Minimalistic multichip DIY LED build

[nettraptor]

Hello guys i need some help...

i got 2 X 40mil 50W Super Actinic Blue Hybrid Led Panel from AC-RC
and one 50DreamLed

and his 100Watt driver with poti. to run the 2 actinics

i started a project with arduino in order to use pwm and dimm the DreamLed..
so i got 36V 9.7A 350W AC/DC Switching Power Supply and 5 10watt drivers to work with it...

I wonder now if there is any pwm driver for the 50W actinics
specifications
DC forward voltage: DC 33V-36V
DC forward current: 1750mA max
Rated power: 63W max

Can i use a LMD18200-DC-Stepper-Motor-Driver as a led driver??

i appreciate any help!

 

[jgwinner]

Well no, motors and LED's are very different

You need an LED driver.

LED's present a constant current, variable voltage source (as they change temperature they will change their voltage, regular power sources will then feed to much current).

Technically, it's more like a constant current power supply, but the easiest thing to remember is you need an LED power supply / LED Driver.

The PWM usually isn't the 'driver' it's usually an input (extra wire) INTO the driver. Check out the Meanwell LDD thread, although those parts only go to 1 amp, did you want to feed yours with 1.75? This thread may have other driver suggestions in it.

== John ==

 

[coralhaven]

wow those units look great, very professional! How long did it take you to put these together?

 

[nettraptor]

Well no, motors and LED's are very different

You need an LED driver.

LED's present a constant current, variable voltage source (as they change temperature they will change their voltage, regular power sources will then feed to much current).

Technically, it's more like a constant current power supply, but the easiest thing to remember is you need an LED power supply / LED Driver.

The PWM usually isn't the 'driver' it's usually an input (extra wire) INTO the driver. Check out the Meanwell LDD thread, although those parts only go to 1 amp, did you want to feed yours with 1.75? This thread may have other driver suggestions in it.

== John ==

thanks John for you reply.
Yes i know the pwm and what exactly does.
I just want to find a driver wich will be connected into the AC/DC Switching Power Supply and work with the pwm over the Arduino.And i need at least 1.5Ampere to drive this LED... as you said the LDD series are fine but the maximum that they can give is 1000mha....

 

[Mike31154]

also what spacing do i need if i use 20 10w led to avoid the spot light effect?

Aside from spacing pattern of the chips themselves, don't forget that height above water & optics are factors that will affect light distribution & the possibility of spot lighting. My 10 watt chip build uses 27 LEDs without optics mounted approximately 6 inches above the water surface. The tank is a 77 gallon Hagen & is 4 foot long, 24 inches tall & 16 inches front to back. The LEDs are on 3 rails of 9 each with the outer rails tilted slightly inward. Spacing between LEDs is about 4 & 1/2 inches on center. I have to look fairly close to see much in the way of spotlighting, it really is barely discernable, probably not at all to the casual observer. I use a mix of Blue, Royal Blue & two flavours of white in an alternating pattern, i.e. one white then one blue on each rail along the length. The center rail alternates between blue & 6-7,000k white (5 blue, 4 white) while each outer rail constists of alternate 10,000k white & royal blue LEDs (5 white, 4 royal blue). If I were to rebuild, I might move the LEDs a half inch closer together, to 4 inches on center & perhaps add some violet or UV LEDs.

 

[TropTrea]

Aside from spacing pattern of the chips themselves, don't forget that height above water & optics are factors that will affect light distribution & the possibility of spot lighting. My 10 watt chip build uses 27 LEDs without optics mounted approximately 6 inches above the water surface. The tank is a 77 gallon Hagen & is 4 foot long, 24 inches tall & 16 inches front to back. The LEDs are on 3 rails of 9 each with the outer rails tilted slightly inward. Spacing between LEDs is about 4 & 1/2 inches on center. I have to look fairly close to see much in the way of spotlighting, it really is barely discernable, probably not at all to the casual observer. I use a mix of Blue, Royal Blue & two flavours of white in an alternating pattern, i.e. one white then one blue on each rail along the length. The center rail alternates between blue & 6-7,000k white (5 blue, 4 white) while each outer rail constists of alternate 10,000k white & royal blue LEDs (5 white, 4 royal blue). If I were to rebuild, I might move the LEDs a half inch closer together, to 4 inches on center & perhaps add some violet or UV LEDs.

I have built rail systems using 3 Watt (XP-E) LED's in the past with 3 1/2 inch spacing on 1" X 2" channels. I'm about to start a new project using the newer 10 Watt LED's. My big question is with your 4 1/2" spacing if your having any heat issues? This is the one big thing I'm afraid of with the 10 Watt LED's where my investment in cooling could surpass the cost of the lighting itself.

 

[Mike31154]

I have built rail systems using 3 Watt (XP-E) LED's in the past with 3 1/2 inch spacing on 1" X 2" channels. I'm about to start a new project using the newer 10 Watt LED's. My big question is with your 4 1/2" spacing if your having any heat issues? This is the one big thing I'm afraid of with the 10 Watt LED's where my investment in cooling could surpass the cost of the lighting itself.

I used individual heatsinks with fan for my 10watt build, so spacing is no issue whatsoever, I could butt the heatsinks against each other if I wished. That would yield a spacing of 2 & 1/4 to 2 & 1/2 inches depending on whether the heatsink is mounted square or diagonally on the aluminum angle. Mine are currently mounted diagonally because that seems to expose more of the heatsink fins to open air. Downside to the individual heatsink with fan/LED is a lot more wiring.

Here are the components I ordered. I decided not to use the optics. The glass is heavy & 27 of these would have added considerably to the weight of the fixture, plus they're a pain to mount. Fixture works great mounted 6 & 1/2 inches above the water without optics.

You'll find that many eBay vendors offering the 10 watt LEDs will also have these heatsinks. They're not that costly & I purchased them in bulk orders of 10 or more. I have to admit though, that they appear to be a bit of a weak link in the build. I've had to replace 3 of them due to failure. Upside is, the heatsink itself, even with a dead fan appears to provide sufficient cooling to keep the LED from burning out. I'm sure a couple of them were lit for hours before I noticed the fan was dead with no apparent ill effects. The fixture has been running since last April with no LED failures. I've also seen heatsinks offered for the 10 watt LEDs that don't have a fan. They're a bit larger & more costly. I haven't tried them personally.

Here's a pic of the rail set up during initial assembly. As per my previous post, I ended up with 3 rails of 9 LEDs each.

This is the almost completed fixture over my tank. Photo take some time ago. I still need to figure out how to put a nice top on the thing to cover the mess of wiring, although open top is handy to keep an eye on the fans. My idea is to use some wooden slats left over from another project, mounted on a couple of semi circular end pieces. I'd leave some space between the slats for air flow. One of the things I like about this build is that the fixture only weighs around 12 lbs.

 

[TropTrea]

Even though I'm doing a 90 Gallon tank I can probably cut back on the LED's since I'm also suplementing with a pair of HO-T5's. I could get way with running 6 of the 10 Watt White LED's using CREE XM-L Led's which would give me roughly 8 inch spacing and keeping my rail design would give me 64 square inches of cooling surface per LED. Then for my Blues and Royal Blues I'd be running 5 watt LED's so I'm looking at 12 to 24 of these. At 5 watts the cooling is not as important since I will have more than enough spacing with 3 rails.

 

[miatawnt2b]

I love those little heat sinks.

 

[Mike31154]

I love those little heat sinks.

Yes, they do the job nicely without needing to worry about having enough metal mass for cooling. Only problem is that with 27 LEDs & 27 fans, there is a lot of wiring and by the time you're done, it's anything but 'minimalistic'. If I were to start over, I'd probably go with fewer, larger multichips mounted higher & using optics.

When I first wired these up I soldered the fan wires to the LED terminals to cut down on the wiring. However this meant that the fans would not spool up when the LEDs were dimmed. This is probably ok since the LEDs don't produce that much heat dimmed down, but in the end, I cut the fan leads off again & ran separate wires to them.

I drilled & tapped the heatsinks to mount the LEDs with small screws & heatsink compound. This photo shows the original work with the fan leads soldered to the LED terminals. I cut the connectors off to do this. The heatsink in the center still has the connector. The solid strand wire I used slides nicely into the connectors, just like a pin on a male connector would.

 

[bishop777]

Help help
I have a 210gal tank and I would like to use the multichip led but not sure on what I would need. And that's where I need help. My tank is 72x24x28 and I would like any suggestion and advise. Thanks

 

[Whitebeam]

Cheated by Chinese LED sellers!

Cheated by Chinese LED sellers!

:furious: I've just been using Spectralworkbench to compare 4 supposedly different blue multichip modules bought from different eBay sellers. One module advertised as 430nm, one as 445nm, one as 455nm and one as 460-470nm. Guess what? All except the 430nm have exactly the same 445nm peak spectrum :furious:

Module 1
<iframe width='500px' height='400px' border='0' src='http://spectralworkbench.org/spectra/embed/3454'></iframe>

Module 2
<iframe width='500px' height='400px' border='0' src='http://spectralworkbench.org/spectra/embed/3455'></iframe>

Module 3
<iframe width='500px' height='400px' border='0' src='http://spectralworkbench.org/spectra/embed/3456'></iframe>

Module 4
<iframe width='500px' height='400px' border='0' src='http://spectralworkbench.org/spectra/embed/3458'></iframe>

That last one, which is supposed to be 430nm is a really peaking at 420nm (but at least I suppose I can say that the 445nm peak is a real 445nm peak!).

Caveat emptor!

Peter

 

[Skinnysloth]

Thanks for sharing your findings, Whitebeam. I can't say I'm all that surprised. I wonder how consistent these peaks are after a few months of usage.

 

[acabgd]

What I would like to know is what were the leds used and from which sellers - without that this info is pretty useless.

 

[TropTrea]

This is not at all supprising. Yhey all probably also fall within legal advertising standards as there band widths do extend into the range they claim. It just is not where there peak is.

Also using any photographic camera will distort these numbers slightly. Photographic cameras do have a nuilt in UV filtering that decreases the amount of light they can detect when you start getting to wave lenghts under true blue. But this should more or less reduce the recorded intensity more than shift the frequency of anything.

:furious: I've just been using Spectralworkbench to compare 4 supposedly different blue multichip modules bought from different eBay sellers. One module advertised as 430nm, one as 445nm, one as 455nm and one as 460-470nm. Guess what? All except the 430nm have exactly the same 445nm peak spectrum :furious:


That last one, which is supposed to be 430nm is a really peaking at 420nm (but at least I suppose I can say that the 445nm peak is a real 445nm peak!).

Caveat emptor!

Peter

 

[megadeth72]

This is not at all supprising. Yhey all probably also fall within legal advertising standards as there band widths do extend into the range they claim. It just is not where there peak is.

Also using any photographic camera will distort these numbers slightly. Photographic cameras do have a nuilt in UV filtering that decreases the amount of light they can detect when you start getting to wave lenghts under true blue. But this should more or less reduce the recorded intensity more than shift the frequency of anything.

Your supposed to remove the uv filter, I did on mine

Sent from my SCH-I535 using Tapatalk 2

 

[Whitebeam]

Your supposed to remove the uv filter, I did on mine

Yeah, but I took photos using my mobile phone and uploaded them - I'm not going to mess with the filter on that. Anyway, a cut off of UV won't affect those readings significantly, as the 420nm peak of the 430nm LED shows.

I do have a video camera that I've stripped and taken the filter out of, but I have yet to get a suitable case to put it in. I also want to source a 'proper' diffraction grating from somewhere

Thanks Peter for these graphs. Do you know the types of LEDs used by EverGrow chinese LEDs exhibit the same patterns?

No idea, sorry. If you have a mobile with a decent camera on it, the 'foldable' spectroscope that can be found on http://publiclaboratory.org/wiki/foldable-spec is actually very easy to make.

Also using any photographic camera will distort these numbers slightly. Photographic cameras do have a nuilt in UV filtering that decreases the amount of light they can detect when you start getting to wave lenghts under true blue. But this should more or less reduce the recorded intensity more than shift the frequency of anything.

As you say, the intensity will be reduced, but the peaks will still fall in the correct position relative to the CFL calibration shot (I took a CFL shot before and after taking the LED shots and made sure that the mercury vapour lines in the second CFL shot were rendered at the correct wavelengths when calibrated by the first (which was used as the cal for the LED shots) - this made sure that nothing had disturbed the spectrometer 'system' whilst I collected the images.

What I would like to know is what were the leds used and from which sellers - without that this info is pretty useless.

I'm wary of posting that on here due to sensitivities of promoting or dissing specific sellers. I should also observe that just because a supplier screwed up my order doesn't mean that he will necessarily screw up yours. It's always worried me that these modules do not have any form of identifying markings on them - honest mistakes could easily be made. If a Mod would like to explicitly tell me that I can post the info you ask for, then I will...

Peter

 

[acabgd]

I'm wary of posting that on here due to sensitivities of promoting or dissing specific sellers. I should also observe that just because a supplier screwed up my order doesn't mean that he will necessarily screw up yours. It's always worried me that these modules do not have any form of identifying markings on them - honest mistakes could easily be made. If a Mod would like to explicitly tell me that I can post the info you ask for, then I will...

Peter

Thanks, I understand the reasoning. Sorry if my initial comment sounded a bit harsh...

 

[Whitebeam]

Thanks, I understand the reasoning. Sorry if my initial comment sounded a bit harsh...

No, your comment was fine. I'm just wary after all the previous problems that this thread has seen with suppliers getting named and/or shamed. :hmm3:

Peter

 

[Mike31154]

So we're raising a flag here saying we're being cheated after using a diy hobby spectrometer?