-
Do you need help identifying a 🌶?
Is your plant suffering from an unknown issue? 🤧
Then ask in Identification and Diagnosis.
Yet Another DIY LED thread
- Thread starter BillyIdle
- Start date
Yup, I think PAR is what we're after, same as with my old coral reef tank...
I did the math on the 300w panel using rough estimates of what equivalent watt & spectrum LED's put out - that's not what the manufacturer claimed Gary. That said I think I did the math right, it's only multiplication... but who knows, I've made worse mistakes in the past
I think the thing to think on here is efficiency of the light - however you want to define efficiency. In this case if you look at the spectrums that chlorophyll use: http://en.wikipedia.org/wiki/File:Chlorophyll_ab_spectra.png
And match it up with LED's in the unit vs the spectrum in HID/Flouro... I think perhaps there is a valid argument that LED's are more efficient as they provide energy in the right spectrum instead of "wasting" energy in the wrong spectrums as HID/flouro's do (and I guess there could be a somewhat valid argument that some of the other spectrums are needed and LED doesn't provide them).
Anyways, I think the calculator Gary used is doing something we aren't aware of - calculating comparable lumen output perhaps? Like how it can be -1 out but with windchill feel like -25? I'm no scientist, just a computer geek with some odd hobbies.
I did the math on the 300w panel using rough estimates of what equivalent watt & spectrum LED's put out - that's not what the manufacturer claimed Gary. That said I think I did the math right, it's only multiplication... but who knows, I've made worse mistakes in the past
I think the thing to think on here is efficiency of the light - however you want to define efficiency. In this case if you look at the spectrums that chlorophyll use: http://en.wikipedia.org/wiki/File:Chlorophyll_ab_spectra.png
And match it up with LED's in the unit vs the spectrum in HID/Flouro... I think perhaps there is a valid argument that LED's are more efficient as they provide energy in the right spectrum instead of "wasting" energy in the wrong spectrums as HID/flouro's do (and I guess there could be a somewhat valid argument that some of the other spectrums are needed and LED doesn't provide them).
Anyways, I think the calculator Gary used is doing something we aren't aware of - calculating comparable lumen output perhaps? Like how it can be -1 out but with windchill feel like -25? I'm no scientist, just a computer geek with some odd hobbies.
The thing is that there is some multiplication involved in the formula. The wider the beam angle is, the more lumens you get. I don't know if I'm able to explain this properly, but I'll try. My theory is that the LED manufacturers does the MCD measuring with a very narrow lens (+/- 3-4 degrees) which gives very high MCD, but the product they sell have a wide angle lens (+/- 20-40 degrees), still using the MCD from the narrow lensed product in the product sheet.
I have only seen these insane high MCD specs on fleabay or other "direct from factory" web sites. I checked the catalogue of a very serious electronic component reseller (this company mostly sells to the professional market), and I could not find any LEDs < 1W that had a MCD spec higher than 2-3000.
I have only seen these insane high MCD specs on fleabay or other "direct from factory" web sites. I checked the catalogue of a very serious electronic component reseller (this company mostly sells to the professional market), and I could not find any LEDs < 1W that had a MCD spec higher than 2-3000.
Pimp my LED
I decided to test run the new setup, now as the competition fataliis should sprout any day now (or should have sprouted a couple of days ago really). As I suspected, the new lamp pattern is unhealthy regarding heat. After maybe 3-4 minutes, I could hardly touch the back side of the board they're mounted at.
Ok, how to cool them? I have multiple PC fans lying around the house. The only problem is that these are not rated to run at the 20V output by the power supply. I looked around and found several candidate powers, but in the end I wasn't too pleased with two power supplies for one panel.
Then I tried to just run one 80mm fan directly on 20V. It worked really, really well, but I was afraid how long it would last at that speed. The panel got more than cool enough. But if the fan malfunctions, I might end up with a ruined panel.
The luck was that I got a pack of resistors together with each pack of LEDs. These are 470 ohms, and some math told me that 6 of these in paralell would do the trick, giving 78 ohms, and each resistor will dissipate safe levels of power for it's rating. The fan now gets approx. 13V, which is also safe. The end result looks like this:
I decided to test run the new setup, now as the competition fataliis should sprout any day now (or should have sprouted a couple of days ago really). As I suspected, the new lamp pattern is unhealthy regarding heat. After maybe 3-4 minutes, I could hardly touch the back side of the board they're mounted at.
Ok, how to cool them? I have multiple PC fans lying around the house. The only problem is that these are not rated to run at the 20V output by the power supply. I looked around and found several candidate powers, but in the end I wasn't too pleased with two power supplies for one panel.
Then I tried to just run one 80mm fan directly on 20V. It worked really, really well, but I was afraid how long it would last at that speed. The panel got more than cool enough. But if the fan malfunctions, I might end up with a ruined panel.
The luck was that I got a pack of resistors together with each pack of LEDs. These are 470 ohms, and some math told me that 6 of these in paralell would do the trick, giving 78 ohms, and each resistor will dissipate safe levels of power for it's rating. The fan now gets approx. 13V, which is also safe. The end result looks like this:
lostmind said:
Thanks! But remember, the picture was taken from the good side. It's actually quite ghetto style with loose cords and stuff. But it works
Hotpeppa said:impressive dude....
wish i stayed in school...
Thanx! School is a very nice thing, but you don't learn this there
Ok, I admit I learned about Ohms law in school, but that was back in the early 90's. Internet is a good thing though.
Gary18 said:
I'm quite sure that yours would produce at least the same amount of heat if you stacked them as tight as I have done. Mine doesn't seem to produce any heat either when spaced properly. All these semi-high power LEDs produce some heat, but in my setup it has nowhere to escape.
Ok, I gotta ask a couple of questions.
You managed to run these LEDs without a current limiting resistor in series with them? Do they have a built-in resistance so they won't create a near short? The ones I'm familiar with can't handle ~5v without some type of external current limiter.
Did you wire the blues and reds separately to meet different ideal current sepcs? Or are they combined in series like my cheap pre-made boards? I wish they hadn't done that. [ooops, just reread the part about having the red and blue separated... nevermind]
Did this end up being cheaper than buying a pre-made board? Yeah, I know the projet was fun, but is it worth the extra work after the fun factor wears out?
You managed to run these LEDs without a current limiting resistor in series with them? Do they have a built-in resistance so they won't create a near short? The ones I'm familiar with can't handle ~5v without some type of external current limiter.
Did you wire the blues and reds separately to meet different ideal current sepcs? Or are they combined in series like my cheap pre-made boards? I wish they hadn't done that. [ooops, just reread the part about having the red and blue separated... nevermind]
Did this end up being cheaper than buying a pre-made board? Yeah, I know the projet was fun, but is it worth the extra work after the fun factor wears out?
caroltlw said:
Yes, I run them without any resistors. It's said to be a bad thing, but it seems works for me. What did was getting my hands on a PSU that matched the LED's typical forward voltage. The PSU is 20V and the red LED's are specified to 1.9V and the blue ones at 3.3V. By connecting 10 red ones in series, each of them will get 2V each, and 6 blue ones gives 3.33V for each of those.
caroltlw said:
Because of the import tax system in Norway, the DIY wins. You have to pay 25% tax of the price of the product and shipping costs combined if the cost of the product itself exceeds 200 norwegian kroner (approx. $29). In addition you'll be charged a fee for the calculation of the tax that rates to about $15. Each of the components used came below this. In addition I got a laptop PSU for free.
But I can absolutely assure yoo that it was the fun factor that carried this project. It was a lot of work! Guess I put something in the ballpark of 15 hours of work into this. Maybe even more.
If I build an upgrade to what I have I think I'll use limiting resistors. My voltage varies from 11.5 to 14.5v, so I can't count on pure math to do all the work.
One thing that really makes me want to build my own next time is looking at how the ones I bought were designed. I mean they're running red and blue in series with each other and there's no way they're both supposed to be optimal at the same current. I also think I can dimprove on the lumens with a bulk purchase of some super-brights. And I still think maybe some of those come with some added internal resistance, thus the increased heat you're getting.
I'll have to look around for exactly what I need. There's so many confusing and contradictory spec sheets out there.
One thing that really makes me want to build my own next time is looking at how the ones I bought were designed. I mean they're running red and blue in series with each other and there's no way they're both supposed to be optimal at the same current. I also think I can dimprove on the lumens with a bulk purchase of some super-brights. And I still think maybe some of those come with some added internal resistance, thus the increased heat you're getting.
I'll have to look around for exactly what I need. There's so many confusing and contradictory spec sheets out there.
I'm not sure that it's a bad thing to run reds and blues mixed in series. It's the current that is the deciding factor here, and the voltage drop over the LED is derived from the current. As long as the recommended forward current is the same for both colors (which is normally the case), you're good to go. If you take my LEDs as an example: They are specified to draw 3.3 and 1.9 V at 25 mA. If I connect 3 blue and 1 red in series, the whole circuit should use 11.8V at 25 mA.
In an environment such as yours where the voltage varies with 3V, I'm not sure how to control the current with one resistor.
In an environment such as yours where the voltage varies with 3V, I'm not sure how to control the current with one resistor.
BillyIdle said:
Could connect a transistor inline and set it to limit current. Link is about balancing current in two parallel loads of LEDs.
http://powerelectronics.com/mag/701PET07.pdf