Lightblaze 400 clone - beginning tests - pics

[ericjwi]

I'm in the process of building a clone of the following using 3w LEDs (more power )
http://www.superled.net/ledgrowlights.html

Initial test pictures for power circuits and heat dissipation. I'm only up to ~13W of light on this test. P4 heatsinks work pretty good

 

[NotBright]

Cool project.
It probably isn't worth it in terms of cost though. That many LED's can get pretty expensive and I believe the efficiency is very close to a HID bulb. I'm interested to know how well this works for growing so please keep us updated. An equipment list including LED's(how many,bin,tint) and driver(s) used would be good too.

 

[jeraldjunkmail]

The advantage of using LED lights to grow plants is that plants only absorb light energy in a limited wavelength between 640 and 720nm. So you are not wasting any energy producing light outside of those frequencies. Ever wonder why a plant is green? Because it reflects all the green light because it is useless for photosynthesis. You are on average producing 80% or more efficient light by using frequency tailored LEDS. The real trick would be to use a laser to grow plants, as I think that they are watt for watt eh most energy efficient light on the planet, though I could be wrong as I never did find that out for sure... A 400 watt laser puts out a LOT of light if you could diffuse it efficiently... just a thought

Cheers, Jerald!

 

[ericjwi]

Lol yeah the entire project $$ kinda got out of hand, probably about $500. Still cheaper than the $650 for the Lightblaze. It was cheaper until I ordered the MCPCB but my results have been a PITA without it and my laser printer PCB idea was going to delay the project a long time.

FYI those 5mm lights are not very efficient for this type of project, it kinda has a bad rep because of people selling those lights that don't work very well.

 

[ericjwi]

Ever wonder why a plant is green? Because it reflects all the green light because it is useless for photosynthesis.

Its hard to tell in the picture but one of those looks very black in person due to such little light getting reflected back.

 

[Oznog]

What LEDs did you end up using?

We talked about grow lights here:
http://www.candlepowerforums.com/vb/showthread.php?t=204092

What I seemed to be seeing from the data is that most red LEDs miss the sharp red peak in the photoactive response, the wavelength is too short. Consequently it won't be very effective.

That LightBlaze thing says it uses Cree XLamps, which reading the charts indicates is not very effective since the peak response is at 670nm and the Xlamp is a sharp 625nm.

The longer wave LEDs "seem" like poor performers because of low lumen/W figures. But the lumen scale is weighted for human eye response via the photopic lumen chart. That scores 670nm much lower per mw of light output than 625nm reds or 615nm red-orange. So you actually look at it terms of PAR/W not visible lumens/W then they score pretty good.

 

[Oznog]

I've been having a hard time understanding the chart though:

See in the middle (greenish) region, Absorption is very very low. Yet it shows significant Photosynthesis Rate there. My question is, does it mean it will absorb very little and make only modest use of what it absorbs?
Or does this mean the plant absorbs very little but responds very strongly to what it absorbs?

Look at 550nm. About 3% Absorption but 25% Photosynthesis Rate. Do I read this as absorbing 3% and at a rate of 25% it's only 0.75% efficient unless we can capture the nonabsorbed reflected light and reflect it back onto the plant a second and third time and so on until it absorbs?

OR, do we say it absorbs 3% but that it reacts 8x stronger, producing a 25% PR? I don't see how, since the % unit would make no sense, I believed the % unit was mw stored chemical energy per mw of light in. One can't absorb 30mw of a 1w source and get 250mw of energy. So this interpretation doesn't seem to pass the sanity check.

Still most growlight manufacturers use the lower PAR chart there, even though I'm unclear on its relevance. I mean if it's showing that there's 25% response at 550nm that's kinda deceptive if the pigments only use 0.75% of the incoming light.

The Wikipedia descriptions are in general terms, but it seems no one wavelength can feed the plant even if the PAR chart shows it to have a top-notch response. Reason being that photosynthesis "Z-scheme" shows 2 steps and I think that 2 different wavelengths are required to complete the cycle. And there are not just 2 different types of chlorophyll but the carotenes and xanthophylls with their own wavelength needs. Two LED wavelengths might not be able to best meet these needs either. I'm not sure. I'm sure it depends on what plant you're trying to grow too. Your plant may have different ratios of cholor a/chloro b/carotenes/xanthophylls.

This could actually lead one to conclude the light requirements are the OPPOSITE of the PAR chart. If you have a photosynthetic step with a poor 5% response to a wavelength and it's essential to completing a cycle needed for growth, nutritional content, or flavor, then maybe you need to add much MORE energy at that wavelength to ensure the process's needs are met. Or maybe this step only needs a trivial amount of energy and the bulk of the process happens at a high peak response? The issue seems more complex than the chart can display.

 

[Oznog]

I think a lot of the effectiveness depends on light capture.

See, the setup you depict shows like 0.1% of the light actually landing on the leaves. So using a reflector or putting the LEDs being closer should raise the efficiency way up, potentially tens or even hundreds of times more efficient. That's significant since we're talking about aiming to be somewhat more efficient in PAR generation than HID or CFL but could be much much much more efficient by revising the light capture configuration.

The difficulty I noted is in directionality. The powerful LED emitters usually have Lambertian distribution. That sucks actually, because putting them closer results in very uneven intensities- leaves underneath the emitter getting far more than those to the side- and it STILL spills a great deal of light off to the side. Ideally we'd use an emitter than creates a plane of parallel light the same size as the planting area, in which case it wouldn't matter how high the source is from each leaf they all get the same angle, though it does matter how obstructed the lower leaves are.

I figured the thing to do is mount LED emitters on flexible stalks so you can rearrange them to point at the plant's leaves in the pattern most appropriate for that specific plant.

The 5mm T1-3/4 LED panels should be surprisingly good though, since they deliver a far more focused beam and are most of the way to being an ideal parallel planar light source. Except of course for all the issues surrounding mass T1-3/4 arrays, and the lack of 670nm reds.

Did you use reflectors to focus the beam? It doesn't really look like it in the pic.

 

[jeraldjunkmail]

A collimator and a mylar enclosure would help. Also, spin the plant on a slow, constant basis, and your efficiency will go way up...

Cheers, Jerald!

 

[ericjwi]

What LEDs did you end up using?

We talked about grow lights here:
http://www.candlepowerforums.com/vb/showthread.php?t=204092

What I seemed to be seeing from the data is that most red LEDs miss the sharp red peak in the photoactive response, the wavelength is too short. Consequently it won't be very effective.

That LightBlaze thing says it uses Cree XLamps, which reading the charts indicates is not very effective since the peak response is at 670nm and the Xlamp is a sharp 625nm.

The other forums I've been on had a lot of success with the Lightblaze, one of the best performers. I'm using Ledengin with about 10 of the reds being "deep red" at 660-670nm so it should perform even better.

www.greenpinelane.com

procyon uses cree but the older less efficient blues

Blooming is WAY less watts than HID.

 

[ericjwi]

The 5mm T1-3/4 LED panels should be surprisingly good though, since they deliver a far more focused beam and are most of the way to being an ideal parallel planar light source. Except of course for all the issues surrounding mass T1-3/4 arrays, and the lack of 670nm reds.

They also put out MUCH less light/watt and much less intensity. I can't remember the terms they used, had to do with the number of photons getting put out.

 

[Oznog]

Ah yes LEDEngin is the only really powerful emitter than sits right there on the 670 nm response peak. That should be very effective- in theory at least.

It's too bad only this odd LEDEngin brand is the only deep red mfg... too used to cheap high powered stuff from Cree, Luxeon, etc coming up on DealExtreme or whatever.

 

[ericjwi]

Beta3 is coming soon... Need smaller than 0.27ohm resistor, get the mounting good and harden the power circuit (will come undone pretty easy... but the LEDS are very secure). Stats:

16 High Power CREE running at about 3W each - This guy is very bright, actually had 50,000LUX reading at 6in at full load (LED+ on wrong side of resistor1). Not accurate for colors but much better than where I was at when comparing to the same wavelengths.

Power Circuit design running 15V currently 1.85A - 6R/4B/6R in parallel each group of leds about 610ma (I want this closer to 1000ma).