DIY LED Project

[Offski]

Thanks BigA
Thanks for your posts.

[Offski]

Im Happy to answer any questions or how to's
(If I'm capable)

[MajorTom]

@Offski: first of all: great job, impressive build!

I would be interested in the type of led drivers you are using. I'm doing a DIY for a fresh water tank (22 cree cm-l cool and neutral white mixed) driven with a max. current around 1 amp. I'm using 2 Meanwell ELN 60-48d for driving (limited cloud simulation support ;-)). Works, but they have this ugly "discharge-flash" issue wenn switched off ("unplug" with the ghl relais EVG-AP HICs) at low dimming levels (night). This can be circumvented by not switching the driver off but dimming them to <<1 Volt, virtually switching the leds off. Do the drivers you use can drive ~10-12 XMLs @ 1Amp and not have this issue? (I've read in some forums that this is somewhat "broken by design" for the 48d's by Meanwell)

Another question is cooling, at which current do you drive the LEDs? Mine are mounted on dedicated passive coolers (6,5x5 cm, 2,5cm high), one for each LED, the LEDs are driven at around 1 Amp. (for XML's max. 3Amp. are possible, although not recommended, if I remember correctly). The coolers get quite "warm to the touch", around ~40-43°C when driven with 1 Amp. for some hours. For me it seems like you have a lot of LEDs close together on a rather small single Cooler, at least it looks like that on the pictures. Do you use active cooling (some fans) or is this no problem for you / do you drive them with lower current?

Thanks and best regards,
Thomas

[Offski]

Hi major Tom.

Before starting this build. I spoke and read a lot of info on the subject. My overall conclusion was that there is still a differing of opinion with regards to the effectiveness of LED's in a marine aquarium. With this in mind and the knowledge I have from working within the lighting industry, I went about this design based on the idea that I will make this work-and work well. I start my post with this statement as my answer to your post is based on my philosophy for this project ( and given also what I already stated above) and that is, I built this led light based on 'over designing which will hopefully give me the opportunity to de-rate'.

So here goes

Firstly the flicker issue.

I do not have any flicker when turning off or dimming up or down. Although I'm not familiar with the Switching device you mentioned. My build is based on dining via the driver and when a 0v signal (off) is sent, a relay switches off the driver primary feed. Although I built the circuit board for this control, it is pretty much the same as the EVG-AP2 (that GHL offer).

Heat and driver

I have been running the light for two weeks via my Profilux 3 unit. I have programmed it to simulate a day cycle. Although that means not all the 162 LEDs are on at the same time, I have set them to rise to 100% when on. On this basis I do not have any heat issues (in addition, I do not have any extra heat loss devices I.e fans etc. I am solely reliant on the heat sink. So my thoughts on this are as follows.
My build involved bolting together 10no heat sink each are approx 650mm x 80mm. Each heat sink is able to control the heat of 20no 3w LEDs. On this basis my build is good for 200no LEDs. (I only have 162no in this project) so on this basis, plus the fact that not all my LEDs are on at once. It is probably running at 50% the heat sink tolerance. Also, the C channel that I've used as seen in the picture is also aluminium which helps a bit a well.

The drivers I am using was chosen after discussion with numerous manufacturers. The driver I eventually choose was a prototype driver. The company only built 100no for test. I persuaded them to give me 10no prior to test. I believe they were designed for new economical street lighting. The drivers are constant current drivers. Each driver has 2x700mA supplies rated to a max of 48v. So in theory, it can run upto 2x11 Cree XML LEDs. For my build, I am running 2x9 LEDs of each driver (approx). The reason I did not max the qty is more to do with wanting the 10no channels (drivers) to give the simulation flexibility.

So, running each led at its rated mA (@700mA) and in fact having an oversized heatsink (due to not all LEDs not being on at once) is resulting in a heat gain no more that a few degrees over ambient temp.

My thoughts on your heat gain (which is only my opinion) is a two fold issue. First of all, although you have individuall heatsinks. The actual surface area of it is not great. Secondly and I think most importantly is the fact that you are running at 1000mA. As you correctly stated xml's are good for upto 3000mA however, like all LEDs the light emmitance gain against mA and heat gain are not constant.

When I was testing the XML. I concluded that the light gain (which I measured in lux at 12") was not a justifiable increase in relation to the additional heat and mA (additional mA is additional running cost). For this reason I decided to keep my LEDs running at 700mA.

I would like to add something at this point which in my opinion is a very overlooked solution to increase light output of an led and that I the use of lenses. LEDs emit light at almost 180' (that is a lot of wasted light. I spent a lot of time testing various lenses. A good quality led lens is priceless. I have used lenses specifically made for XML LEDs an on test my readings went from 320lux to over 1200lux at 12" per led. That lux increase is far greater than the effect of cranking the mA to your LEDs!

I will pm you later the drivers that I used but I will say that I think that there are many drivers on The market that will offer similar results.

I will post some photos of the internals of the drivers, controls located inside the grey box later.

[MajorTom]

Thanks for all your input, offski.

Heat: okay, so you are driving your LEDs at 700mAmp. I agree that the higher current might be the "top reason" (along with the small surface area) for my coolers getting warmer than yours (although I think that ~40°C is no big issue yet). According to the led/light gurus at the planted tank forums, around 700-800 mAmps should be enough for my tank to achieve a PAR value of around 40-50 at the substrate, so maybe I can reduce the light intensity later, if I'm cultivating algae in my tank ;-). For me it is amazing though, how much more light a marine tank requires compared to fresh water tanks.

Optics: agreed, i've outfitted mine with 40° optics each (had to buy them from the US at LED group buy, XMLs and fitting supplies where hard to find in Germany, at the time I bought mine), because the fixture will be around 20-25 inches above the tank

Drivers: see PM

Thanks again!