JorgeGonzales
Well-Known Member
When BML became Fluence, their indoor spectrum basically evolved into 4000K white leds + 660nm, which ought to sound familiar to you.
Why are COBs better, and whats so great about the 3590?
- Thread starter Smoke-A-Cola
- Start date
Abiqua
Well-Known Member
because we are still in the price fixing stage? the classic bus101 archeotypical j curve of a marketing trend? Beta max is used by who anymore?
Its nice to be right when the argument fits the timeline, but in a year or two even, it might not be anywhere near the case...back to the regularly scheduled programming....
Abiqua
Well-Known Member
Fixed it ^ .....they could be outputting 3w's if the vF matches the Current
That's the only way a 3w diode exists in my mindchurchhaze
Well-Known Member
Have fun soldering all those tiny little diodes together.
CobKits
Well-Known Member
well the solution is to buy packaged boards from cutter, timber, that place in germany who's name escapes me, etc.
still for example top bin osrams are a few bucks and are damn near $1/watt.
great cobs can be had for $0.10-$0.25/watt. i think manufacturing costs favor cobs in the long run
still for example top bin osrams are a few bucks and are damn near $1/watt.
great cobs can be had for $0.10-$0.25/watt. i think manufacturing costs favor cobs in the long run
PhotonFUD
Well-Known Member
I tried some new terms lately and found out quite a few members don't know basic science. But that is to be expected in a public forum.
White led does make it simple to avoid photomorphogenesis issues as you point out, but the majority of radiant power I think should still be from the red region. It might be time to start dropping references to P680 and P700 photosystems but that may get the ignorant all upset and howling at the moon. Their heads may even explode once they find out chlorophyll is capable of luminescence and that a plant can use that as method to transfer energy.
Anyhow - 10% white balanced across PAR/90% 600-700nm blend has been my mantra for a good starting point to find optimal spectrum to best assist plant efficiency. Also I am promoting uniformity of light being delivered to the entire plant with a target of ~400 umoles hitting exposed leaf surfaces to achieve optimal photosynthetic efficiency. This is based on the research that the saturation point of photosynthesis is ~400 umoles and anything over is wasted energy.
PhotonFUD
Well-Known Member
This. CoBs cut down on labor.
And thanks for providing the additional information.
I guess the next step would to pick a ppfd/w as the target efficiency and figure out what balance of chips run at what wattage would be most cost effective.
For example, say I picked 11.5umoles/w. Why? Because I care for the environment, have lots of time on my hands to build, trying to hit a gpw record, whatever.
It looks like all of the chips will hit that mark when run at 8-15w. Factor to determine the total umoles per chip and divide by cost to find which would be the most cost effective.
Let's amp it up another notch and put coverage into the picture. More chips allows for more uniformity and better thermal management. Sure, upfront cost is more but it may return more in production for the least amount of power. With the expected drop in prices coming it may make even more sense.
CobKits
Well-Known Member
well thats the point of why i do those graphs. while its certainly an arbitrary, simplified technique, you can do good relative comparisons between chips. esp knowing the cost.
lets make some assumptions:
-once you get into cob arrays, more chips dont buy you *that* much difference in coverage, gains are mostly in efficiency. for example, going from 16 to 25 to 36 chips in a 5x5 isnt going to be that great of a difference as the 16 chips still do a great job of coverage and we are into diminishing returns above that point.
-lets discount the cost of optics and assume someone would solder up chips without reflectors onto the same bar heatsinks. its a different analysis once you get into cost of individual heatsinks, holders, reflectors. etc. it can be done, but initially lets skip it for simplicity and we'll throw some variables in later
-the measure above is somewhat arbitrary but lets use the CXB as a benchmark to estimate efficiency numbers. at 35W that DB bin 3590 is about 60% efficient, so lets use a factor of 6.4 to normalize all PPFD/W measurements to % efficiency. We can generate the following from the same data:
zooming in to our favorite 55-70% efficiency range:
lets make some assumptions:
-once you get into cob arrays, more chips dont buy you *that* much difference in coverage, gains are mostly in efficiency. for example, going from 16 to 25 to 36 chips in a 5x5 isnt going to be that great of a difference as the 16 chips still do a great job of coverage and we are into diminishing returns above that point.
-lets discount the cost of optics and assume someone would solder up chips without reflectors onto the same bar heatsinks. its a different analysis once you get into cost of individual heatsinks, holders, reflectors. etc. it can be done, but initially lets skip it for simplicity and we'll throw some variables in later
-the measure above is somewhat arbitrary but lets use the CXB as a benchmark to estimate efficiency numbers. at 35W that DB bin 3590 is about 60% efficient, so lets use a factor of 6.4 to normalize all PPFD/W measurements to % efficiency. We can generate the following from the same data:
zooming in to our favorite 55-70% efficiency range:
Last edited:
rob333
Well-Known Member
in the long run mate and after being in the world of hydro store this is what most of us hydro store owners are doing we will not take any led systems at all none zero reason being is they are just to hard to deal with and in the long run they just don't work we have done side bysides we have tested to see how long they run they all turn out good but the main issue is the ease of replaceing now if a cob shits itself halfway thru a grow what are u gunna do if u built it u are kinda ok but a store brought one u are fucked in the ass why here were i work all led systems are trash we take hps mh t5 and plasma that is all and some lec lights
rob333
Well-Known Member
i no they are power saving produce a good light source but they are just to hard for the normal grower to mess around with and the price is just stupid i can fit some1 out with a whole hps and mh system nutes coco pots and tent for the price of 1 light lol
CobKits
Well-Known Member
but that doesnt tell the whole tale. basically with one exception that list is almost entirely ranked by chip size (flagship chips by any manufacturer are more expensive per unit cost, just like buying the latest and greatest hard drive vs say the second size down -ok maybe im dating myself. iphone 8 vs iphone 7 or whatever... think of it as 3590 vs 3070, or 3618 vs 1212, etc)
to level the playing field lets add $5 in holders and reflectors to each cob:
lets assume youre a passive pinfin guy, add $20 per cob :
different story now, huh? and if your head is already spinning lets look at some other factors we havent considered:
-at a given efficiency, smaller chips like a 1212 are running at lower wattage than say a larger chip like an 1825. So in a build out of similar total wattage, the smaller chips could use smaller heatsinks, say $10 heatsinks instead of $20 heatsinks, which would skew the curve above
-does not factor in labor to install, mount, and wire more, smaller COBs (again, fundamental to our argument of why COBs beat individual diodes
-in large buildouts, you may get a better deal on 100pc of smaller chips vs 10 pc of larger chips which would skew the economics
-in an array of many smaller chips vs fewer larger chips, one may elect to use or not use optics accordingly, again skewing the economics of a given build
-in the zoomed in charts above we are snobs looking at the 60-70% efficiency range. What about the average joe who is on a budget, doesnt have heat issues, and just wants to get better coverage, better spectrum, reduced maintenance, and a little efficiency bump over his current SE HPS. That guy could be totally happy with a 45% efficient rig, but move on down the chart and we see a whole different set of variables. Suddenly, the chips that were the best value at 70% efficiency arent the leaders at 45% efficiency. And the difference in cost effectiveness from top to bottom is now not a factor of two but a factor of 3.5 between the best and worst chips.
etc etc. many many factors. I guess the takeaway is know your operating range, and choose wisely because across the board at any efficiency, the most cost effective chip is getting you 2-3.5X the light per dollar of the ones at the bottom
to level the playing field lets add $5 in holders and reflectors to each cob:
lets assume youre a passive pinfin guy, add $20 per cob :
different story now, huh? and if your head is already spinning lets look at some other factors we havent considered:
-at a given efficiency, smaller chips like a 1212 are running at lower wattage than say a larger chip like an 1825. So in a build out of similar total wattage, the smaller chips could use smaller heatsinks, say $10 heatsinks instead of $20 heatsinks, which would skew the curve above
-does not factor in labor to install, mount, and wire more, smaller COBs (again, fundamental to our argument of why COBs beat individual diodes
-in large buildouts, you may get a better deal on 100pc of smaller chips vs 10 pc of larger chips which would skew the economics
-in an array of many smaller chips vs fewer larger chips, one may elect to use or not use optics accordingly, again skewing the economics of a given build
-in the zoomed in charts above we are snobs looking at the 60-70% efficiency range. What about the average joe who is on a budget, doesnt have heat issues, and just wants to get better coverage, better spectrum, reduced maintenance, and a little efficiency bump over his current SE HPS. That guy could be totally happy with a 45% efficient rig, but move on down the chart and we see a whole different set of variables. Suddenly, the chips that were the best value at 70% efficiency arent the leaders at 45% efficiency. And the difference in cost effectiveness from top to bottom is now not a factor of two but a factor of 3.5 between the best and worst chips.
etc etc. many many factors. I guess the takeaway is know your operating range, and choose wisely because across the board at any efficiency, the most cost effective chip is getting you 2-3.5X the light per dollar of the ones at the bottom
Last edited:
bri77
Well-Known Member
For the beginner Leds are extremely attractive for their simplicity. You just buy one item, the light, and plug it in. With HPS you have to pick a bulb, a hood type and shape, ballast. Its extremely intimidating. If you've been growing for years you may not understand how important that is.
PhotonFUD
Well-Known Member
More great info.
And regarding the snobs to the average joe comment, keep in mind that the average joe reaps the benefits from the pursuit of perfection. Those last few charts would be a real downer to a sales guy working on volume however.
From the first graph, there shows a rapid decline in efficiency from first reading to ~40w. If one was to run the chips at only 10w each, which ones would be the most cost effective? It kind of looks like it is still Cree or the Vero coming out on top based o the last charts or maybe I am missing something.
40 emitters putting out ~100umoles (~10ppfd/w) each using a total of ~400w electricity costing ~$1000 (1212s) or ~$2000 (3590s) might sound a little crazy for some but looking at the retail prices for a lot of offerings these days, there is definitely a market that can pay.
And regarding the snobs to the average joe comment, keep in mind that the average joe reaps the benefits from the pursuit of perfection. Those last few charts would be a real downer to a sales guy working on volume however.
From the first graph, there shows a rapid decline in efficiency from first reading to ~40w. If one was to run the chips at only 10w each, which ones would be the most cost effective? It kind of looks like it is still Cree or the Vero coming out on top based o the last charts or maybe I am missing something.
40 emitters putting out ~100umoles (~10ppfd/w) each using a total of ~400w electricity costing ~$1000 (1212s) or ~$2000 (3590s) might sound a little crazy for some but looking at the retail prices for a lot of offerings these days, there is definitely a market that can pay.
CobKits
Well-Known Member
its a lot of data to digest.
the ones on top are more efficient for example go back to the first 2 graphs for bare chips and lets look at the uber efficient side of the curve
the most cost effective in $/light produced is the $12 1212:
it has a respectable ~13.5W of output at 65% and about 25W @ 60%
where the least cost effective is the $40 cxb which has about 15W @ 65% and 34W@ 60%
so lets peg 60% as our target, we want 1000 watts.
1000W of 1212s at 60% = 40 chips@ 25W x $12 =$480 = $0.48/W
1000W of CXB @ 60% = 30 chips @ 34W x $40 = $1200 = $1.20/W
you gotta take low level readings with a grain of salt as there is maybe 5 umol of ambient lighting not adjusted for, as well as low current measurements are less accurate. Data above 10W/100 umol should be fairly accurate tho
the ones on top are more efficient for example go back to the first 2 graphs for bare chips and lets look at the uber efficient side of the curve
the most cost effective in $/light produced is the $12 1212:
it has a respectable ~13.5W of output at 65% and about 25W @ 60%
where the least cost effective is the $40 cxb which has about 15W @ 65% and 34W@ 60%
so lets peg 60% as our target, we want 1000 watts.
1000W of 1212s at 60% = 40 chips@ 25W x $12 =$480 = $0.48/W
1000W of CXB @ 60% = 30 chips @ 34W x $40 = $1200 = $1.20/W
you gotta take low level readings with a grain of salt as there is maybe 5 umol of ambient lighting not adjusted for, as well as low current measurements are less accurate. Data above 10W/100 umol should be fairly accurate tho
wietefras
Well-Known Member
It's really the same as the other CXB's. The thing is it's bigger with more diodes on the blue die which means that if you test all CXB COBs at a certain current the individual diodes on the CXB3590 will be driven softer and therefore run more efficient.
The CXB3590 is around 30% bigger than the CXB3070. So you should really give it more current than the smaller ones to keep it comparable.
If you drive the CXB3590 @ 1750mA it's about the same efficiency as the CXB3070 @ 1400mA both around 2.4umol/s/W. At those currents the CXB3590 ($35) produces 4.22umol/s/$ and the CXB3070 ($2
4.17umol/s/$. So when compared at appropriate currents, the price difference is tiny. Although you also might have to factor in extra cooling costs for more COBs.It's best to consider both types and see which of the two best allows you to distribute the light uniformly over your grow surface. For instance you can use 8 CXB3070 or 6 CXB3590 for around the same amount of light at the same price level and efficiency (ie electricity costs). However, those 8 COBs CXB3070 give a lot better uniformity over a square 100x100cm surface than 6 CXB3590.
rob333
Well-Known Member
are you serious lol so tell me if the led u just brought say one or two of the lights go out mid grow ?? are u gunna go buy a hole new led light or just unscrew the bulb of a hps walk down to the nearest hydro store buy a bulb and screw it in how hard is that ? unless u don't have hand well then u are fucked
rob333
Well-Known Member
not saying the are not good and u can pick up some nice cheap ones off ebay heap or chinese factorys are building these that are ok i have found a chinese factory that will punch out cob cree's for half the price of what timber lighting or any of these american made cree's just search made in china.com.org chinese light are good i have had test lights shit themselfs be4 the chinese tai one have
Resinhound
Well-Known Member
Stephenj37826
Well-Known Member
wietefras
Well-Known Member
On the other hand, for DIY, the COBs are easier to use and probably cheaper too in the end.
Imagine soldering hundreds of 3W or 5W leds or only wire up five 100W COBs (all run at half or even less of max power)
Commercial horticulture fixtures from the bigger corporate sellers go for 3 to 5 times the cost of the parts. If not more. You see China burples rebadged and sold for thousands (dollars or euro's) when they can be bought direct from China for hundreds.
Either way we can use high quality parts and still come out cheaper. Commercial fixtures would prices themselves completely out of the market if they used quality parts.
Imagine soldering hundreds of 3W or 5W leds or only wire up five 100W COBs (all run at half or even less of max power)
Commercial horticulture fixtures from the bigger corporate sellers go for 3 to 5 times the cost of the parts. If not more. You see China burples rebadged and sold for thousands (dollars or euro's) when they can be bought direct from China for hundreds.
Either way we can use high quality parts and still come out cheaper. Commercial fixtures would prices themselves completely out of the market if they used quality parts.