Inverse Square Law

We have all kinds of threads on spectrum, cooling and bins, but seems no one has been talking about one of the biggest issues in lighting. The Inverse Square Law(ISL). All radiant source output follows the ISL. including the Sun. Except that from the Sun, the top of your plant isn't very much closer than the bottom of your plant. So outdoors, height of the plant is unimportant. Indoors, that is very much not the case. No matter the lamp you chose, it also must obey the ISL

Quick tangent, thinking about choosing a lamp, I want to talk about intensity. If we define intensity as output vs surface area, we can make several observations about popular lamp options. Floro tubes are not terribly intense because their output is dispersed over a large surface area. CFLs have similar intensity issues as they simply trade linear surface area for spiraled surface area. HID lamps will all have similar surface area within class and so intensity will increase with power. At 1k the small actual surface area of the cylindrical active part of the lamp means significant system intensity. However, looking at COBS, we have a fraction of the output of the large HID, but the output surface is only a circle of 22.5mm (3070), which means that the high lumen output is coming from not only a very small spot, but is also all being output towards the canopy.

Now back to our regularly scheduled rant....

So as the previous generation lamp tech required large hoods and ventilation systems, and gave off massive amounts of heat, lamps were installed high up in the room. This mounting position also allows the point source light to illuminate a large area, up to 5'x5' with decent performance to the edges. But consider what this high above the canopy mounting position means for intensity at plant level.

LED lamps came along and the only way to get anything useful out of them was to line up hundreds of them. But small lamps with little intensity, even hundreds of them, just can't get it done. Especially when mounted high above in a small square, claiming much larger footprints than realistic.

COB however are game changers. The low heat profile means they can be close to canopy without burning. While not as intense individually as as 1k HID, the COB chips can be placed a fraction of the distance from the canopy and as such will deliver a far greater portion of its intensity to the plants than the HID.

If we could define optimum intensity at the canopy level and then map the output of various chips at various intensities, we could work out the required density of each model of chip at the desired drive levels and height from canopy to maximize efficiency of photon delivery.

When I was testing the heatsinks I noticed that even using my hand as a reflector increased the light at the sensor more than I would have expected. Even when I tilted my hand at angles that did not make sense, the readings increased significantly. So the next thing I want to examine more closely is reflectors for COBs and at what distance do we need them/not need them. Sounds like your thinking is heading the same direction?

I run mine at 800mA (28W) and a distance to canopy of about 8". But for those running at 1.4A (52W) with no reflectors, I am wondering how much they could gain with very simple DIY reflectors for starters. I added one to a CXA3590 that is on vegging duty (270mA 18W). Because I do not need a high intensity, I increased the vertical height to 12-15" and the DIY reflector seems to be bouncing quite a lot of light because it is glowing so bright I cannot look at it from the side. None of that light would have made it to the canopy without the reflector, just a piece of a soda bottle painted with titanium dioxide white paint.

If we know the optimum photon density at the canopy and can measure the output of the chips, we should be able to determine the optimum distance from the canopy for a given chip at a given power level. Adding a reflector to the system would change the output pattern and likely raise the chip level a bit, but probably not all that much. Altho, would love to be proven wrong with some real numbers. Reflectors are cheap and easy to add.

It should be simple enough to catalog the optimum placement characteristics for all the common chips that could be used to design custom systems at maximum efficiency. I think we want to get these things way down low and spread out quite a bit.