'Redline's LED - DiY - Build' aka"GROW LIGHT"

**veggii** · 04-20-2008, 03:13 AM

Hi everyone I am veggii

veggii Reviewed by veggii on 04-20-2008. 'Redline's LED - DiY - Build' aka"GROW LIGHT" Hi everyone I am veggii ;) I have been working on LEDs and have been over in the Led Theory Thread and A little bit ago REDLINE joined up and posted a awesome LED BUILD I thought it deserved it own Thread :thumbsup: I hope to keep working with Redline as his style is good. His builds are simple $ cheap & practical you will like them too. I am going to post a summary of redline's post's from the Theory thread. That should get everyone started and up to speed. Hopefully Redline will get Rating: 5

**veggii** · 04-21-2008, 04:27 PM

Redline , so you left off @ you were still in the middle of the build process for the 25watt 660nm Led Grow light.
Have you gotton it done ? how is the heat!?
I sure hope everything is working out as haven't seen you in over a week. :jointsmile:

I don't know why the plant chart is not posting proper

**veggii** · 04-21-2008, 05:25 PM

Hi,
I have a bit of free time today so I was going too try to
repost the thread better. But it seems the pics won't show up
on the post's.
You can see from the above plantchart.bmp pic that it is not posting. I am going too test a few more pics rite know.
If they don't post proper I will end this thread as it needs pic's

ok the pics are working its just the above plantchart that is'nt posting proper

**veggii** · 04-21-2008, 05:46 PM

pg14
led info
The current through an LED is established by the voltage drop across it. But here is the tricky part.. the relationship is not a predictable constant and it is not a linear relationship. The voltage drop required to produce the desired current will vary from LED to LED even in the same bin. It will also change with operating temperature and age of the LED.

I have been designing,hand building and using high power LEDs to grow for almost a year. When I build an array, I burn it in for two weeks powering it by a bench constant current power supply. I then measure each series string and set the current limiting resistor (or LM-type regulator) to the exact value needed for that string.

Another tip: If you are trying to achieve energy efficiency you should be driving them at 350 ma instead of 700 ma. If you are trying to get the most light output for your money, then go the 700 ma. The high end commericial lights all drive at 700 ma in order to cut material costs. If you double up on LEDS used, drive at 350 ma, you will pick up about 20% better energy conversion efficiency. Plus you will have twice as many points of light.

I ordered up 20 of the Ledengin 660s to integrate into my setup as soon as I can find the time. I will only be using them on a portion of the crop so I can get an idea of the difference if any. I will be removing an equal wattage of 630s to get a fair comparison.

If you are trying to pick up some 660 energy,you might also try using the K2s instead of Crees for your 630s. They have a much wider bandwidth and only have about a 60%-70% drop from peak at 660. Looking at the SPD charts, the Crees ain't got nothin at 660. Also you might try using warm whites since most of them put out ok power at 660. Probably the best one at this time is the SSC P4, but the K2s aren't bad.
pg15
Thanks for the links, found some really useful info.
If you want to build your own drivers, you got to check out National Semiconductor, High-Performance Analog for Energy-Efficient PowerWise Designs. They got some really slick design tools, you can even order up a prototype kit for your design.

i-- let me know what you want to do. I can help you with any electronic design issues and step by step construction (with pics) and parts sourcing.

I got lots of practical hands-on experience and have done mucho research. The 24 volt drivers used in the sign and lighting industry are a good way to go but can get a bit pricey.

I am not much of a computer guy, but I heard somewhere that some of the older Macs had a 24 volt 3 amp source on their power supply. Anybody got any info on this?

I took the time to read your post a little bit closer.
If you mean the "Advanced Transformer" 60 watt Xitanium driver, that is a very good choice.
I strongly recommend running the LEDs at 3 watt instead of 5. You will get better energy efficiency, longer life, and your thermal issues will be much easier to deal with. 1 watt would be optimal, but those little buggers are just to damn expensive to run at that power level. I personally run the 660nm LEDengins at around 3 watts.
First, you are going to need 9 square inches per watt of heat sink exposure to air. 36 watts is going to need 324 square inches. You count both sides of all the fins and top and bottom of platform when making this calculation.
There is a guy on ebay that sells them cut to your size at very reasonable prices. I would go with about a 4.5 "x 12" and use a to a couple of computer fans on top.
Unless you want to order in another 4 LEDs, which would be a better choice, you are stuck at running 12 volt instead of 24. If running 12 volt, just buy a PC power supply for $10 to $20 bucks. Assuming 2.9 volt drop for 1 amp of current, you will need to run 3 parallel chains of 4 LEDs. You will want to use a .4 ohm resistor in each leg with minimum 2 watt minimum rating.
If I was only running 12 LEDs, I would consider making 3 small units instead of one, so I could be more flexible where I put the light

**veggii** · 04-21-2008, 05:55 PM

pg16
--Those are good prices on either the Xitanium or the GE. It wouldn't surprise me if they were made at the same place with different labels. Either one is a good bet and the prices are very good... jump on them. I use a 25 watt Xitanium as one of my test supplies.

While you can physically layout the leds in a single string, they will actually have to be electrically connected in 2 parallel strings of 6 to 8 LEDs depending on desired wattage. I would have a positive feed point in the middle of the layout and the negative on each end.

You can put the 2 resistors on either end or both in the middle.
5 watts is the max rating and gives you no margin of error on your voltage control plus you must have superb thermal management, plus you really shouldn't let your grow area temp get over 75 degrees. In order to run at 5 watts you will need a $180 digital bench power supply in order to get the 2 line resistors set up right on the money. You will then have to burn in each LED string over a period of 3 to 4 days while monitoring voltage changes at a constant current and also monitoring temperature. After the string has stabilized and you take measurements, you will probably have to make a custom resistor combo in order to get the precise value needed.

If you decide to run in the 3 to 4 watt range it will be much simpler to set up without the danger of wiping out $120 worth of LEDS. In reality you are probably only going to get about 10% more light at 5 watts then you would at 4 watts and will avoid the need for the expensive test equipment and testing.

Your first step is to get your hands on a heat sink. What is the desired and maximum length you want your array? I will send over the heat sink link from Ebay after I get back from dinner.

tetra info
The best configuration for the Tetra is 3 parallel strings of 8 LEDs.
Giving you about 72 watts to the array (3 watt per light).

If you want to run the lights at a higher rating of 4 watts you would be limited to 2 strings of 7 lamps. Giving you approx. 56 watts total for the assembly

Here's some pics of a 25 watter I just knocked out using K2s. I will be using different nm combos in seperate areas of the grow to try and find out what wavelengths are most effective.

------ I am starting on a 660nm array. If you want I can post pics and details of construction of progress, if you think that will be of help to you.

Both supplies are 120 ac in and 24vdc out. The big one is 180 watt.

The link for the heat sink is HEATSINK ALUMINUM EXTRUSION 8 1/2 WIDE PROFILE - eBay (item 170209048370 end time Apr-15-08 11:05:34 PDT)

So many theories floating around on color combos. I have even read one paper, that says plants will readily adapt to the wavelengths that are available to them. I have no personal opinion on what is going to work the best, so I am going to start experimenting, trying this and that.. It will probably take about a year or two before we start getting it really dialed in.

-----, I will start on the 660 nm strip array this weekend and post pictures of progress.

I got some 660s and am starting on some lamps this weekend.
I am able to set up 4 different grow sections to try different light combos.

Anyone know where to get high power 435s. I am currently running cool whites in order to pick up some light in that area.

Don't worry about replacing individual LEDs. By the time they burn out they will be obsolete. 2 or 3 years from now we will be getting our hands on the good stuff. No more HPS.

**veggii** · 04-21-2008, 05:59 PM

pg17
Here's the catch 22. If you have the background to swap LEDs, you will probably be building your own units instead of using commercial ones.

Remember each kind of LED has a different voltage drop/forward current curve. The V-drop can even vary in the same model LED depending on the mfg. bin. Also good heat sinking does not readily lend itself to easily switching out LEDs. Plus you need fast and good soldering skills.

I looked at the possibility of being able to change LEDs when I first starting building. The only practical idea I could think of was to mount an entire series string ( or a portion of the string) on a strip of copper or aluminum and fasten it to the bottom of the heat sink. I didn't think it was worth the effort especially considering it would be introducing another thermal junction

I have been using mostly K2s since efficiency is better then the IIIs and most Rebels. I do use SSCs for my cool whites. I have noticed a price drop of 30% this month from Future Elec.

To my knowledge most of the medium to large LED drivers are primarily voltage regulated at 24 V. I am only seen current regulated in the smaller DC to DC bucks and boosts. If you look at the Xitanium part numbers, the 4th character from the right designates voltage or current control. That is why I try to leave about 1 to 2 volts to drop in each series string and use a resistor.
I don't know if that is overkill or not, but at least it gives me a convenient way to monitor individual current to each string without going to the hassle of inserting an ma meter.
------- found some 60 watt drivers on Ebay for $20 plus shipping.

you might want to recheck your physics texts regarding
"Closer is not actually good because light only drops off with distance if the inside of the room is absorbing light".
Everytime you double the distance from a light source, it diminishes by a factor of four, whether it is in a grow room or outer space. Just play around with a light meter to verify.

At this stage of LED development, I believe the only significant edge LEDS have over HID is the ability to get very close to the plant. The only disadantage to getting closer is your area of coverage decreases. I put mine 4 to 6 inches max above the scrog.

If you run at 4.5 inches that gives you a 16X advantage over an equivalant light source run at 18" which is typical of HID.

more corrections.
Dont mean to beat up on you, but If I don't Physicsnole will.
The top of the UV range is 400nm. 435 is some of kind of blue and it is one of the peak absorption points along with 660nm for Chlorphyll (either A or B, I can't rememberwhich). That is why everyone is trying to hit those wavelengths.
The big question will it be worth the effort in practical results?
By the way, I have heard the same thing from good sources that UV helps the quality but not the quantity.

Something everyone needs to keep in mind is why are we messing around with LEDS. I think the obvious reason is to get more light energy per watt of household current. I can understand why everyone wants to drive them at max currents because they are so expensive but that is really defeating the purpose of what we are trying to achieve.

If you want to make better units then the Pycron or the Megablaster 2000, you can't drive LEDs at their max rated currents. Look at the LEDengin red 660nm for example. When you increase current from 700 ma to 1000 ma, power is increased by 56%. However, output only increases 25%. That means that the additional watt output is only 50% as efficient as the first two watts. If you used that extra watt instead to power 1/2 of an additional LED, you would be getting twice the light output from that watt.
And from a practical standpoint it becomes much harder to keep the LEDS cooled to a normal level when you push them towards max rating, which causes the output efficiency to drop a bit more.

I'm not buying that about the reflectors . Maybe Physicsnole can offer an understandable explanation, I'm not that swift with physics.
My thinking is that the reflected light energy has to travel an additional distance to the relective floor or wall with its energy decreasing by the square, take a really big hit with the relective inefficiency of whatever reflective material is used and then lose more energy traveling to the plant. Plus it is going to be tough getting the angles of reflection where they bounce right back at the plant. Most of the reflected light is probably going to miss the plant and hit another reflective surface.
I don't use focus lenses on the LEDs, so I have a typical 140 degree view angle, I do use a small hood on the edge lights in order to keep a consistent pattern.

If I had enough leds to give me good coverage, I would have them almost touching the tops

Yes, those are the two trade offs: cost or energy.
Each person will have their own criteria. I lean more towards the energy efficient, not so much for saving energy but to maintain the a good profile with my electric bill. Also in the spirit of research, since the goal most people are working on is trying to make them more energy efficient then HIDs.

Trying to make LEDs $$$ efficient is a losing battle at this time. HPS is so much cheaper. Of course in a year or two or three, it will be a different story. We are all just playing around now trying to figure out the best configurations until we can get our hands on the "good stuff" that is hopefully in the pipe line.
However, I do see your point. I am driving the LEDengin 660nms at
over 4 watts instead of 2.5 because they are so damn expensive.
When I am spending $2 to $3 on K2s, not that big a deal. On the K2 blues if you run 100 at one watt instead of 50 at 2 watts like the Procyn, you will put out about 25% more light from the array, with the added benefits of twice the light sources and easier to manage temps.

If you have a small grow it is not worth the expense of doubling up on your LEDs to get the extra efficiency. If you are running over a KW, energy efficiency starts to be an issue.

I am going to play around with some water cooling designs in a few weeks and see if that can prove useful.

Thanks for the link: I was going to solder some copper tubing to copper plate screw it to the bottom of a heat sink and mount the LEDs right on the copper tubing, but those are cooling plates are muy bueno.
I bet those puppies are expensive, they sure look nice. Have you seen any pricing? I guess you still have to run a bong cooler in order to get below ambient temp or have some other way to cool your water.
However, on my 25 watt arrays. They only run 2 or 3 degrees hotter then ambient temp with the fans on, so I am not really feeling much pressure to water cool.

If you are using K2s, dont sweat the PCB. I just attach the bare emitter directly to the heat sink using Artic Alumina Adhesive. It is not electrically conductive. It looks a little funky having connecting wires on the bottom of the light, but I really didn't want to screw around with PCBS and you get a very direct thermal path.

I only use the 25 watt Xitanium for test purposes. I got a bunch of
180 watt, 24 volt drivers used for the sign and lighting industry. I posted a picture of one a day or two ago. I run six 25 watt arrays per drivers. I do have some extras that I will be disposing of soon.

I still don't get it about the reflective room and the distance. You guys are saying that if I had a perfectly reflective room I could turn the light upside down and it wouldn't make any difference or move it an infinite distance away as long as it was inside a perfectly reflecting sphere. How come you are not taking into account the extra distance the reflected light has to travel?
Lets say you got a bulb that emits light evenly in all directions and it is in the exact middle of a sphere coated with a material that reflects with 100% efficiency. Lets put the target at the out edge of the sphere. All the light that does not hit the target directly is going to travel a longer distance before it hits the target bouncing off at least one surface but most likely multiple surfaces. That light will be diminished by the inverse square law.

That is why you want to get those lights right on frikken top of the plants, not 18 inches away.

I presume you are using rebel stars instead of the bare rebel emitter?
If you are using the bare emitters you can get a pcb from Asia Signals to mount them on. I believe you do have to have a hot air setup to solder them, I don't know if an iron would work. I have been wanting to try some Rebel emitters with the Asia Signals PCB, since the emitters are cheap and some colors have better efficiency then the K2s. The Rebel stars are just too darn expensive at this time.

**veggii** · 04-21-2008, 06:05 PM

pg18
think I will stick to the electronics end of things, the light wave physics is starting to hurt my head.

Re Procyn power supply. Ryan is smart enough to design one from scratch. My guess is that it is a one-off design. However, if someone with a Procyn is able to pop the hood and take a look, please let us know.
Physics, you are a brave man working with Rebel emitters. I would suggest not reinventing the wheel by making your own PCB. Check out AsianSignals Controllers. Please order some, I need a guinea pig to see how good they work.

I will be finished with the 660nm setup later today and will post construction pics.

didn't quite get finished on the 660 nm array today, but have some pics that might be of interest to anyone interested in building one. I always like to scrounge through junk that I accummulate at swap meets and had a bunch of these cheapy 2x2 inch computer fans kicking around. I decided to mount them on a piece of copper that I will cut to accomodate 7 LEDengin 5 watt 660nms.

The bottom of the heat sinks had a lot of scratchs plus I suspected they weren't perfectly flat. Having a good smooth surface with out high and low spots is essential to good heat transfer across a junction. I used a magic marker to color the bottom of the heat sink. I then laid fine emery on top of the cut copper bar. Lightly rubbing the bottom of the heat sink across the emery I was able to see the high spots.
I then leveled them out using the emery paper and polished with crocus cloth. I did a final check of the heat sink surface by coating it once again with Magic Marker and then applying a very light coat of fine automotive valve grinding coupound. I then rubbed directly on the copper plate in order to get a good mating surface.
After both the copper and heat sink were free of flat spots I eliminated the fine scratches with metal polishing compound.

This is the most tedious part of the construction, but is probably well worth the effort to keep things cool

This is the fun part. I did a layout to see how the components would fit and then drilled and threaded the mounting holes.
I decided not to mount the fans with screws since I was in a hurry but to fasten them on with Artic Alumina Adhesive. I don't know if this was a good idea or not. I will know when I start doing thermal testing.
Caution, if you are working with copper, go real slow with the small taps and back them out every quarter turn. Copper likes to grap taps. I managed to break off two 4-40 taps.

Everthing is now mounted and ready for wiring and LED installation.
I decided to go with a couple of aluminum side reflectors to keep the pattern tight. I prefer this to using focusing lenses.

Question: Does anyone know if the bottom of the LEDengin stars is electrically isolated. Their emitters are not isolated but it seems like the base on their stars are isolated. Nothing ohmed out. I could find no info in their data sheets. Would like a confirmation from someone.

I will post the final steps, along with testing in a few days

Thanks for the link. I had read them but couldn't find anything regarding stars. I am pretty sure that they are not conductive. I am going test them after I mount two of them. I will use grease and screws instead of insulating adhesive. Just hope they don't explode in my face.
No plans for EBay, they take too darn long to make. Plus shortly there will be a flood of Chinese knockoffs hitting the market at low prices. Someone is already copying the Procyn even before they have proven successful.
I'm just having fun playing around and experimenting I have made a few for friends. I try to make each one different and improve upon previous designs. I save a lot of metal scraps and electronics parts and try to design with what I have on hand.

Did you already buy your 660nms?

Physics, I have no idea what it is going to dissapate. My design criteria is to have at least 9 square inches per watt for passive heat sinking. But I like to run fans regardless, just as a back up for that extra edge and to add to room circulation.
Since these LEDs are so damn expensive, I will try to run at plus 4 watts each and will test all the way to 5 watts. I will give complete details on my thermal testing methods once I get to that stage.
To be honest, I probably screwed up by mounting the fans with adhesive instead of screws. I could not get the surfaces to mate as good as I wanted. I should have used 6 to 8 screws on each fan and torqued them down. Plus grease (Artic Silver) is a better thermal conductor then the adhesive.
I was just not looking forward to drilling and tapping 30 to 40 more holes. I spend so much time making these things, I am always looking for shortcuts. Since I was concerned, I did use aluminum instead of my usual galvanized steel for the reflectors in order to help with the heat sinking. Picked up an extra 90 square surface inches. I will probably keep using aluminum.

Lots of questions. I looked at the high power LED drivers and would be curious about price. I noticed they are 12 volt which is a little weird. The lighting industry is pretty much gravitating towards 24 volts. The more you can string in series the better.
Holy Crap, that 500 watter can belt out 40 amps. You are going to need power line cable to hook all your arrays and some heavy duty connectors. I really don't see how it would be workable.

For most people 100 to 250 watt power supplys running at 24 volts are probably the most practical answer.

Veggi, yes, I got some extras I want to get rid of. I had to buy a complete lot. Lets figure out a way to hook up and I will get you details. I don't want to get black balled for selling stuff on the forum.
The nice thing about LEDs is they lend themselves to a wide variety of configurations to fit a variety of situations. That is why I have been custom building them.

Next question,
Are 660nms worth the expense and effort? I will let you know in about 3 to4 months. I will be starting some comparative tests next week. As you have probably figured out, I am not much for theory, but more of a hands on guy.
I visited a lot of forums and read a lot of papers about plants and light waves. So many different opinions floating around but no one is testing them. It is probably going to take at least a year before we figure out what are the most efficient light combos.

Yes, I do use the red Luxeon K2 at 625 nm. I chose it over other brands because of the wide bandwidth. Check out the SPD curve, it does a far better job of hitting 660 then any other red.
At this time I have pretty much been covering my bases with lights, even using the different whites. As time goes on, I will see what can be eliminated and added.

The copper strip is 2.25" x 12". I would't want to run more then 7 on it. I think it is better to build 2 units with 7 each. I will dig up the url of my online metal supplier if interested.
If you go with the LEDengins buy the stars. I usually use emitters, but they aren't that much extra. I think they were about $11.50 from Mouser.

K2s vs. Rebels: All things being equal, I would go with the K2s. Simply because I can paste those babies right on a heat sink and hand solder them (I made some very tiny clamp-on heat sinks to clip on the wires).
I am going to try some Rebel emitters very soon because some colors have better efficiency then K2s and are a bit cheaper. The Rebel Stars are a rip-off, that is why I have avoided them. I will get some of the boards from Asia Signals and try to mount some emitters. I do have a hot air reflow soldering system and some good soldering paste in the frig next to the ketchup, so I probably won't screw the Rebels up too badly.

I do like the idea of a new string that is focused on practical building techniques, where we can share ideas, sources, experimental results and see how we can make better units then ones commercially available.
And keeping the theory to a minimum.

Veggii How do you start a new string?

**veggii** · 04-21-2008, 06:07 PM

picd from post 434

**veggii** · 04-21-2008, 06:11 PM

pics from post 435

**veggii** · 04-21-2008, 06:13 PM

pics from post 436

Parts of site failed to load...
If you are using an ad blocker addon, you should to disable it (it blocks more than ads and causes parts of the site to not work).

'Redline's LED - DiY - Build' aka"GROW LIGHT"

Thread Tools

Search Thread

Display

'Redline's LED - DiY - Build' aka"GROW LIGHT"

Advertisements

'Redline's LED - DiY - Build' aka"GROW LIGHT"

'Redline's LED - DiY - Build' aka"GROW LIGHT"

'Redline's LED - DiY - Build' aka"GROW LIGHT"

'Redline's LED - DiY - Build' aka"GROW LIGHT"

'Redline's LED - DiY - Build' aka"GROW LIGHT"

'Redline's LED - DiY - Build' aka"GROW LIGHT"

'Redline's LED - DiY - Build' aka"GROW LIGHT"

'Redline's LED - DiY - Build' aka"GROW LIGHT"

'Redline's LED - DiY - Build' aka"GROW LIGHT"

Similar Threads

need advice where to find someone to build me led light

Got the green light, time to build, suggestions/advice please!

It's getting hot!!! How can I build a light damper for my window ac unit?

6 bucket recirculating DWC closet grow...build and grow!

Build your own light by Latewood

Posting Permissions

Parts of site failed to load...If you are using an ad blocker addon, you should to disable it (it blocks more than ads and causes parts of the site to not work).

'Redline's LED - DiY - Build' aka"GROW LIGHT"

Advertisements

'Redline's LED - DiY - Build' aka"GROW LIGHT"

'Redline's LED - DiY - Build' aka"GROW LIGHT"

'Redline's LED - DiY - Build' aka"GROW LIGHT"

'Redline's LED - DiY - Build' aka"GROW LIGHT"

'Redline's LED - DiY - Build' aka"GROW LIGHT"

'Redline's LED - DiY - Build' aka"GROW LIGHT"

'Redline's LED - DiY - Build' aka"GROW LIGHT"

'Redline's LED - DiY - Build' aka"GROW LIGHT"

'Redline's LED - DiY - Build' aka"GROW LIGHT"

Similar Threads

Posting Permissions

Parts of site failed to load...
If you are using an ad blocker addon, you should to disable it (it blocks more than ads and causes parts of the site to not work).