My upcoming LED DIY, and found some 660nm red led's for you DIY'ers

Big Perm

Well-Known Member
I've been doing a lot of research on led lighting technology lately, and decided to make some led light panels to test, but have been having trouble finding decent 660nm red led's.
I'm planning on doing some controlled tests on clones with different ratio's and wavelengths, to see for myself where the sweet spots are, and what the real-life effects on the plants are.
When I'm done I'll make a complete DIY post with all of the information I come up with on what worked and what didn't. Since I am unbiased really in the semi-hot topic on if they are good or not. This will be my first DIY with the Ladies, but I used to do a lot of them on motorcycle and truck forums, since my profession is a mechanic.

Anyway, the blue's are easy to find and are a dime a dozen, along with the other light wavelengths I plan on using (I'm throwing in some 525nm's and some UV led's as well based on some research I did on the effects of a very small wavelength width of UV light and THC, as well as how the different wavelengths effect Chlorophyll A&B),I'm even reading research notes from some NASA horticulture tests using led's, lol.
660nm on the otherhand isn't as easy to get ahold of short of buying some for $.008 including resistors with a 3-8 week delivery time from China. Honestly, I believe there is some false advertising going on there based on the leds they have for sale. Now that 660nm led are gaining popularity in horticulture and skin therapy, I actually saw one company change their description on some red led's from 627nm to the more sought after 660nm with the same part number, lol.

I'm wanting to stay in the 5mm 20-30ma range, but most manufacturer websites send you in an hour long loop in search of a supplier, with the ones that I did find having a delivery time of 3-8 weeks from China.

Ebay is little to no help, but I did find one auction that caught my eye. If you do a search for 660 red led, there is a pic of a clear LED with something about US on the pic, lol. So I emailed them for more information and didn't get anything for a few days, mainly I wanted more numbers than what they had for sale, since I will be building multiple lights for multiple tests. They did send the information a few days later, and said that after Thanksgiving their in-stock numbers will be higher, so that's a plus. A little pricey, but with free 2 day shipping, it's decent IMO, plus my dad always talked about getting what you pay for, and all that.

Long story a little shorter, yet another huge DIY is in progress for those interrrested in it, although I plan on hitting some topics that have been left out, mainly the science on why I'm doing what, not just what I'm doing. So, I'm loooking forward to building my lights and starting what I hope will be a good grow, I'll keep you guys updated.
:leaf:
 

ilikecheetoes

Well-Known Member
I'm in. Keep the info coming. I'm just starting my research. Getting a joint and the soldering iron warmed up. With LEDs costing 14cents theres no way in hell Im spending $400 on a light.
I'm in the middle of a custom grow cab and want to use LEDs for the clone area to save space and heat. And my stoned mind thinks they will fit just perfectly.
The DIY on my opensource enviro controls is about 3 months away but I'm getting there.:joint: Damn I shouldv'e gotten into electronics when I was 12...
 

Big Perm

Well-Known Member
Almost all of the LED's I'm going to use in this newest project have shown up, still waiting on the circuit boards. Once they do, hours of soldering will follow, and the pics will start popping up here.
 

KenWood

Active Member
Hope you take alot of pics, we have been wanting to make one ourselfs but we dont know about how to go about it.
 

withano

Active Member
been wanting to do an LED grow for a while. this will be very interesting. keep us updated.
 

SupraSPL

Well-Known Member
I think efficiency is a consideration when deciding between the 20mA leds and the 350-700 mA ones. Also, if you are using a resistor instead of a constant current driver the efficiency will be very poor (30%?). A good constant current driver will convert 118v AC to variable DC at 80-85% efficiency. Without these advantages you would probably be better off using fluoro tubes.

There is also the consideration of penetration. The higher power LEDs can be lensed to 30 or 60 degrees and will have enough penetrating power to be useful at a reasonable distances. AFAIK the best deals on LED lights that actually work well are from LEDgirl and Prosource. You can expect to pay about $4 per watt. I am a tight wad and even at these prices I am attracted because of the longevity. HID bulbs dim out so damn fast and the magnetic ballasted ones have a terrible power factor (50%?) which is not too good for the environment.

All that said I would like to at least calculate how much it would cost to make our own lights. Meanwell drivers at 700mA can put out 36 watts for $25. Consider the cost of the 750mA leds, the heatsinks and the thermal compound and see if you can beat $4 per watt. Even if we could, I believe that LEDgirl Hydro Grow and Prosource lights run their LEDs at 350mA which is more efficient in some cases. They are also more compact because they are actively cooled.

Long story short I'm not sure if we can beat their prices even with DIY.
 

Big Perm

Well-Known Member
I'm starting the build now. Have everything I need, made my layout on paper, and am building a prototype. I'm going to use it on one plant from seed (KC Brains, Brain Damage Fem) just to see if it is even possible to get it to flower, and if so what the results will be. I have just 4 of the 200 660's installed, and right now it's too bright to look at directly, my eyes hurt after just a couple seconds. All said and done it'll have a combined total of 576 LED's, and will pull 28.8 watts.
I'm using resistors, and am testing it with a motorcycle battery I have on the trickle charger for the winter. When I'm done I'll use a 12v 3A AC Adapter.
It's going to take some time to make, I estimate a couple weeks to a month. I only work on it in the spare time of my spare time.
I'll update here with pictures. Not too many of the build probably, maybe one halfway done, and another with it done. Once the seed hits the miniature hydro setup I made just for this project I'll post about a picture a week, and we can all see what $40 of LED's will give us when it's done.
If it works, sweet. If not, I'll throw the unflowered fem in my HPS hydro to finish it off. I can probably get half my money back for the thing on eBay and aim for a better project.
I understand what you are saying Supra, and agree. No doubt bigger and more powerful is better, but I have to make the cheap version though, and do a test for a proper R&D project, and at least see what the RL results will be. I haven't came across anything about anyone making one of these (that knew what they were doing), so we have nothing to compare anything to.
 

KenWood

Active Member
I'm starting the build now. Have everything I need, made my layout on paper, and am building a prototype. I'm going to use it on one plant from seed (KC Brains, Brain Damage Fem) just to see if it is even possible to get it to flower, and if so what the results will be. I have just 4 of the 200 660's installed, and right now it's too bright to look at directly, my eyes hurt after just a couple seconds. All said and done it'll have a combined total of 576 LED's, and will pull 28.8 watts.
I'm using resistors, and am testing it with a motorcycle battery I have on the trickle charger for the winter. When I'm done I'll use a 12v 3A AC Adapter.
It's going to take some time to make, I estimate a couple weeks to a month. I only work on it in the spare time of my spare time.
I'll update here with pictures. Not too many of the build probably, maybe one halfway done, and another with it done. Once the seed hits the miniature hydro setup I made just for this project I'll post about a picture a week, and we can all see what $40 of LED's will give us when it's done.
If it works, sweet. If not, I'll throw the unflowered fem in my HPS hydro to finish it off. I can probably get half my money back for the thing on eBay and aim for a better project.
I understand what you are saying Supra, and agree. No doubt bigger and more powerful is better, but I have to make the cheap version though, and do a test for a proper R&D project, and at least see what the RL results will be. I haven't came across anything about anyone making one of these (that knew what they were doing), so we have nothing to compare anything to.
Just keep us informed. My ? is how to come up with circuit board layout and power.
 

Big Perm

Well-Known Member
Just keep us informed. My ? is how to come up with circuit board layout and power.
This is one way of doing it, not the easiest, but the cheapest, and safest IMO. Again, all of this is my opinion, and everyone has one. If you don't agree, or think I'm doing it all wrong, make your own thread about how you are making yours, not 'how you would do it'. Thank you. If you are using this as a walkthrough, you might have to read it a couple times.

Well, LED's hooked up in series is the only way to go. The LED's hooked up in series means your linking a few LED's in a row (all the exact kind/ type of LED) so that when you are done you have a positive wire on one side of the string, and a negative wire on the other side of the string. Each LED is hooked up to the opposite polarity of the other, so if linking 3 LED's together you would have a positive wire hooked up to nothing as of now from the first LED, the negative wire of the first LED hooked up to the positive wire of the second LED, the negative wire of the second LED hooked up to the positive wire of the third LED, and that will leave you with the negative wire from the third LED hooked up to nothing as of now. So now we have a positive wire free, and a negative wire free, with 3 LED's in the middle on the example above.

Before you hook this string up to a power source you need to limit the current that will pass through the string you just made, or else they will just blow when you hook them up.
Also note that when you hook them up in series, the mA will remain the same and the voltage will add up.
To find out what resistor you need, you first need to know the power demands of your LED. Here is an example:

For my prototype, I went with 5mm LED's, but these same rules apply for any LED.
Lets say you use a 2.0v 25mA LED (you can insert whatever numbers your LED's are in place of these numbers) different color LED's use different voltages, as well as different kinds of LED's of the same color.
Our supply voltage will be 12v
I like to leave about a 2v cushion in the series, so we can get away with using 5 of these LED's in series. 2 volts x 5 LED's = 10 volts with a 2 volt cushion.
To find out what resistor to use you would take your supply voltage (12v) subtract the volts the LED string is going to use (10v), which will leave you with 2 volts left over.
Now that we know all of these numbers, we can do the math for the resistor needed. The way we would do that is to use ohm's law, but in a nutshell it goes like this:
You have 2 volts left over, so take 2 volts and divide it by the mA the LED's are going to use. In series the mA stays the same, so it's still 25mA, or 0.025 on a calculator. 2/0.025= 80. so you would want to use an 80 ohm resistor for each 5 LED string in this example.
Resistors come in standard sizes, and resistor prices are all over the place based on how common they are. For example 100, 330, and 470 ohm resistors are really common.
You can change the math to meet the mA demands of your LED's based on the cheapest resistors you can find. I bought my resistors on eBay 100 of them for $1. Just run the formula backwards to see if a certain resistor will work, you'll have to change the number of LED's you use in series to change the resistor needed.

You want to use 1/4 watt resistors (really common).
So lets say you get a great deal on a huge package of 330 ohm resistors...what can we hook up to it given all of the information I just talked about, using the LED's in the example?
Let's try 2 LED's.
2 LED's x 2 volts = 4 volts.
12v (power supply) - 4 volts (LED's) = 8 volts.
8 volts divided by 330 (the resistor)= .0242424.....
So if you used a 330 ohm resistor on (2) 2 volt LED's with a 12 volt supply, it would push the LED's at 24mA....close enough for me.

You can also run LED's a little hot and get away with it. I wouldn't push a 20mA LED past 25mA, nor a 25mA past 30mA. You don't want to go lower, because the light output will also be lower. If you run them hotter the light output will also be higher. The LED's probably won't last as long (maybe 80,000 hours instead of 100,000 hours lifespan....80,000 hours is still a long damn time).

Now that I covered what resistor to use, and how to find out which resistor works for you, you can install it in your string of LED's.
The resistor goes on the positive side of the string so that your power supply connects to the resistor, the other wire of the resistor hooks up to the positive side of the LED string, and then the negative side of the LED string hooks up to the negative wire of the power supply.
When you turn the power on, the LED's will light up.
-You can hook the resistor up either way, there is no positive/ negative on a resistor until you hook it up, so it is impossible to hook it up backwards.

A mistake people use when figuring up how many watts they are using when they are done is that just because they are using 1/4 watt resistors, doesn't mean they are using 1/4 watts per resistor. "A 1/4 watt resistor, I am using 100 of them, so I'm using 25 watts, right?" No.
You have to go back to your formula you used earlier to find out how many watts you are pulling, and it is a lot less than 1/4 watt. In the last example we found out that if you use those 2 LED's with a 330 ohm resistor you're using 24mA per string. Also, we're only pulling 4 volts.
You multiply the volts times the amps (in this case milliamps, mA) to get the watts.
4 volts x .024mA = 0.096 watts
So if we round up to make it a little simpler, we're only using 1/10th of a watt per string, not 1/4. So every 10 strings, you're using 1 watt.
20 LED's running on 1 watt is not too shabby, and is a LOT brighter than you would think 1 watt at 12v is capable of producing.
4 running on .2 of a watt is so bright I can't look directly at the beam.

When you are all done, you'll need a power supply. Since we are regulating the current very accurately (if all of your math was right), we just need a power supply that is big enough to provite the volts and amps. The one I am making will pull 28.8 watts, and I want to use 12 volts. 28 watts divided by 12 volts = 2.4 amps. So a 12 volt 3 amp (36 watts) AC adapter will work fine, the excess will be regulated by the resistors.
Chances are you have a few of these laying around the house already. Old cell phone chargers, toy chargers, anything with a black box you plug into the wall is what you are looking for. On the adapter itself, it should say it's rating.

Keep in mind you have to do this for each kind of LED type you are using. Red and Blue LED's use different voltages.

You might get a good deal on some resistors that comes out to you only using 1 LED per resistor. Check the prices in your formulas and go with whatever is cheaper, or easier. In my setup i'm using some with 3 in a series, another kind of LED with 2 in a series, and others with only 1 per LED. I'm using 6 different spectrums, so mine are all over the place.

Hopefully this answers a lot more questions than it creates.
I had to sober up to write this, and now that I'm done it's time to roll a J.
:joint:
 
Hi there,

I'm really excited to see what you are coming up with. I too have been doing a ton of research into making my own LED lights.

What leds did you finally end up going with? I am interested to know so that I can see how they compare to what I think. I was going to use a 430nm, a 450nm, a 640nm, and a 660nm. Is this what you are going with, or something close.

Any more details on what parts you ended up using would be great. Also, thanks for the detailed posts so far.

M
 
Top