I'm starting this thread to one share knowledge, and two gleam some light on an alternative method of lighting our grows that isn't talked about much here on these forums.
So "What" you say "Lighting a grow with Off the shelf Screw in Led Bulbs that's non sense, i'm just gonna stick to my expensive panels and custom build COB arrays"
Now hold on a sec and hear me out, as i'm sure questions about heat generation, light efficiency, light output, and penetration is running through your heads.
So to start off i will state again Screw in led bulbs are a perfect and viable option for medium to small sized grows, and are cheaper to do than most comparable options and here is why.
First let use diy COB setups as a comparable example, which i have built a few of myself. Most people on here focus on the Citizen CLU048's or the Cree 3590, or what i've liked for bang for the buck the Bridgelux Vero Gen7 Series, they are all great options for building a COB array or setups but lets jump into the meat of it and make some comparisons.
All figures i'm gonna point out come from manufacturer spec sheets and i'll use Lumen, Wattage, and Lumens/watt at 2700k and 5000k color temperatures at Typical values with junction temperatures at 85degrees C, as i'll say almost no one is running their stuff at 25 degrees C. To note i'll get into Underdriven stuff later, but not yet.
So "Hold On" you say "Lumens are for humans, shouldn't you be comparing moles of energy or PAR values" and i will say yes, but this is just for comparison purposes and lumens is an easily available figure to do it with as manufactures don't regularly publish par ratings. So because of that we cannot make accurate comparisons to red-blue-mix led panels but that's not my goal.
Output
Alright let's begin by taking a Citizen CLU048 1212 for example its typical output is as follows for each color temperature.
2700k produces 5320 Lumens at 37.4watts, with an efficiency of 142Lu/watt
5000k produces 5815 Lumens at 37.4watts, with an efficiency of 155Lu/watt
how about a CLU048 1812
2700k produces 7790 Lumens at 56.1watts, with an efficiency of 139Lu/watt
5000k produces 8510 Lumens at 56.1watts, with an efficiency of 152Lu/watt
or a CLU048 1818
2700k produces 11370 Lumens at 84.2watts, with an efficiency of 135Lu/watt
5000k produces 12420 Lumens at 84.2watts, with an efficiency of 147Lu/watt
Which you could say are all pretty comparable
Let's move on to a Cree 3590
2700k produces 11000 Lumens at 86.4watts, with an efficiency of 127Lu/watt
5000k produces 12000 Lumens at 86.4watts, with an efficiency of 139Lu/watt
Or how about a Bridgelux Vero29
2700k produces 9720 Lumens at 79.1watts, with an efficiency of 134Lu/watt
5000k produces 11800 Lumens at 79.1watts, with an efficiency of 149Lu/watt
Bridgelux Vero 18
2700k produces 4060 Lumens at 30watts, with an efficiency of 123Lu/watt
5000k produces 4540 Lumens at 30watts, with an efficiency of 138Lu/watt
Now let's move on over to what we are here about i present to you a standard Off the shelf led bulb
A 800 lumen Philips bulb
And here's some different broken down bulbs, two two being list 8w bulbs and the bottom a 14w bulb.
Using a 8w Philips bulb as an example, their package rated output is 800lumen at 8w, but to note here that with the ballast included and is the minimum they have to be cause anything else is false advertising.
But let's jump to what led's they are using with this example being a Philips bulb they are most likely using Luxeon 3030 SMD led's because of the size, current, forward voltage, and because philips partially owns lumiled which makes luxeon, Spec wise in their documentation they are rated at a 25 degree C rating and don't offer at 85 degree C rating but if you do the math of any of the other leds they all work out to be a roughly 10% drop in luminous flux at 85 degrees, so i use that to compare like to like.
Now will your retail bulbs have the top bin leds "definitely not" or the latest generation "possibly" but here are the figures for a luxeon 3030 smd led which there are 10+ of in each bulb depending on what it is.
Using a 10% drop in comparsion of a 25 degree C vs 85 degree C junction temp we would get the following
2700k produces 93 Lumens at 0.73watts, with an efficiency of 126Lu/watt
5000k produces 103.5 Lumens at 0.73watts, with an efficiency of 142Lu/watt
Thats pretty damn comparable to the afore mention COBs, but we aren't necessarily running at that as i will describe below.
Now through some latest measurement and research, and to note this is with the latest spec sheets as that's what i could find easily, the "8w" daylight bulbs are using 14 smd chips and what seem to be Osram Duris 2835, but they are running at 6.9w after the ballast at 116ma and 59.5v in a 7 series x 2 parallel array which would make them actually underdriven but they would be putting out around 61-76lm each which puts it right over the 800lumen mark which make sense, so a lumen efficiency of around 123-153lu/w but i think were using some older chips.
The "8w" soft white bulbs are most likely using the Luxon 3030 Led's but are running in a straight series array at 59v at 126ma total of 7.43w but the new ones would be putting out 93lm each at 85 degrees for a 125lu/w effiency, but i think we have a bit older model smd chips. Now the bigger "14w" softwhite is also most likely using the Luxon 3030 but is running a 15 series x 2 parallel array at 86.5v at 148ma after the ballast so 12.8w but again these will be under driven, and off the newer spec sheet would be running around 64lm each so 150lu/w, but again i think we're running an older version, so i'll have to hunt down a older spec sheet.
Now with all those specs and measurements aside through my own past measurement with the Ballast included and going by the package rated lumens we would be getting around 110-125lu/watt efficiency which make perfect sense as that's their minimum they much achieve.
So for like to like they are not far off efficiency wise, with the COB's being more efficient obviously, now with diy COB setups you can underdrive your leds to get higher efficiency but that trade off come at an increased expense of more COB's, ballasts, and heatsinks to get the same light output.
COST
Now let's jump into cost comparisons using lumens as the comparable measurement, and prices are done in USD and taken from mouser, digikey, and big reno stores like home depot.
Let start with a Citizen CLU048 1212
They typically run around $30 with an associated Meanwell ballast setting you back another $14, and you still will need a heatsink such as an old used CPU cooler $5, and thermal compound and wiring so let's say $50 even and that's not including any shipping charges you may incur for 5320 lumens at 2700k or 5815 Lu at 5000k
What about some "Off the Shelf" Philips bulbs at package rated minimum lumens using a "8w" bulb at a rate "800lumens" as an example, They cost $8 for a pack of four, so to equal the same output we would need 6.65-7.27 bulbs respectfully, but let's just jump that up to 8 bulbs for good measure,
So $16 Total vs $50 for the same thing that's 70% cheaper, what if we include some keyless sockets or cleat sockets for the bulbs at $1.5-3 a piece, Were still looking at $28-40 so 44% to 20% cheaper and that's still not equating in shipping vs walking into your local store.
Say you went with a Cree CXB 3590 instead, that price difference is about the same.
So what if you don't want to run so many bulbs, well bump up to the "100w equivalent bulbs" that put out twice the light, unfortunately though they are about 50% more expensive than the "60w equivalent" but at $12 for 2 bulbs, you're still looking at $24 vs $50 or with sockets $30-36 vs $50 still without taking into consideration any shipping charges.
Kinda show we have a clear winner in the cost aspect.
Electricity Cost
But "What about potential increased electricity costs" well let's try to compare but i'll add back in later with actual measurements after the ballast, but i'll try my best
Let use a Meanwell LPC-60-1050 Ballast as example to power out a Citizen CLU048 1212 COB, which at typical consumption is 37.4watts, now because nothing is 100% efficient our ballasts will lose some energy in the conversion process, so let's consult the spec sheets, Meanwell lists that specific ballast at 87% efficiency under typical conditions, do the math and that should mean will be consuming right around 43w from the wall, where as "package rated" led bulbs will pull 64watts for the same amount of light.
Now my past measurements have those bulbs actually pulling less in the 6-7w range for the "8w bulbs", but let's go with the bigger number of those of 56w total draw till i take some readings and we can compare after the ballast like for like, or i hook a bunch of bulbs up in parallel and see their total draw.
So 43w vs 56w means a 13w difference, now let's say our energy cost is $0.20 a kw/h, mine is less than half that but let's use a greater figure as not everyone has cheap electricity. So per cob/bulb comparable unit at $0.20 a kw/h running at 18/6 for us auto guys and gals, that works out to be 234 watt/h more a day or 0.234 kw/h
The cost difference between going to a Citizen CLU048 1212 vs bulbs per my previous calculations, is $14-22 respectively with which route you go, so how many days will it take to break even with the additional electricity cost vs up front cost ?
$14 / (0.234 kw/h x $0.20 a kw/h) = 299 days
$22/ (0.234 kw/h x $0.20 a kw/h) = 470 days
And again that's not including shipping costs of your COB's, Ballasts, and Heatsinks, which is usually $10+ an order for me anyways, or if your under driving them and need more of all that. But if we factor that in look at tacking on an extra 213 days to break even.
Myself at the underside of $0.09 a kw/h im looking at 3-4 Years to pay off the difference.
So if you have cheaper electricity bulbs are the way to go, if it's more expensive DIY COB's can be advantageous so long as you use them for over 1-1.5 years or less depending on how high your kw/h rate is
Heat and Penetration
Now let's get into heat generation and penetration.
Heat is our byproduct of taking electricity and turning it into light, now if side by side the efficiency is comparable, our heat generation will also be comparable, and it's not much simpler that that.
As far as penetration, if you are supplying the same amount of light to a given enclosed area the plants will use that, now if we have a singular source of higher intensity light we will have to have our light further away from out plants in comparison so things don't get burned heat and or light burn related, but if we supply our plants and canopy from multiple angles with less intense sources but more of them, equaling the total light output as a singular source, our plants won't care how they receive it, sure they will grow differently in say denser flatter canopies vs single cola tee structures, but bud is bud.
But look at any commercial led panel and that's the same principle, just spread out a little further in our case.
Now even still some people on other forums are getting plenty of penetration with good dense bud up to 36"or 92cm from the lights "100w equivalent bulbs", how much more space do you really need with some minor training techniques, unless your wanting to grow big trees you could still side supplement light where you plant is getting equally lit vs only the top getting light and everything trying to race towards that light.
Only thing left to discuss for this starter would be that you can also if you want break down these Off the shelf led bulbs and make your own arrays from them as a cheap alternative, which i have for my 5 gal bucket build, as i wanted to go simple in that application.
And here's a link to an album on how to break down the philips bulbs, which are pretty much the same regardless of their color temp or wattage.
So for now i'll leave this here, and i will contribute my own grow using Off the Shelf bulbs, and will look at getting permission to use some photos of other peoples grows on other forums using Off the Shelf bulbs in their own grows.
So feel free to contribute, add, show off your own grows, and or discussions about this alternative which i believe and know is a perfectly acceptable alternative to other current ways.
So "What" you say "Lighting a grow with Off the shelf Screw in Led Bulbs that's non sense, i'm just gonna stick to my expensive panels and custom build COB arrays"
Now hold on a sec and hear me out, as i'm sure questions about heat generation, light efficiency, light output, and penetration is running through your heads.
So to start off i will state again Screw in led bulbs are a perfect and viable option for medium to small sized grows, and are cheaper to do than most comparable options and here is why.
First let use diy COB setups as a comparable example, which i have built a few of myself. Most people on here focus on the Citizen CLU048's or the Cree 3590, or what i've liked for bang for the buck the Bridgelux Vero Gen7 Series, they are all great options for building a COB array or setups but lets jump into the meat of it and make some comparisons.
All figures i'm gonna point out come from manufacturer spec sheets and i'll use Lumen, Wattage, and Lumens/watt at 2700k and 5000k color temperatures at Typical values with junction temperatures at 85degrees C, as i'll say almost no one is running their stuff at 25 degrees C. To note i'll get into Underdriven stuff later, but not yet.
So "Hold On" you say "Lumens are for humans, shouldn't you be comparing moles of energy or PAR values" and i will say yes, but this is just for comparison purposes and lumens is an easily available figure to do it with as manufactures don't regularly publish par ratings. So because of that we cannot make accurate comparisons to red-blue-mix led panels but that's not my goal.
Output
Alright let's begin by taking a Citizen CLU048 1212 for example its typical output is as follows for each color temperature.
2700k produces 5320 Lumens at 37.4watts, with an efficiency of 142Lu/watt
5000k produces 5815 Lumens at 37.4watts, with an efficiency of 155Lu/watt
how about a CLU048 1812
2700k produces 7790 Lumens at 56.1watts, with an efficiency of 139Lu/watt
5000k produces 8510 Lumens at 56.1watts, with an efficiency of 152Lu/watt
or a CLU048 1818
2700k produces 11370 Lumens at 84.2watts, with an efficiency of 135Lu/watt
5000k produces 12420 Lumens at 84.2watts, with an efficiency of 147Lu/watt
Which you could say are all pretty comparable
Let's move on to a Cree 3590
2700k produces 11000 Lumens at 86.4watts, with an efficiency of 127Lu/watt
5000k produces 12000 Lumens at 86.4watts, with an efficiency of 139Lu/watt
Or how about a Bridgelux Vero29
2700k produces 9720 Lumens at 79.1watts, with an efficiency of 134Lu/watt
5000k produces 11800 Lumens at 79.1watts, with an efficiency of 149Lu/watt
Bridgelux Vero 18
2700k produces 4060 Lumens at 30watts, with an efficiency of 123Lu/watt
5000k produces 4540 Lumens at 30watts, with an efficiency of 138Lu/watt
Now let's move on over to what we are here about i present to you a standard Off the shelf led bulb
A 800 lumen Philips bulb
And here's some different broken down bulbs, two two being list 8w bulbs and the bottom a 14w bulb.
Using a 8w Philips bulb as an example, their package rated output is 800lumen at 8w, but to note here that with the ballast included and is the minimum they have to be cause anything else is false advertising.
But let's jump to what led's they are using with this example being a Philips bulb they are most likely using Luxeon 3030 SMD led's because of the size, current, forward voltage, and because philips partially owns lumiled which makes luxeon, Spec wise in their documentation they are rated at a 25 degree C rating and don't offer at 85 degree C rating but if you do the math of any of the other leds they all work out to be a roughly 10% drop in luminous flux at 85 degrees, so i use that to compare like to like.
Now will your retail bulbs have the top bin leds "definitely not" or the latest generation "possibly" but here are the figures for a luxeon 3030 smd led which there are 10+ of in each bulb depending on what it is.
Using a 10% drop in comparsion of a 25 degree C vs 85 degree C junction temp we would get the following
2700k produces 93 Lumens at 0.73watts, with an efficiency of 126Lu/watt
5000k produces 103.5 Lumens at 0.73watts, with an efficiency of 142Lu/watt
Thats pretty damn comparable to the afore mention COBs, but we aren't necessarily running at that as i will describe below.
Now through some latest measurement and research, and to note this is with the latest spec sheets as that's what i could find easily, the "8w" daylight bulbs are using 14 smd chips and what seem to be Osram Duris 2835, but they are running at 6.9w after the ballast at 116ma and 59.5v in a 7 series x 2 parallel array which would make them actually underdriven but they would be putting out around 61-76lm each which puts it right over the 800lumen mark which make sense, so a lumen efficiency of around 123-153lu/w but i think were using some older chips.
The "8w" soft white bulbs are most likely using the Luxon 3030 Led's but are running in a straight series array at 59v at 126ma total of 7.43w but the new ones would be putting out 93lm each at 85 degrees for a 125lu/w effiency, but i think we have a bit older model smd chips. Now the bigger "14w" softwhite is also most likely using the Luxon 3030 but is running a 15 series x 2 parallel array at 86.5v at 148ma after the ballast so 12.8w but again these will be under driven, and off the newer spec sheet would be running around 64lm each so 150lu/w, but again i think we're running an older version, so i'll have to hunt down a older spec sheet.
Now with all those specs and measurements aside through my own past measurement with the Ballast included and going by the package rated lumens we would be getting around 110-125lu/watt efficiency which make perfect sense as that's their minimum they much achieve.
So for like to like they are not far off efficiency wise, with the COB's being more efficient obviously, now with diy COB setups you can underdrive your leds to get higher efficiency but that trade off come at an increased expense of more COB's, ballasts, and heatsinks to get the same light output.
COST
Now let's jump into cost comparisons using lumens as the comparable measurement, and prices are done in USD and taken from mouser, digikey, and big reno stores like home depot.
Let start with a Citizen CLU048 1212
They typically run around $30 with an associated Meanwell ballast setting you back another $14, and you still will need a heatsink such as an old used CPU cooler $5, and thermal compound and wiring so let's say $50 even and that's not including any shipping charges you may incur for 5320 lumens at 2700k or 5815 Lu at 5000k
What about some "Off the Shelf" Philips bulbs at package rated minimum lumens using a "8w" bulb at a rate "800lumens" as an example, They cost $8 for a pack of four, so to equal the same output we would need 6.65-7.27 bulbs respectfully, but let's just jump that up to 8 bulbs for good measure,
So $16 Total vs $50 for the same thing that's 70% cheaper, what if we include some keyless sockets or cleat sockets for the bulbs at $1.5-3 a piece, Were still looking at $28-40 so 44% to 20% cheaper and that's still not equating in shipping vs walking into your local store.
Say you went with a Cree CXB 3590 instead, that price difference is about the same.
So what if you don't want to run so many bulbs, well bump up to the "100w equivalent bulbs" that put out twice the light, unfortunately though they are about 50% more expensive than the "60w equivalent" but at $12 for 2 bulbs, you're still looking at $24 vs $50 or with sockets $30-36 vs $50 still without taking into consideration any shipping charges.
Kinda show we have a clear winner in the cost aspect.
Electricity Cost
But "What about potential increased electricity costs" well let's try to compare but i'll add back in later with actual measurements after the ballast, but i'll try my best
Let use a Meanwell LPC-60-1050 Ballast as example to power out a Citizen CLU048 1212 COB, which at typical consumption is 37.4watts, now because nothing is 100% efficient our ballasts will lose some energy in the conversion process, so let's consult the spec sheets, Meanwell lists that specific ballast at 87% efficiency under typical conditions, do the math and that should mean will be consuming right around 43w from the wall, where as "package rated" led bulbs will pull 64watts for the same amount of light.
Now my past measurements have those bulbs actually pulling less in the 6-7w range for the "8w bulbs", but let's go with the bigger number of those of 56w total draw till i take some readings and we can compare after the ballast like for like, or i hook a bunch of bulbs up in parallel and see their total draw.
So 43w vs 56w means a 13w difference, now let's say our energy cost is $0.20 a kw/h, mine is less than half that but let's use a greater figure as not everyone has cheap electricity. So per cob/bulb comparable unit at $0.20 a kw/h running at 18/6 for us auto guys and gals, that works out to be 234 watt/h more a day or 0.234 kw/h
The cost difference between going to a Citizen CLU048 1212 vs bulbs per my previous calculations, is $14-22 respectively with which route you go, so how many days will it take to break even with the additional electricity cost vs up front cost ?
$14 / (0.234 kw/h x $0.20 a kw/h) = 299 days
$22/ (0.234 kw/h x $0.20 a kw/h) = 470 days
And again that's not including shipping costs of your COB's, Ballasts, and Heatsinks, which is usually $10+ an order for me anyways, or if your under driving them and need more of all that. But if we factor that in look at tacking on an extra 213 days to break even.
Myself at the underside of $0.09 a kw/h im looking at 3-4 Years to pay off the difference.
So if you have cheaper electricity bulbs are the way to go, if it's more expensive DIY COB's can be advantageous so long as you use them for over 1-1.5 years or less depending on how high your kw/h rate is
Heat and Penetration
Now let's get into heat generation and penetration.
Heat is our byproduct of taking electricity and turning it into light, now if side by side the efficiency is comparable, our heat generation will also be comparable, and it's not much simpler that that.
As far as penetration, if you are supplying the same amount of light to a given enclosed area the plants will use that, now if we have a singular source of higher intensity light we will have to have our light further away from out plants in comparison so things don't get burned heat and or light burn related, but if we supply our plants and canopy from multiple angles with less intense sources but more of them, equaling the total light output as a singular source, our plants won't care how they receive it, sure they will grow differently in say denser flatter canopies vs single cola tee structures, but bud is bud.
But look at any commercial led panel and that's the same principle, just spread out a little further in our case.
Now even still some people on other forums are getting plenty of penetration with good dense bud up to 36"or 92cm from the lights "100w equivalent bulbs", how much more space do you really need with some minor training techniques, unless your wanting to grow big trees you could still side supplement light where you plant is getting equally lit vs only the top getting light and everything trying to race towards that light.
Only thing left to discuss for this starter would be that you can also if you want break down these Off the shelf led bulbs and make your own arrays from them as a cheap alternative, which i have for my 5 gal bucket build, as i wanted to go simple in that application.
And here's a link to an album on how to break down the philips bulbs, which are pretty much the same regardless of their color temp or wattage.
So for now i'll leave this here, and i will contribute my own grow using Off the Shelf bulbs, and will look at getting permission to use some photos of other peoples grows on other forums using Off the Shelf bulbs in their own grows.
So feel free to contribute, add, show off your own grows, and or discussions about this alternative which i believe and know is a perfectly acceptable alternative to other current ways.
Last edited:
......lol
