DIY LEDs - The write-up

[Volkster]

FWHM - Full Width Half Maximum, which basically means that a specific lens has lost 50%of the light at that point, measured in cd.
When the FWHM of an optic is 30 degrees, everything outside that angle is less than 50% of the maximum at the center.
It doesn't say much about the "total" degrees of the lens but generally one can assume that the FWHM it's roughly half the total width. So, I expected an angel of ~40 degrees with bright light from an 80 degree lens, but that turned out to be less than ~20 degrees, the same angle used for a flash light which I used for reference.

Anyway; the new 80 degree optics from the US are at customs at the airport in Amsterdam,
so maybe next week all LEDs have optics again, and then we'll know.

Here's a post that illustrates this very well:
http://www.reefcentral.com/forums/showthread.php?t=1952442&page=26 - Check post 640, 641, 650 & 651.

I think I might just extend my fixture, will add 2 more rows for a total of 80 LED's just to be on the safe side. Now need to check my spread length wise :hmm1:

 

[porksmash]

Thanks for those links. Good information in there. After reading a few I came up with a plan, if someone doesn't mind doing a spot check. I have a 100 gal tank (60"L x 16"W x 24" H). I want to keep SPS and LPS corals. I would also like to make sure I am able to use this fixture over a different (probably wider) tank sometime in the future.

My plan is to use 72 LEDs total. 32 XP-E Royal blue, XP-E 16 Blue, XR-E 18 cool white, and 6 XP-G neutral white. I have a custom controller that can only provide two 0-10V signals for dimming, so I need to limit myself to 2 drivers. The blue LEDs would be arranged into 4 parallel strings of 12 and use a Mean Well HLG-185-42B (pdf) to drive them. The white LEDs would be organized in 2 parallel strings of 12 driven by a Mean Well HLG-100-42B (pdf). Each string would have a 1A fuse for protection.

I think I did my math correct to size the power supplies. Each blue string's forward voltage totals to 42V. The white strings are 41.4V due to the NW LEDs. All of the strings should be OK with the 42V drivers then. The RB/B LEDs can do 1A max, so the blue driver needs to put out 4A for 4 strings. The neutral white LEDs can do 1.5A but they'll be in series with the cool whites which max out at 1A, so the driver needs to do 2A for 2 strings.

I know color preference is very subjective, but any opinions on the LED selection? I could change out the XP-G neutral whites for XP-E warm whites to match the voltage and max current specs of the cool white, but I don't think I want too much red/yellow light. I am also thinking about finding some 420nm LEDs to replace some of the regular blues. Any other suggestions or things I missed?

 

[brassmonkyballs]

Power Consumtion

Power Consumtion

Question for Kcress and other guru's:

I was recently asked on another thread if I had measured power consumption on my fixture because they had a friend who only had 72W of LEDs but measured 125W between the drivers and the wall. My response was not knowing the specifics of what LEDs or how hard they were being driven that it was entirely possible mathematically that 24 LEDs could use over 100W but I couldn't get to 125.

The question though is how does the driver behave? In other words if we use an ELN 60-48 does it use 60w regardless of how many LEDs or how we drive them? Same with an HLG? I asked but never got a response if his friend used a killawatt or similar measuring device...I assume that was the case.

This is an interesting question just from the standpoint of calculated savings folks are going through measuring strings and calculating W consumption....may not be the best way to determine this.

Interested to hear some thoughts/facts.

 

[trggerman]

if i use meanwell HLG-240H-48 (48v 5A), does it mean that I have to connect in parrallel 5 strings of 12 3W crees. NO MORE, NO LESS. Anything less will cause the leds to be damaged.

right?

 

[brassmonkyballs]

if i use meanwell HLG-240H-48 (48v 5A), does it mean that I have to connect in parrallel 5 strings of 12 3W crees. NO MORE, NO LESS. Anything less will cause the leds to be damaged.

right?

No....wrong.

With current XPE and XPG Crees my experience is you can run 14 in a string easily on a 48V driver.

The number of parallel strings simply determines what current you can run at. If you ran 4 strings off a 5A driver that is 5/4 or 1.25A per string....that won't kill todays Cree's but the payback of running them at higher currents is more heat and not as much additional light output for that heat. However, running 6 strings is 5/6 or 0.833A per string...a more preferred range to run them in.

I encourage you to read through this thread and others to gain an understanding of the basic relationship between drivers and LEDs. Or use the search function for parallel string calculation. We can give you answers for a specific question but its only valid for that specific question.....such as this case...come back with another driver and the answers are different.

 

[porksmash]

if i use meanwell HLG-240H-48 (48v 5A), does it mean that I have to connect in parrallel 5 strings of 12 3W crees. NO MORE, NO LESS. Anything less will cause the leds to be damaged.

right?

The spec sheet for that one says it can be adjusted down to 2.5A, so 3 strings.

 

[trggerman]

well ... my rig calls for 196 leds and 16 pcs eln-60-48. i'm just looking for an alternative driver that can let me add strings on the fly. meaning, that i'll buy a bigger wattage driver initialy and just add leds strings as needed. but it seems not to be the case, the driver required is pretty much attached to the current numbers of leds. anything less that the miniumun, its going to fry the whole setup. right?

 

[jcgd]

Question for Kcress and other guru's:

I was recently asked on another thread if I had measured power consumption on my fixture because they had a friend who only had 72W of LEDs but measured 125W between the drivers and the wall. My response was not knowing the specifics of what LEDs or how hard they were being driven that it was entirely possible mathematically that 24 LEDs could use over 100W but I couldn't get to 125.

The question though is how does the driver behave? In other words if we use an ELN 60-48 does it use 60w regardless of how many LEDs or how we drive them? Same with an HLG? I asked but never got a response if his friend used a killawatt or similar measuring device...I assume that was the case.

This is an interesting question just from the standpoint of calculated savings folks are going through measuring strings and calculating W consumption....may not be the best way to determine this.

Interested to hear some thoughts/facts.

Looking at the cree product characterization tool, a 3 watt NW XPG at 1000mA is using 3.397 watts. So 24x 3.397 is 81.00 watts. A driver could be say 75% effecient. 81/.75= 108. Getting close to 125 watts. Factor in loses from the wire (a few watts) and heat you could easily use 125 watts.

A XML running at 2200mA, waaay under it's max of 3000mA, uses 7.198 watts.

They are 3 watt LEDs, but they don't max out at 3 watts. An XPG for example maxes out at 5.321 watts at 1500mA. An XML maxes at 10 watts at 3000mA. :uzi:

 

[jcgd]

The spec sheet for that one says it can be adjusted down to 2.5A, so 3 strings.

You could run two strings if you are running XPG. Max current is 1500mA so 1250mA could work. You just wont get your full 12 years outta them. But lets face it, LED technology is gonna be killer 6 years from now so XPG will be pretty obsolete. I doubt many people wont have upgraded by then.

 

[kcress]

brassmonkyballs; The only measure that really matters is the power cord measurement. There are many other variables that can mess with the results that "calculation" rarely works. Really it's what you have to pay for that matters.

Adding up things like the number of nonsensical/mythical "3W" LEDs will always lead you astray.

Since HLGs are PF corrected you should have no issues with what a KillaWatt or some other wattmeter reports for consumption.

There could be some issue though with reading a pack of ELNs for the same reason you don't want to run scads of them in one place. If the wattmeter is really poor it will have issues reading the highly distorted waveforms feeding a slew of ELNs

 

[brassmonkyballs]

Yeah I get all that and the math. I don't have a killawatt so the question is really if you plug in a driver with no load on it.....ie open circuit on the output end, what's the killawatt going to read. The rating of the driver such as 60w or just the power to run the driver. I think the answer is the rating of the driver.

 

[der_wille_zur_macht]

With an open circuit, you'll get a really, really low reading, if any perceptible reading at all. The driver's not going to dissipate power without a load, and certainly not it's nameplate power rating. Keep in mind that the nameplate power rating on the driver is essentially the power it'll deliver to an LED array, not the power that the driver itself or the complete circuit "uses."

 

[kcress]

That's correct.

The reading of an un-loaded driver would be exactly what the un-loaded driver consumes. Probably a few watts. That reading will have almost nothing to do with the driver's loaded reading. This is because the actual consumption by the driver to run the driver changes radically when the driver starts to drive a load. The power demand increases to run the LEDs and increases because the driver itself consumes more, to provide for the control of more current.

 

[patmack]

Kcress, How did you find out that the ELNs have bad harmonics? did you look at them on a scope? If I understand correctly the ELNs put a weird harmonic back on the line?

 

[Volkster]

Originally Posted by Jimmy54
FWHM - Full Width Half Maximum, which basically means that a specific lens has lost 50%of the light at that point, measured in cd.
When the FWHM of an optic is 30 degrees, everything outside that angle is less than 50% of the maximum at the center.
It doesn't say much about the "total" degrees of the lens but generally one can assume that the FWHM it's roughly half the total width. So, I expected an angel of ~40 degrees with bright light from an 80 degree lens, but that turned out to be less than ~20 degrees, the same angle used for a flash light which I used for reference.

Anyway; the new 80 degree optics from the US are at customs at the airport in Amsterdam,
so maybe next week all LEDs have optics again, and then we'll know.
Here's a post that illustrates this very well:
http://www.reefcentral.com/forums/showthread.php?t=1952442&page=26 - Check post 640, 641, 650 & 651.

I think I might just extend my fixture, will add 2 more rows for a total of 80 LED's just to be on the safe side. Now need to check my spread length wise :hmm1:

Thanks Jimmy, very important information I needed to know there.


So to help me visualise this I've added this to my sketches and now you can see where the "hot spots" are. Doing this I can see now that I definately need more LED's on the end if I want "quality" coverage.

FYI: LED's are placed ~2.5" apart

See below:

I'm now concidering using 80º optics and lowering my fixture to 12" (max) to achieve a wider spread; as I only have 4 rows. This should also help with my width coverage.

An interesting observation is with 60º optics and @ midway coverage, only 4 LED's (at most) appear to be mixing whereas with 80º optics I can see 6/7 LED's mixing. So the theory is that lateral coverage (PAR) could be on par with or higher than what the 60º optics produce. (this can be seen in these PAR readings: http://www.reefcentral.com/forums/showthread.php?t=1761942&page=14 Post 638) mid & lower coverage is higher that that of the 60º PAR lateral readings)

As for my width coverage:

Realy tight here and probably not going to get "quality" coverage...

Much better suited (if I choose to stay with 4 rows - which I probably will do for now)

So I've now decided to extend my fixture to 80 LED's, adding 2 extra LED's to each row and use 80º optics. What you think?

What I've learned in these recent post is that there is alot more to optics that meets the eye Optics introduce more than just light intensity: spacing, placement, type & especially spread mixing needs to be given some indepth thought... :reading: there is light coverage and then there is quality light coverage :thumbsup:

Thanks guys, I think I'm just about ready to place my order :bounce1: - Now just need to know what is the US$ going to do... :eek1:

 

[kcress]

Kcress, How did you find out that the ELNs have bad harmonics? did you look at them on a scope? If I understand correctly the ELNs put a weird harmonic back on the line?

They have a rectifier front-end and no power factor correction. Q.E.D. they have a weird current input waveform. They suck current in gulps. It's like asking if a road-killed skunk smells. Or if water flows downhill

That's why they have the HLG series. A lot of places in the world take a dim view of bulk lighting with only rectifier front-ends.

If you need only a few though don't worry about it. If you get much over five you should be thinking about a larger driver anyway (HLG etc.)

 

[porksmash]

That's good to know. Reinforces my decision to use the HLG series. Originally I was only going with them because they accepted all 3 types of dimming input in one model.

 

[Amoore311]

I'm eye'ing a fixture made locally with Cree's, however it does not have a controllable driver.

With very basic soldering skills can one get full dimming control by simply swapping to a controllable driver?

 

[TheFishMan65]

Yes, assuming it is a compatible driver. What is in the existing one?

 

[JoshuaG]

Neat little LED on the market, Cutter stocks them.

http://www.illumitex.com/horticulture-lighting.php