The Creator of Light and Programmer of LEDs.
Showing off a prototype LED grow light.
Problem
Recently I started to grow a plant of a Japanese origin called Shiso. A plant which thrives in the scorching heat of the Japanese climate. But experiences a shock when exposed to UK’s short, gray, sunless winter days. Despite having placed it in a big pot next to a huge window, the plant started to wither.
Knowing it had plenty of water, and my house was suitably warm, I knew the missing ingredient was likely more sun. Especially in the long dark night when the sun set early.
Component quality and plant needs
Using components I had on laying around, I decided to make an LED grow light. Having done some basic research, I quickly discovered the LEDs I had are falling short from what professional grow lights can supply. The key issue is the wavelength of a given light colour that the light emits, and the range which the plant is able to absorb it. My lights are individually addressable WS2812B LEDs. This is designed for decorative lighting.
| WS2812B LED Color | Peak Wavelength (nm) | Plant Absorption Effectiveness | Notes |
|---|---|---|---|
| Red | ~620–625 nm | Medium | Useful for photosynthesis but slightly below the deep-red (~660 nm) peak most efficient for plants. |
| Green | ~522–525 nm | Low | Plants hardly absorb green. |
| Blue | ~465–467 nm | High | Very close to chlorophyll’s strong blue absorption band (~430–470 nm). |
What some papers also suggest is that different light colours are more beneficial during different grow cycles. Blue being more beneficial as leaves first develop and plant grows. Red being more beneficial at the fruiting stages. I am no expert here however. The details likely change for different plants.
Looking at the specs a few things became obvious:
- Switching off the green spectrum will save me 1/3 of the energy, while compromising little on the effectiveness.
- WS2812B are individually addressable. This is overkill. Adjusting the value of a whole led strip is what we are after regardless.
- For optimal growth I should try to set red and blue LED to maximum values. Giving our grow light a distinct pink colour.
The prototype
The build is simple.
The shell is cardboard.
- I wanted the light to operate at night without illuminating the house.
- The inside of the light is lined with tinfoil. This reflects as much light back into the plant as possible.
The lights are programmed with Arduino Elegoo Uno R3
- I set my lights to either full blue, or full red & blue configuration. Super simple stuff.
#include <FastLED.h>
#define NUM_LEDS 216
#define DATA_PIN 7
CRGB leds[NUM_LEDS];
void setup() {
delay(2000); // Allow power to stabilize
FastLED.addLeds<WS2812B, DATA_PIN, GRB>(leds, NUM_LEDS);
FastLED.clear();
// Set all LEDs to purple (red + blue)
fill_solid(leds, NUM_LEDS, CRGB(255, 0, 255));
FastLED.show(); // Display the color
}
void loop() {
// Nothing to do — LEDs stay purple
}
Improvements and lessons learned
This is obviously a prototype.
First, for anyone trying to replicate this, I recommend you avoid using tinfoil and cardboard with anything electrical. Do as I say, not as I do haha. The original prototype is realistically a bit of a tinder box. Having soldered the led connections and cables it is instantly apparent that the tinfoil can short the circuit due to it being conductive. So a more robust light will isolate these components.
Second, it would be really interesting to use more lights that are purpose built for plant growth. Saying that, as a small-scale experiment, the decorative leds worked just fine.
The results so far
I have been somewhat surprised by the effects of the light. While no miracles occured, the plant went from being on the verge of death, to steadily growing back up. It looks so much better than it did before. I wish I had some before/after pictures, but I thought of this too late. Here are some more general images for reference.
An outside view of light plugged into a power bank.
An inside view.
