DIY LEDs - The write-up

[dipan]

Are the Meanwells starting to become available again, if so where is everyone getting them?

Yes. I got 2 from nanotuners. Another 40+ on the way, but probably sometime next month.

 

[laverda]

Another big reason to anodize your heat sinks is it will help keep the aluminum from oxidizing, which it will do fairly rapidly around salt water. Salt spray will cause oxidation of anodized or natural aluminum but it Will happen faster and worse on natural aluminum. and as some one pointed out, oxidized aluminum transfers heat very poorly.

 

[dipan]

Another big reason to anodize your heat sinks is it will help keep the aluminum from oxidizing, which it will do fairly rapidly around salt water. Salt spray will cause oxidation of anodized or natural aluminum but it Will happen faster and worse on natural aluminum. and as some one pointed out, oxidized aluminum transfers heat very poorly.

Anodization is an intentional thicker oxidization layer (usually with added dye). As kcress pointed out bare aluminum oxidizes spontaneously, but this is a very thin oxidation layer. Anodization applies a thicker oxidation layer.

 

[g8gxp]

I looked into this several months ago. Unfortunately, most places do not anodize small pieces for cheap. Quite often they have a minimum charge around $100 or so. i have a 4 foot heatsink that still fell under the minimum charge of $125. For a smaller build IMO it's not cost efficient. Better to buy a couple of $4 comp fans.

Plus it would be a PITA to cover the spots where the LED stars would go.

 

[kcress]

Thanks widmer and DWZM on the fan feedback. Interesting. I guess I can add that scheme to designs. Nice.

 

[DFason]

I found a place that will anodize quick and cheap. I will be sending mine in for a light gray color.

He quoted me $3.50 per heatsink ( 5x5 ) and $3.00 per heatsink to glass bead them.

http://www.six-2.com/Home/

-Dave

 

[der_wille_zur_macht]

If it were me, and the sinks would normally be hidden from view, I'd do lime green or hot pink or something. Then when you have other reefers over to check out your sweet LED rig, you can play a game to see how long they'll listen to you ramble about how that specific color helps dissipate gamma rays left over from combustion of the lubricants inside the LEDs. . .

 

[HndMafia]

if it were me, and the sinks would normally be hidden from view, i'd do lime green or hot pink or something. Then when you have other reefers over to check out your sweet led rig, you can play a game to see how long they'll listen to you ramble about how that specific color helps dissipate gamma rays left over from combustion of the lubricants inside the leds. . . :d

lmao!

 

[TheFishMan65]

Food for thought

Food for thought

I read all the threads on LED builds here (ok any I could find), and I know there is a lot more information out there. One of them (maybe this one) asked if PWM dimming had any affect on coral growth. I never saw an answer. So I have an idea and need some help disproving or proving it. Thanks in advance.

I think I read that the Arduino has a PWM at 400Khz. This means that it turns on and off once every 1/400,000 of a second or 2.5 microseconds (10 raised to the minus 6th). Someone please correct me if I am wrong here, I have never worked with the arduino. I found slower speeds online and they would be worse by my theory.

One of the threads (I have been reading a lot) linked to:
http://reefcentral.com/forums/showthread.php?t=1496092
From my interpretation some of the the photosynthesis reactions in coral are taking place in the nanosecond(-9) and picosecond (-12). Also quoting greenbean36191 (who sure sounds like knowledgable):

Photosynthesis is essentially instantaneous, so the photoperiod has nothing to do with saturation.

This means that it does not matter how long the light is on, saturation [bleaching] is dependent on how bright it is.

So with PWMs are we creating lots of little days for the coral since it can do 1,250 to 1,250,000 photosynthesis operations in the mini day at a 50% duty cycle. It may look dimmer to us, but maybe not the coral.

So the question is. If you had a bleaching problem with your corals when you went to LEDs AND you turned down your LEDs did it help and what dimming control did you use.

My theory is that those with PWM saw no or little affect by turning them down, but those that actually dimmed by limiting the current would see a change.

I hope this was clear, becuase I think this could be an important step in creating the best environment for our livestock.

So did any lights come on?:idea:

 

[der_wille_zur_macht]

Very interesting, thanks for the research.

Though, the Arduino's PWM frequency is more like 400hz, not 400khz - so 1/400 second, not 1/400,000.

FWIW, for my 360g tank, I'm going to use a lower-than-typical drive current (probably 500mA) and hopefully only dim to create visual effects, not to prevent bleaching.

(Edit - just looked it up on arduino.cc and the frequency is 488hz, but you can change it within a narrow window by editing the wiring.c file in your Arduino environment.)

 

[stugray]

If there was a concern about PWM dimming, there is a way around that.

If you were to dim the LED current through a MOSFET that was designed to be capable of "avalanche current" and intentionally put an inductor in series with the LEDs, then during the OFF phase of the PWM, the current will not go all the way to zero.

During "avalanche" the MOSFET continues to conduct current even though it was told to turn Off. The Inductor provides "momentum" for the current.

So the current waveform would look more like a sinusoid than a square wave.

By adjusting the value of the inductor, you could tune it to the PWM frequency and "smooth" out the waveform.

Sorry for the technical bit.

Stu

 

[der_wille_zur_macht]

Here's an interesting article:

http://ncr101.montana.edu/Light1994Conf/5_11_Bula/Bula text.htm

photosynthetic rates of tomato leaves were equivalent when the light was provided as a pulse of 5000 μmol.m-2.s-1 when on 1 % of the time (1.5 μs on and 148.5 μ s off) compared with a continuous photon flux of 50 μmol.m-2.s-1

If I'm reading that correctly, it's arguing against your conclusion - a pulsed, high intensity light gave the same results as a constant, low intensity light.

I'll PM greenbean and see if we can get his comments again. . .

 

[der_wille_zur_macht]

If there was a concern about PWM dimming, there is a way around that.

If you were to dim the LED current through a MOSFET that was designed to be capable of "avalanche current" and intentionally put an inductor in series with the LEDs, then during the OFF phase of the PWM, the current will not go all the way to zero.

During "avalanche" the MOSFET continues to conduct current even though it was told to turn Off. The Inductor provides "momentum" for the current.

So the current waveform would look more like a sinusoid than a square wave.

By adjusting the value of the inductor, you could tune it to the PWM frequency and "smooth" out the waveform.

Sorry for the technical bit.

Stu

Isn't that essentially how a constant current buck regulator works?

 

[JOHN4AVIATION]

I looked into this several months ago. Unfortunately, most places do not anodize small pieces for cheap. Quite often they have a minimum charge around $100 or so. i have a 4 foot heatsink that still fell under the minimum charge of $125. For a smaller build IMO it's not cost efficient. Better to buy a couple of $4 comp fans.

Plus it would be a PITA to cover the spots where the LED stars would go.

Why dont you just Alodine the aluminum for corrosion protection. A Qt of Alodine will cost between $20 to $30. and a bottle of Alumiprep 33 another $30. That is what we use in Aviation to Prep the Aircraft skin for corrosion prevention.

 

[TheFishMan65]

der_wille_zur_macht I understand your point on tomatoes, but we would need to know if the leaves reached the saturation point to know if it was equivalent. I think it did not. If a tomato needs and can use 5000 μmol.m-2.s-1 (units) a seconds it won't matter how they get it. As one burst or trickled in - at least that is my understanding. To answer this we would need to find out the out how many units it takes to burn the tomato. Then apply 1% of that rate to the tomato on a PWM. If my theory holds it will then burn the tomato.

Thanks for the response - I sort of thought people might think I was off the wall.

 

[MontyJack]

Hi Folks!

I just won't to verify my plans and before I order the LEDs I will ask you which Bins I should order?? I'm not sure if it should be Cree XR-E or XP-G. I'm somehow lost in the amount of information now! I now that the XP-G R5 are the best out now, but still expensive. Also I didn't find an equivalent in Royal Blue. So if I mix them, the Whites will be to strong while the Blues will be inefficient in comparison. So what do you recommend? Which Bins should I mix together and will be the best in W/Lumen/price??

Thanks a lot Monty

 

[laverda]

Anodization is an intentional thicker oxidization layer (usually with added dye). As kcress pointed out bare aluminum oxidizes spontaneously, but this is a very thin oxidation layer. Anodization applies a thicker oxidation layer.

Anodizing is a very thin layer, easily scraped or sanded off. Oxidation around salt water especially from splash can get very thick, greatly reducing cooling and also create pitting. I have seen it 1/4" thick or more.

 

[der_wille_zur_macht]

Hi Folks!

I just won't to verify my plans and before I order the LEDs I will ask you which Bins I should order?? I'm not sure if it should be Cree XR-E or XP-G. I'm somehow lost in the amount of information now! I now that the XP-G R5 are the best out now, but still expensive. Also I didn't find an equivalent in Royal Blue. So if I mix them, the Whites will be to strong while the Blues will be inefficient in comparison. So what do you recommend? Which Bins should I mix together and will be the best in W/Lumen/price??

Thanks a lot Monty

Monty,

FWIW, you'd be fine with either XR-E or XP-G. The best XP-G (R5 bin) are around 30% more efficient than the best XR-E (Q5 bin). That means you'll get that much light at the same wattage, or you can run them at lower wattage and get the same light. It's not a DIRECT swap, because the XP-G pull slightly different voltages than the XR-E at given currents, so make sure that you consider that when planning for drivers. The voltages are all in the datasheets.

You're right on colors, too. There is no blue or royal blue XP-G. So you have to use other LEDs for blue - XR-E, XP-E, Luxeon Rebels, whatever- they're all pretty much equivalent. Most people go with a 1:1 ratio of cool white to blue, so if you mix XP-G and something else for blue, you might want to load up on blues a bit more if you want that "typical" blend.

For price, I've seen the XP-Gs quoted as low as just under $6 (the nano-reef group buy that ends soon) to over $7. So they're proportionately more expensive than XR-Es, which in cool white, you can get for like $4.50 from dealextreme. It's personal choice - would you rather pay a LITTLE mroe up front to save a LITTLE every month on your power bill? In smaller tanks, the difference would be small enough that accidentally leaving a light or two on overnight in your home would make up the difference on the power bill. In a bigger tank, it becomes a bit more significant.

 

[sbleile]

Has anyone used the XP-E for the Royal Blues?

 

[dipan]

Can 12v DC fan wires be run with the LED control wires, or will there be interference? I was thinking about using a multiconductor cable from multiple Meanwell's to the LED's. Would 22g suffice?