DIY LEDs - The write-up

[MosMike]

chris023
I have encountered the same problem as you. I wanted a long string of LEDs. After browsing the web I understood that Mean Well drivers are actually poorly designed.
I found a manufacturer of better LED drivers, here is the datasheet for their 100W model: http://www.inventronics-co.com/uploads/DS-EUC-100SxxxST 20100531 E.pdf
For your setup you will need 2 such drivers, 1 long string per driver.
I have contacted the manufacturer and I am waiting for their response.

 

[TheFishMan65]

MosMike, looks interesting any idea on the price? The only issue is that is does not support dimming so there is no easy way to adjust color. Of course I set all mine to 500 ma turned it on and liked it, but someone else might not

 

[MosMike]

MosMike, looks interesting any idea on the price?

I think they are more expensive than MeanWells.
Also they have models with dimming, but in other wattages. I plan to order samples and have them dissected. :smokin:

 

[TheFishMan65]

I came across an old (2006), but IMHO interesting article while trying to figure out how many watts of LEDs equal how many watts of MH. The article deals with the Solaris - does anyone know if they ran a a true 3 watts?
Product Review: A New Horizon in Lighting: PFO's Solaris LED System

 

[der_wille_zur_macht]

I don't think we can draw any conclusions based on the Solaris units. From what I've seen, they didn't even use the same MODEL of LED from unit to unit, much less the same bin.

PS I think your link is broken.

 

[TheFishMan65]

Try this Product Review: A New Horizon in Lighting: PFO's Solaris LED System

I also found this article (still reading) that talks about CREE. Anyone know what generation CREE the Tropic Marine Center AquaRay or AquaBeam use?

It is still easy to make assumptions from the raw data based on this study with plants that a 12 Watt LED can at least replace a 100 watt MH of equal Kelvin ratings in aquarium applications (the TMC AquaBeam 1000; 30 Watt LED should easily replace one 175 Watt Metal Halide of similar rating (14,000K for example) for marine applications.

 

[TheFishMan65]

This is cool also (if true)

Use of LED to prevent Red Slime

Another positive attribute of LED Aquarium lights as per a recent study (August of 2009) is that LED used in marine aquariums that suffer with Marine Red Slime Algae (Cyanobacteria) can immediately eradicate Red Slime algae when used in a full spectrum lighting configuration. These “immediate” results were just two weeks all the while other lighting configurations (such as switching to 6400 K CFL from 10,000K CFL) show slight improvement, the LED lights showed much more dramatic results. I do not know (as of this update) the exact reasons, but my suspicions is that CFL (including VHO & SHO) still produce more of the yellow/green nanometer range of light which encourages Cyanobacteria and discourages more discourage competing green algae.

 

[mseepman]

So, in taking my first "bite" of this elephant and given the math that I've seen all through this thread (thanks so much Kcress), would the following setup work...
96 XP-E RB hooked up in 8 parallel strings of 12 running on an HLG-320H-48B at 800mA
72 XP-G Whites hooked up in 6 strings of 12 running on another HLG-320H-48B at 1000mA
Please assume that I will get the fuses, resistors terminal blocks and cooling done correctly (as you can see, I believe in miracles, lol)
In looking at the data-sheet for this driver (http://www.cdiweb.com/datasheets/meanwell/HLG-320H-spec.pdf) can I assume a controller with the right 1-10V module will be able to dim these?

 

[chris023]

:furious::furious::furious::furious:

I spent 20 minutes on this answer and it vaporized.

So will give the budget repeat.

LED drivers are categorized by wattage. You need to come up with a ball park wattage to know where in the families to look.

Watts = 3.3V x 0.8A x the total number of LEDs.

Once this is known you look at the 60, 100, 120, 150, 200, 250, 320 watt models.

The parallel isn't real hard.

Here's the basic setup:

The resistors are all the same value so you don't have to think about 'different values'.

You'll use them to shift around a few LEDs to get the strings matched current wise.

Hey, let me just say "Thanks". I really appreciate your time and effort in helping me out. I know you guys (and gals) take a lot of your time up answering many many questions, and in some cases over and over again. So just wanted to let you know that it is very appreciated.

Okay, so wattage is clear, parallel setup is clear.

Driver voltage, and calculating which resistors and fuses to use is still fuzzy.

Driver Voltage:
Based on the setup I described earlier, wouldn't each string need ~60 volts in this case (18 LEDs per string x 3.3 = 59.4? Is that correct?).
I think I read that the driver would equally distribute the required voltage to each string, so if all strings are equal, I only need to calculate required voltage for 1 string and ensure the driver can handle that single string voltage (~60 in this case), yet all strings in parallel will receive required voltage?

 

[chris023]

chris023
I have encountered the same problem as you. I wanted a long string of LEDs. After browsing the web I understood that Mean Well drivers are actually poorly designed.
I found a manufacturer of better LED drivers, here is the datasheet for their 100W model: http://www.inventronics-co.com/uploads/DS-EUC-100SxxxST 20100531 E.pdf
For your setup you will need 2 such drivers, 1 long string per driver.
I have contacted the manufacturer and I am waiting for their response.

I think i'm looking at ~250 watts for 108 LEDs at 700ma. So probably 3 of those drivers, right?

 

[kcress]

So, in taking my first "bite" of this elephant and given the math that I've seen all through this thread (thanks so much Kcress), would the following setup work...
96 XP-E RB hooked up in 8 parallel strings of 12 running on an HLG-320H-48B at 800mA
72 XP-G Whites hooked up in 6 strings of 12 running on another HLG-320H-48B at 1000mA
Please assume that I will get the fuses, resistors terminal blocks and cooling done correctly (as you can see, I believe in miracles, lol)
In looking at the data-sheet for this driver (http://www.cdiweb.com/datasheets/meanwell/HLG-320H-spec.pdf) can I assume a controller with the right 1-10V module will be able to dim these?

Precisely so! A+

 

[kcress]

Hey, let me just say "Thanks". I really appreciate your time and effort in helping me out. I know you guys (and gals) take a lot of your time up answering many many questions, and in some cases over and over again. So just wanted to let you know that it is very appreciated.

Indeed.. We do. Thanks for all of us.

Okay, so wattage is clear, parallel setup is clear.

Driver voltage, and calculating which resistors and fuses to use is still fuzzy.

Driver Voltage:
Based on the setup I described earlier, wouldn't each string need ~60 volts in this case (18 LEDs per string x 3.3 = 59.4? Is that correct?).
I think I read that the driver would equally distribute the required voltage to each string, so if all strings are equal, I only need to calculate required voltage for 1 string and ensure the driver can handle that single string voltage (~60 in this case), yet all strings in parallel will receive required voltage?

Umm.. Sort of.
For resistors just use 1 ohm, 2 or 3watt, 1%. We can get you that part number. It's all thru this thread up above.

In parallel all strings will have the same voltage across them so the current thru them will be different based on each string's total Vf. Unfortunately the Vfs of individual LEDs are ALL different. So when you make up strings they are each and every one a crapshoot. This means ultimately some strings will have a high total Vf and some will have a relatively low Vf. This means the ones with a high Vf will pass relatively less current and the low Vfs will pass relatively more current. So, you WILL have to measure all the string currents - (the reason for the resistors) - and probably move a few LEDs around your strings to get all the strings passing about the same current.

Because you will have to move a few LEDs around you don't want to glue your LEDs down OR you need to be willing to leave them glued where they are and run longer wires around to electrically re-locate them into different strings. This re-wire solution might preclude using glued LEDs on something like U-Channel because you'd likely have a few wires crossing all the U-Channel gaps - if you cared.

 

[kcress]

I think i'm looking at ~250 watts for 108 LEDs at 700ma. So probably 3 of those drivers, right?

I caution you NOT to use the drivers MosMike is recommending. They send lethal voltages across your strings.

Our LED rigs have a huge number of bare electrical exposures. Your's will have more than a thousand bare-wire sites. Each and every one of them waiting for an errant finger, hand, or forearm to kickyourass.

Every large build seems to have at least one lead shorted to the heatsinks. Do you want some huge area hooked to a dangerous voltage?

Use drivers that limit their voltages to below 50V.

 

[kcress]

chris; If you want to use 108 LEDs, I'm guessing you want about 36 XP-G whites.

That's about 3 strings of 12. This means a driver that provides a little over 40V. Call it 42 or 48V.

Run them at about 800mA times 3(strings).
That's 800mA x 3 = 2.4Amps.

48V x 2.4A = 115W so we need a driver about this power rating.



And about 72 XP-E royal blues?

Which is this same setup - twice more.

What driver? Lets look...... co.... late... ing.

Check out the HLG-120-48B "B" for dimming.
Here's the data sheet: http://www.meanwell.com/search/hlg-120/default.htm

 

[TBR701]

MSEEPMAN, have you found anyone that has the HLG-320H-48B available? I was looking for the same driver and could not find it available on line. I was not sure if it was out yet.

 

[chris023]

I caution you NOT to use the drivers MosMike is recommending. They send lethal voltages across your strings.

Our LED rigs have a huge number of bare electrical exposures. Your's will have more than a thousand bare-wire sites. Each and every one of them waiting for an errant finger, hand, or forearm to kickyourass.

Every large build seems to have at least one lead shorted to the heatsinks. Do you want some huge area hooked to a dangerous voltage?

Use drivers that limit their voltages to below 50V.

Roger that...

 

[mseepman]

MSEEPMAN, have you found anyone that has the HLG-320H-48B available? I was looking for the same driver and could not find it available on line. I was not sure if it was out yet.

I just spoke to Component Distributors and they are going to get back to me Monday to let me know availability and price. They have not stocked it before and so the factory has to get them the info. Best case scenario, someone would be waiting 2-3 weeks for them to arrive.

I'll post what I hear...though remember that I was working with the Canadian side of CDI.

 

[mseepman]

If the HLG-320H-48B is not available can someone suggest a different string setup for me that would work?

 

[chris023]

Indeed.. We do. Thanks for all of us.

Okay, so wattage is clear, parallel setup is clear.



Umm.. Sort of.
For resistors just use 1 ohm, 2 or 3watt, 1%. We can get you that part number. It's all thru this thread up above.

In parallel all strings will have the same voltage across them so the current thru them will be different based on each string's total Vf. Unfortunately the Vfs of individual LEDs are ALL different. So when you make up strings they are each and every one a crapshoot. This means ultimately some strings will have a high total Vf and some will have a relatively low Vf. This means the ones with a high Vf will pass relatively less current and the low Vfs will pass relatively more current. So, you WILL have to measure all the string currents - (the reason for the resistors) - and probably move a few LEDs around your strings to get all the strings passing about the same current.

Because you will have to move a few LEDs around you don't want to glue your LEDs down OR you need to be willing to leave them glued where they are and run longer wires around to electrically re-locate them into different strings. This re-wire solution might preclude using glued LEDs on something like U-Channel because you'd likely have a few wires crossing all the U-Channel gaps - if you cared.

Understood. Will read up more on measuring individual LED current to make the strings as close as possible. Wondering if it is better to try to measure each LED before stringing them together and capture individual numbers, then make strings based on results. Or make strings and swap out here and there... Thanks...

chris; If you want to use 108 LEDs, I'm guessing you want about 36 XP-G whites.

That's about 3 strings of 12. This means a driver that provides a little over 40V. Call it 42 or 48V.

Run them at about 800mA times 3(strings).
That's 800mA x 3 = 2.4Amps.

48V x 2.4A = 115W so we need a driver about this power rating.



And about 72 XP-E royal blues?

Which is this same setup - twice more.

What driver? Lets look...... co.... late... ing.

Check out the HLG-120-48B "B" for dimming.
Here's the data sheet: http://www.meanwell.com/search/hlg-120/default.htm

It's all making much more sense now... You are correct (at this time) on the CW & RB numbers. Thank you so much. Few final questions (for now, more will probably come up later):

1) Is 12 LEDs per string pretty much the standard? Was looking to put 18 on each u-channel 3 inches apart. May squeeze 24 per channel making 2 strings per channel, instead of trying to have a string cross channels (bumping up to a driver with appropriate wattage).

2) When adjusting the current in the driver, do I adjust to 800ma and each string will get 800ma, or do I adjust the driver to 2.4amps and each string will get 800ma? I am thinking it is the ladder, just want to make sure.

3) Any advice on fuses? Maybe something that is rated at 1amp? Are they rated on amps only, or is it volts and amps?

Once again...Thanks!

 

[mseepman]

I just spoke to Component Distributors and they are going to get back to me Monday to let me know availability and price. They have not stocked it before and so the factory has to get them the info. Best case scenario, someone would be waiting 2-3 weeks for them to arrive.

I'll post what I hear...though remember that I was working with the Canadian side of CDI.

Well, got in touch with Meanwell directly and here is what they told me about the HLG-320H

"Hello Mark,

This series is still under developing, will take another 3--6 months to release.

So sorry for the delay.

Kind regards,
Natalie Chen
Mean Well USA, Inc."

I guess I need to come up with another option for how I plan to do this.