DiY LED - Cree CXA3070

[alesh]

50% efficiency.... That's just crazy talk my friend.... Either way I have a friend whose gonna sell me some z2's for 25 a piece, soooo for that price that'll hold me Over and I can just be late to the party lol

$25 a piece is a nice price. But I think that Vero29 for $28.21 would do better.

 

[Lighterfighter]

COOL INFO ON DRiVERS AND DIMMERS.

3.3 ELV Conclusion: Improper pairing of dimmers, ELV transformers, drivers, and LED loads can lead to any or all of the following problems:
• Providing insufficient resistive-type load to the ELV transformer output can lead to output instability and light output shimmer.
• If the ELV transformer appears too capacitive to the dimmer, the result can be poor dimming curves or even flicker in the loads.
• Using forward phase controls on ELV transformers can lead to high repetitive peak currents, potentially damaging the ELV load or dimmer control.

3.4 ELV Best Practices For ELV dimmers to operate on ELV transformers with LED loads, the LED load should meet all of the following criteria:
• Provide sufficient resistive-type load in the lamp to meet the minimum wattage requirements of ELV transformers.
• Minimize the use of integral capacitors on the input end of ELV transformers to prevent dimming curve problems.
• Minimize the input capacitance of the LED load as seen by the transformer. Large capacitance values may cause high repetitive peak current from the transformer to the load due to the high frequency switching of the transformer. These high currents can cause stress in the lamp, or falsely trip overcurrent protection mechanisms in the transformer. Besides ensuring that the dimmer is properly rated for ELV loads, selecting a dimmer with adjustable low end and high end trim can help minimize regions of poor performance. It is possible to meet these criteria with proper design of an ELV transformer and low-voltage LED load. However, like the MLV transformer case, it is nearly impossible to predict if these are met by either a particular LED product or an existing transformer in a retrofit application. The only way to ensure the dimming will be smooth and continuous, especially at low end, is to test all the devices together as a complete system.


http://www.lutron.com/TechnicalDocumentLibrary/3683586_Challenges_of_Dimming_Whitepaper.pdf full article.

 

[LeanMcsheen]

Still working analyzing the datasheets. Have to re-do all the tedious parts, pixel counting the Vf curve and droop curves for both CXB3070 and CXB3590 at each common current drive level. In short, the numbers are awesome

Same recommendations as far as drive currents

CXA3070 3K AB at 1.4A (52W) 43.3% $1.77/PAR W (aliex)
CXB3070 3K AB at 1.4A (50W) 45.4% $2.03/PAR W (digikey)
CXB3590 3K CB at 0.7A (49W) 51.7% $2.38/PAR W (digikey) (dang! )

You could run 4 CXB3070s at 1.4A on the HLG-185H-C1400 or
you could run 4 CXB3590s at 0.7A on the HLG-185H-C700

perfect match for our favorite drivers makes you wonder if Cree is watching

@SupraSPL What is the math behind the efficiency numbers? I'm sure i've seen it somewhere on here before but i can't for the life of me remember where. There's just too much info crammed into single threads over here. We could do with some stickies that have just data in them, or i suppose i could start bookmarking pages that have the specifics i want and catalog them...

 

[SupraSPL]

So the goal is to convert the data from the Cree PDF to efficiency % so we can make easy comparisons with other leds and other drive currents. So I estimate how many lumens at each current with estimated junction temp of 50C, and I estimate the dissipation wattage. The way I get the numbers is by pixel counting the graphs on the PDF and entering the data into a spreadsheet.

For the Cree CXA, they gives us a Vf curve and a combined temp droop/current curve. I also take into account the Vf shift for Tj50C. The spreadsheet parses all that and we get a "lumens/W" figure for each drive current. Once we have that, divide by the LER to convert into PAR Watts. Divide PAR Watts by dissipation watts and we get % efficiency. That is a very interesting number because it allows us to see how the LEDs are progressing and as we approach 100% the gains will be come more slowly. Currently with CXA we can get about 65% without getting into obscene cost/PAR W and that number will move a bit higher once we get top bins in CXB3590 5000K, maybe 70% efficient.

You can imagine the drastic difference in heat between a 25% efficient LED and a 70% LED. When you scale it up that is a miraculous difference. And as heat drops we will be able to use smaller and smaller heatsinks and still maintain very low junction temps near ambient, eliminating temp droop, which is already very low in most of our DIY lamps, 1-3%

 

[Lighterfighter]

Is the dimmer part of the Unit or is it a chip add-on you need to plug in? I think I will getting 2 of the Hgl185c, but i would like to see lumens at draw with the driver. I am a bit confused, I think what I learned from what I am reading is the amp that the driver runs affects the efficiency of the led, thus affecting the lumen output? Iff and run 2 or 3 cobs per driver how many w per light? I would really like to see Lumen output at each amp. rating for the driver. It seems to be the biggest AC driver in thier dimmmable series. Also then how the dimmable so what ill pump out with the dimmer at 50% and 75%

P.s I do understand how difficult it is to extrapolate data from that curve. the curve being ploted and maped with high res for accuract, thn cross ref ect prolly a few hours work. Thanks much for your info, Youll dig my light when its done lol.

 

[SupraSPL]

LF, your CXBs are 72V class, so at 700mA drive current they will each dissipate 49W, making 8200 lumens, 168 lumens/W, 51.7% efficient. Unfortunately that is the lowest the data sheet shows, but if we extrapolate from the CXA3590 data, if you went down to 350mA, you would bump up to 58% efficient or 189 lumens/W and almost doubling cost.

Here are the numbers for the higher currents. You can see the law of diminishing returns at work by looking at the $/PAR W column. As you increase current there is less and less of a cost savings.

 

[pirg420]

are we expecting that the newest bins will be avail , cb, cd, db and dd? those flux numbers look UNREAL.

 

[Lighterfighter]

@pirg420 I Remember reading Supra saying Cb were the best of the cxa model to be released, so that is what I ordered of the cxb It was highest bin I saw listed for release. Im a bit confused, but it seems I should pick highest v possible to reduce current and resistace to increase efficiancy as thats the whole model of led, right? but also want to get the most bang for my actual dollar spent by increasing lumen output, right? So i must pick between 2 or 3 currents? Im ganna do a dab and start watching some electrical videos on youtube. need a lil more guidance.

 

[Lighterfighter]

What I am drawing after relaxing and clearing my head( I always Understand my readings better high, even in college) is I kinda have two options I noticed you ran the numbers for the low currents for the Hlg185 I was looking at. Lol perfect, so now I think that either the 1050 or 700 are psuhing the limit of whats been done one here and based upon the prefered outcome; higher lumens with the 1050 and higher efficiency with the 700. scenerios c is 1400 I would venture to say psuhing those hard towards the 1.4 range listed as the next model up. would the led be more likely to fail due to heat? it would provide an extra 20% lumens at that rate. I did review the models people are working with now and found them to be ranging, but topping out at like 2.02 and the high end was 1.5 for $/parwatt, so I find the balance may be 1050 @ 1.75 $parwatt with a lm per watt of 148, that was not attainable form the previous charts.

This is all just technical though. to get to the nitty gritty, the implications that we are refering to are coverage area of my 6 lights. Do we shoot for what most top led companies do and run for highes lumen output, disregarding led lifespan and head issues (thow a bigger heat sink on it! blow another fan through it!)jk but you get it.

Scenerio A)
What is the cost in coverage area between the 700 and the 1050? If ran at the lower current, 700 my leds will produce 8k lumens at 48w. I believe it would be safe to say this could covera 1.25x1.25ft square or 1x1ft with ease. (following the assumption that 35w/per sqft is where people wanted to be before)

Scenerio C)
The second scenerios would conclude that I get 75w per led and perhaps could place them farther appart, sat 2ft? these would have depper canopy penetration, with the loss of a closer, coooler more even par coverage light. If any of this is off because I just dabbed and may have went totally stupid, please let me know. but I think I can also see how a mix of the two could be good(maybe scenerio D and save an led for later?). but the difference really comes down to how much additional penetration of the canopy would the 1050amp provide? Does the additional penetration compensated by the heat issue, meaning will the extra lumens and wattage cause me to really loose lumens because I have to move the light farther away thus depreciating the value of this option? All shit I would test if I had the resources lol.

Boom! hope i was making sense and on point this time. works been killing me. Its the weekend though. Happy saturday!

 

[Lighterfighter]

Interesting, the potentiometer on the mean well will change the color of the light?

 

[Scotch089]

What I am drawing after relaxing and clearing my head( I always Understand my readings better high, even in college) is I kinda have two options I noticed you ran the numbers for the low currents for the Hlg185 I was looking at. Lol perfect, so now I think that either the 1050 or 700 are psuhing the limit of whats been done one here and based upon the prefered outcome; higher lumens with the 1050 and higher efficiency with the 700. scenerios c is 1400 I would venture to say psuhing those hard towards the 1.4 range listed as the next model up. would the led be more likely to fail due to heat? it would provide an extra 20% lumens at that rate. I did review the models people are working with now and found them to be ranging, but topping out at like 2.02 and the high end was 1.5 for $/parwatt, so I find the balance may be 1050 @ 1.75 $parwatt with a lm per watt of 148, that was not attainable form the previous charts.

This is all just technical though. to get to the nitty gritty, the implications that we are refering to are coverage area of my 6 lights. Do we shoot for what most top led companies do and run for highes lumen output, disregarding led lifespan and head issues (thow a bigger heat sink on it! blow another fan through it!)jk but you get it.

Scenerio A)
What is the cost in coverage area between the 700 and the 1050? If ran at the lower current, 700 my leds will produce 8k lumens at 48w. I believe it would be safe to say this could covera 1.25x1.25ft square or 1x1ft with ease. (following the assumption that 35w/per sqft is where people wanted to be before)

Scenerio C)
The second scenerios would conclude that I get 75w per led and perhaps could place them farther appart, sat 2ft? these would have depper canopy penetration, with the loss of a closer, coooler more even par coverage light. If any of this is off because I just dabbed and may have went totally stupid, please let me know. but I think I can also see how a mix of the two could be good(maybe scenerio D and save an led for later?). but the difference really comes down to how much additional penetration of the canopy would the 1050amp provide? Does the additional penetration compensated by the heat issue, meaning will the extra lumens and wattage cause me to really loose lumens because I have to move the light farther away thus depreciating the value of this option? All shit I would test if I had the resources lol.

Boom! hope i was making sense and on point this time. works been killing me. Its the weekend though. Happy saturday!

Higher efficiency all day every day. I came in on this late so I don't know what you're trying to do, but the only perk personally of running them harder is less up front cost. Life span of LED, within their given range (current and temps), is nothing to worry about whether at 700 or 1400mA. as long as every things in check and rigged right. I would forget about them ever going out. I realllllly like that quality of LEDs.

Anyways, It sounds like your trying to figure what current to drive em at? Dono the given space, but inside my 2x3 with 272w (8 ABS and some xml2s and xpes) all at 700mA I am emitting in the center:

@18"- 857umols
@12"- 1495umols
@6"- > 3000 umols (meter maxes at 3k)

@500mA ~192
18- 626 umols
12- 1069 umols
6- 2287 umols

@350mA ~119w
18- 395
12- 636
6- 1401

Basically if you want to run half power, drop your height half and save on heat and gain efficiency.

I can throw pics up if you really want.

Edit: the farthest apart a cob is from the next is about 4"

 

[Lighterfighter]

okay man I am here now, down to chat! lol you even got ratings I was hopeing to find form your lights. thats results. thats sick! do you change patterns of the lights to test for highest efficiency and most level canopy? I am trying to cover a 2x4 space to 3x4. I just bought the new cxb top bin version of cxa. So Yea, I was asking about science of what you just gave me. lol so sick!

 

[SupraSPL]

Yep as long as you use sufficient cooling, heat and reliability will be no problem at any current you choose. Downside though, the HLG-185H-C1050 is not a great match for the CXB3590 due to voltage range (190) so it would increase driver cost and slightly reduce driver efficiency.

Also, to give fair warning you might want to hold off buying the driver until Digikey confirms the shipment. It is possible that they cancel the order if it turns out they cannot source them.

Dont mean to be discouraging though, you grasp these concepts correctly, especially after the dabs LOL. We will get you set up with a sick light one way or another.

 

[Lighterfighter]

Just for scale real quick whats lumen output of your abs in theory? Il comp then solve for A. I think the higher amount will lead to light poisoning lol the depreciation rate of these lights seems PRetty high, only downfall. I wants more info on optics.

@SupraSPL They wont charge me till it ships and it is a 30 day order. So if it doesnt pop within like 10 days after expected date the agreement defaults, theyll prolly ask me to extend, but I wont. Additionally they stated I can call to get lead time at any point. I also can call and cancel if I change my mind. I fell like The logic is me just jumping in line to make sure I get in on first order... know what im sayin!! lol I didnt find any cost in signing up. I asked a rep.

 

[Scotch089]

okay man I am here now, down to chat! lol you even got ratings I was hopeing to find form your lights. thats results. thats sick! do you change patterns of the lights to test for highest efficiency and most level canopy? I am trying to cover a 2x4 space to 3x4. I just bought the new cxb top bin version of cxa. So Yea, I was asking about science of what you just gave me. lol so sick!

No the pattern stays, I scrog so I wanted a nice even spread as much as I can but at the same time I wanted a higher center reading to make my outside rise a bit more than the middle (similar to a coliseum SOG style) with you trying to cover such a similar footprint it sounds like you and are pretty close to the same layout/page, maybe just some more distance between the center emitters or a couple more on each side.

I don't know about lumens, I ditched that theory a longgg time ago. But I do have a lux meter outside...

 

[Lighterfighter]

It also seems that dimmable would be irrelevant, and I am worried about color change that I have read in a few places. no bennifit if I jsut go efficient I wont ant to go any smaller on the lummens at that rate. I believe there is a scientific measure for to many lumens. Im starting to get it all now. these are all questions that ARe really hard to find clear cut answers to, honstly. Thank you!

 

[Scotch089]

And yea you need to listen to supe about drivers, I can set down and figure out what I'm looking for if I know the total V's and shit but that guy probably has the shit memorized part for part without him and a couple others I'm sure many of us wouldn't have even fucked with it.

 

[SupraSPL]

Ya at 50W I highly recommend reflectors or lenses. No worries about color change with dimming it will be minimal.

 

[Lighterfighter]

I am fuckin w it cause I care to understand and make it better, so it take initial thought and ongoing builiding upon that base. I am working okay so I def could say that the best is probably the 6 on an aluminum plate (would this work as reflector) at 7a, LIKE SUPRA SAID initially. jsut had to throw all the good questions up there.... lol in all fairness I have shared your wisdom with others here to share. what do you think about any supplemental color. Uva Uvb IR or REd? I can do maybe a 3x3 or 2x4 with the 7A? do you have another recommendation for driver since I WILL NOT use a dimmer.

Ya at 50W I highly recommend reflectors or lenses. No worries about color change with dimming it will be minimal.

What about the reflector any resources? would an aluminum panel work?

 

[Scotch089]

No I totally understand moving them around. If I could have just would've. I recently asked another member @Greengenes707 to record his readings off a single cob so I could estimate better for my next build. I Think I Just Got REAL Lucky On This On. Why is my phone typing caps?? Anyways. I came out and measured some more and I'm pretty sure my lux meter is bunk. So. But I took some edge readings and the falloff was not horrible. At 18" at the center I'm @857 at the edge it is @607, at 112w (350mA, not 119w) I got from 395 down to 242 umol at 18"