COB LED Lighting


Plants primarily need light which corresponds to the absorbance spectra of the major photosynthetic pigments - the basis for PPFD metrics.

Generating significant light in frequencies which do not fit the absorbance curves usually equates to wasted electricity.

See all the green light these “white” LEDs produce (520nm-560nm):

As far as I know, PAR meters generally measure the light according to a predefined curve - an approximation of the above curves for chlorophylls - and this predetermined curve may or may not fit your application. That is to say, you may want additional energy in certain frequencies not well represented, or less energy in others. For instance, light in the UVA-band seems to stimulate production of some secondary metabolites in many living things; some green can apparently be of use to some plants at some times…

I think highly of 7- or 9-band “full spectrum” LED grow fixtures, as they cover a wide area (not single point source), but tend to be pretty expensive.
If you’re feeling the DIY spirit, full-spectrum driverless grow COBs are hard to beat, thanks to low cost and simplicity, and usually offer a nice distribution of energy that generally fits the needs of plants well…

Could be faked data, but gives you an idea of what to expect, vs HPS:

I made this 50W 400nm-840nm “full-spectrum” grow COB light fixture for ~$20, all told:

Solder two wires, slap a CPU heatsink on it and mount!


Nice build! How were growing results with it?


Thanks! I actually haven’t grown anything with this driverless one just yet, but am about to put it to the test…


@Webb.Peter Great work on the MVP btw, lots of great work there! In terms of output vs upfront costs, those bright stiks look like a good solution. I’ve used GE Slimline t5 fixtures, and while more efficient, they’re considerably more expensive to purchase (almost twice the cost for a ~52 watt setup). The only benefit I see of the enclosed fixture is increased safety, but that’s off topic. I’d love to see a head-to-head comparison without reflectors using a lab grade spectroradiometer. I’m wondering if the bright stiks would deliver more or less usable power to the grow surface as their design may emit more indirect light.


a lot of informations, like the topic


So as a followup to my earlier comment, i got the RGBW COB led mentioned above. It’s pretty bright, though i have my doubts it is as powerful as it was stated.

The Ebay listing said rated for 700mA on each color channel. Found the the Red, Green, and Blue to use 400mA. And the White to use 200mA. Also the Red one required 20v DC while the rest required 30V. This is a bit annoying that they don’t all run the same. Put a large power resistor so i can run all of them on 30VDC.

The WS2811 chips i ordered from Poland did not work as expected. Maybe i’m just dumb. I couldn’t get them to work right in a simple test circuit with smaller LEDs. Was hoping to use the WS2811 with some transistors or relays to control the RGBW color mixing and basically make a giant NeoPixel. sigh

If using these COB LEDs to just grow anything i feel like getting an all-in-one white led to be better than a specific purple COB grow LED. But the reason i wanted to use a RGBW LED in the first place was for the added benefit of being able to independently control the blue-red light ratio. This would have come in VERY handy for daylength sensitive crops from mexico or south america where they are not adapted to flower in my temperate climate until the day-length gets into late fall with more red light. These adjustable lighting requirements may be required or even crucial to get exotic plants to grow or flower in an artificial environment.

Honestly though, from testing this cob led vs the small 5W led i got from Poland, i actually find the smaller 5W LED to pack quite a punch. I think someone in this thread said they still prefer making their own small LED arrays rather than the COB LEDs. Honestly i don’t think you could go wrong going that route. But i’m not counting these COB leds out just yet. Just need to do more experimenting.

Ebay does have some other powerful COB LEDS that say they are color controllable, and they may already have easy control circuits built in, though i think they are more expensive.


@BioLumo Way to take a swing at it! I sincerely appreciate the followup and hope your initial results don’t slow you down.

Are you driving the blue/green/white at 30V or 28V? You mention 30V but your post-it reads 28V. As I understand, under-driving the forward voltage across the array will reduce the power dissipated at the diode, resulting in less light intensity. A few volts can make a big difference. Take the Cree CXA3070 mentioned above. Looking at its data sheet (pg 2), we find a published maximum VF of 41V, a typical VF of 36.2V, and unlisted minimum VF. Skip to page 7 and check out the electrical characteristics graphing current against VF. Take note that driving the COB at 35V, just one volt below the published typical VF, results in over 20% less forward current drawn across the array junctions.

Secondly, what are you driving the COBs with? Perhaps you mentioned that somewhere, my apologies if I missed that. I use two types of drivers, adjustable constant current drivers to prototype and fixed constant current supplies+drivers in finished builds. Adjustable cc drivers like the meanwell ELN series allow you to adjust both voltage and current limits on the circuit using simple trim pots located on the driver board. I use the adjustment features to find realistic performance characteristics of an array and dial in VF/IF combinations. Some of these drivers (the ‘P’ or ‘D’ variants) also include a PWM driven dimming feature which might be useful to you. After I’ve settled on a load profile for the array, I move to a switched PS and fixed cc driver setup using Meanwell LDD drivers because I’ve found it cheaper to build and easier to manage several channels with different power requirements.

It looks like your COB is rated to 31V/33V on the green/blue/white, I say feed those puppies and see what happens :zap:!

In the pic below, the fixed setup populated with LDD drivers is on the left and the ELN adjustable supplies are on the right. BTW, I am not an EE guy… far from it. If you or anyone else sees misleading info in my post please let me know.


Thanks @Drew, i was testing with 28 at first, and then cranked it up to 30 later. 30 is what is in the photos. I think the green was still slightly brighter than the other colors so color mixing was still not perfect. I was just using this old power supply for testing. You can’t control current with it, but it does have a current meter, which is nice. I too know barely anything about electronics. I know enough to take on more than i can chew and to get myself into trouble.

On my blog i have an Open Source Mosfet speed controller i designed myself. It is meant for large motors like what one uses in Competition Robots. I used to participate in a FIRST Robotics team in High School and was the head electrical guy since i knew just a little bit more than anyone else. lol. The motor controller is now on github, but it still does not work quite right. If you or anyone has knowledge and interest to help me fix and finish it please help and check out my blog.


My home work :grin: DIY spectrometry for calculating PAR of cheap LEDs?

And these is spectrum of my 2nd LED lamps

consists of 5 LEDs 462nm + W5000(445nm) + 635nm + 660nm + 66% of 740nm. 34% of current is flowing throu 9ohm resistor.

These days I’m starting of lettuce and Komatcuna comparison grow with these 2 lamps.

3th days of the experiment, full spectrum lamp at right. The days before were under the weaker blue light.

HP Aeroponics exact spray without tank

Few years ago I bought extremely cheap red and blue 10W COB with bad lumen. Now I have DIY spectrometer and saw, that these COB is horrible, some is 10 times less efficient than 3W star base.


nice. Good to keep in mind that what you buy is not always what you get. Interesting but not surprising that your 3W star leds were better.

I found this today which looks interesting, but expensive.

I particularly like the ability to switch from extreme purple to extreme magenta. But it would be nice if it also could switch to White light for human eyes when a human is in the room. Still, pretty cool.


You add a white LED of some 30% of the power of the existing grow light and the light will be less disturbing. Maybe on a timer.


@wsnook @mason @catbarpH @BioLumo

Interesting that this review doesn’t really cover any white COBs. That being said, it’s extremely detailed and certainly worth taking a look if you’re looking into buying grow lights.


Has anyone had success finding 12V COBs that are able to provide 250 umol at 18"?