Minimalistic multichip DIY LED build

[Lassef]

Respectfully, your calculations simply do not account for several realities and the fact that the rest of us don't get better PAR in water than in air....

As stated, if we take into account the refractive index of seawater and air (not a seawater and a vacuum), couple that with the losses incured by the density and clarity of the water and the fact that a partially collimated light source is already present (a focused beam) then account for the small diameter of the sensor reading the PAR it is hard to fathom a gain.


Look at it this way:
A large portion of this thread has been spent talking about the color of light that reaches the coral and what is absent, but presnt in the sunshine that strikes the water. That light travels 83 million miles from the sun (subject to the Law of Inverse Squares) and then passes through about 600 miles of "air" with most of its wavelengths still intact when it strikes the water surface (even on a cloudy day), only to have MOST of those wavelengths stripped away in the first few feet of water.


This is not something your spreadsheet calculations are going to model very well


If you are seeing an increase in PAR in some areas during in water measurments (as compared to the in air measurements) they are likely due to the reflectance of the glass, and to that end the more focused the initial beam, the less collimating effect you from the water and/or reflectance from the glass. Hope that makes sense

I have done a lot of tests i different areas in my aquarium with one glass window (120*65*60 cm). The same results everywhere. We are talking about a distance of 60 cm in water and a concentration of the lighting area with around 40 % through the differences of the refractive index between air and water. I do not think you loss 40 % of the intensity in 60 cm of water. There is other people I know that have the same experiences when working with lenses and water depths around 60 cm. And I´m using lenses - not reflectors. For reflectors - I do not know the result - I have not test. I was myself surprised over the result with my calculations of difference of the beam area in water vs air, so if you find any major mistake in my calculations you are welcome to correct me.

I found 44 % lesser area - it means for me somewhere between 30 - 45 % in reality because my formulas is based on a average of the refractive index and a calm surface. It is also important to stress that I use ozone in this aquarium so I do not have the normal amount of yellow humus in my water. My water is very clear, some particle but not colored water

I know that reflection can have a huge importance on PAR readings, I have seen examples with bare glass aquariums how you can changes the PAR readings just through not paint your backside of the aquarium or to chose a colored film as a background. This is not the case with my aquarium. The whole backside is covered by artificial rocks

However - I´ll do some experience later on when my Dream Chips are up and running - and I´ll try to find a way to do the measurements without any chance of reflection.

Sincerely Lasse

 

[BeanAnimal]

Lasse,

Thank you for the response. I would be very interested in seeing your measurement results when your new setup is in place.

 

[TropTrea]

from what ive read earlier it is my understanding that supplementing a 100w chip with 10 or 20w blues is not going to work out well. also what ive read is that the color the sellers state is not necessary correct, so if you buy a 10k chip you might end up with 6.5k, so i would buy 2 20k kelvin chips and one with say 16000 kelvin or 12000k

greetings

Actualy if they following industry standards a 10,000K chip would probably anywhere in the range of 9,000K to 12,000K if the CRI rating was at 85 or higher. For it to be in the 6,500K range the CRI rating would have to be extremly low. However there is nothing that says that Chineese chips must conform to the US industry standards since they are not regulated by any body with athourity.

Now when you get up the so called 20,000K chips the range that could fall in is much much higher since there is realy no way of rating these on the CRI scale. Soi it would be surprising if you found one chip actualy around 16,000K and another another at 28,000K. If you look at a color plot for K temps you will see there is very little difference between 20,000K and infinity especialy compared to a 10,000K light source.

 

[TropTrea]

Until they're readily available to purchase individually at the group buy price, they will be niche and less adopted. Its been a waiting game for months now for this to happen, and I still cannot purchase one without going to another forum and participating in a group buy in which I am not comfortable with.

People like myself are not comfortable with paying $100+ through a group buy to basically beta test a chip... and the whole process in itself is banned from reef central. For those who do purchase them, your experiences with them are greatly desired to hear about.

I really do want to like these, but they still aren't really ready yet until they won't get you banned here to purchase them.


Sent from my SGH-T999 using Tapatalk 2

I had worked for years with various companies developing electronic products. Waiting months for a good product is not unusual as in many cases it takes years for it to go from initial development to final production as a sellable product.

But the big difference between what I'm used to and what I'm hearing in in your note is very explainable. Preproduction and Beta products are usualy given to the top customers free for evaluation. When initial evaluation reports come in the designs are readdressed and modified untill the reports from from these evaluators meet the expectations of the marketing departments. At this point an initial production run is made and the product is introduced to the market on a limited bases. Watching the reactions of the market to this new product can then push it to normal production, scrape the idea, or even bring it back to redisign for later rerelease under an upgrades Rev.

but I will agree Betta products are general very hard to come by and when they are available they are usualy free rather than offered at a premium price. When you work with a Betta run product your actualy taking a risk where you are the guine pig for the company that wants to eventualy make money on a new idea. I would say that at least 80% of the time this Betta production item will never see the market.

 

[TropTrea]

Looked at the ratings on this chip. Basicly with everything turned up full you have a 165 watt multi chip with 100 LED's in it running an average of 1.65 Watts each. As far as the wavelenght distribution some people can argue it is great other will say it is not ideal.

Looking at the equivelent in new single chips your talking roughly 16 XML chips totaling $104.00 or 32 XPG chips totaling $160.00. So the price range of the "dream multichip" is in the same rough range. Performances though I would say the individdual chips would give you a little more PAR dependent upon the area your trying to cover.

There is the shadowing question that I'm still very leary about. In say a 6' long tank you would probably go with 3 of the mutichips and have three point sources of light. While the colors will not create the disco effect since they are so close together there will be black shadows under items with the 3 light sources. On the contrary with the 48 XML you would have a minimium of 16 specific light points hitting any specific point so the shaddows under items would be considerably softened.

this does again reflect back to personal choice in your desired effect. Do you like a strong hard lined shadow effect that emmphisizes texture and detail or do you prefer a more constant equaly distributed light that is softer. Another way of looking at it is if you like a softer HO Floresent look then go with many individual LED's but if prefer a sharper MH look then the fewer multi chips are your better option.

 

[acabgd]

which 20w hybrid did you use? there's quite a few combos out there

Ebay link removed~dc

 

[acabgd]

A lot of us would love to see pics and impressions.

If you're talking about the 20W chip, here is one good example - a small SPS frag, growth within 20 days, this coral was placed the highest, nearest to the 20W led.



As for the new DreamChip, I've started a new thread with photos and impressions: http://www.reefcentral.com/forums/showthread.php?t=2227680

 

[Whitebeam]

One piece of data I seem to be missing everywhere is how much heat do these modules actually generate? I have just bought ac-rc 20W 'EPILED' modules in 10,000K/453nm hybrid, 455nm & 430nm variants to play with. To be able to design a test setup properly I need an idea of how much heat (in Watts) each module will generate. For my playing I'll drive them with ac-rc's "3 x 20W panel" manually dimmable bare-board driver.

Can anyone pass on any actual data on this please?

Peter

 

[acabgd]

One piece of data I seem to be missing everywhere is how much heat do these modules actually generate?

Don't know exactly the numbers, but I successfully cooled it with an old-style Pentium I or AMD CPU cooler.

 

[Lassef]

One piece of data I seem to be missing everywhere is how much heat do these modules actually generate? I have just bought ac-rc 20W 'EPILED' modules in 10,000K/453nm hybrid, 455nm & 430nm variants to play with. To be able to design a test setup properly I need an idea of how much heat (in Watts) each module will generate. For my playing I'll drive them with ac-rc's "3 x 20W panel" manually dimmable bare-board driver.

Can anyone pass on any actual data on this please?

Peter

Hi

The energy efficiency of MH, T5 bulbs and Power LED:s is nearly the same - around 25 - 35 %. The diference betwee MH/T5 bulbs and power LED´s is that most of the "waste" heat from MH and T5 bulbs is IR radiation (heating the water), Maybe only 20 - 30 % of the heat waste will be as convection heat and has to be "chilled" out. For the power LED.s nearly all of the "waste" heat will be as convection heat and has to be chilled out. If your power LED is rated 10 watt - probably around 7 watts has to be cooled away. When I calculate for the large chip (30 - 250 watt) i normaly calculate that I have to be rid of the same ammunt of heat as the effect of the chip - just to be sure.

10 and 20 watts chip can be cooled rather simple. I normaly use a square tube and a fan in the midle - blowing down in the tube and out through the sides. 1 ten watts each 12 cm and 1 twenty watts each 20 cm. If I have a longe tube - I use two fans. the AC-RC chip normaly handle around 55 - 60 degree Celsius at the backplate but it is not a bad idea to run them at a lower temperature.

Sincerely Lasse

 

[Whitebeam]

... i normaly calculate that I have to be rid of the same ammunt of heat as the effect of the chip - just to be sure.

In the absence of any better information I too was assuming that all of the power in would be converted to heat as the worst possible case. I'm looking at passive heatsinks with a TR of just under 2C/W for my experiments - I'll then think a little more of how to implement these in a hood for my AO Reef 200 - probably look to one of the smaller Pentium coolers like acabgd suggested.

Tack så mycket Lasse!

Peter

 

[megadeth72]

One piece of data I seem to be missing everywhere is how much heat do these modules actually generate? I have just bought ac-rc 20W 'EPILED' modules in 10,000K/453nm hybrid, 455nm & 430nm variants to play with. To be able to design a test setup properly I need an idea of how much heat (in Watts) each module will generate. For my playing I'll drive them with ac-rc's "3 x 20W panel" manually dimmable bare-board driver.

Can anyone pass on any actual data on this please?

Peter

20w of heat lol, or for 3x it's 60w of heat

 

[Lassef]

20w of heat lol, or for 3x it's 60w of heat

Probably more around 42 watt of heat but just to be sure - its good to have some marginals.

Many people here around (me included) use graphic card cooler because they are normally rated to manage higher wattage compared with the processor coolers. But I think that it is some miss thinking here (my thinking included) because a graphic card can bee much more hot than a processor so you can´t compare the cooling capacity in watts between this two types of coolers. Many of the top end processor coolers works better than the graphic cards cooler with chips between 100 - 200 watts.

Sincerely Lasse

Ps Whitebeam - det var så lite så

 

[megadeth72]

I use intel p4 or 775 heatsinks with good results, I've not tried over 100w on one but they probably can handle that

you can get them for under 10 $ brand new

break the mounting tabs off and they fit into a 4" pvc coupler, I glue in some old clear tubing to center it, instant DIY pendant

I run the fan blowing down towards the water, wrecked enough fans to not want to draw salty air up into the fan

I still need to paint it black and shrink wrap the wires

 

[zoaracer]

I have use 10 watts (of the AC-RC type but the old 38 mil version - both RB and whites at 14 000 - 16 000K) for nearly 2 years without any problems. There is differences in quality out there. I have a total amount of 50 pcs of 10 watts chip running in different aquariums. 25 of them has been running nearly 2 years - the others for 1 year. Normally I run this chip at 900 mA. temperature at the backside from 35 to 45 degree C.

The last year I have also run 5 cree XM-L at 2000 mA (5 watts) and 10 Cree XP-E RB at 900 mA. 2 of the XP-E has been broken.

Sincerely Lasse

thank you lasse,

have you tried any DIY drivers for these LED's like NCP3066 ??

regards
zoaracer

 

[cubanazo83]

If you're talking about the 20W chip, here is one good example - a small SPS frag, growth within 20 days, this coral was placed the highest, nearest to the 20W led.



As for the new DreamChip, I've started a new thread with photos and impressions: http://www.reefcentral.com/forums/showthread.php?t=2227680

Hello i have a 12g nano cube similar to your tank. I want to put a LED 20w like yours. Can you tell me what driver did you use? did you only used a 20 w LED or you used other LED as well? Thanks

 

[Skinnysloth]

I use intel p4 or 775 heatsinks with good results, I've not tried over 100w on one but they probably can handle that

you can get them for under 10 $ brand new

break the mounting tabs off and they fit into a 4" pvc coupler, I glue in some old clear tubing to center it, instant DIY pendant

I run the fan blowing down towards the water, wrecked enough fans to not want to draw salty air up into the fan

I still need to paint it black and shrink wrap the wires

That's a pretty sweet looking and low cost setup you got there. Gives me another option to think about.

 

[acabgd]

Hello i have a 12g nano cube similar to your tank. I want to put a LED 20w like yours. Can you tell me what driver did you use? did you only used a 20 w LED or you used other LED as well? Thanks

I used a laptop power supply with a home-made driver to get the exact power needed, no dimmer. However, you can get many ready-made drivers for the 20W on Fleabay for a reasonable price. The chip draws almost 2.1A at 12V.

On my 7g I was using only the 20W hybrid LED and it's the best light I've had up to now, if looking at coral growth. On your 12g you might want to go for a slightly larger LED or add another 10W Blue/RoyalBlue chip, but I would say only if you're having SPS corals.

 

[Lassef]

thank you lasse,

have you tried any DIY drivers for these LED's like NCP3066 ??

regards
zoaracer

Yes - a guy here in Sweden build driver boards based on CAT 4101 drivers. Somethings like the board you can get at Steve´s LED. The driver can give 1000 mA but if you connect two drivers (CAT 4101) in parallel you can get 2000 mA (at least with that construction my friend has) I always connect two pcs of 10 or 20 watts (of AC-RC type) in a daisy chain and adjust my ingoing voltage as close to the combined FV as possible (total FV + 0.5 V)

Sincerely Lasse

 

[Lassef]

I used a laptop power supply with a home-made driver to get the exact power needed, no dimmer. However, you can get many ready-made drivers for the 20W on Fleabay for a reasonable price. The chip draws almost 2.1A at 12V.

On my 7g I was using only the 20W hybrid LED and it's the best light I've had up to now, if looking at coral growth. On your 12g you might want to go for a slightly larger LED or add another 10W Blue/RoyalBlue chip, but I would say only if you're having SPS corals.

You are using a constant voltage driver. Normally - it is not recommended to use this because of that the relationship between the voltage and the amperage is not linear with a LED. A very small change of the voltage cause a rather high change of the current. 2100 mA to this chip is also a little to much. It is recommended to use a constant current source instead - when if you choose 2000 mA - the driver will never exceed that number - it will adjust the voltage instead.

Sincerely Lasse