25% more yield with a diffused LED ?

[Cerathule]

In theory an approximation of the diminishing over distance could be formulated if the beam angle of the diode is known. The light disperses in the shape of a cone. Both volume of sphere and cone can be calculated, and a co-factor derived by dividing the sphere's volume by the cone's.
For 90° beam angle COBs that should be about 6.
But ofc, the hardware doesn't emit the light so homogenously, and the cone isn't round-shaped at its bottom, which would be the correct geometric form to use.
Still usefull to give an idea how different LEDs deliver light than HIDs in an open reflector.

 

[goingrey]

In theory an approximation of the diminishing over distance could be formulated if the beam angle of the diode is known. The light disperses in the shape of a cone. Both volume of sphere and cone can be calculated, and a co-factor derived by dividing the sphere's volume by the cone's.
For 90° beam angle COBs that should be about 6.
But ofc, the hardware doesn't emit the light so homogenously, and the cone isn't round-shaped at its bottom, which would be the correct geometric form to use.
Still usefull to give an idea how different LEDs deliver light than HIDs in an open reflector.

Samsung LM301H beam angle is 120 degrees. But for the reds Osram, Epistar etc many angles are available. Since there are less of them it would make sense to use the widest 150-180 degree beam diodes, but from the pattern photos seems like the narrow 50-60 degree diodes might be used. Anyway adds to the complexity a little bit.

 

[Cerathule]

How would a potplant grow and flower out if placed in an experimental glass chamber that has equal light coming from all 6 sides. Maybe with the light from above being a chunk more strong? Thousands and thousands of low power diodes everywhere, savely separated by glass to prevent any locale leaf burns. Is there a point when the ambient diffuse light is at a "saturation point" and any increase won't matter?

 

[Cerathule]

Samsung LM301H beam angle is 120 degrees. But for the reds Osram, Epistar etc many angles are available. Since there are less of them it would make sense to use the widest 150-180 degree beam diodes, but from the pattern photos seems like the narrow 50-60 degree diodes might be used. Anyway adds to the complexity a little bit.

Some fixture mabufacturer use reflectors and lenses/collimators to influence this spread. There's pros and cons to do so but what is proven to be optimal?

 

[exploziv]

I think not covering the leds as to let them cool better and also to not lose valuable light won in the end.

 

[GoatCheese]

Interesting that the Gavita 1650e Pro LED (older model) came with diffusers over the diodes but Gavita removed them from the latest 1700e LED Pro's............ there was a reason but I can't seem to find it now.

Maybe they trapped too much heat?

640w from the wall will generate alot of radiant heat out of Samsungs/Osrams.

Yea, the problem with diffuser covers is that they will trap alot of heat to warm up the led chip, which prolly will shorten it’s life span - and so their efficiency will start going down faster too, i’d imagine.

..at least my Samsungs put out good amount of heat radiation when really pushed, so i can see air tight covers could be a problem in high-power led lights.



I have a fan blowing on my Cree cobs all the time cause they have lenses on them. I can only use about 30 watts out of each the three cobs cause my tent isn’t very tall. Each can do 75 watts max so i’m running them well under 50% of they maximum so the chips wont warm up so much. ..Maybe i’d run them with out the lenses if i’d run them full power cause Cree CBX3590 will put out a lot of heat in the 75-100 watt range.

So it would be better if the diffuser set up would allow air passing thru to keep the chips cooler.

..My 35mm wide sheet strips won’t trap so much heat. Haven’t got any on my veg lights yet – too much shoulder and arm problems atm,so i can’t be bothered just now - but i have been thinking of hanging them around 10-13mm away from the led chip

:

The curvature of the diffuser cover/lens will affect the viewing angle of the led – if it has too much curvature on it will make the beam very narrow = see the curved cover on the right.

 

[sshz]

The thing with Gavita is, they state even though their unit have over 3300 diodes, they don’t need to push them as hard as others do because there’s so many of them. So they actually run cooler and last longer than competitors.

Im still looking for the thread as to why they dropped their diffusers, it’s out there somewhere.

 

[Ca++]

In theory an approximation of the diminishing over distance could be formulated if the beam angle of the diode is known. The light disperses in the shape of a cone. Both volume of sphere and cone can be calculated, and a co-factor derived by dividing the sphere's volume by the cone's.
For 90° beam angle COBs that should be about 6.
But ofc, the hardware doesn't emit the light so homogenously, and the cone isn't round-shaped at its bottom, which would be the correct geometric form to use.
Still usefull to give an idea how different LEDs deliver light than HIDs in an open reflector.

I'm not sure the beam angle matters.
Say you are the infinately small in size, emitter of the light. Look out the window and imagine an infinately thin line from you, to the next town in that direction. The law gives us deteriation along that infinately thin line. They try and explain it with pictures of cones, but we are measuring along an axis so thin it has no dimension. Your light from this infinately small emitter, spreads evenly in all directions. Meaning no two beams of light are truly parallel. With these fixed facts, we can measure levels of light anywhere along the line, and calculate the light at any other point along that line. Presuming you are the only light source.

Beam angle isn't really involved, except for it coming from a tiny spot and heading out equally. We are only measuring along one line, so the others are not important for this rule.


I must say, I have not studied this. I have never seen it as applicable, nor risen willingly to this level of maths. Perhaps the formula gives the space now occupied by an amount of matter, but I believe it's how much matter is now accupying a space. The question being one of density change, not space.


My covers (which are not diffuse yet) hang below my lights. Well.. they use stand-offs rather than just hang. It's easy enough to hang some flat sheeting from most lights. Leaving a useful air gap that you can get your fingers in to clean the sheeting sometimes.


Edit: An LED with a 120 beam angle, still puts out most of it's light in about 60. It's not an even cone of light. It's just like your car headlights. With most down the middle. We had an image of two beams beside each other. One had a beam cross-section that looked like a ball. Showing the light in the middle reached out further.

Fetched it..
That's perhaps a 160 degree beam, but most falls within 80 degrees. The middle 40 really getting the lions share

 

[Cerathule]

I'm not sure the beam angle matters.
Say you are the infinately small in size, emitter of the light. Look out the window and imagine an infinately thin line from you, to the next town in that direction. The law gives us deteriation along that infinately thin line. They try and explain it with pictures of cones, but we are measuring along an axis so thin it has no dimension. Your light from this infinately small emitter, spreads evenly in all directions. Meaning no two beams of light are truly parallel. With these fixed facts, we can measure levels of light anywhere along the line, and calculate the light at any other point along that line. Presuming you are the only light source.

Beam angle isn't really involved, except for it coming from a tiny spot and heading out equally. We are only measuring along one line, so the others are not important for this rule.


I must say, I have not studied this. I have never seen it as applicable, nor risen willingly to this level of maths. Perhaps the formula gives the space now occupied by an amount of matter, but I believe it's how much matter is now accupying a space. The question being one of density change, not space.


My covers (which are not diffuse yet) hang below my lights. Well.. they use stand-offs rather than just hang. It's easy enough to hang some flat sheeting from most lights. Leaving a useful air gap that you can get your fingers in to clean the sheeting sometimes.


Edit: An LED with a 120 beam angle, still puts out most of it's light in about 60. It's not an even cone of light. It's just like your car headlights. With most down the middle. We had an image of two beams beside each other. One had a beam cross-section that looked like a ball. Showing the light in the middle reached out further.

Fetched it..
That's perhaps a 160 degree beam, but most falls within 80 degrees. The middle 40 really getting the lions share

This is why I wrote "approximation", it's not precise but a more precise/realistic result would require a way more complex formulae.

You can divide the cone's bottom area ("the footprint" - either the tentfloor or canopy of leaf foliage) into a grid, then take irradiance measurements of all inner, middle & outer points. Then add the numbers, divide to arrive at an average, which should be approximately true.

In this scenario, even if we put a collimator diffusion lense over the emitter so have an equalized light spread the inner grid points are still going to read higher as they are nearer (less diagonal) to the LED.

Other reasons why there is a numbers mismatch is reflective loss, diode binning, measuring instruments imprecision, amongst others.

The LED generates a specific number of photons each second, and the beam angle dictates how strong these spread out as these flee from their source. The lower the angle, the more dense they will be at a specific value in comparison to 360° which has the biggest dispersion, thus, the "1" within the inverse-square laws formulation. To use this formular is only going to arrive at true results if compared to measurements of a point-source spherical light generator/spreader. For other beam angles, the "1" needs to change to a higher number to reflect the fact a lower area is used.

I believe there're multiple ways to calculate the average approx irradiance by extending the 360° inverse-square law via either a 2 or 3-dimensional approach:

Spherical sector - Wikipedia

en.wikipedia.org

Circular sector - Wikipedia

en.wikipedia.org

 

[Ca++]

Beam angle isn't in the law. The 1 is a distance. The D is then multiples of that distance. From which we can calculate spread

 

[Cerathule]

The 1 is a distance.

No the "r" is the distance, the "1/" is part of the inverse-square.

 

[Cerathule]

Here is a recent test done by Migro on growlights, growtents and the inverse-square-law

Inverse square law is WRONG for grow lighting : MIGRO : Free Download, Borrow, and Streaming : Internet Archive

The inverse square law does not apply to grow lights. For the inverse square law to work the radiation source must be a point or sphere and emit radiate evenly...

archive.org

1/20cm^2=0,0025
1/40cm^2=0,000625
0,0025/0,000625=4
>>double the distance, reduces fluence by the power of 4:

a relationship like between "r" & "2r"

But that's not what Migro's measurement did confirm. The drop in intensity is less. This must be due to light reflection that otherwise would be lost but is reflected and falls into the sensor.

And a growtent essentially is highly reflective which trap/encases the light particles to conserve them until a leaf or other pigments absorb them.

 

[Koondense]

The larger the slice of bread, the less peanut butter you get with every bite Assuming a finite amount of peanut butter

Cheers

 

[exploziv]

I must admit I wasn't expecting reflectivity to count for so much in those measurements.

 

[X15]

Good photos f-e
..those shows exactly what i mean with concentrated beams

Even thou the light from a led light fixture appears to be fairly even to a human eye when youre looking at your plants from above, it’s still made out of fairly concentrated beams, which then create these hot spots, i’m thinking. You can’t feel the hotspots on your skin but it seems the plants can = Grow lights are wave length-tuned to get most out of the plants.

HPS/MH light is much “softer” and “even” compared to many plug n play led light fixtures



Imo, it’s these hot spots that, when they hit the leaf surface, warm the tissue up - the plants start to sweat and so the tissue gets dehydrated, something like that. This is why we get the dry feeling yellowing leaves under high intensity grow lights, esp. when the enviro conditions are off (temps, humidity) and/or the light is too close to the canopy.
...And then we need CalMag and Epsom salt to help grow tougher plant tissue that holds moisture better because of these led hot spots.

Hydro grown plants can handle led heat stress better than plants in soil, cause hydro plants have much more water in their tissue.

:::

The better diffuser materials pass over 90% of the light thru.


Ofcourse lenses and covers do trap heat which may affect negatively your led chips, so it could be wise to keep this in mind when your thinking about the cooling/heat sink size of your covered leds and high power use. ... esp. with COBs cause they heat up so much

Hi Goat! May I ask how a light mover might effect things?
I just started reading about how you guys have been researching light diffusion and making your own builds and it’s trippin me out. I dig it!
Thank you guys for this info!!

 

[X15]

The old school way of dealing with hot spots and poorly designed reflectors was to use light movers. with a little planning you can actually benefit from the extra canopy penetration the trick is to move the light but not too much be creative there are gains to be made.
Have the light mover on a separate timer that switches on with the lights then runs 5 to 10 minutes after lights out.

Yeah I’m with ya there. Just curious why would you want the mover operating after lights out? Just curious.

 

[Ca++]

Here is a recent test done by Migro on growlights, growtents and the inverse-square-law

Inverse square law is WRONG for grow lighting : MIGRO : Free Download, Borrow, and Streaming : Internet Archive

The inverse square law does not apply to grow lights. For the inverse square law to work the radiation source must be a point or sphere and emit radiate evenly...

archive.org

View attachment 18731345

1/20cm^2=0,0025
1/40cm^2=0,000625
0,0025/0,000625=4
>>double the distance, reduces fluence by the power of 4:
View attachment 18731364

a relationship like between "r" & "2r"

But that's not what Migro's measurement did confirm. The drop in intensity is less. This must be due to light reflection that otherwise would be lost but is reflected and falls into the sensor.

And a growtent essentially is highly reflective which trap/encases the light particles to conserve them until a leaf or other pigments absorb them.

Notice in the wiki you are quoting, they give the formula after this explanation. In which, there is no r being used as D. That would be confusing because r is radius and while it could be incorporated in a trig based equation, this law is boiled down to losses along a line.
If you have a distance (1) and you move multiples of that distance, the answer is that multiple, squared. It's really quite neat.


Are you bored of this yet? Shall we move on. Get your last word in if you like. I won't reply, unless it's outrageous


As neat as this law is, it's for calculating depreciation moving away from a point source.
If moving away from one light, moves you into the beam of another, this law knows nothing of a second beam. You might be able to use the law for each beam and add them, but that is something else.


There is always some room for movement. If we had a 288 LED QB and moved a light meter from 50mm to 100mm we wouldn't see 75% less light as the law states. We might not see 25% less. Because as the beam from one led spread out, another beam spread enough to land on the light meter.
If we took the same board, and moved from 1 meter ta 2 meters, we might see more like a 75% drop. From a meter, that QB is more or a single point, than it was at 50mm. If we move a mile away, it's looking a lot like a single point. The math will never come perfect, but it's getting a lot better.

I can't think of any use for this. Only if you knew the radiant pattern of your lights, could you use computer modelling to place them, with this formula. Though it might be better to just lay them down, lighting the sky, while you shuffle them for optimum spacing using a light meter (at maybe waist height)

 

[GoatCheese]

Hi Goat! May I ask how a light mover might effect things?
I just started reading about how you guys have been researching light diffusion and making your own builds and it’s trippin me out. I dig it!
Thank you guys for this info!!

Light movers would just move the hot light beams around- they would still be there but not warming the same spot as long - so the hot spots (beams) would just move to fry some other spot for awhile.

The problem with light movers is that many of the modern led-bar lights are rather large in size so many home growers don’t have much room to move their lights in their small growing space, many times just a 100cmx100cm tent or even smaller than that – like me.


The diffuser sheets for two my veg lights were around 15 euros including shipping, i wonder how much would two light mover units cost + the electricity to run them?

 

[Ipotato]

Debunking the Inverse Square Law | parabolix-light

The Inverse Square Law only applies to light that radiates omni-directionally. Let's understand how focused parabolic lighting works and why the Inverse Square Law does not apply.

www.parabolixlight.com

 

[Cerathule]

Notice in the wiki you are quoting, they give the formula after this explanation. In which, there is no r being used as D. That would be confusing because r is radius and while it could be incorporated in a trig based equation, this law is boiled down to losses along a line.
If you have a distance (1) and you move multiples of that distance, the answer is that multiple, squared. It's really quite neat.

The wiki uses a squared pyramid for its geometrical model, so it would be inappropriate to use "r". A radius is the distance of the center point of a circle to its outer circle.

The "1" isn't the distance, as you can see it's above the fraction bar, the distance is below that.
The distance gets squared, the 1 not. Thus, the result gets lower the higher the distance is, inversely proportional.