A Guide to LED lighting for Plants

This post is meant as a rough guide to building your own high powered LED lights. These lights are fairly versatile, not too complicated and provide a reasonably cheap option for a number of different plant growing situations.

This guide will focus on a light fixture that has three strings of LED’s: Blue, Red and a combo of White. They will be on/off with no dimming available – but I am planning to write up a guide to dimmable LED’s shortly (I’m waiting on new LED drivers to be delivered.).

Once built, we can use these lights by adding a main switch and turning on all the LEDs, make a switch for each string of LEDs or hook all or individual strings up to a set of relays connected to a microcontroller like an arduino. In the next lighting guide, we’ll get into the wiring and code to connect the lighting fixture to an ESP32 for PWM (Pulse Width Modulation) dimming.

This project will require you to work with mains voltage – which is the power that comes from your home wall outlets. In North America, that is 110-120 V. This can kill or seriously harm you.

Being careful, triple check your work, use a functional multimeter, and not drinking BL Limes while working will save you a lot of pain. Please be careful.

What you will need:

  • 1,3 or 5w LED mounted on starboard aluminum heat sinks
    • I purchased 3w LED’s in 4500k (warm white), 6500k (Cool White), Deep Red, and Blue. I purchased them from AliExpress. 20 of each came to about $25 CAD.
  • 3 Constant Current Drivers of various forward voltages.
  • 22 AWG Hook-up wire
  • Thermal Paste
  • Nylon screws, nuts and rubber washers
  • Aluminum angle stock
  • Hacksaw
  • 1/2″ bolts + nuts
  • Soldering iron, solder and associated tools
  • Drill with bits for metal work

Concepts to know:

  1. Your LED’s should have a datasheet available somewhere. For me, I found it on the AliExpress page where I purchased the LED’s.
    Check the datasheet for two important numbers:

    1. The current requirements. It should be something like 600mA or 1000mA. This is the maximum amount of current that can be handled by the LED. The higher the number the brighter the light.
    2. The forward voltage. This is often expressed as something like 3.4vdc.
  2. Drivers are specialized power units that ‘drive’ the LED’s. When selecting a driver we need to balance our voltage needs with the current needs. Our LED’s want a constant current, so that means that the driver will be constantly compensating the voltage in order to supply a consistent current across an unspecified number of LEDs.
    1. We should look for a Constant Current driver with a current rating appropriate for our LEDs. My LED’s are rated for 700mA, so I purchased a driver that was good for up to 600mA, but 700mA would have worked.
    2. Secondly, the driver will have a max voltage rating. Let’s say the max voltage is 60Vdc – well, we don’t want to run it at full power because that’s unsafe, so let’s say 55vdc. Now we can divide 55vdc by the forward voltage of the individual LED, which for the warm white, cool white and blue LED’s is 3.4vdc: 55/3.4 = 16.18
    3. So we can have 16 LED’s running on one driver.
    4. The Red LED’s have a lower forward voltage, their vdc is 2.4, so we can again divide 55 by 2.4 for a total of 22.91, so 22 LEDs. This is why you need to check the datasheet of the LED’s.
  3. Running LEDs in series vs parallel.
    1. We are going to be running our LED’s in a series. This means that we feed the positive lead to the positive side of the LED, the negative side of the LED then goes to the positive contact of the next LED. The last negative contact of the last LED in the string will connect to the negative lead of the driver. It should make one big loop starting and ending at the driver.
  4. Heat Diffusion:
    1. LED’s are very efficient, but they will lose energy through heat. The extent to which they generate heat depends on a few factors, but regardless – you should figure out a way to dissipate the heat because this is a main reason why LED’s burn out or operate at low efficiencies.
    2. This is why we are going to use aluminum angle stock – aluminum is a good conductor of heat. We’re also going to use thermal paste to make a highly conductive bridge between the LED and the Aluminum.
    3. You can also buy aluminum or copper heat sinks that have numerous skinny tines to dissipate the heat more effectively but I found these prohibitively expensive and now, having built the fixture, realize they were unnecessary anyway. More LED’s, a tight fitting enclosure or other factors may necessitate a proper heat sink.
      1. I am however, constantly on the look-out now for scrap aluminum or copper that could be used as a heat sink.

Plant Science:

  • In the past, aquarists used watts per gallon to figure out their lighting needs. As technology has progressed, this has become more and more outdated and ineffective. For one, lighting has become more efficient, so 1 watt of light from an incandescent bulb literally pales in comparison to a watt of light from an LED. For two, wattage only tells how much power is running through the light – not what spectrum of light you are getting. An older incandescent bulb will burn warm yellow, while some florescents will appear clinical blue/white. Not only will the amount of visible light differ between these technologies, the light spectrum being outputted will also differ.

  • Many people now focus on PAR, which is an acronym for Photosynthetic Available Radiation – which means, how much light is available for plants to use in photosynthesis. You can purchase a PAR sensor, but they are very expensive. We can also use Lumens as a unit – lumens should be included on the LED datasheet.
  • Plants use light to photosynthesize energy to grow. For most of history, plants grew outside and under the sun. Our sun gives off a massive amount of light…obviously. It is full spectrum white light – which means, it is comprised of all of the colours of the visible light spectrum, but to various amounts. Plants use different parts of the light spectrum for different things and use some parts more than others. This also changes depending on species.
    • For instance – plants tend to use a lot of the blue and red parts of the spectrum. They use very little green – this is why a lot of foliage is perceived as green. Because the leaves don’t absorb much of the green spectrum, it is bounced back and absorbed by the cones in our eyes.
    • This is also why high tech indoor growing facilities often use fusia or pink-ish lights (a combination of blue and red spectrums). Since the cost of energy is such a big factor in indoor growing, they prioritize the light that has the most impact on plant grow and leave out parts of the spectrum that add little benefit.
  • In our light fixture, we’re going to use red and blue LED’s to make sure these spectrums are strongly present, while using warm white and cool white to round out the spectrum and provide something that doesn’t look like a grade 12 acid trip disco lounge.


‘We are all broken, that’s how the light gets in.’ – Ernest Hemingway

New lights are in Titan Phase.

I’ve built a mostly working version of the networked high powered LED lights.

The Lights:

  • Three arrays of 3W LED’s.
  • There are an assortment of 25 or so white LED’s ranging from 4500-12000 kelvin
  • Eight Red at 660nm
  • Eight Blue at 445nm
  • Each array can be powered individually
  • Led’s mounted on aluminum one inch angle stock for heat dissipation.


  • ESP32 hooked up to 4 relay module
  • Programmed with micropython. Connects to a mqtt broker on a raspberry pi running Node-Red to schedule the lights (This might sound like overkill if you’re not home-automation adjacent like myself, but if you are at all interested in IoT stuff, running Node-Red on a Pi is a gamechanger and I wish I did this way earlier).

Obviously the lights are not yet ready for Phase SmokeShow, but they are a proof of concept, of sorts. I need to clean up the wiring and I need to replace the drivers. The drivers, though they seem to be working fine, are fairly cheap and I’d like to replace them with something more substantial. At that point, I’m hoping to have a better handle on dimming with constant current arrays and have the ESP32 handle that as well.

This is a photo of the setup with the old lights – primarily 5050smd led strips and a couple of NeoPixel 12” led’s off amazon. In hindsight, I see the folly of that setup now, but its been a cool process of learning not only about the mechanical/electrical aspect of light as well as the way that plants utilize light.

Here’s another shot of the new rig:

The comparison is a little unfair because of the way the camera on my phone calibrates automatically – but the difference is drastic.

I’ve noticed a few things in the two weeks or so I’ve been running this on a 12 hour light cycle.

  • There’s been an all round shift in plant growth. Not only in volume, but in character.
  • Almost all plants have more red in their stems. The Hygrophilia Stricta that I have growing above the surface in basically growing deep red/purple leaves. The stems of the Creeping Jenny have turned red, while their leaves have a very nice tender green colouring.
  • There is a small clustering ground cover plant that I unearthed from my garden. Outside it is often deep purple, but previously in the tank it had turned a healthy-ish green colour. It has now returned to its dark purple colouration.
  • My Morning Glories flowered almost immediately with the new lights. The leaves that are directly under one of the red LEDs is growing with these cool purple veins through the leaf structure.
  • The aquatic plants are reacting more slowly. I’m noticing some steady growth, but nothing spectacular. My thinking is that the plant growth is now most limited by my water quality/paramaters.  My water is fairly hard from the tap, and with the amount of volume the tank loses to evaporation… its causing problems, to say the least.
  • The Sagittaria Subulata seems to be enjoying the increased light.
  • There’s been an increase in Algae. Nothing dramatic, but I wouldn’t be surprised if I came home one day soon to some green water.

Aliexpress is amazing.

I just received these powerful little LED’s. They are 5 watts each and run at about 1500mA each. I’ve hooked one up just to see it in action and it was blinding.

LG 3535 LED
-Power: 5w
– Color: White
-LED brand: LG(from South Korea)
-Luminous flux: 533 lm (IF Curren:1500mA )
-Conditional current:350-1500mA
-Forward voltage: 2.8-3.6V
-View angle:115 degree
-Working temperature:-20–70celsius degree

My plan it to hook up 8 of these LEDs over the tank for about 3500-4000 Lumens.

These LED’s go for about 11$ each at my favourite local electronics shop – I love this store and will continue to shop there (I bought my arduino Mega 2560 there) – but I got 10 of these little starboard LED’s for $6.03 Canadian, which included shipping and they arrived in about a month. I also purchased an LED driver for them, which cost about 9 dollars all in.

PS: I took this photo with a super old tablet that I’m using in lieu of my phone (which I lost on a bus coming home from an aquarium store). I figure the Renoir painting kind of compliments the blurry photo of the LEDs.


Goddamn shit. Tank Update:

  1. Project DEAL BREAKER is going smoothly. It was a little harder than I thought it would be to implement, but that’s because I have switched some of the hardware around and some of the libraries didn’t talk well with one another. It is a reminder that I still have a lot to learn.
    At this point, I have almost everything working in concept. The relays are all working and (in theory can be) triggered by toggles as well as the DS3231 RTC at specific times. I have temps from three different temperature probes, the DHT22 temp/humidity sensor working well and I think I’m very close to have the Flow Rate sensor working. The code that I have for the Flow sensor compiles fine, but gives a funky number. I think its electrical noise, I need to add a resistor into the circuit, but I haven’t had time to do detailed electrical work recently… mostly just been rewatching the Sopranos and mulling over the code. When DEAL BREAKER is fully implemented, I will post the code and a more complete list of components to a post.
  2. After the main DEAL BREAKER stuff is done, the next step will be interfacing it with a ESP01 Wifi Module and start logging inputs on a MySQL database here – ie: PROJECT REAL FAKER.
  3. The above video is the first thing I ever made with an Arduino Uno and my introduction to this stuff. I had a rough idea of what Arduino was – a simple tiny computer that could prompt an output based on an input. If this, do this. So obviously its more complicated than that (or at least, it often is), but I ordered a kit off Amazon for $50 or so and its expanded exponentially since. Two years later and I basically need a barn where I can have an electronics lab, a grow/tank room and a painting/fabrication studio…
  4. Goddamn. I lost my phone. I went to a suburban fish store last week and bought some fish, a couple plants and another thing or two. I was writing an in-depth post for this blog while on the long bus ride back into Toronto proper, when I nearly missed my stop. I grabbed my fish friends and inadvertently left my phone on the seat beside me. In the large scheme of things in the universe – this is insignificant. In the tiny and self involved scheme of the universe- this is a giant pain in the ass. Not only do I not have a phone to appease my nervous fidgeting, I also can’t take photos of the tank as it rapidly progresses in this early phase. The dawning irony that my trip to fish store was about to double or triple in cost due to the replacement of my phone was also not welcome.
  5. I bought five more Panda Corydoras catfish for the tank. A giant and tragic flaw in my system: I recently found three Cory corpses in various stages of the sump weir. I thought that I had protected the opening enough, but it would appear that I had not. It was unfortunate, but I’ve temporarily over-covered the weir. The entire opening is stuff with polyfill. I think in the near future I may order a laser-cut acrylic insert that can be placed over the opening, allowing the full flow of water, but nothing big enough for a fish to get through.
  6. I also bought 8 Ember Tetras (Hyphessobrycon amandae) on the same trip. They seem to be doing well. They are colouring up and I’m looking forward to setting up a brine shrimp hatchery for them. Occasionally they will school together and its very cool – but I think there’s not quite enough of them to display shoaling behaviour all the time.
  7. I think once the tank chills out for a bit and I can get the water parameters under more control (more on that in another post), I would like to get a predatory fish – probably a pair of Dwarf Pike Cichlids so start regulating the Guppy population, and ideally, keep it at an absolute minimum. The guppies are already too numerous and I have a funny feeling that their random swimming behaviour disrupts the Ember Tetras shoaling. I think also just have a calmer tank with more substantial fish personalities would feel more resolved – hopefully, keep the water cleaner too.
  8. I also got a few more plants – a couple bunch of Sagittaria subulata and some Java Fern. Underwater garden is looking good, but they need some time to grow in. My Cryptocoryne Usteriana (I think) is doing pretty well. From what I have read, upon replanting they typically lose a lot of their leaves and send up new shoots and leaves once re-established. 2 of the 3 plants are sending up prominent new leaves, which is great. I would like to start Co2 injection soon, but apparently… this “hobby” is becoming “an obsession” and I have to “manage” my “time” and “money” more “responsibly“.
  9. Experiment update:  A little while ago I mentioned that I was trying to grow some aquatic plants in my hydroton sump tank emersed from the water. I have some Alternanthera reineckii, a rosebud of the same crypts I mentioned above as well as a year old, stunted and struggling Anubias in the sump. The anubias was a last ditch effort to see if the plant was worth keeping at all whatsoever.
    I am happy to report that all three plants are showing emersed growth. The Reineckii is growing the slowest, but you can see it has corrected its direction a couple of times and has tiny new leaves. The crypt is sending up a new leaf just like its buds upstairs and the Anubias is growing brand new, electric green leaves.All in all, the idea of have a sump with a “laborartory” in it is paying off. Any new sump design in the future will emphasize a space to work with plants (not to mention, grow food plants) within the system.