DIY 12V LED

[GrowQs]

Can I get say a bunch of 3V LED chips, put them in series of 4 per strip for 12v, then parallell multiple 12v strips, and run them directly off of a 12V battery bank?

Lets say I have an off-grid style 12v battery backup power supply that I can pull over 2000 Watts of 12V off of continuously 24/7 without it breaking a sweat..

Could I make some LED grow lights that I could drive straight off of that cheaply?

 

[Old Salt]

You'd need to locate a driver with a 12VDC input. You may be able to use a boost/buck power supply instead of a regular driver. This would let you adjust the voltage and depending on the type, current to the LEDs. A 12VDC light can be made, but when all is said and done it will probably cost far more than a manufactured solution. My advice would be to use an inverter, and run at 120VAC or 240VAC. You can get inverters with 95% efficiency. I think you'll find it a more cost effective solution.

 

[GrowQs]

Like this?



Would this also work as a dimmer by adjusting the volts lower than and up to the rated volts of the LED chips or is that not how it works?

This would work like a "constant voltage" LED driver rather than a "constant current" LED driver, right?

 

[Old Salt]

That's the general idea. Notice that the input current for this device is limited to 40A. At a 12VDC input, the maximum power output drops to 12V X 40A X 0.98(efficiency) or 470W. If you battery bank can provide a higher voltage, you'll be able to get more power out of it. The output is limited to 22A, so you may need to take that into consideration. I'd order one and bench test it to see if it fits your needs.

As an example, my COBs are 37V, 1.7A max...
22A/1.7A = 12 Cobs maximum in parallel by current capacity
470w/(37V * 1.7A) = 7 COBs max by power
Use the lower of the two.

Test the converter, before moving on to the LEDs or COBs.

 

[GrowQs]

Seems like it would be pretty easy to blow LEDs with something like that, easily overdrive them if they get thermal runaway or I bumped the volts a little too high (depending on adjustment sensitivity), but maybe I could put an inline fuse to each LED chip to stop it incase it starts drawing too much current..

For example if I was to try to run CREE CXB3590 chips which are..
Typical Forward Voltage : 36V
Max Drive Current : 3600mA

I could put a 3.5 or maybe 4 amp fuse on each one and turn them on, and up, by turning the volts up close to 36 volts, with the chips wired in parallell to the power source? And could run plenty of those chips with that power source..

And I could use holders like this @$3 each to hold them to my heat sync/syncs with 2 studs..

And make home made liquid cooled heat syncs! Or one big liquid cooled heat sync to hold them all in an array..


Would need about 10 of them to replace a 1kw HPS @ $30 each is $300, plus holders, and converter, and another $100 to make heatsyncs so like $500 to build that alone..

Worth $500?

What would be better chips/strips to run on a constant voltage driver like this with plenty of current available?

 

[GrowQs]

That's the general idea. Notice that the input current for this device is limited to 40A. At a 12VDC input, the maximum power output drops to 12V X 40A X 0.98(efficiency) or 470W.

Oh, so that wouldn't be big enough then.. It's not really going to do 1800 watts..
So I'd need like 2 of them and run 2 circuits or one even bigger..

Hold up, if those cobs are 3600ma each max, or 3.6A, and the box will only put out 22A, then it would only power 6 of the chips..

But sheesh 3.6 amps at 36 volts is almost 130 watts each.. I thought they were only rated 75 watts each..
Confused..

Oh.. It won't even do that because it is more limited by intake power @ 12V..

But whatever, basically it would work but might need more than one boost converter or find a bigger one..

 

[Old Salt]

...if I bumped the volts a little too high (depending on adjustment sensitivity), but maybe I could put an inline fuse to each LED chip to stop it incase it starts drawing too much current..

These converters look to have constant current outputs. It is a good idea to incorporate overcurrent protection for the LEDs/COBs. The COBs I use have a maximum rating more than double the figure I quoted. The normal drive current is 1400mA. I like the extra headroom 1700mA gives me.

I think it would only take two or three of the converters to replace your HID. My COB array is overpowered for my 2' X 4' tents. It will go up above 500w. I have the COBs on 1' centers, so there are eight in the array. I run the light at less than 350w in flower. I'd be using one converter for four COBs, arranged in series parallel. That requires 74VDC / 3.4A or a little more than 250w.

Check out the specifications of the lights some manufacturers using the same COBs or LEDs you plan on using. You'll fins all of them use less than 1/2 the maximum power the devices are rated for. This is for higher efficiency. You may be able to use Quantum Boards, or the Samsung light strips in your design as well as COBs.

 

[GrowQs]

What would you think of running like 20 cheap ebay 100w cobs @ like 10-20 watts each?
They are $1.90 each and I read they are decently efficient when driven very softly..

Being run that soft you think just screwing them to a steel plate would be enough heat sync? Put em in 4 rows of 5 on about 3.5'X3.5' plate..
20 @ 20 watts would be 400 watts, and would be a lower voltage so less work on the boost converter trying to kick it up from 12..

It would be like $65 for 400 watts of LED including boost converter, plus some wires and a steel plate from the junkyard..

It's not like I don't 110V AC but I just have so much spare 12VDC sitting here, on all the time anyway, might as well pull on it since I've already lost it's base efficiency just being on and have gone to DC from AC already, it's severely underutilized..
And a good excuse to upgrade my battery bank ya know..

 

[Old Salt]

What would you think of running like 20 cheap ebay 100w cobs @ like 10-20 watts each?
They are $1.90 each and I read they are decently efficient when driven very softly..

Being run that soft you think just screwing them to a steel plate would be enough heat sync? Put em in 4 rows of 5 on about 3.5'X3.5' plate..
20 @ 20 watts would be 400 watts, and would be a lower voltage so less work on the boost converter trying to kick it up from 12..

All can do is go by the specifications I found. I can't recommend them when I compared them to the Citizen CLU048-1216s I'm using. Notice that at 1/2 the power the Citizens produce 85% of the light that the e-Bay chips do. The Citizens cost $16 each, but you'll easily get that back in power savings (5 years 12hrs/day @ 50w = 1095KWh. Savings @ 0.16/KWh = $175 saved PER COB)

e-Bay Chip:
LED Lamp Chip Specification:
Output Power: 10W-100W
Light Color: Warm White 3000k-3200k Cool White 6000-6500k
Beam Angle: 120-140
LED Support Material：Al
Working Life: 50000h

Wattage: 100
Voltage: DC 28-36V
Current: 3000mA
Lumen: 8700-9000
Lumens/Watt: 90 (Calculated)

Citizen CLU048-1216:
Voltage: 31.2-36.8 VDC
Current: 1440mA (max: 3680mA)
Lumen: 7000-8030 (Calculated @ 34V, 1440mA - 48w)
Lumens/Watt: 143-164 (depends on °K)

 

[GrowQs]

All can do is go by the specifications I found. I can't recommend them when I compared them to the Citizen CLU048-1216s I'm using. Notice that at 1/2 the power the Citizens produce 85% of the light that the e-Bay chips do. The Citizens cost $16 each, but you'll easily get that back in power savings (5 years 12hrs/day @ 50w = 1095KWh. Savings @ 0.16/KWh = $175 saved PER COB)

I am absolutely sure your citizens blow them out of the water hands down, especially moreso the higher you run them..

Isn't the l/w quite non linear in all LEDs though? The higher you turn them up the less l/w you get out of any of them correct?
And basically, the higher quality the LED the more linear their l/w output is, with the lesser ones efficency dropping off more rapidly the higher they are ran.. No?
So the citizens keep up with the crees at half power but the better (more expencive) chips start walking away with efficiency when turned up higher..

And that basically, say to replace 1kw of HID, you actually need to buy about 1500 watts worth of LED, and then turn them down almost half way, because that is how they get their efficiency advantage..

You have over 500W in a 2x4, therefor over 1000w to a 4x4, and turn them down almost half way to under 350w, so would need to be actually running at 700w in a 4x4..

It's like they state "This engine will put out a constant 500HP and get 15MPG!"
But in all actuality it can run at a constant 500HP, but will only get 15mpg at 100HP..


I see you are a machinist, or atleast are interested in machining and metalworking.. #Respect and thank you for your help understanding LEDs..

My father retired after 34 years as a tool and die maker and did performance engine work on the side, mostly harleys, building twin plug heads, building cranks and such, so I grew up around the "you can make anything" lifestyle and I make all sorts of custom stuff.. I have quite some tools of my own and do a lot of actual custom work, performance related and otherwise, but not too much on the electronics side other than building and running some standalone engine management stuff.. I don't naturally know much about circuit boards kind of stuff..

 

[GrowQs]

"but you'll easily get that back in power savings (5 years 12hrs/day @ 50w = 1095KWh. Savings @ 0.16/KWh = $175 saved PER COB) "

I also want to address this part of your post so we understand eachother here..

I think it is futile for me to try to replace my 1kw HPS with LEDs, and will not be doing that for the following reasons.. (but I may build an LED for veg)

1. I will be focusing my efforts on outdoor growing as a priority, and that means that when the season is right for plants to be outside, all of my legal plant count is going to be outside in the ground, therefore I may only flower once or twice indoors in the winter while I'm bored waiting for spring to come, so all of this long term cost savings calculations doesn't really apply to me, because I'm only going to run indoors for 5-6 months a year..

2. A flippin HPS puts down 145,000 lumens @ 145 lumens per watt..
You are getting 8000 lumens @ 48W for 167 lumens per watt, only beating the HPS by a mere 22 lumens per watt, 15.2% efficiency gain..
At 8,000 lumens you would need over 18 COBs (running efficently to get the l/w savings, not maxed out) to match the lumens of the HPS...
You say your 8 COB setup costs $560 USD, therfore 18 COBs would logically cost $1260..

So I would have to spend $1260 to gain a 22 lumens per watt savings advantage, for only part of the year, while maintaining the same performance..

I think not sadly..
The numbers may make sense for a 5 year plan running all year, but not for me..

But I do need a better veg light, so I might play with some LEDs for that..

 

[Old Salt]

And basically, the higher quality the LED the more linear their l/w output is

Not necessarily. The efficiacy (l/m) is generally higher, as is the quality of the light. The curve for efficiacy needs to be consulted if this is a design consideration. Usually it wouldn't matter, as a power supply can easily be programmed to get a linear response to a mechanical or electrical input.

So the citizens keep up with the crees at half power but the better (more expencive) chips start walking away with efficiency when turned up higher..

Yes the efficiency of the CREEs is higher, as is the cost. I computed the cost of ownership of the two before buying. At my power rates that was equal. I chose to pay the power company over time, rather than pay up front.

And that basically, say to replace 1kw of HID, you actually need to buy about 1500 watts worth of LED, and then turn them down almost half way, because that is how they get their efficiency advantage..

It's just the LEDs/COBs that have the higher maximum power rating. The light is usually designed so the LEDs/COBs run at 1/6 - 1/3 of their maximums. The drivers and any other electronics is chosen for the power used, so the power is not turned down, it's just not available. The high power LEDs/COBs only increase the manufacturers' cost by a few percent.

I see you are a machinist, or atleast are interested in machining and metalworking.. #Respect and thank you for your help understanding LEDs..

I spent over 20 years in the Navy as an Electronic Technologist. After that I worked as a systems and network administrator. When I was forced into retirement for health reasons, I wrote a major part of TurboCNC 4, which is where the machining comes in. Dad was a Certified Tool and Die Maker, as well as a draftsman. My grandfather was a blacksmith in the Netherlands. I guess I do have a bit of metal along with some free electrons in my blood.

 

[Old Salt]

"but you'll easily get that back in power savings (5 years 12hrs/day @ 50w = 1095KWh. Savings @ 0.16/KWh = $175 saved PER COB) "

I also want to address this part of your post so we understand eachother here..

I think it is futile for me to try to replace my 1kw HPS with LEDs, and will not be doing that for the following reasons.. (but I may build an LED for veg)

1. I will be focusing my efforts on outdoor growing as a priority, and that means that when the season is right for plants to be outside, all of my legal plant count is going to be outside in the ground, therefore I may only flower once or twice indoors in the winter while I'm bored waiting for spring to come, so all of this long term cost savings calculations doesn't really apply to me, because I'm only going to run indoors for 5-6 months a year..

2. A flippin HPS puts down 145,000 lumens @ 145 lumens per watt..
You are getting 8000 lumens @ 48W for 167 lumens per watt, only beating the HPS by a mere 22 lumens per watt, 15.2% efficiency gain..
At 8,000 lumens you would need over 18 COBs (running efficently to get the l/w savings, not maxed out) to match the lumens of the HPS...
You say your 8 COB setup costs $560 USD, therfore 18 COBs would logically cost $1260..

So I would have to spend $1260 to gain a 22 lumens per watt savings advantage, for only part of the year, while maintaining the same performance..

I think not sadly..
The numbers may make sense for a 5 year plan running all year, but not for me..

But I do need a better veg light, so I might play with some LEDs for that..

I went with LEDs due to the heat put out by HID lighting. Yes, I know it can be addressed with a cool tube, and separate air supply. Those are extra components that need to be moved when raising or lowering the light. I'm lazy, and work hard at keeping it that way.

 

[GrowQs]

I'm lazy, and work hard at keeping it that way.

Yeah I can tell...
Your one rope compound LED fixture raiser is excellent work in that aspect!!