Cree chip efficiency charts?

I am looking to have another light made, similar to spectrum king. The one I have right now is doing very well, but is using 288 XBD's to get 494 watts. The data sheet shows that they produce 309 lumens @ 3w. The heatsink can only cool 500 watts @ 85c (from what I was told). I have been looking at a few of the newer chips, including the MHB-A, MHD-E, MHD-G, XHP50, and XHP70. All are showing a minimum of 118 lumen per watt to 134 lumen per watt @ max power. These chips are also able to run reliably @ 105c (I think). I am hoping to run the light at around 600 to 650 watts and achieve over 140 lumens per watt. Since I am not a math wiz and haven't been able to find an efficiency chart, I looking for help here before I try to contact Cree again. So far I have talked to 2 customer service reps and they haven't been able to help me. I am looking for the most penetration as possible.
So far I haven't gotten a full run in with this light, but so far they are looking very good. This is a couple weeks ago.

I would like to see a comparison of the XBD to the XHP50.

I thought that light design creates significant inconsistency in light intensity from center to edge of grow space. Why wouldn't you distribute the watts around the tent with more less-powerful fixtures? Some COBs spread around, or panels made with Cree diodes?

Seems like that would be an area of improvement that would yield more benefit than upgrading the diodes?

The problem with spectrum king, in my opinion, is the reflector he uses. With an HPS hood, I can still cover a 4'x4' area @ less than 24". I am still going to change reflectors for a better spread, but this hood is working good for now. These are my PAR testing # with and without a reflector.

Seeing that my #'s are still a lot less in the center with no hood, should mean the right reflector will increase my edge # and allow it to be lower if needed. These PAR # might be a little low, since it usually has a fan blowing on it and was quite hot how I had it sitting.

Am I reading that right? It says 782 (or 500 without hood) on center. 215 (or 152) 24" to the side?

Doesn't that support my question that it would be better to distribute the watts to create that amount of light so it more uniformly fills a 4x4 space? That seems like pretty steep fall off. I don't understand why you'd do another one of those. Maybe different lenses or reflectors would help. But, it seems like the long way to get light to the edges (subject to the inverse square law).

My reasons for this test was to determine if I could improve on my hood. And I can tell it can be improved. I don't have the time to build my own and my cost on this one was less than $2 a watt. When I use it for vegging I ended up with much bigger plants than I wanted. That is why I ended up with 13 50" tall plants crammed into a 4x4 canopy. I just can't throw away a good plant.
found this-

@SupraSPL- Is it possible to do an efficiency chart for the XHP50?
I just found a monster driver they might be able to use instead of 2 HLG 240h48b? http://www.meanwellusa.com/new_product/image/pic4.jpg

Long story short, COBs would give you higher efficiency at much lower cost, especially if you are looking to drive them at high power. I am sure you already know that and I understand the prefab manufacturers are not set up for that yet, so light designers are limited to choosing from certain models and want the best of those models. It may be difficult to know what you are getting though because they do not seem to be able to provide the brightness bin numbers, so in this case I would assume the lowest bin. What color temp are you aiming at and what price are they charging?

TJ 50C:

Vero29 V1.2 4000K @ 2.1A (79W) = 125 lm/W = 38% = $0.93/PAR W

Vero29 V2.0 4000K @ 2.1A (79W) = 132 lm/W = 41% = $0.88/PAR W

Cree XHP70 4500K M4 @ 2.4A (31W) = 101.6 lm/W = 31% = $0.78/PAR W

Vero 29 V2.0 4000K @ 3.6A (142W) = 115 lm/W = 35.6% = $0.56/PAR W

This is why I'm asking. What I see on the specs. is 4022 lumen @ 32 watts (125 lm/w). The XBD's are running around 1.7w, so I would think they would run the XHP70 @ about 18 watts. On Monday I will be getting word from both Cree and the engineers at Semlight if I can run 600w on this heatsink and if they can get the efficiency where I want it. I am pretty sure I will also be able to choose the bin I want, but will also check that. Thanks SupraSPL.
I know I over feed them once or twice, but still looking and smelling good.

Which chart did you see those specs on?

On the this page. The XHP50 is even higher.http://www.cree.com/LED-Components-and-Modules/Products/XLamp/Arrays-Directional/XLamp-XHP70

From there description, shouldn't I be able to run more power on this heatsink?

Not sure where they derived that number from but it is unfortunately misleading. According to their data sheet the top bin 4000K, typical figures, Tj 85C, at full power is 3152 lumens, 30.5W, 103.3lm/W, 32% efficient.

Good question regarding the heatsink. I have only seen pictures, lots of heat pipes and fins. Best way to test might be to do a temp droop test and see how much output you lose from room temp startup to warmed up temp. Have to monitor power consumption with a kill-a-watt at the same time to get a general idea of temp droop.

So if they are ran at half power I am only using 34 compared to 288 XBD's. The efficiency should be better, right?

Taking the example of XHP70 4500K M4 bin, run at half power (1.2A), Tj 85C:

122.2lm/W, 37.6% efficient, 14.1W dissipation
* 38 emitters = 535.8 dissipation W or 334W of heat

Which color temp are the XBDs or did they give you any bin info or part numbers. Do you know what current they are running at?

When I had this made, I didn't know how important binning was. I had it made with all the color temps. Do you see any chips in there lineup that would work better. I am obviously still confused on the math to find the best fit. The XPG looks like it has the best warm, the XHP50 looks more efficient than the XHP70, and the MKR's cool and neutral white peak at different nm in the 400 range, and the XML's, not the XML2, also peak at different nm. I'm so confused. Thanks again, I have learned a lot from you, but obviously not enough yet.

I have finally found the perfect hood for this light. It's an agrotec hood that fits perfectly over the heat pipes and leds. Now the light can sit at 18" and still cover a 4'x4' area. The edges do look stronger, but I will have to check with the PAR meter later. The lux meter now reads lower in the center compared to the old hood. This is what I was looking for. So far I have cut down the top 12" to 18" and will let the rest go another week.

Fix the light leak/gap in the hood ^^^^^ and your ouput #s will improve

good luck