New DIY LED light

Discussion in 'LED and other Lighting' started by nfhiggs, Jun 17, 2017.


    nfhiggs Well-Known Member

    Wrong. Again.

    Quite uniform:

    and all in parallel. Please get off of the "parallel is bad" train. It works fine and has been demonstrated over and over.
    Budzbuddha and Discreetdigrowman like this.

    nfhiggs Well-Known Member

    blah, blah blah.... 100's of people have done it but they are all wrong, blah blah blah. You're a broken record.
    Budley Doright and wietefras like this.

    Dave455 Well-Known Member

    LUX meter does not read cxb3590 3000k accurately......
    GrowLightResearch likes this.

    wietefras Well-Known Member

    No, that's how it's supposed to be done. The amount of light is determined by the amount of light produced by the leds/cobs/styrips or whatever. The height is determined by the desired level of uniformity.

    A lux meter works perfectly fine for checking uniformity.
    RandomHero8913 and Randomblame like this.

    GrowLightResearch Well-Known Member

    You just keep digging deeper and deeper. Lux has NO place in horticulture. None, zip, ziltch, nada.

    I said:

    You respond with:

    Let me be clear I have not disagreed with using a Lux meter to measure uniformity.

    WTF do you not understand about uniformity and the number of photons reaching a plant??? I keep saying that Lux cannot help with measuring the number of photons that reach the canopy with respect to . Your response is about uniformity. Uniformity is NOT the topic. The topic is about using a PAR meter or spectrometer to measure the photons reaching the plant and the reason for doing so is to gauge whether the grow fixture can be lowered closer to the canopy or not.


    To simplify where you may understand. Lets say this
    1. The Lux meter found the uniformity to be perfect at 12 inches
    2. The Lux meter found the uniformity to be perfect at 24 inches
    3. At 24" the Lux meter read 10,000 Lux.
    4. At 12" the Lux meter read 40,000 Lux.

    Now the question is: Using a Lux meter how do you determine whether the fixture should be at 12" or 24"?

    Using the "simple conversion" formula to convert from Lux to PPFD that you previously mentioned answer this:

    • How many µmol/m²/s are reaching the canopy at 24"?
    • How many µmol/m²/s are reaching the canopy at 12"?


    To show you another reason Lux do not work in horticulture compare the luminous response of the human eye with photon photosynthesis.

    In the response curve of the human eye shown below the majority of the response is between 500nm and 600nm. Photosynthesis wavelengths are from 400nm to 700nm. Far red (720nm) plays a role in horticulture so I like to measure from 400 to a little beyond 720. Many scientists like to measure up to 800nm.

    So if we compare the wavelengths important to plant physiology with human eye luminosity, Lux is missing two thirds of the plant's bandwidth.

    Not only is that horrific, take a look at how much distortion Lux is adding to the picture.

    One watt of photons at 400nm equates to 0.27 lux.
    One watt of photons at 720nm equates to 0.72 lux.
    One watt of photons at 555nm equates to 683 lux.

    One lux at 400nm equates to 2.6 watts and 12.3 µmols/s

    One lux at 720nm equates to 1.4 watts and 8.4 µmols/s
    One lux at 555nm equates to 0.001 watts and 0.007 µmols/s

    Are you seeing the magnitude of distortion?
    Lux is emphasizing the green yellow and orange photons under the luminosity curve. Lux completely ignores the blue and red wavelengths.

    Furthermore McCree has shown that the blue and red wavelength ignored by the lux meter are the most relevant to photosynthesis.

    This means the photons most important to plant physiology are ignored by the lux meter.

    So how is it that you can convert lux to PPFD?



    wietefras Well-Known Member

    So a lux meter does have it's place in horticulture ... *ba dum tsss*

    Yeah I don't know why you don't seem to be able to grasp this. You tell me?

    Distance between light and plant canopy is only relevant for uniformity. You should not try to alter light intensity with the height of the fixture. That's just a waste of photons on the walls. Although some people are under the incorrect impression that inverse square law applies. It doesn't in our case since the light reflects back and/or overlaps with other COBs, strips, smds etc.

    So the only thing you really need light meters for is to determine the uniformity and with it the correct height for the fixture.

    Desired uniformity is usually: "the darkest regions should still receive 80% of the average light intensity". It then normally does not matter if you use a PAR meter or LUX meter for that. Most people don't even care about that. They just go with rule of thumb values. Uniformity isn't that important anyway.

    Once you have corrected the height of the fixture above the plants for the desired uniformity then indeed you might want to know how much PAR on average your plants are getting, but then who gives a shit about accuracy there?

    Light insity is what it is and the plants are fine with anything between 250umol/s/m2 and 1500umol/s/m2. If you can get a calculated indication within 10% or even 20% of what the PAR intensity is with a LUX meter then that's more than enough. No need to spend a bundle of money on improving the accuracy by a few percent for a measurement where the plants can deal with a 600% range anyway.

    The only case where you really need a PAR meter is when you compare light sources with different SPDs. You might end up with compounded inaccuracies in the comparison then. But hardly anyone has that issue and even then I wouldn't care personally.

    Ultimately the only things we really care about is how much watts go into the lights and how much harvest comes out. The plants are ultimately the real PAR measuring tool. Anything in between is just theoretical bullshit anyway. A PAR meter is going to give you 0% advantage over a grower who does not have one.

    BTW you can convert from LUX/lumen to PPFD/PPFD quite easily. You can simply calculate a conversion factor from the SPD and divide or multiply the lux measurement.
    Randomblame likes this.

    GrowLightResearch Well-Known Member

    Except the lux meter only measures between 500nm and 600nm. It cannot even see 400-500nm and 600-700nm. You are such an idiot.

    Only you only need to worry about uniformity because you're the one handicapped with a shitty fixture. Let me guess a fixture with CoBs wired in parallel? CoBs are a poor choice to begin with. I do not care about your problems with uniformity. Not everyone has your issues.

    Wow. Laws of physics do not apply to you. Hard to counter that. Fantasy land. Delusional. Good luck. Good riddance.

    CobKits Well-Known Member Rollitup Advertiser

    in a given spectrum even a $20 meter is highly repeatable

    imma put agro guy on ignore now since hes taken to slinging insults. so i wont be around to correct his near-constant false statements, youre on your own guys
    wietefras likes this.

    Randomblame Well-Known Member

    A luxmeter can not be used to meassure par but it can be used to make estimations.
    No one cares it's accuracy here, we are all hobbyists.
    Some of the meters use also conversion factors to convert par to lux, why one could not do it manually in the other direction to close from lux on par? It's not accurate but it gives you an idea and nothing more is expected.

    You just ride on principles, it's the same as with the parallel circuit ...
    It's not about whether it's correct or accurate, it works in a certain range and that's enough...

    wietefras Well-Known Member

    First of all, that's not true and second, it doesn't matter. That's what the lumen par conversion factor is for. Even if you only measure the middle you can still extrapolate that to the whole chart since both the lumen and SPD are known. Check out the Apogee PAR conversion factor site link I gave. That does work.

    Coming from you I will take that as a compliment.

    Your guesses are as wrong as your "knowledge". If you care about hot spots then that's a matter of uniformity.

    Laws of physics do apply. Difference is that I actually understand them and know how and when to apply which. One of those physics principles at play here is called "reflection". Perhaps you could look it up and see how that counters the "inverse square law"?

    Anyway, I agree with CobKits and Randomblame. You are the lastest new kid on the block who thinks he got it all figured out and has a long way too go still. You probably will understand one day how wrong you are now, but you will have wasted some people's money on an expensive tool they never needed to begin with. So sad that newcomers always have to be like this. Why not first learn the facts and practicalities before you start shouting?
    PSUAGRO. and Randomblame like this.

    GrowLightResearch Well-Known Member

    Yeah, good luck with that. Correction: I actually BELIEVE I understand them. When someone believes something to be true like they can defy the Laws of Physics, that's referred to as delusional. What are you smoking? You sound like you are suffering from chemically induced psychosis. And you are a presumptuous nit wit. The world does not revolve around you. The world does not do things the same as you so many of the things you say only apply to you. You keep bringing up uniformity. I HAVE ZERO PROBLEMS with uniformity. You keep bringing it up like you will get a different response. Is that not the definition of insanity?

    And please don't tell me the wattage (wall) of a fixture means anything. I believe I have seen you defend that watts idiocy too.

    You know it might behoove you to cite some science rather than expect anyone to believe your BS. Like show us how reflection is the anti-inverse sq. law.
    I missed that day in physics class. When I take PPFD measurements with my worthless (as you say) spectrometer, I do not notice much of a difference measured when in a open space or in a Gorilla tent. I do notice a huge difference when the sensor is a few inches away vs. a few feet. I think that's due to the real Inverse Square Law. Why is that?

    To summarize I measure the PPFD at 12" and at 24" in a tent with reflective sides (Gorilla tent) and my spectrometer measurements match the calculated measurements for the 12" difference using the inverse square law. Does a Gorilla tent not have your mystical reflexive powers? Please 'splain that to me.

    And I need PPFD not PPF. PPF is of no use. PPF is an isotropic value that is intended to be an exaggerated PPFD same as lumens where the grow lighting marketeers are trying to bamboozle you. What is the photometric luminous equivalent for PPFD and PPF? Can I find those equivalents in the datasheets or Apogee website or some handy dandy spreadsheet?

    How you can say you understand and say the above in the same post is insane!

    The luminous curve I showed you for the Ambient Light Sensor (ALS, used in all lux meters) is the same curve as the Photopic Relative Sensitivity Curve for the C.I.E. Standard Observer. Every lux meter uses a luminous ALS. If a lux meter were to include any wavelengths outside that curve it would be inaccurate. That is why every lux meter uses a ALS adjusted for human luminosity. A lux meter will not detect ANY deep red (660nm), it will not detect any deep blue (450nm). The way an ALS does this is by filtering out all the deep blue and red. Just the human eye does.

    So if I use a lux meter to measure a Red Blue fixture (and I do NOT give two shits if you do not use BR) the lux meter is going to read ZERO.

    Now back to your whinny "I use white, not burple". A corporation spending millions on a grow facility is going to listen to science not YOUR psychotic ramblings. Do a search on vertical farming images. What color shows up magenta or white. (Not including lettuce).

    Okay so I use a lux meter to measure a white trash fixture.

    True but PAR to Lux is possible if enough is known about the wavelengths. The other way is not possible...

    Below is what a lux will detect in the highlighted green area. How does that even relate to an SPD?? That my friend is why lux to PPF conversions do not work.

    LED SPDs are usually radiometric. The lumens in the datasheet are not. So when you use the SPD to do your conversion you must filter the non-luminous wavelengths before multiplying it by the lm/w.

    Don't you find it particular when the spreadsheets, you so love to use, come up with 5 µmol/s when the most efficient LEDs can only squeak out 2.5? Case in point the Cree XP-G3 (more efficient than any CoB) specs out at 2.58µmol/s and that's the 451nm blue before the phosphors cover the blue LED to make white and thereby reducing the PPF.


    The below lux SPD is from the Texas Instruments OPT3001 ALS. but you probably understand physics more than the TI engineers (in your world).


    Don't lump @Randomblame with you two. He has intelligence and an open mind.
    Last edited: Dec 7, 2017

    nfhiggs Well-Known Member

    LOL - The Inverse Square law is a mathematical function that describes a point source radiating 360 degrees in 3D infinite space. Last time I looked I was growing in an enclosed, reflective very finite space with a light distributed across a 2 ft by 4 ft area.

    No, it does not apply, it never has applied and it never will apply in any meaningful manner.

    ganjamystic Well-Known Member


    Randomblame Well-Known Member

    Sorry! But you can not override laws of nature, not without magic!
    Of course it also finds application in a tent with hundreds of point light sources. The problem is overlapping and reflection makes the computation so complex that you need probably a small super computer to include all the factors.

    nfhiggs Well-Known Member

    Its not overriding anything. The inverse square law is simply the mathematical expression of one extreme - a point source emitter with maximum radiative spread. The opposite extreme is a 100 percent reflective tube of infinite length with the same diameter as the emitter. Such a tube would have uniform intensity all the way down its length - a fibre optic cable almost achieves this but not quite, due mostly to internal absorption. Our lights fall somewhere in between those extremes. And you're correct, the factors are too complex to calculate to any realistic degree. The best we can hope for is to measure it and see how we are doing.
    graying.geek and Randomblame like this.

    GrowLightResearch Well-Known Member

    Remember these guy for some no good reason just like to blow smoke up my ass. The above is 100% bullshit. why?

    No I, did it with with less than 10 lines of code

    He's confusing a point source with a A lambertian surface which is defined as as surface reflects or emits equal (isotropic) luminance in every direction. Looks similar to this. where ISL does not work.


    Look again. Verify your source. Show me I'm wrong.

    The light source does not have to be a isotropic source i.e. 360°.
    It makes zero sense for it to have to radiate 360°. or could you please 'plain why I am wrong.
    It does NOT radiate into 3D infinite space. It would have very little value if that were true. So are you saying Inverse Square has no practical use on this planet? Or could you give me direction to this infinite 3-D spot on this planet where it works.

    Now me. this is how I think ISL works.

    As a surface is illuminated by a light source and the surface moves away from the light source the surface appears dimmer. Getting dimmer faster than it moves away from the light source. As if the dimming rate distance is associated with the distance squared.

    The light source only needs to approximate a point source. The distance need only be 5 times greater than the largest dimension of the light source. That is know as the "Five Times Rule". In other words you need to be a few miles away from the sun or it won't work well with the sun.

    Now I am confused. The problem is when I say stuff, I do not just pull it out of my ass like you have apparently just done.

    Inverse Square Law: inverseSquareLaw.jpg

    To confirm ISL actually works, I decided to measure a strip of 16 LEDs with a spectrometer at six different heights. The procedure was to measure one height, then calculate the five other heights using ISL, then measure at the five heights.

    The first attempt had too much error like 5-10%
    The I made the calculations as 16 separate point sources for each LED on the strip
    I thought about it and I decided to try calculating using each individual LED's spacial radiation characteristics, (e.g. the angles and their intensity at each angle with respect to the spectrometer's sensor position) and each distance being the hypotenuse of the right angle associated with each LED. Essentially 16 separate trigonometric equations to use as the distance in the ISL formula.
    I also used wire shelving (as shown) to make the measurements at each distance repeatable.

    My calculation then match the spectrometer measurements with ≈1% error.

    The math
    ISL Formula Highlighted
    intensity[$angle] was entered from the intensity and angles in the datasheet's spacial radiation characteristics.

    $ref = 3.93

    for($i = 1; $i < 7;$i++){

    $height = $i * .7857142857 + $ref;

    for($j = 0; $j < 8;$j++){

    $offset = $j * .7;

    $angle = rad2deg(atan( $offset/$height));

    $distance= $height / cos(deg2rad($angle)) ;

    $lux = pow($ref/$distance,2) * $intensity[$angle];

    $total += $lux;



    datasheet's spacial radiation characteristics.

    Results with Error in %
    Total Error 1.38%

    Columns: Distance, Measured, Calculated, Error
    0 3.93 440 440 -----
    1 4.71 347 350 0.83%
    2 5.50 270 283 4.82%
    3 6.28 230 233 1.22%
    4 7.07 193 196 1.34%
    5 7.85 164 166 1.09%
    SUMS 1644 1667 1.38%

    Notice Measurements were taken in a reflective very finite space with a light distributed across a 2 ft by 2.5 ft area.



    Last edited: Dec 7, 2017
    Randomblame likes this.

    Randomblame Well-Known Member

    Unfortunately that is mathematical nonsense! A calculation can be checked "always" backwards. If that does not work, the first result "must" be wrong. That means, if you have the necessary data and they are correct, you can calculate it in both directions.

    Thank you for the flowers..!
    I will always try to keep my mind open and to look at all aspects. I'm not satisfied before I understand a thing and I'm always open for something new! If I'm wrong and I'm disproved, well, and as long as I'm ready to accept it I can only get better!

    And to your meassurements...
    Can I ask how big your tent is and how much LED's was tested at once and at what point in the tent(corner or in the middle)?

    If you meassure only one LED/COB in the center of an 3x 3' or 4x 4' tent at 1' and 2' heights with the sensor placed at the ground, the meter will probably not recocnize much of the reflected light.
    Due to the emission angle of about 150°, the light will only hit the ground around the sensor and will then be reflected upwards, away from the sensor.
    The little bit of light that hits the sensor after another reflection hardly plays a role in the measurement.
    It would be different if you did not measure the LEDs 30-60cm above the ground, use a smaller tent, or insert and measure several LED's together.
    If you place multiple LED strips 60" above the ground for instance and hold the sensor underneath, the results would look completely different due to overlapping and reflection!

    GrowLightResearch Well-Known Member

    I'm saying it is possible to measure Lux with a simple meter and get fairly accurate results. But the measurement is just the number of photons that strike the sensors surface with no discrimination for wavelength. Wavelengths are not needed. With no wavelength intensities there is no possible way to even estimate PPF.

    The lux sensors are very similar to a solar panels photo-voltic cell. The photon excites the cell which has a "memory" effect. The energy created (e.g. voltage level) are sampled, measured, and converted to a lux value.

    The PAR sensors are photdiodes that are only excited by a specific wavelength.

    On the other hand to measure PAR you MUST have wavelength intensities. Most PAR Ambient Light Sensors also give lux. Tomorrow I will be receiving a development kit from ams for their AS7262 which is an ALS/6-Channel Visible Spectral ID Device. See attached datasheet

    I have another demo kit one being delivered next week from SparksFun. It's just a PCB with the chip on it that requires another development board. I think I can make it work with a rigged serial port cable. If so it will be a very inexpensive "PAR meter" that plugs into a laptop. I will likely have to write an app for it. It should be better than the $150-$300 PAR meters and this only costs $25. I would bet the PAR meter on the market today only have 3 or 4 sensors. But they get by.

    I don't think the AS7282 will be accurate enough. If not I have plans to add their TCS3400 (datasheet attached). Dependents on how much trouble it is to merge the two I²C ports to one USB port. The cool thing is the 3 TCS wavelengths does not duplicate but rather fills in between the wavelength gaps of the AS7262 plus adding a white 2700K sensor. I would then have a 120 channel ALS. The AS7262 is supposed to be a kick ass PAR meter chip. It is very simple to design a mini-spectrometer with just the AS262. I hope to have it up and running tomorrow by noon.

    This past weekend I posted the wavelength of the AS7262 over an SPD. since then I thought about adding the TCS3400 so I updated the image by adding the new 3 color channels to the previous 6. I just wish there was a 480nm photodiode and the PAR values would be very high quality. 480nm is an important wavelength for white LEDs. I'm hoping I can get around it by creating profiles with the 9 wavelengths plus the 2700K sensor.

    These are very new parts just released recently.


    The tents I have are Gorilla Lite 2.5' x 2'. But for the test I descried for testing the ISL the measurements were between 3.93" and 7.85" in 0.78" (20mm) increments (the distance between the wires of the shelving).

    Attached Files:

    Last edited: Dec 7, 2017
    Randomblame likes this.

    nfhiggs Well-Known Member

    Its not 2ft by 2.5 ft distributed light. It may be shining down into a 2 x 2.5 ft area, but that's not the same thing.

    Its diodes in a single line so its more like 2 ft by nothing. Big difference.
    And only 16 (near) point source emitters. My single row strips have 40.
    Now fill that ceiling with nine 22 inch bars with 440 diodes and "do the math" again. You'll have one half of my light.

    Note that your image link has "point source" in bold. That part might be important. Key even. Pretty sure I mentioned that as well. Stressed it even.

    Infinite space was a bad choice of words. Point taken. "Point source unbounded by reflective surfaces" would have been better. You may be in a "reflective finite space" but your measurements are too close to the light source for the reflections to make much difference.

    Thing is, you just made my point for me and you don't even realize it. You had to do so much mathematical manipulation to "make it fit" that your final equation is not even remotely recognizable as a derivation of "E equals I divided by D squared". 90% of people could not even begin to contemplate the method of figuring that out, much less actually doing it. Kudos to you for that.

    Thank you for eloquently demonstrating my point for me. Myself, I simply don't have the time or inclination to do such complex math.
    wietefras and Randomblame like this.

    Randomblame Well-Known Member

    Whow, I thought it would be more complicated to compute multiple light sources inside a tent.
    Deviations of only 1.4% are very impressive for a calculation like this!
    You have a lot of useful things to contribute here if you just refrain from your animosity against some users.
    A construction manual for a $ 25 PAR meter, for example, would be very popular!!!
    Avoid comments like, "you're such an idiot" and just ignore it when others disagree. I like the way you base your arguments with facts and I'm sure we can all learn something from an old hand like you.
    I would hate to see you banished from here just because you are too impulsive sometimes and say what you think.
    Trust me, sometimes it's better to swallow it and just keep going on... ;-)
    Even your two "favorite enemies" have done a lot of good things for the community in the past, even if it does not look like that for you.
    I really want to see how you build the PAR meter, even if it costs $ 50 at the end. It though belongs more in the field of electrical engineering, but nonetheless it is extremely interesting!!!
    An own thread about that device would certainly get many more likes than posting your research and development into other ones threads. Even though I think likes do not interest you...
    graying.geek, Serva and ttystikk like this.

Share This Page