AvidLerner
- 296
- 63
AvidLerner
- 296
- 63
I mounted the fixture in the tent and it works great, air flow and cooling with just one exhaust fan, I was already using. The kit has two ducts, that slide inside of each other so you can make two of these and it could also work for cob fixtures to remove heat from your tent wioth round holes. The duct is 3" deep and 6" wide.
AvidLerner
- 296
- 63
There is a second round of DIY CO-OP pm me if you are interested
Schkronis
- 77
- 33
This looks great! Is there any reliability info on QB's? Wonder what will happen if single diodes burns out?
AvidLerner
- 296
- 63
The board is designed as constant current. If a diode burns out in a given string that string will fail to continue functioning. The other parallel strings will continue drawing the same current draw as the non working string shows as an open circuit fulfilling its role and the rest of the light will function
AvidLerner
- 296
- 63
I have been running four of these lights now for some time at full power in four2x2 areas and no heat problems in tents or open areas very pleased with the performance of these boards.
Well built and no heat issues. At 100w they generate very little heat at 200w active cooling is needed
Well built and no heat issues. At 100w they generate very little heat at 200w active cooling is needed
AvidLerner
- 296
- 63
Three lights running at 200w for more than 30 days no issues
Dan789
- 2,954
- 263
Very ingenious.
AvidLerner
- 296
- 63
Thanks other folks enjoy them as well. 200w replaces four cobs and a lot less heat and half the kilowatt cost daily 200w 14kwh
AvidLerner
- 296
- 63
The current boards have solderless connectors
AvidLerner
- 296
- 63
Feel free to DM me for any particulars
AvidLerner
- 296
- 63
I will be happy to share the design process as well as gerber files source of fabrication and all costs.
AvidLerner
- 296
- 63
looks nice and healthy
AvidLerner
- 296
- 63
I have been using cobs for a couple of years aqnd decided it was time for change. I was watching various folks with their designs and I decided I wanted to make the move to Samsung s6 bin diodes capable of 207lm/w max if driven properly.
I started research. I learned two things for Constant Current design. One, you decide what power you want the board to deliver, and two what drive current you want to use for the board. The third item is how to drive the board economically and efficiently.
Starting from item three economy and efficiency. Samsung diodes are most efficient when driven between 150mA and 165mA a very narrow band given the diode can be driven at 200mA max. Keeping this in mind on to item one.
I decided I wanted a board to deliver 200w to replace existing 200w cob rigs I have. I wanted to reuse as much of my equipment as possible from heat sinks to drivers etc. everything. The majority of my drivers are HLG-185H-C1400B drivers. I decided I was not going to use 1400B drivers because the only other companion would be a C700B driver. More on that thought later.
I decided a better driver would be to go with a HLG-185H-C1050A/B driver. I had a few of these so that would also support the reuse approach. The C1050 driver has a drive current of 1050mA and a max Forward Voltage of 185v. Those are my two targets now to focus my design parameters on.
I now have a target of 200w, 1050mA, 185v, and a current range of 150-165mA per diode.
First step. Decide how many diodes to use in the design. I want6ed to use as many diodes as I couldwhile be economical and efficient. I contacted a manufacturer in ShenZhen China and discussed my parameters. Their primary point is there are 2500 diodes on a reel of Samsung S6 LM5661C diodes and a reel cost about $150 per reel. I decided on two reels as I wanted to use two Spectrum 5000k 80cri and 3000k 90cri to be used in unison or separate. I would decide that later.
With this info I did some calculations. I decided I would go with 384 diodes on each board for a total of thirteen boards for 2496x2 leaving 8 diodes for the manufacturer to keep. Now the trick is to drive 384 diodes at 150mA.
I decided to accomplish this easily with two colors was to go with two channels with 192 diodes on each channel. Now comes the hard part, how to configure 192 diodes correctly fort the desired outcome.
I developed a strategy. A circuit in series has additive voltage, while a parallel circuit has additive amps. Each diode has a forward voltage of 2.9v and a target of 150mA.
So I looked at how Samsung does it. They use a circuit of 8 diodes in series with three parallel rows to get their hard stips to work at 200mA. SO I decided again I would work with that same rationale to start with, but I think bigger, So I multiplied the 8 by 4 and came up with 32 diodes in series that gave me 92vf; forward voltage. excellent.
Now comes the harder part creating a balanced parallel circuit. Looking at the target drive current I took 1050/150 = 7. I wanted to be closer to 160mA so I went with six. 1050/6 = 159mA. excellent again
Now comes her science. I have designed circuit that has 32 diodes in series with six rows connected in parallel. What I have now is six rows of 32 diodes or 192 diodes with one channel at 92v, and 1050mA. I add a second channel just like the first and I now have 384 diodes in two channels each with 92v and 1050mA.
As it stands I can run each channel with a HLG-120H-C1050A/B driver and control each channel independent of each other or, I can connect both channels in series with a total Vf of 185v and 1050mA, because volts add up in series and amps add up in parallel.
Each channel now runs at 100w max or 200w per board. I can use either a 120H-C1050B or 185H-C1050B driver and get 200w from the board. I can also connect a 240H-C1050B and get an additional 20 watts out of the board due to voltage differences with the two drivers. I can connect the channels separate or in series a choice, or I can connect channels from two boards in series, channel 1 to channel 1 and channel 2 to channel 2 with two drivers for two boards controlling separate spectrum that way, as well. flexibility.
I mounted my new boards on my old 5.886x18"; 200w dissipation heat sinks; from heatsinks usa because I used those instead of pin sinks. When I need a new heat sink it cost me $35 plus S&H. cheap.
I used EasyEda.com for circuit analysis and I used Shenzhen Hansion Technologies Jeff Wei https://tophspcb.en.alibaba.com/?spm=a2700.8443308.0.0.KQatRJ for my production facility. They have my gerber file on file there called the GrowGreen 100 board. I also have have tghe4 dxf design file and I will share that as well./ I have a gothub page called GrowGreen AvidLerner with files there and code for my controllers I share public domain. I told Jeff my friends can use my gerber file and it will save $100 for Engineering fees. They charge $100 fabrication fee, shipping $300 for reels, production $150, shipping and fees $100 total cost around $800 for 13 boards that deliver 200w each.
That is 2600w for $800 which comes to $0.30/watt!
Only one drawback you have to buy all 13 boards and pay 50% up front and 50% upon completion prior to shipping. So I can up with the CO-OP concept. Community working together for a common purpose.
So if you want to know about the CO-OP DM me.
If you want to build your own boards lets do that too. I love helping people help people. You can post your ideas designs everything here, so we can all learn and grow together creating affordable lighting.
I started research. I learned two things for Constant Current design. One, you decide what power you want the board to deliver, and two what drive current you want to use for the board. The third item is how to drive the board economically and efficiently.
Starting from item three economy and efficiency. Samsung diodes are most efficient when driven between 150mA and 165mA a very narrow band given the diode can be driven at 200mA max. Keeping this in mind on to item one.
I decided I wanted a board to deliver 200w to replace existing 200w cob rigs I have. I wanted to reuse as much of my equipment as possible from heat sinks to drivers etc. everything. The majority of my drivers are HLG-185H-C1400B drivers. I decided I was not going to use 1400B drivers because the only other companion would be a C700B driver. More on that thought later.
I decided a better driver would be to go with a HLG-185H-C1050A/B driver. I had a few of these so that would also support the reuse approach. The C1050 driver has a drive current of 1050mA and a max Forward Voltage of 185v. Those are my two targets now to focus my design parameters on.
I now have a target of 200w, 1050mA, 185v, and a current range of 150-165mA per diode.
First step. Decide how many diodes to use in the design. I want6ed to use as many diodes as I couldwhile be economical and efficient. I contacted a manufacturer in ShenZhen China and discussed my parameters. Their primary point is there are 2500 diodes on a reel of Samsung S6 LM5661C diodes and a reel cost about $150 per reel. I decided on two reels as I wanted to use two Spectrum 5000k 80cri and 3000k 90cri to be used in unison or separate. I would decide that later.
With this info I did some calculations. I decided I would go with 384 diodes on each board for a total of thirteen boards for 2496x2 leaving 8 diodes for the manufacturer to keep. Now the trick is to drive 384 diodes at 150mA.
I decided to accomplish this easily with two colors was to go with two channels with 192 diodes on each channel. Now comes the hard part, how to configure 192 diodes correctly fort the desired outcome.
I developed a strategy. A circuit in series has additive voltage, while a parallel circuit has additive amps. Each diode has a forward voltage of 2.9v and a target of 150mA.
So I looked at how Samsung does it. They use a circuit of 8 diodes in series with three parallel rows to get their hard stips to work at 200mA. SO I decided again I would work with that same rationale to start with, but I think bigger, So I multiplied the 8 by 4 and came up with 32 diodes in series that gave me 92vf; forward voltage. excellent.
Now comes the harder part creating a balanced parallel circuit. Looking at the target drive current I took 1050/150 = 7. I wanted to be closer to 160mA so I went with six. 1050/6 = 159mA. excellent again
Now comes her science. I have designed circuit that has 32 diodes in series with six rows connected in parallel. What I have now is six rows of 32 diodes or 192 diodes with one channel at 92v, and 1050mA. I add a second channel just like the first and I now have 384 diodes in two channels each with 92v and 1050mA.
As it stands I can run each channel with a HLG-120H-C1050A/B driver and control each channel independent of each other or, I can connect both channels in series with a total Vf of 185v and 1050mA, because volts add up in series and amps add up in parallel.
Each channel now runs at 100w max or 200w per board. I can use either a 120H-C1050B or 185H-C1050B driver and get 200w from the board. I can also connect a 240H-C1050B and get an additional 20 watts out of the board due to voltage differences with the two drivers. I can connect the channels separate or in series a choice, or I can connect channels from two boards in series, channel 1 to channel 1 and channel 2 to channel 2 with two drivers for two boards controlling separate spectrum that way, as well. flexibility.
I mounted my new boards on my old 5.886x18"; 200w dissipation heat sinks; from heatsinks usa because I used those instead of pin sinks. When I need a new heat sink it cost me $35 plus S&H. cheap.
I used EasyEda.com for circuit analysis and I used Shenzhen Hansion Technologies Jeff Wei https://tophspcb.en.alibaba.com/?spm=a2700.8443308.0.0.KQatRJ for my production facility. They have my gerber file on file there called the GrowGreen 100 board. I also have have tghe4 dxf design file and I will share that as well./ I have a gothub page called GrowGreen AvidLerner with files there and code for my controllers I share public domain. I told Jeff my friends can use my gerber file and it will save $100 for Engineering fees. They charge $100 fabrication fee, shipping $300 for reels, production $150, shipping and fees $100 total cost around $800 for 13 boards that deliver 200w each.
That is 2600w for $800 which comes to $0.30/watt!
Only one drawback you have to buy all 13 boards and pay 50% up front and 50% upon completion prior to shipping. So I can up with the CO-OP concept. Community working together for a common purpose.
So if you want to know about the CO-OP DM me.
If you want to build your own boards lets do that too. I love helping people help people. You can post your ideas designs everything here, so we can all learn and grow together creating affordable lighting.
AvidLerner
- 296
- 63
AvidLerner
- 296
- 63
After I retired all my cob's I had all of these HLG-185H-C1400B drivers at least three or four.
In decided to experiment before I would commit to building PCB's to save a lot of money and aggravation.
I decided to use flxible strips over hard strips as teh flexible strips are easier to create various formulae such as 32S6P. I practiced with 49S8P which yields the same outcome as the previous model, with more current.
I will spare you the math.
I order flexible strips of Samsung LM561C diodes from here ->https://mufue.en.alibaba.com/?spm=a2700.8443308.0.0.NoIvyo
Shenzhen Mufue technologies Roget Zheng. I purchased 20m 10m 5000k 80cri and 10m 3000k 90cri. These are all 24v Constant Current strips which will behave identical to any PCB built, only the diodes are more spread out so the PAR numbers will be lower due to Physics. I will save you that as well. Just know it is correct.
Flexible strips come in precut segments. Each segment has seven diodes in series, so they are 200mA with 2.9x7 = 20.3vf per segment.
math time:
143Vf / 20.3 = 7.04 so lets say 7 segments using 49 diodes.
7 x 20.3 = 142.1Vf demand / 143 = 98% demand. quite efficient.
Now we have Vf resolved with a set segment length now it is time to decide how hard to run the diodes. I want to run them below 200mA and above 150mA if possible.
1400/165 = 8.48. So If I use 8 rows or parallel strings I will get 175 mA per row. I can lower that by adding an extra row say nine rows now so 1400/9 = 155.55mA not bad.
math over: back to design
based on this I can now build a light from flexible strips using 7 segments long by 9 rows wide
So I can cut nine seven segment strips and connect therm all in parallel to create a 200w fixture.
I have a total of 441 diodes operating around 195-200lm/w the driver is driving close to 200w so we can estimate that at 200w the flexible strip fixture is delivering 40,000lumens in a 18"X12" area.
Not bad.
It was from here I used up all my cob drivers 1400mA for the flexible strips and experimented to get the best effect from a PCB fixture light engine.
So here is another activity to try for a couple hundred bucks total cost. I mount these flexible strips on cookie sheets, 1/8" th x 3/4" wide aluminum flat stock with no heat issues or any other light gauge you can think of. Cookie sheets work best and cheapest. I build these for cloning vegging and they can bloom too at 200w.
You can build these just talk to Roget Zheng, great gentleman and he even makes rigid strips too, every imaginable color or spectrum you could think of he has all in Samsung LM561C format even specific spectrum like far Red Deep Red Royal Blue etc.
enjoy
In decided to experiment before I would commit to building PCB's to save a lot of money and aggravation.
I decided to use flxible strips over hard strips as teh flexible strips are easier to create various formulae such as 32S6P. I practiced with 49S8P which yields the same outcome as the previous model, with more current.
I will spare you the math.
I order flexible strips of Samsung LM561C diodes from here ->https://mufue.en.alibaba.com/?spm=a2700.8443308.0.0.NoIvyo
Shenzhen Mufue technologies Roget Zheng. I purchased 20m 10m 5000k 80cri and 10m 3000k 90cri. These are all 24v Constant Current strips which will behave identical to any PCB built, only the diodes are more spread out so the PAR numbers will be lower due to Physics. I will save you that as well. Just know it is correct.
Flexible strips come in precut segments. Each segment has seven diodes in series, so they are 200mA with 2.9x7 = 20.3vf per segment.
math time:
143Vf / 20.3 = 7.04 so lets say 7 segments using 49 diodes.
7 x 20.3 = 142.1Vf demand / 143 = 98% demand. quite efficient.
Now we have Vf resolved with a set segment length now it is time to decide how hard to run the diodes. I want to run them below 200mA and above 150mA if possible.
1400/165 = 8.48. So If I use 8 rows or parallel strings I will get 175 mA per row. I can lower that by adding an extra row say nine rows now so 1400/9 = 155.55mA not bad.
math over: back to design
based on this I can now build a light from flexible strips using 7 segments long by 9 rows wide
So I can cut nine seven segment strips and connect therm all in parallel to create a 200w fixture.
I have a total of 441 diodes operating around 195-200lm/w the driver is driving close to 200w so we can estimate that at 200w the flexible strip fixture is delivering 40,000lumens in a 18"X12" area.
Not bad.
It was from here I used up all my cob drivers 1400mA for the flexible strips and experimented to get the best effect from a PCB fixture light engine.
So here is another activity to try for a couple hundred bucks total cost. I mount these flexible strips on cookie sheets, 1/8" th x 3/4" wide aluminum flat stock with no heat issues or any other light gauge you can think of. Cookie sheets work best and cheapest. I build these for cloning vegging and they can bloom too at 200w.
You can build these just talk to Roget Zheng, great gentleman and he even makes rigid strips too, every imaginable color or spectrum you could think of he has all in Samsung LM561C format even specific spectrum like far Red Deep Red Royal Blue etc.
enjoy
AvidLerner
- 296
- 63
Here are two strip lights i built
AvidLerner
- 296
- 63
This is a 2x6 area with two 200w boards on the bottom level and two 200w flexible strips on the upper level. I can stack my growth because there is no rising heat issues with either light design. Efficient use of space and no heat issues. In a tent I just install the pcb in a dryer duct connect to my exhaust and no heat
sixstring
- 7,079
- 313
vegas i like your approach and your open sourcing,co op stuff.but we gotta talk about that spectrum lol. i know 5000k is good to drive numbers up and decent for veg but with these diodes the spectrum is so full and balanced i feel its a step in the wrong direction mixing 5000k and 3000k.now if you mixed 2700k and 3500k or maybe even 4000k you would have my complete support. but even with vero cobs im finding 3000k 80 cri veg pot plants better than 6500k t5 lamps. my plants are just as tightly spaced node wise and grow faster under 3000k vero at 100w then they ever did under 6500k t5 at 320w in the same space.
but i have another question.is there a way to mix these colors without needing seperate channels? kinda like photon phantom does with the wavey boards just put 2 or 3 colors on the thing all run off one driver.i just dont think too many peope are going to want to have or need seperate control over colors on a board like this.most growers just want full power for 12 hours with some of us wanting dimming at sunrise and sunset.i have many other questions but ill wait till we get past these. be nice if you could space out the diodes and get away from heatsinking like photon does with his boards also. i know this is more like chilled boards and less like photon or hlg,but chilled and these boards being so packed together sure seems silly to have all that power on such a skinny board instead of spreading them out .
but i have another question.is there a way to mix these colors without needing seperate channels? kinda like photon phantom does with the wavey boards just put 2 or 3 colors on the thing all run off one driver.i just dont think too many peope are going to want to have or need seperate control over colors on a board like this.most growers just want full power for 12 hours with some of us wanting dimming at sunrise and sunset.i have many other questions but ill wait till we get past these. be nice if you could space out the diodes and get away from heatsinking like photon does with his boards also. i know this is more like chilled boards and less like photon or hlg,but chilled and these boards being so packed together sure seems silly to have all that power on such a skinny board instead of spreading them out .
AvidLerner
- 296
- 63
You can mix the spectrum as many ways as you can desire.
I use channels because it gives more flexibility if you want it.
The industry uses 5000k/3000k called natural daylight. You can find the mix at sporting arenas and other public spaced that use leds now. That is what led me in that direction.
I am testing the difference between 5000k/3500k and 5000k/3000k. The blue from the 5000k really increases secondary growth.
I love my chilled boards they motivated me. I wanted to spread my footprint out but I also wanted to use a good 200w dissipation heat sink so the 5.886x18" works great like have a whole bar of cobs. The spread is three feet on the long direction and on the narrow direction I get a foot each way with manageable drop in ppfd.
I have one in a 2x2 tent diagonally and the ladies are having a party with 200w that is 50w psf with a ppfd of 400 at 24".
Thanks for checking in. I appreciate the attention.
namaste
I use channels because it gives more flexibility if you want it.
The industry uses 5000k/3000k called natural daylight. You can find the mix at sporting arenas and other public spaced that use leds now. That is what led me in that direction.
I am testing the difference between 5000k/3500k and 5000k/3000k. The blue from the 5000k really increases secondary growth.
I love my chilled boards they motivated me. I wanted to spread my footprint out but I also wanted to use a good 200w dissipation heat sink so the 5.886x18" works great like have a whole bar of cobs. The spread is three feet on the long direction and on the narrow direction I get a foot each way with manageable drop in ppfd.
I have one in a 2x2 tent diagonally and the ladies are having a party with 200w that is 50w psf with a ppfd of 400 at 24".
Thanks for checking in. I appreciate the attention.
namaste
Latest posts
-
light stress or just a stupid grower ?- Imzzaudae posted
-
Back and (slightly) better than ever- Fromunda506 posted
-
Hi! New grower looking to learn- Jmaes Mabley posted
-
It All Starts From Seed- MEGA956 posted
-
Weird leaves growing on buds- pulchripes posted