1. We are currently experiencing issues with viewing and uploading images, our team is working on the issue.
    Dismiss Notice

King Cob Led grow light

Discussion in 'LED and other Lighting' started by VegasWinner, Dec 1, 2017.

  1.  
    Randomblame

    Randomblame Well-Known Member

    Haha, you're good!
    Everyone deserves a 2nd chance...
    (Psssssst! Looked furtively right and left and wispers)
    "I already had a suspicion after 10 minutes of discussion, who I have in front of me, LOL! Hope you keep it a little longer this time. ;-)"
     
  2.  
    nc208

    nc208 Well-Known Member

    Sorry but making statements like this doesn't go away with a comment like "I recant and apologize." Perhaps you should learn to converse with your peers instead of trying to belittle them.



    Actually they have just started to enter the horticulture game with their sunplus line. These are the first leds I know of which were created specifically for our applications. The current emitters you refered to are re-purposed for our applications and not originally intended for. Their sunplus 20, 35 sunplus cobs are the future and I'm waiting for them to start building fixtures utilizing it. They are very slow at doing this though. It took phillips over 5 years after leds hit the aquarium trade and were proven to work before they made their first fixture. Now that Leds are becoming mainstream and widely accepted for horticulture some of the big players will start pumping lights out like what Phillips did here.
    https://www.philips.co.uk/c-m-li/coralcare

    As per the bridgelux decor - they rock I am testing out vero 10 2000k's and they have shortened my flowering time by a week it looks like.
     
    Randomblame likes this.
  3.  
    GrowLightResearch

    GrowLightResearch Active Member

    Their EB Series strips are kick ass too. I don't keep up with CoBs, but I didn't know there was a Vero 10 2000K. I have played with the Decor 1750K 97 CRI. I use CoBs for thermal heatsink experiments so I just need a heat source, but I choose them to be used with my horticulture research as well.

    That's not what I recanted or apologized for. Your reading comprehension is lacking if I don't say so myself.

    You do make a good point about belittling others. In my mind the idiots and morons was a shot at two specific idiots and morons that have been using pseudo science and BS to defend parallel wiring. It was insensitive of me to take a shot at them and not consider it might offend others. I'm sorry, I was wrong, please forgive me.

    Okay. Lumiled's SunPlus. The 35 is newish (≈1 year) but still it's thermal resistance sucks. It's what they think are the proper BR ratios. More likely what a big client that essentially funded SunPlus thinks are the correct ratios. The SunPlus 20 is just a re-branding of their Color C Line that has been around a long time. As has their Luxeon Rebel Color which is targeted to horticulture and also has been around a long time. I like the Color C/SunPlus20. They are just too small for me to work with. I love the footprint of the Rebel Color. It's (IMHO) the best thermal footprint far superior to any other. Luxeon is not keeping up with efficacy. I no longer buy Rebels. The Color C Line has a mini Cree XP foot print with the thermal pad between the anode and cathode.

    The SunPlus CoB line is of no interest. CoBs are inferior due to their substantial thermal management costs and poor uniformity. Today, Samsung's F-Series strips are the best off the shelf product for the hobbyist market. And Samsung's 218 lm/w (173 lm/w CRI 90) LM301B LEDs are almost available now. Expect strips with them to be out soon.

    Any LED manufacturer that has a 450nm blue, a 660 deep red, and especially a 720nm far red is in the horticulture game. That would be Lumiled, OSRAM, and Cree. OSRAM is increasing their horticulture position and Cree may be backing off. Lumiled will always be in the game as they supply Phillips Horticulture. Today it is my opinion that OSRAM is the leader in horticulture LEDs. They have a kick ass deep red (hyper red) and a deep blue that is very competitive with Cree's XP-G3 deep blue. The new (not yet on store shelves) OSRAM Square deep blue may beat Cree's XP-G3.
     
    Last edited: Dec 7, 2017
    Randomblame and VegasWinner like this.
  4.  
    nc208

    nc208 Well-Known Member

    I should clarify my response, I think the sunplus 35 line is a great accent for a flower boost, 2700k-3500k will still be golden area for main lighting but the purple 12.5% blue and 2.5% blue hit the 450, 660, and 720 all in one. Im not a fan of mono colors but prefer to use a blend to give a better spectrum like the 2000k veros vs a 660nm deep red. I think Im going to make some accent strips to run vertical and horizontal instead of the veros next run to see which performs better.
     
    Randomblame likes this.
  5.  
    GrowLightResearch

    GrowLightResearch Active Member


    I saw your pic of your setup. I love that yellow glow. You may like my post which is more about yellow than it appears. http://rollitup.org/p/13942003/

    The SunPlus 20 are specifically designed for color mixing. But given that, plants do not have eyes, they do not care. They just want photons. They love deep red the most. Deep red is the most efficient photosynthetic color to produce with LEDs and the wavelength used most efficiently by plants. If you look at my post I referred to above you will better understand my stand on color. It's not just about photosynthesis. Yellow is the most inefficient color to produce and very temperature sensitive (can lose 80% of intensity if it gets hot).

    If you look at the radiometric power of a blue photon vs. a red photon the blue carries a lot more energy than red. As wavelength increases energy decreases. So you'd think blue photons are better because they have more energy. For plants blue can have too much energy. Red is like weed and blue is like crack to plants. See diagram.

    absorptionAndEmmisionSpectra.jpg

    When it comes to the energy of photons plants do not care. A photon is a photon and the amount of energy does not matter. That works out very well for red. Because red carries less energy, it takes less energy to create a red photon. 1 watt of radiometric power can be contained in 3.75 µmols of 450nm blue photons. Whereas 1 watt of deep red 660nm equates to 7.55 µmols of photons.

    Blue and red are the most efficient because the material used allows the narrowest junction band gap. The color emitted from the band gap is determined by the width of the band gap. As the band gap gets wider the efficiency goes down because it takes more energy to bridge the gap. So orange is not as efficient as red, but more efficient than yellow. Moving up from deep blue to cyan is more efficient than green and green 555nm (being exactly in the middle between 450 and 660) is the most inefficient color to produce and also the least efficient for a plant to use for photosynthesis. Keeping in mind photosynthesis is not the only concern.

    LED datasheets and their statistics are very deceptive. If look compare deep blue and deep red in a line of LEDs which is more efficient? Example OSRAM Olson SSL at 350 mA blue=763mW and red= 560mW. But wait blue is 3V and red is 2V, so blue takes over 30% more electricity. If we reduce blue by 30% 762mW X 0.7 = 534mW. So blue and red are in the same ballpark. That's until we look at the quantum number of photons.

    Red: .560 W x 5.77µmol/W = 3.23 µmols of photons
    Blue: .534 W x 3.75µmol/W = 2.00 µmols of photons

    So that deep red isn't looking too bad.

    I saw a post of yours where you said "...your retarded" I'm sorry too. It just makes you look retarded when you used "your".

    Bottom line you are okay. Or your okay.
     
    nc208 likes this.
  6.  
    PSUAGRO.

    PSUAGRO. Well-Known Member

    More like three chances now, but who's counting?....... oh yeah I am, he's a psychotic egomaniac that questioned alesh's math and drove him the fuck out with his repetitive babble. Limped the community with his bs.

    What's his new online persona now? Boeing test pilot?nuclear submarine technician? white Truffle forager?
     
  7.  
    GrowLightResearch

    GrowLightResearch Active Member

    Old, dying, and trying to do what he can with the little time he has left to make the world a better place.

    White truffles? Why not psilocybin mushroom forager? Remember, world, better place.

    I can understand that. I am often misunderstood.

    I questioned Alesh's math at his request.

    I questioned his math because my calculations came out with much lower values of PPF than his spreadsheet. I completely spelled out how I arrived at my results. No one pointed out any errors in my math. No one, not even Alesh, could answer my questions regarding his math. So I reverse engineered his math.
    I do not believe Alesh created that spreadsheet or understood the math.

    I spelled it all out. If you can't follow the math, I do not expect you to understand.

    What did not help is you and your CoB snobs buddies talking shit the whole time when none of you knew Jack Shit.

    And you are still talking shit. Care to try and back it up Mr. Smart Guy?


    And I still say the spreadsheet does not do what you guys think it does.
    Do you understand that you cannot take radiometric PAR bandwidth and multiply it by photometric lm/w to get a quantum result. Think measurement geometry.

    Yeah drove you guys crazy when I questioned how you get 5+µmol/s when the most efficient LED only gets 2.77 µmol/s on it's best day at a temperature of 25°C and under other unrealistic parameters. And at the end of the day you do not have what is needed and that is PPFD µmol/m²/s. The whole concept of that spreadsheet is flawed and no one could explain otherwise. Feel free to step up and 'plain it to me. If not STFU.

    I'll make real easy for you. If you understand any of this stuff you should have no problem answering this: The column in Alesh's spreadsheet is label "energy per mole", what should it be labeled? Answer that and I will STFU.

    The flaw in the spreadsheet has to do with crisscrossing geometries. I attached a PDF that has the answer. Start with the table 3.1. HINT: Which quantities can be accurately converted to the other, which can be estimated, and which cannot be converted? Example, one of the common errors is when confusing lm/sr and lm/m² with one another. When is m² spherical and when is it planar?

    Once you understand all the measurement geometries then in the spreadsheet add the units of measure to each column.


    This is how I learned about it:
    I first read, no studied extensively, the attached PDF. I studied it until I thought I could explain it accurately to others. Lotta good that did.

    I have a spectrometer that measures Lumens, lux, irradiance, radiance, luminance, illuminance and etc.
    So what I did was measured lux, irradiance, and PPFD. I then learned enough about them to convert from one to the other. I did the conversion calculations and compared the measured values against the calculated. I converted from each measured value to the two other values. The hardest thing is doing calculations and have no way to know if the answers are correct. At that time there was no help to be found anywhere on the Internet. It's out there now, not easy to find.
    The results of my calculations are posted here:http://www.growlightresearch.com/ppfd/convert.html

    For anyone that wants to learn:
    You need to know how to convert irradiance to photon flux and vise versa.
    1. Use Plancks constant to calc the distinct energy quanta at each wavelength.
    2. With this quantum energy, calc number of photons per watt
    3. Use Avogadro to get photon flux
    You've just converted irradiance to photon flux.
    After you fully understand that, the rest of the conversions are easy.

    The conversion between irradiance and illuminance is a piece of cake.
    You just need to know which and Radiometric, Photometric, and Quantum geometries and quantities can be accurately converted.
    You may want to look into the industry standards the datasheets are supposed to be following to measure lumens, SPD, and Spectral Intensity Distribution curves.
    If you have any problems go back to my original post where I reverse engineered the spreadsheet doing the above.
    If you find any errors in my stuff 'splain them to me please.
     

    Attached Files:

    Last edited: Dec 7, 2017
  8.  
    nfhiggs

    nfhiggs Well-Known Member

    Actually if you compare apples to apples - 1 QB 288 quantum board to Samsung 288 LED strip, its not nearly as big of a price difference - $79 (or thereabouts) to $48. Samsung is a better value, but you still have to mount them and power them - comparing the price of a fully finished fixture to the price of raw strips was a bit unfair.
     
  9.  
    nfhiggs

    nfhiggs Well-Known Member

    Oh - and you can get the QB 120 board for $29 - 120 LEDs on a larger footprint board.
     
    Stephenj37826, key4 and Randomblame like this.
  10.  
    CobKits

    CobKits Well-Known Member Rollitup Advertiser

    i havent test the 120s yet but when i tested QBs against 1212 gen6 it took twice as many diodes for the cob to match it

    so 240 diodes closest match would be a citi 1812. chip is obv cheaper but with a heatsink its about a wash
     
    Stephenj37826 likes this.
  11.  
    nfhiggs

    nfhiggs Well-Known Member

    That's something we can agree on.
     
  12.  
    alesh

    alesh Well-Known Member

    Regarding 5+µmol/s (you actually mean 5+ µmol/J I guess): This is a quantity of a spectrum, not an LED. It is photon flux [µmol/s] per radiated power [W] - µmol/s/W which is equivalent to µmol/J. You have to multiply it by an actual LED's radiometric efficiency to get the said LED's photon flux (quantum) efficacy. Compare to the "From Radiometric" column in your link.
    Yeah this one is incorrectly labeled and I'm sorry for that. It's a remnant of an older version. It's actually SPD/energy per micromole or rather photon flux.
    The spreadsheet is about calculating LER and "QER" of a spectrum in order to convert photometric, radiometric and quantum units of the same geometry between each other. That means luminous flux (lm) to radiometric flux (W) to photon flux (µmol/s). Or illuminance (lx) to irradiance (W/m^2) to photon flux density(?) (µmol/s/m^2). Or luminous intensity (lm/sr) to radiant intensity (W/sr) to photon flux intensity(?) (µmol/J/sr).
    Most common use there probably would be luminous flux (lm - from a LED's data sheet) to photon flux (µmol/s).

    If you want to be a stickler about units (I suggest you start in your own posts - they're full of such mistakes), then it's (if we take the columns as a whole):
    SPD [W]
    luminosity function [lm/W]
    SPD*luminosity [lm]
    e̶n̶e̶r̶g̶y̶ ̶p̶e̶r̶ ̶m̶o̶l̶e̶ SPD/energy per micromole [µmol/s]
    each row is obviously per nm

    This is basically the same thing the spreadsheet does. Only the spreadsheet is one or two steps ahead and gives values for complex spectra as well. You can confirm this if you put a single wavelength SPD in the spreadsheet (slight discrepancy is because of rounding on your side).

    It's not that difficult confirm results as these are simple conversions of well defined units. All you need to know is luminosity function, speed of light, Avogadro's and Planck's constants. Even Wikipedia has the topic well covered for many years.

    BTW a spectroradiometer has to be coupled with an integrating sphere to be able to measure luminous flux [lumens] - or radiometric flux or photon flux for that matter.
    I also can't help myself but point out that columns in your link are incorrectly labeled as well. Should be lx, W/m^2 and PPFD (or µmol/s/m^2).
    After you fully understand that simple conversion between irradiance and photon flux is not possible we can talk some more. This is actually an example of units of different geometry (total power per surface area vs total power).
    So you did reverse engineered a spreadsheet that had been attached to a post explaining every single action done in that spreadsheet? I call that reading TBH.

    So let me sum it up:
    1) The column "energy per mole" in the spreadsheet is incorrectly labeled, calculations are correct, though.

    2) LER and "QER" values the spreadsheet produces allow you to convert:
    lm <-> W <-> µmol/s
    lx <-> W/m^2 <-> µmol/s/m^2
    cd <-> W/sr <-> µmol/s/sr
    It cannot be used to convert ie:
    lx <-> W
    cd <-> W
    lm <-> µmol/s/m^2 (the quantity is called PPFD)
    as these conversion (amongst others) would require quite complex simulation

    3) For the know-it-all attitude you're displaying, there's quite a lot wrong in your posts.
     
  13.  
    Stephenj37826

    Stephenj37826 Well-Known Member Rollitup Advertiser


    Phillips doesn't own Lumileds and they haven't for a while.
     
    PSUAGRO. likes this.
  14.  
    VegasWinner

    VegasWinner Well-Known Member

    I have been using Limileds Luxeron SunPlus 20 diodes for the reds and blues.

    I find some folks here are not worth the time and dialog lead nowhere.

    Take this thread. it was started about a DIY idea. Initially it was this is a BAD light. BAD LIGHT. Than it was about each other, and now it is about engineering with no context.

    I like the SunPlus 20 line. they are small, but my supplier has no problems with the size of the diode. They make a nice compact led baord. 1"x12" with two channels of red and blue total of 36 diodes 24 red/12 blue.

    BTW, I design boards and use series/parallel to create a balanced circuit without any other electronics, after all a diode is resistor, an infrared collector device and an emitter device. A diode is a very versatile device.

    I was experimenting with one of my GG 200w boards. I accidentally dropped a tool on the board and broke a diode on two separate parallel strings. I checked the light and the the current draw was rthe same, but since the board has a specific number of parallel strings to maintain lower amps, the board was driving harder than it should. I placed a dab of solder at each broken diode base shorting the circuit across the pad, and repaired the two circuits. The boards works great, resistance is almost identical, as the solder worked to create adequate resistance to rebalance the circuit again. Board works great just two diodes short of 384 is all. Current draw remains the same as does the forward voltage.
     
    Moflow likes this.
  15.  
    Randomblame

    Randomblame Well-Known Member

    Really?
    Are not the "shorter" strings brighter with one LED less?
    Maybe you need to dim the board a bit down to "see" an effect. The solder can increase the resistance, yes, but does it reduce the voltage going thru the remaining string so much?
    Normally, if the same voltage runs through all the strings, then the "damaged" ones need to burn brighter as the rest. If the 32 LED's in one string run normally at 2,9v, the 31 diodes in the "defective" string would probably run at 2,99v.
    Can a blotch of solder actually compensate that?
    Is there a way to check the voltage going thru each string?
     
  16.  
    OLD MOTHER SATIVA

    OLD MOTHER SATIVA Well-Known Member



    ok..i hook my cobs /pcbs to drivers in parallel if i have the correct matching vf for them...
    i hook them up in series if i need to use drivers that do not match but can fit by hooking up in series

    you are saying that parallel is not as good because??

    better performance with series hookup??

    please try to make it eay to understand..because i am not a tech..i am a grower
     
    Last edited: Dec 8, 2017
    Rolla J likes this.
  17.  
    Randomblame

    Randomblame Well-Known Member

    The top most reason is because imbalance between the parallel strings.
    When there are differences in Vf of the used COB's some COB shine brighter as others. In a worst case scenario they could run into thermal runaways. This would mean, while one or more COB's are getting brighter and brighter(they become hotter and take more current) the others gets lower and lower. At the end the brightest ones can burn out, the current is shared between the remaining COB's and if there is even more imbalance more COB's could burn out.

    But like I said, this is the worst case and you will probably never see it when the COB's are
    a) proper cooled and
    b) there is no difference in Vf of more than a few tenth of volts.

    0,5v can be already enough to see a notable difference!
    That's the main reason, but to be honest, I have not noticed that much difference in the Vf.
    Only one of my COB's had previously a higher Vf, but in this case two or three strings were broken.
     
  18.  
    GrowLightResearch

    GrowLightResearch Active Member

    I was not sure. Last I heard there was some legal entanglements and the sale did not go through. That was awhile back, I have no concept of time.
    I did hesitantly
    The patent was written by Phillips in early 2016.
    It was in response to someone saying something like no one on this site uses Philips CoBs.
    I just looked it up. A year goes so quickly when you're old.
    Phillips divested 80% in December 2016 and retained 20%.
    So I think it's okay to say someone on this site is using a Phillips CoB.

    You are 80% correct. Not sure if I am 80% wrong.
     
  19.  
    GrowLightResearch

    GrowLightResearch Active Member

    I really wanted @PSUAGRO to attempt to answer. I doubt he knew the correct heading for that column. Or anything else on this topic for that matter.

    I know. It's difficult to communicate when someone takes your stance.

    I do make mistakes and this site does not allow edits for much time. If I write something at night, get tired, go back to proof it when I wake, I find then I cannot edit any mistakes I find.

    If you could, try to pretend I know what I am doing. Humor me please. I have no animosity towards you. I appreciate what you are trying to do. I'm just not sure what you are trying to do.

    I call reverse engineers you spreadsheet trying to figure WTF you are trying to do. And I still do not know. I know what you think you are doing. I'm not so sure.


    The "own post" is not a public page. The only way to get to that page is if I give the link which I had never done until two days ago. It was a spot where I could write a script to test my calculations against measured values to verify the calculations were correct. The column headers were only for my eyes and I needed short easy titles. I know Watts means W/m² but Watts was easier for Me to understand with a very quick glance.

    You are correct the column should be PPFD not PPF. I removed the D because I didn't like how it widened the column. That was some time ago. I often found PPF and PPFD used backwards. I still never have found any authority on the definition of PPF and PPFD. It would be nice if they would be made SI units or CIE would put them in a spec.

    PPFD If anything it is a "defacto" standard not a standard SI unit. Seeing we are being technically correct, PPFD can only be used when the wavelengths are from 400nm to 700nm whereas µmols/m²•s applies to all wavelengths.

    Refer to my conversion below.

    Lumens are representative of how the human eye perceives radiometric brightness as defined by the CIE. It's not a law of physics. It was determined by ask people if when the brightness of a 555 nm source was reduced until the observer felt that the two sources were equal in brightness. The fraction by which the 555 nm source was reduced, became the luminous sensitivity with respect to the second observed wavelength aka Luminous Efficacy. After enough of these observations were recorded, the CIE formalized it in a spec titled Relative Sensitivity Curve for the C.I.E. Standard Observer. However there are multiple types of vision Photopic , Scotopic , and Mesopic.

    The point being it is a simple single factor conversion between photometric and radiometric.




    So I take that to mean you are saying W/m² cannot be converted to µmols/m²•s without making estimated conversions when crossing geometries.

    The only issue I see with a conversion from watts/m² is the wavelengths from 360nm to 830 nm will stimulate the human visual system and elicit a response from an observer. Whereas PAR (PPF &, PPFD) are confined to 400-700nm. Otherwise the conversion between Lux, W/m², and µmols/m²•s all follow the laws of physics and do not require any estimations. Straight forward calculations.

    So yes, I believe it is a simple conversion. Please comment on the following:


    This is how I make a simple (accurate) conversion from irradiance in W/m² to quantum in µmols/m²•s.

    A photon has a distinct energy quanta Ep which is defined by (Note: the m value for λ is used): Ep= h•f = h•(c/λ)

    (with Plancks constant h=6,63•10-34 [Js]; Speed of light c=2.998•108 [m/s]; Frequency f [1/s]; Wavelength λ [m])

    The number of photons Np can be calculated by (Note: the nm value for λ is used):

    Np= E/Ep= E•((λ•10-9)/h•c) = E [W/m²]•λ10-9[m]•/ (1.988•10-25) [J/s•m/s] = E•λ5.031015 [1/(m²s)] (with Irradiance E [W/m²])



    The photon flux can be determined by:

    EQF = Np/NA (with Avogadro number NA=6,022•1023 mol-1)

    Taken together this leads to the following equation for converting irradiance [W/m2] into quantum flux [µE]:

    EQF = Np/NA = (E•λ•5.031015[1/(m²•s)])/(6.021017[1/µmol]) = E•λ•0.83610^-2 [µmol/(m²•s)]


    In your spreadsheet you factor in the efficacy when you multiply by your "LER" in lm/W. The watts in lm/W are electrical watts. The ratio between lm and watts is the efficacy. You seem to be saying your lm/W the W is radiometric. Otherwise it is the efficacy factor. Your efficacy is very high. The best I know of is the LM301B at about 218 lm/W.

    I see we both have the same problem with quantum intensity and density For intensity I go with cubic meter: mols/m^3
    Also illuminance and irradiance do not convert to photon density. The term is photon Irradiance or quantum Irradiance.


    I believe we are getting some where with this.

    Before I continue, I need you to clarify the "Most common use" conversion you refer to above, is what you are doing or trying to do in your spreadsheet.

    And you disagree my conversion from irradiance to photon flux is more complex as if maybe I am missing the point?

    And you have no use for my "simple" and "not possible" conversion in your spreadsheet?

    Answer just these two points and I think we can clear this all up post haste. I am now thinking the problem is we did not on the same road. I am just about done with what I have to say.









     
  20.  
    GrowLightResearch

    GrowLightResearch Active Member

    The only way to measure the voltage in the strings as individual strings is to measure them separately. Once they are wired together, they all must be at the same voltage BECAUSE they are wired together.

    BTW voltage does not flow throw the string. Electrons flow through (aka current) , The number of electrons that are flowing in the string build up an electromotive potential measured in volts. Basically if you could count the number of electrons contained within the string at any given instant (coulombs) and you could calculate the voltage. A coulombs is the charge transported by a constant current of one ampere in one second.

    Unlikely. A resistor could. That is only given if the system conditions remain Steady-State e.g.current and thermal equilibrium.

    Adding a resistor to the string with the lower Vf will help. A CCR would do a better job. A CCR is like a dynamic resistor. Although a couple of transistors does well also.

    See the second answer by Jack Jack Creasey (ignore Tony, he's way off) here: https://electronics.stackexchange.c...forward-voltage-higher-than-indvidual-leds-vf
    Also look at Jack's comment below his post. Jack knows his shit. He says the excess voltage is due to the imbalance. And when the current raises in the one CoB it increases the Vf to more then when the forward voltage was measured individually.

    When the string with the lower forward voltage (Vf) is forced to the higher Vf in the other sting it conducts more current. Vf would decrease due to the increase in temperature. But it is tied to the other string and so both strings must be at the same voltage causing further stress and therefore altering its normal IV curve. The other sting is going to have a slightly different IV curve. The IV curve being the reason for the difference in Vf from one LED to another. As the string with the lower Vf is going to begin drawing more current it gets warmer. This raise in temperature decrease the Vf. As it draws more current the Vf increases due to the increased current. What can happen is they never reach a steady state. Depending where they are on the IV curve (how much current) determines if a steady state can be reached. The most problematic is when the LED is operating in the steeper part the curve which is higher current. In the steep part of the curve the struggle between the two strings is most volatile because a smaller change in current changes the Vf more then when operating in the flatter part of the curve with lower current flowing. Sometimes the two strings will seesaw back and forth, each time getting a tiny bit closer to thermal runaway.

    When my parallel strings went into thermal runaway I never saw it. It took hours to happen and when it did I was not there in the lab. I could repeat the thermal runaway every time I connected them in parallel. Furthermore I could make either one of the strips fail. I would power up one and it would get all the current and get very hot. Before the temperate did any damage I would connect the other strip in parallel. The board I connect first was the one that always failed. The thing is I could not tell by looking at the there was any problem. Had the strips been in a grow tent being turned off after 12 hours, the thermal runaway may never have be discovered.

    I have never used mid-power LEDs like the Samsung LM561. So I did not know if they had a problem like the strips I had manufactured with high Powered LEs (e.g. Cree XP, OSRAM Olson SSL, and Luxeon Rebel. That is why I bought some this week. I have not yet measured the currents but the Vf of each strip is very close like 10mV difference. Although the Vf changes with the amount of current flowing and the temperature. At 700mA the temperature of the boards reached steady state equilibrium at about 40°C. My boards hit 40°C at 80mA. The forward voltages can differ more than 2V. The CoBs I tested also had too much difference in Vf. The LEDs I'm using max out at an amp so a separate buck step down driver for each strip powered with a CV power supply is not that costly and keeps things simple. The two strips when I put them in use I will drive then in series with a HLG-60H-54A. 60W is almost enough to drive the Bridgelux EBs at max 1.4 amp.

    It appears that the Samsung Series-F and Bridgelux EB strips will likely not have much of an imbalance. I would like to know how much the QB board's Vf differs. I spent $50 on the two pair of strips. I just cannot get myself to buy a QB. I hate the rectangle form factor. The 4000K Samsung strips I can mount under the kitchen cabinets. And warmer 3000K Bridgelux can be used for lighting in the living room. I also planning for a green wall in the living room. I've been making a list of which plants to use. I'm thinking of putting strips on thin cables attached to a stepper motor then time them to turn on and to slowly have the go up and down when no one is around. When not in use the strips would be hidden under a valance. Someone was asking me to design something the would pop the LEDs out of the green wall like a lawn sprinkler pops up. I could not wrap my head around the mechanics for that one.
     
    Last edited: Dec 9, 2017 at 6:24 PM
    Randomblame likes this.

Share This Page