lumens, again within the Q rank rating.


CM

[/ QUOTE ]

Actually for the Lux IIIs, the "S" bin is 60-67.2 lm so this would equate to near the top of the "Q" bin for the 1W if you use the midpoint of about 64 lm.

 

[Free]

Guy's forgive me if I am not too technical, but I have a 3X3 watt Tristar that Wayne made and I posted comparison beam shots with my 3X1 watt Tristar HERE

What I am seeing is that the 3 watt and 1 watt only differ in color and not brightness. I would like to know the best batteries to use to get the most out of my 3X3 light.

Thanks /ubbthreads/images/graemlins/smile.gif

 

[Double_A]

Is the Luxeon III a three watt device?

or is it a one watt device that because of better manufacturing processes and heatsinking (heavier leadframe) can be safely run at a light output?

Can someone clear this up for me?

GregR

 

[RussH]

Double A, let me take a stab at answering your questions. Yes, to both. Except that It doesn't appear to have a heavier lead frame. To date, I don't think anyone has seen (identified) any actual difference in the construction
-Russ

 

[Double_A]

Thanks Russ


GregR

 

[McGizmo]

CM,

I took a 1W star and measured lux at various current levels and didn't see the percentage gain you have shown in your graph. I wonder if this is due to real world temperatures VS the 25C of the lumiled basis?

- Don

 

[Doug S]

[ QUOTE ]
McGizmo said:
CM,

I took a 1W star and measured lux at various current levels and didn't see the percentage gain you have shown in your graph. I wonder if this is due to real world temperatures VS the 25C of the lumiled basis?

- Don

[/ QUOTE ]
I would think so. Without knowing your actual data, let me correct CM's factors for temperature and then we can see how they compare with your real world data. I am going to making the following assumptions: 1) You are having a heat wave in Berkeley and the ambient temp is 25C; 2) your test setup heavily heat sinks the star boards and you take the measurements quickly enough that the heatsink does not warm appreciably; 3) your die to heatsink thermal resistance is 20 C/W [the star itself is 17 C/W and I added 3 C/W for the star to heatsink to get a nice round 20 C/W; 4) the test LED exhibits "typical" Vf vs I characteristics of:
<ul type="square"> 3.4V x .35A = 1.19W > Tj = 49C
3.7V x .70A = 2.6W > Tj = 77C
3.9V x 1.00A = 3.9W > Tj = 102C [/list]
Now from the Luxeon datasheet, the correction factor for these temperatures are:
<ul type="square"> 49C = .91
77C = .82
102C = .74 [/list]
Now if we take CMs 25C factors for .35A, .70A and 1A and multiply by the above temperature correction factors we get:
<ul type="square"> At .35A (.91)(1.00) = .91
At .70A (.82)(1.67) = 1.37
At 1.0A (.74)(2.17) = 1.61 [/list]
Now if we renormalize to 1.0 at .35A we get:
<ul type="square"> .35A = 1.0
.70A = 1.50
1.0A = 1.77 [/list]
Now something worth noting, If we compare the 1.50 at .70A to the 1.77 at 1.0A this represents only an 18% gain even under the very favorable thermal assumptions. It will be worse for typical steady state conditions.
OK Don, your turn. Show us the real world data.

 

[Clifton Arnold]

So what you guys are saying is in essence that these luxion
III's are nothing but a one watt with high bin numbers. I
have been reading posts on this subject like the gentleman
with the 3X3 tri-star and I have ordered in 6 3 watt stars
and I'm wondering if I have wasted my money! Well this is
what CPF is all about finding out what works and what don't.
Do we or do we not get more light from the luxeonIII in real
world conditions does anybody know for sure yet! when I get
mine in I will try them and see if I can tell any difference.

 

[Doug S]

[ QUOTE ]
Clifton Arnold said:
I'm wondering if I have wasted my money

[/ QUOTE ]
I think that the proper way of looking at it is to view it as paying a small premium to be assured of receiving what at a minimum is the equivalent of a high flux bin 1W. Folks here have often said they were willing to pay such a premium and now they can.

 

[evan9162]

In addition, it is a guarantee from Lumileds that it is okay to run these luxeons at 700mA or 1A with sufficient heat sinking.

That may not mean much to a modder, but to a large scale manufacturer wishing for a brighter light source, this assurance from Lumileds is a very nice thing to have.

 

[McGizmo]

Doug S.

As I recall, I think there was a less than 15% increase in Lux from 700 to 1000 mA on the 1W I checked. I did find a quick study I did where I looked at a 1W star at current levels from 250 to 700 mA, in increments of 50 mA. I had the star at a fixed distance from the light meter and as I recall, I took the lux reading at a five count after powering up the star. The lux reading only has relative value. I was primarily interested in delta lux/ delta mA. A table is below:

To read the table, for instance at 300 mA the delta mA is 20%. That is we went from 250 mA to 300 mA or a percent change of (300-250)/250=.20 . Each delta figure relates to the value in its column as increased from the previous column. I don't know if this makes sense as presented or not.

- Don