Build Your Own LED Grow Light


Terms you should know, and some you should know to disregard:



Another term that needs to make the list is 'Efficacy,' which is the amount of light put out per watt input to the COB. This makes the term watts relevant as well. The light you get out of the COB is then watts * Efficacy so a Citizen CLU048-1212C4-303M2M2-F1 puts out 5335 lm at 35W, while driving them at 50W produces less than 7650 lm. The output needs to be de-rated for the higher power.

If you buy a kit from Kingbrite, don't get the 1212 COB, they're cheap but hardly much more efficient than HID @ 50w.

KingBrite is currently selling the Series 6 Citizens, not the Series 5 shown in the graph.

FYI, I was putting together the data before I found this thread...

3000K
CLU028-1201C4-30*M2M2-F1 153lm/w CRI 80 2.8W 34.0V 90mA
CLU028-1201C4-30*H5M3-F1 131lm/w CRI 90 2.8W 34.0V 90mA
CLU048-1212C4-303M2M2-F1 153lm/w CRI 80 37W 34.0V 1080mA
CLU048-1212C4-30*H5M3-F1 131lm/w CRI 90 37W 34.0V 1080mA
CLU048-1216C4-30*M2M2-F1 149lm/w CRI 80 49W 34.0V 1440mA
CLU048-1216C4-30*H5M3-F1 128lm/w CRI 90 49W 34.0V 1440mA
CLU048-1812C4-30*M2M2-F1 149lm/w CRI 80 55W 51.1V 1080mA
CLU048-1812C4-30*H5M3-F1 127lm/w CRI 90 55W 51.1V 1080mA

3500K:
CLU028-1201C4-35*M2M2-F1 156lm/w CRI 80 2.8W 34.0V 90mA
CLU028-1201C4-35*H5M3-F1 133lm/w CRI 90 2.8W 34.0V 90mA
CLU048-1212C4-35*M2M2-F1 155lm/w CRI 80 37W 34.0V 1080mA
CLU048-1212C4-35*H5M3-F1 133lm/w CRI 90 37W 34.0V 1080mA
CLU048-1216C4-35*M2M2-F1 151lm/w CRI 80 49W 34.0V 1440mA
CLU048-1216C4-35*H5M3-F1 129lm/w CRI 90 49W 34.0V 1440mA
CLU048-1812C4-35*M2M2-F1 151lm/w CRI 80 55W 51.1V 1080mA
CLU048-1812C4-35*H5M3-F1 129lm/w CRI 90 55W 51.1V 1080mA


Running COB's hard makes no sense, and even 1,4 amps is too high for my liking, 0,7 or 1,05 will be more efficient, but of course require more COB's.

Running COBs hard makes no sense - true. The current cannot be used for COBs in general to state that it's running hard. See the table above. The second pair of numbers in the second group for Citizen COBs is an indication of the number of LEDs in parallel on the COB. The higher the number the more current is required. A 1201C4 runs at only 90mA, while the 1216C4 needs sixteen times the current at 1440mA.

My current design is two banks of four Citizen COBs, each bank driven by a Mean Well driver. This is meant to provide coverage for a 2' X 4' area. Multiples of this setup would be used for larger spaces. I hadn't considered the Samsung strip lighting, as I was thinking of concentrating the various wavelengths into point sources, and distributing these over the canopy at one foot intervals. I'll look into the light strips, as it will be a while before I'll get my tax refund :winkyface:
 
Thanks for that Old Salt, I wish I could edit the first post, it was a lot to put down in one go and a lot more could be added to get a bigger picture and to make it easier to understand ;)

It's easier to talk watt when determining COB efficiency for sure, I think 25w - 37.5w is the sweetspot, but the 50w many put in is certainly a serious upgrade from HID and blurple ;)


Ordered some strips from Digikey this weekend, just got a text, they should be here(northern europe) tomorrow :slide:
 
I'm of the opinion that running the Vero 29's at 1400 watts is awesome my first grow with them has been stellar you can see the results in my signature the only thing I would do different is replace the 4000K with 2700K cobs that would give me 8 3500K cobs and 4 2700K cobs in my array. You could run the Vero's up to 2.1 amps and gain even more output. Would you gain more yield? Not sure I need to check Buckshot 317's journal to see if he ever did a comparison.

TO Each His Own what woks for me doesn't work for others Either way you do it COB's ROCK!
 
Thanks for that Old Salt, I wish I could edit the first post, it was a lot to put down in one go and a lot more could be added to get a bigger picture and to make it easier to understand ;)

It's easier to talk watt when determining COB efficiency for sure, I think 25w - 37.5w is the sweetspot, but the 50w many put in is certainly a serious upgrade from HID and blurple ;)


Ordered some strips from Digikey this weekend, just got a text, they should be here(northern europe) tomorrow :slide:
U can have it edited bud

Report the 1st post with what u want changed, added or a completely new post

Teddy will hook ya up
 
I suggest requesting an edit once all is said and done. Just keep notes until that time, and have all the additions and corrections done at once. The moderators will probably thank you for not asking too often.
 
I'm of the opinion that running the Vero 29's at 1400 watts is awesome

That must be one helluva COB :laughtwo:



U can have it edited bud

Report the 1st post with what u want changed, added or a completely new post

Teddy will hook ya up

Thanks, that's good to know ;)


I suggest requesting an edit once all is said and done. Just keep notes until that time, and have all the additions and corrections done at once. The moderators will probably thank you for not asking too often.

Yeah, Ill try to piece it together in the coming months, any input is appreciated ;)
 
Here it is

IMG_313617.JPG
 
People keep touting one COB or lightstrip over others, so I prepared comparison tables for the color temp and CRIs of interest to me. The tables will help in selecting what I'll use for my lights, and the drivers required.

ManufacturerPart No.EfficiacyCurrentVoltageWattslm @ 50WCost
BridgeLux Gen6BXRC-30E10K0-L-2x137lm/W2100mA38V79.86850Obsolete
BridgeLux Gen7BXRC-30E10K0-D-7X146lm/W2100mA36.6V76.87300$31.89
CREE CXB3590CXB3590-0000-000R0HCD30H179lm/W2050mA36V73.88950$58.15
CitizenCLU048-1216C4-30*M2M2-F1149lm/w1440mA34V497450$13.65
SamsungLT-FB22B168lm/W1120mA46V51.58400$26.51
Luminus XNovaCXM-22-30-80-36-AC00-F2-3133lm/W1280mA35V456667$15.01
Table 1: Comparison of 3000K 80CRI LED Sources


ManufacturerPart No.EfficiacyCurrentVoltageWattslm @ 50WCost
BridgeLux Gen6BXRC-30G10K0-L-2x114lm/W2100mA38V79.85700Obsolete
BridgeLux Gen7BXRC-30G10K0-D-7X123lm/W2100mA36.6V76.86150$31.89
CREE CXB3590CXB3590-0000-000R0UBD30H149lm/W2050mA36V73.87450$48.60
CitizenCLU048-1216C4-30*H5M3-F1128lm/W1440mA34V496400$13.65
SamsungLT-M562H (3 req for 50W)117lm/W700mA2416.85850Obsolete
Luminus XNovaCXM-22-30-90-36-AC00-F2-3113lm/W1280mA35V455667$15.01
Table 2: Comparison of 3000K 90CRI LED Sources

OK! I heard that! Yes I know, Lumens and Lux for people, PAR and PPFD for plants. So then, why did I include Lumen data in the tables? It's simple really. Manufacturers rate their lamps that way, and there is a method of determining PAR and PPFD from their spectrums which are also published by manufacturers. It's a long process involving picking relative intensities from images and plugging the values into an equation. That's next on the agenda for a couple of the lamps.

It seems reasonable that PAR and PPFD will track with Lumens and Lux. If you can accept that, then we can use the data above to select lamps for further investigation. It's readily apparent that the CREE CXB3590s outperform all contenders in efficiacy measurements. Perhaps a better measurement would be power required for a standard of illumination. At 5000lm this is 27.9watts for the CREE CRI 80, and 33.5 watts for the Citizen. This translates to an extra 5.6 watts of heat per COB that may have to be removed from the grow space if the Citizen were chosen.

Next we might want to compare the cost of ownership for the COBs, including initial cost and electrical cost over their lifetimes.

CO = Initial cost + (life expectancy in hours * power used / 1000) * cost/kwh
CO = $13.65 + (50,000 * 33.5 / 1000) * 0.15 = $264.90 (Citizen at 15¢/kwh)
CO = $58.15 + (50,000 * 27.9 / 1000) * 0.15 = $267.40 (CREE at 15¢/kwh)

As you can see, at 15¢/kwh they pretty much cost the same. Just for giggles, at 25¢/kwh let's see what happens:
CO = Initial cost + (life expectancy in hours * power used / 1000) * cost/kwh
CO = $13.65 + (50,000 * 33.5 / 1000) * 0.25 = $432.4 (Citizen at 25¢/kwh)
CO = $58.15 + (50,000 * 27.9 / 1000) * 0.25 = $406.90 (CREE at 25¢/kwh)

Surprisingly there is not much difference. It would take two to three years before the CREE pulls ahead of the Citizen at the higher power cost.

There's more I need to check. First there's those PAR and PPFD conversions. Then we have the drivers, heatsinks, mounts, and other parts needed to make a useful grow light out of the COBs.

Hold your horses! Relax, and have a little of your product. I'll get there, hopefully sooner rather than later.
 
Super helpful comparison SALT, I dont get all of the technical information, but I will. Thank you for putting it somewhere that is easily accessible for me and for everybody else wondering about a new-build.
 
Here's the instructions I found for computing PPFD if you have the spectral power distribution of a light source:

From AGi32 WebHelp

The mathematical basis for the calculation of PPFD:

If the spectral power distribution (SPD) of a light source is known across the relevant wavelengths (400-700 nm), then the amount of photosynthetic energy available to plants can be determined. Based on its SPD, a light source will have a conversion factor that can be used to translate luminous flux density (illuminance) received by the plant into photosynthetic photon flux density (PPFD), in μmol/s-m2.

One watt of radiant power at 555 nm is by definition equal to 683 lumens. Given the CIE 1931 luminous efficiency function V(λ), we can calculate the spectral radiant flux Φ(λ) for plants in watts per nanometer for each lumen as:

Φ(λ)/lumen = [Wrel(λ)] / [683 * Σ(400-700) [V(λ) Wrel(λ) Δλ]]

where Wrel(λ) is the relative spectral power distribution and V(λ) is the luminous efficiency function at wavelength λ.

With this, the photosynthetic photon flux (PPF) per nanometer in micromoles per second per nanometer can be calculated:

PPF /nm = (10-3) * [λ Φ(λ)] / (Nahc),

where:

Na = Avogadro's constant, 6.022 x1023

h = Planck's constant (6.626 x 10-34 joule-seconds)

c = speed of light, 2.998 x 108 m/s

λ = wavelength in meters.

Summing over the range of 400-700 nm yields the photosynthetic photon flux (PPF) per lumen for the given light source:

PPF » 8.359 * 10-3 * Σ(400-700) [λ Φ(λ) Δλ]

Given an illuminance value (lux or footcandles), we can similarly calculate the photosynthetic photon flux density (PPFD) in micromoles per second per square meter (μmol/s-m2) for the given light source.

SPD graphs are relatively easy to come by, but finding the same information in tabular form, needed for the above equations, is more difficult. One source is CIE 15:4, Colorimetry (2004).

Since we can get the spectral power distribution of the COBs and light strips, all we need to do is pick off the data, decipher the above, and perform the calculations. Easier said than done.

I'm still plugging away at it. But first a little :yikes::rollit:
 
Hi Guys,
I have not been keeping up with grow LED lighting last grow was in 2014.
I've dusted things off now not much to do but wait, so while my mind is in this space I thought I'd look at what is going on at 420.
Why is everyone so keen on white LEDS, looks like people here are hell bent on the most efficient white LED as the solution. What have I missed?
 
...Why is everyone so keen on white LEDS, looks like people here are hell bent on the most efficient white LED as the solution. What have I missed?

Modern LEDs have higher efficiacy (lumens / watt) than HPS or MH incandescent lamps, so you get more light for your power buck.

PAR is fine and dandy, as far as it goes but it's not the be all end all light measurement for plants. Yes it captures the wavelengths of photosyntheticly active radiation, but take a look at the LED grow light offerings available commercially. You'll soon note that they all add LEDs outside of those wavelengths in the UV, deep blue, far red, and infared areas of the spectrum. I've done a lot of reading lately and discovered that while green and yellow light while not contributing to photosynthesis much, do increase the rate of photosynthesis. Mix it all up, and you have a white light. The color will vary, and it has been found that 2700K - 3500K are good for flowering, while 3500K - 5000K work well in the vegetative stage. Best of all, you don't have that bluish purple light that advertises to the world what you are doing when it escapes.

Many thanks for asking. I ran into a wall trying to figure out how to convert between lumens and PAR. Trying to find one of my references back led me to this: Photosynthetically active radiation (PAR), Lumen, and Radiant Flux conversion Calculator. Hopefully it will kick me out of the rut I'm in.
 
Thanks old Salt.
I have done a lot of reading at least 6 years back now and have been growing with leds for many years(when I do). The early years most people were focused on the 4 chlorophyll a and b peaks, as per NASA experiments(actually only two wavelengths for NASA). Then manufactures of grow lights widened the spectrum a bit then there was much discussion on other wavelengths required for optimum plant growth, both accessory pigments and plant signaling. Classic example of signaling been the use of IR. But the research at the time was based on other plant species so the action spectrum peaks quoted can only be used as approximations.
I figured that with cannabis now emerging as a legit commercial reality there may be specific science on PAR as it relates to cannabis.

Anyway I have always been in the DIY space for LED's as that's half the fun for me with growing.
Currently flowering with 2X100W 7band COB chips (with additional 660nm on the side) that are many years old now, (not they have seen much use).
I'm a bit motivated at the moment (why I poking around the DIY side of this site) and am thinking a grow journal might help me and others on thinking about different approaches to cold weather veg cabinet. I have some new bits and pieces that I have not seen anyone try yet. Surprised that exclusive white grows are so popular.

My only journal to date.
 
Back
Top Bottom