LED Lighting Project 150-1000W

[vari]

Here is a little DIY project i almost completed today. It is a general LED lighting driver that supports 150 to 1000 watts wort of LED power.

The design is extremely simple, modular and made mostly from parts rescued from going to the garbage dump.

Only a few steps are necessary to get this big boi working.

See the in-depth details at my blog at: LED Lighting Project 150-1000W | Varikonniemi.

Comments, questions and critique welcomed

 

[weeba2kv]

wow 1 kW hehe cool

 

[Cyparagon]

Neat. How did you calculate 100VDC from 86VAC? What values are the caps? Why the WW, CW, and actinic?

 

[vari]

I have not made the precise calculations yet, but it should end up in the ~100V range (86V*1.4=120V). The series resistor is one of the components used to drop the voltage somewhat.

The LEDs were chosen kinda random. This project has gone through so many redesigns that these are the components i ended up with.

The first 150W LED panel will be used as a growlight, so i thought a mixture of these three different leds would bring the best possible spectra.

 

[Cyparagon]

By my calculations, the voltage should be 86*√2-1.4(from bridge rectifier)= 120V (minus whatever ripple you get). The resistor will only drop voltage when under load. In this case you need a 4A load for it to drop down to 100V. A 1.5A load will only drop ~8V.

You'll want to double check, but I'm pretty sure plants do best with blue and red - specifically ~665nm red as opposed to the ~620nm red in most common red LEDs. It'll work, but all the green/yellow generated by the white dies goes to waste.

 

[vari]

Thanks for your input, i am well aware of these issues.

1. The design was draft up around the standard high-powered LEDs with a forward voltage of 32-34V. This would put the rectified DC voltage + LM350k voltage drop pretty much on the spot directly with 3xnormal LEDs. Very few watts would be wasted. The problem is here the one hybrid led with a smaller forward voltage, so depending on the situation i probably must waste some volts into a resistor connected with the LEDs in series.

2. I am aware of the chlorophyl absorption efficiency spectra. Red and blue light at the correct wavelength are absolutely most effective. However, it is a misconception only chlorophyl acts as the energy absorbing part of the plants, and they actually can utilize a much wider spectrum of light due to many mechanisms. Wikipedia will tell you about this in good detail. Different plants also like different color spectra for optimal growth. I want this light to be a general purpose grow light, so thats why i started out with this combination of LEDs. Also, the leds should be re-usable outside of this application, if the plan changes. Red or blue LEDs are not that useful in house lighting etc. whereas the white ones are perfect.

3. The luminous flux of colored diodes is severely smaller than white diodes. I THINK the non-optimal spectra could easily be outweighted by the larger luminous flux/W of the white LEDs compared to red/blue ones. It is not out of question i will connect a red and blue LED series to this setup later, after all it is totally modular

 

[vari]

Today i decided to mount active cooling to the LED heatsink. I decided to use an old Antec 3-speed 12V computer fan. It should keep the LEDs considerably cooler.

 

[vari]

Today i connected the components, and with a test load of 140W the DC voltage was 108V. I also connected an inductor to further smoothen the current peaks. I may even replace the series resistor with the inductor to get rid of the voltage to load dependency.

So for now everything is looking good. Even great. The ripple was less than i had expected, even though i was using only the smaller capacitor.

Tomorrow i will start to mount the components to the casing and the biggest obstacle to actually test out the light is to get thermal paste for the LEDs.

 

[vari]

Now the AC side and DC output of the lighting driver is completed and mounted to the casing. The biggest changes so far have been switching over to a radial fan for cooling, dropping the bigger capacitor and replacing the series resistor with a coil, which increased the output voltage somewhat, but makes it independent of load. It also increases overall system efficiency.

The ripple is surprisingly small even with only one capacitor. The test load i used was a 89 ohm resistor, and the output voltage under load was 114V. This gives us a 1.3A current resulting in a 145W load.

I also applied thermal paste to the LEDs, so now they are ready for use.

TODO: Connecting and mounting the current regulators and connecting the LEDs.

 

[vari]

If you have not been following my blog, then here is a quick update on the current status of my LED lighting project.

I finished the first 150W module (running at 124W) and made some comparison shots with a 110W high-powered fluorescent light. The pictures are taken at manual setting with a DSLR so they represent the real difference in brightness quite well.

Currently i am working on fitting a 12V voltage regulator into the casing, so i could drive all the cooling fans from the same lighting driver without an external wall-mounted 12V adaptor.

I have also ordered 3x 100W LED emitters, that i will connect with the current 50W ones to get to a total of 450W of LED power

 

[vari]

Now i have finished the project. The full description and set of build pics can be found at my blog at LED Lighting Project 150-900W | Varikonniemi.

 

[BShanahan14rulz]

I like the last picture. It shows the different tints from the different arrays you used quite nicely.