Great convo going here, thanks all for the info and resources. I’d like to share my experience first, then explain my take on applying semiconductor light sources to horticulture using market ready technology. I’ve built several types of LED and hybrid fluorescent+LED fixtures and it appears we all agree spectrum and photon concentration (delivered to the canopy) are key variables.
COB arrays with single/mixed spectrum emitters are manufactured in many flavors. In my experience, the packages sold form Chinese distributors (ebay, AliExpress) work but they’re prone to manufacturing and design defects that reduce efficiency and lifespan. COBs manufactured by known manufacturers and sold by reputable distributors cost more, especially driver-less versions. Mixed spectrum COBs usually do not allow for independent channel control so you get a set spectrum output with the only control being overall intensity of that set spectrum.
Great cheap COB comparison
Mixed emitter COBs with independent channel control (mouser)
Like most commercially available lighting fixtures, COBs pack as much output as they can into a small package to reduce manufacturing costs. This makes sense from the manufacturing side but single source fixtures result in a greater concentration of heat and a greater dependency on a single point of failure in a single package setup.
Although I see great potential in COBs, I’m still rocking 3W/5W combos in my arrays. The two main reasons are:
Due to the inverse square property of light intensity and the potential benefits of oblique spectral radiation (side lighting), I like my light spread out and close to the canopy. COBs mean concentrating radiant energy from fewer sources which are on average farther from the average vegetation. Implementing light in a more modular way can also increase the reliability and fault tolerance of the array - you avoid a single point of failure.
Because I’m isolating the color spectrum through independent emitters, I get to control the spectrum completely by swapping fixture plates with a different mix of emitters. In theory this has limited adverse effects on total output or efficiency because I’m not forcibly dimming emitters that remain in the array. When you invest in a fixture that concentrates high power light sources, you’re stuck with that spectrum and in a concentration that imo is difficult to optimize for plants. Compounding this potential issue, full spectrum emitters (like the Cree above) emit good light energy in an arguably less useful spectrum (green/yellow).
I’ll continue to integrate COBs into my arrays as the price for low-watt packages (10-30W) decrease and the quality increases. As a tool for in-depth experimentation I see COBs as playing a supporting role to other spectrum specific components.