DIY LEDs - The write-up

[dnahas]

With the updated specs from cree on the XP-G's, 1.5A rather than 1.0A, should i adjust the meanwell's to run them at 1.5 or stay with 1A?

 

[der_wille_zur_macht]

If it were me, I'd stay at a lower current and use more LEDs if you want more light. Even though they'll take 1.5A, they're a lot less efficient at that point.

 

[koensayr]

Hi all,

Thanks for this great informative thread. I've been reading it for hours, but it is just too much to cover everything. I'm planning on building a LED lighting for my small tank and perhaps the veterans here can give me some comments and suggestions.

My tank is a ~5gallon tank that's approx 12'' x 12'' x 14'' (depth). I'm just estimating the size right now cause I am at work and cannot measure the tank. Right now, I only have a frogfish, some green star polyps, and one yellow polyp in there. I don't think those polyps need much light, but I'm thinking of possibly adding more corals in the future. Also, the lighting right now is too weak, I think, even for the polyps, as they are not really expanding.

Here is what I have in mind for the lighting at this moment:

2 XP-G CW, 2 XR-E RB, 2 XR-E Blue + 1 Buckpuck driver (500mA)

My idea is to get those up and running first. Then I can add a PIC or Arduino controller and perhaps other things in the future (such as another driver to be able to control the different colors separately).

Now, if that is perhaps too much light for my tiny tank, maybe I can also scale it down to 4 LEDs total, but I'm not sure how I should choose the colors in that case.

The only other thing that I'm not sure how to approach at this moment is how to build the fixture. Since the tank is very small, I am thinking that the easiest way might be to make something that straddles two edges of the tank. But I'm not sure what's the best way to go about building something like that.

The other thing is enclosure. I read a bit earlier where DWZM mentioned something about corrosion on the solder joints. If I cover up the joints with say, hot glue from a hot glue gun, would the rest be ok with not being behind an enclosure?

Hope you guys can provide some insights to help me narrow down the design.

 

[dnahas]

I would think 1 acrylic sheet mounted to the fixture would be a much better way to go than hot glue.

 

[dnahas]

Now for a stupid electrical question to make sure I am not doing something wrong. If the 48d meanwell is a 1.3A driver, why am I reading 1.7 with the current set all they way up?

 

[HeneryH]

Real specs for all the Carclo optics can be found here.

http://carclo-optics.co.uk/catalog/documents/Cree-01-10.pdf?osCsid=639fae1a16e0ce3a4fbe7858cedd4dbc

Wow, the Cutter specs are totally screwed up. Both the angles and efficiency are way off.

 

[der_wille_zur_macht]

I would think 1 acrylic sheet mounted to the fixture would be a much better way to go than hot glue.

Agreed - no glue or other substance is required on the solder, just protect the WHOLE ARRAY from salt spray, etc. Put an acrylic shield on it, but mount it in such a way that spray or warm, salty air won't constantly get sucked or blown past the LEDs themselves.

I think your plans for 6 LEDs on that tank are fine BTW.

Now for a stupid electrical question to make sure I am not doing something wrong. If the 48d meanwell is a 1.3A driver, why am I reading 1.7 with the current set all they way up?

How are you reading the current? You must connect the meter in series with a load, otherwise you won't get accurate numbers and/or the driver won't be properly regulating itself.

 

[koensayr]

Well, I was thinking of a way to get away from having the light go through a sheet of acrylic because acrylic has about a 10% absorption across the visible spectrum.

 

[dnahas]

How are you reading the current?

Basically interrupting the v+ from the meanwell with the DMM.

Cutting into the v+ output from the meanwell into the + probe of the DMM and then the - probe of the DMM continuing to the start of the diode series . The - out of the meanwell connected to the other end of the series as normally operational.

 

[der_wille_zur_macht]

Ok that strikes me as correct, but getting a reading of 1.7A scares me for two reasons: 1) the LEDs should have let out the magic smoke at that point, and 2) the meanwell shouldn't be capable of generating that much current.

What are you doing for a dimming signal?

Also, do you have any way of measuring a known standard with that multimeter? i.e. measuring the voltage drop across a resistor or something?

 

[dnahas]

Scares me too for both reasons, that's why I posted the question !!

I am dimming with a 10V wall wart, until I finish and use the apex.

I guess I could run to rat shack and get a resistor to test.

 

[kcress]

Well, I was thinking of a way to get away from having the light go through a sheet of acrylic because acrylic has about a 10% absorption across the visible spectrum.

As far as I'm concerned the only good way to lick this is with distance. Then go for optics. Unless your LEDs are less than 3" from the water you are going to have tons of spillage. Light growing algae on your glass, lighting the floor in front and behind your tank, etc., etc.

Move the lighting up a foot, or two, or three, and run optics. The optics keep dust off the LED lenses. The height keeps all non optical heat out of your tank. The optics align the light rays into columns that go where you want the light to go - into the tank.

Otherwise... You need a glass or plexy shield. And, you are inheriting the headache of weekly or monthly cleaning of any splatters off of this shield.

 

[der_wille_zur_macht]

I was going to suggest optics and lots of height as a way to get around having a splash shield, too. Then I realized - with optics, the light is going through a hunk of acrylic, too. :lol: So you win one battle and lose the other.

It's arguably all going to come out in the wash though from an efficiency perspective. And not having to clean so frequently is a nice potential benefit.

 

[jimlundborg]

I've been without a tank now for a few years (because of a move) and I'm now in the planning stages of a new nano build. I am really intrigued by these LEDs but I'm not too confidant in my DIY electronic skills. One question I have is how do you adjust the current from the drivers. I've been reading this thread and always see comments about running at 700ma and / or 1A and I don't have experience in electronics.

 

[TheFishMan65]

You have to make sure you get an adjustable driver. Then there is a potentiometer (screw head) that you use to adjust the current (and in some cases voltage). If you decide on a driver (usually Meanwell or Buckpuck) someone will help you decide which one to get.

 

[der_wille_zur_macht]

One question I have is how do you adjust the current from the drivers. I've been reading this thread and always see comments about running at 700ma and / or 1A and I don't have experience in electronics.

Jim,

I'll give you a three part answer.

First, it depends on the driver. Most/all drivers come with a preset "max" limit. For instance, you can buy a 700mA buckpuck. It'll run at 700mA out of the box. No adjustment or thought required.

Second, some drivers have an internal adjustment that allows you to set a lower max current. For instance, the meanwell ELN-60-48 has a max current of 1.3A, which is really too high for our needs. But there's a trimpot inside that lets you dial it down to something more reasonable. These internal adjustments are meant to be an initial configuration item, i.e. you set it when you build the rig then you "forget about it."

Thirdly, many drivers have an external dimming capability. This lets you hook up a potentiometer (knob) to adjust the current "on the fly" as the driver is in operation. Or, you can hook up a microcontroller to dim the LEDs on some preset schedule, i.e. to imitate sunrise, sunset, cloud cover in the afternoon, etc. These external dimming controls are meant to be a "normal operation" adjustment, i.e. they are provided for use on a regular basis during normal operation of the driver.

 

[dnahas]

In the interest of learning and informing others I'll divulge my idiot mistake on the current reading.

It seems I did not notice the pot for the wall wart was set to max, 12V. It seems that I am lucky and did not let the smoke out of the LED's or meanwells.

I forget where but a few pages back someone had mentioned running 12 to the dimmer circuit on the meanwells and not sure what the out come would be. Here's my ignorant opinion, I assume it is what caused the current to go over the 1.3A rating of the meanwell. That doesn't make a lot of since, I figure it would max @ 1.3A.

Instead of using the wall wart while testing the rig I am going to get the apex up and running and use the apex dimming circuits to ensure 10V going to the meanwell.

I sure am happy I didn't let the smoke out!

 

[The Grim Reefer]

If your current readings were accurate it says a lot about the quality of the Cree LED's. I wouldn't have been surprised if you had let the magic fire out, let alone the magic smoke.

 

[der_wille_zur_macht]

I'm still wondering about the accuracy of the current measurements. Not that I doubt your findings, but it would surprise me if providing a too-high signal on a dimming circuit resulted in a higher-than-normal output. Typically, the maximum output is hardwired in the unit's guts, and the dimming circuit is used to restrict the driver to a smaller output, i.e. there's literally no way for the dimming signal to overdrive the LEDs.

Then again there are a few ICs out there for driving LEDs that would allow this to occur if in a "poorly" designed circuit, so I guess it is a possibility.

 

[jimlundborg]

Ok thanks all that clears it up for me. I'm building a 10 nano and I'm now leaning toward either one of these two kits
http://www.rapidled.com/servlet/the-38/06-Premium-LED-Do-dsh-It-dsh-Yourself/Detail
or
http://www.rapidled.com/servlet/the-36/12-Premium-LED-DIY/Detail