Powering LEDs without drivers/buckpucks

[mtcs]

I found this article in the net that would help understand the need of very constant voltage in LED powering (it does not mean that it is reliable but as far as I know it seems ok):
http://tinkerlog.com/2009/04/05/driv...ut-a-resistor/

" ... There are two ways to bypass the resistor. One way is to lower the input voltage. If you are able to run your complete circuit with the same voltage as forward voltage of the LED, perfect. No resistor needed. ... "

So, what those LED drivers do is simply use the second way to power a LED without using the resistor. They regulate voltage on the fly using a PWM or some other methods keeping the current stable. That is ok and is efficient, but if you already have to input a DC power source, you would be doing voltage regulating twice... So if you have a power source that can give you some voltage flexibility and you know exactly the voltage you would need for your LEDs you do not need a driver, bulkpuck, etc..

Using only your DC power supply you rely on the power supply voltage control effectiveness, if you power supply doe not do the job, your leds could fry... But considering that a switching power supply should have overload/over voltage/short circuit protections and the power supply is well sized for your LEDs, you should have no problems at long therm.

Honestly, I think those led drivers are simply a way for securing that someone with no electronic knowledge would not mess the system up with wrong voltage and burn the LEDS. Because, the same micro-controlled voltage correction/trimming that there is in the LED driver should be present in the Switching Power Supply. An great example is the PC power supply, that are the same type of power supply as this one on ebay and need much more accurate voltage control than any leds. This power source here has a protection on 135% rated output power overload protection. So, if you had 300W of LEDs running on it it would protect from a 470W overload. Depending on the overdrive you are doing on the LED and the heat dissipation, even 200% of overload is not enough to burn a hight power led.

http://cgi.ebay.com/13-5V-DC-25-8A-3...item563cd23464

Imagine how much LEDs you can power with one of this 12v 30A power supply with only U$ 39,99(free shipping)! Maaan, LED drivers are expensive!

Those led driver is something we call "Colombo's egg" or the "golden egg goose" for electronic suppliers.

But in order to use this kind of power supply, you need to know EXACTLY the LED voltage for the current you LED handles. Knowing that you can configure a setup that your LEDs voltage fits a certain voltage interval. For example, I will use several modules with 4 XP-G leds in series at 1,2 A. Thanks to the Internet and some datasheets I know exactly the voltage that my LEDs would need at that current(http://www.candlepowerforums.com/vb/...&postcount=354), that is

4 x 3,3 V = 13,2 V

Since the Power supply I bought on e-bay can be adjusted from 10v - 14v ( I measured it) I can even change my system max power manually... It could give me more efficiency or more light as I please.

I would also recommend to buy one power supply to the white LEDs and another to the blue ones so you can match voltage separately because the blue LEDs have different specs than the white ones (XPG to XPE).

After scaling the power supply voltage, you wold need also to know how much current your setup would draw to buy the wright power source. For instance, I would use 12-16 white LEDs, so 3-4 12v modules, so a 6 A power supply would be good.

Ps.1: You also find 24v power sources that fits those PCBs with 7 LEDs in series. With this boards you can put much more LEDs in much less space. Like 42 Leds in less than 10 cm^2. Of course you would need a huge heatsink or watercooler(that is a great idea for cooling aquarium leds)...

Ps.2: And a last thing, ALWAYS use a multimeter to measure the voltage or the current to avoid unnecessary smoke

I hope I have helped more than I confused
Please forgive my Brazilian English.

[der_wille_zur_macht]

Originally Posted by mtcs

I found this tutorial in the net ( it does not mean that it is reliable but as far as I know it seems ok ):
http://tinkerlog.com/2009/04/05/driv...ut-a-resistor/

" ... There are two ways to bypass the resistor. One way is to lower the input voltage. If you are able to run your complete circuit with the same voltage as forward voltage of the LED, perfect. No resistor needed. ... "

Please note that the tutorial you are linking to is aimed at plain old low-current LEDs. While the principles are the same as HP LEDs, everything (including the margin of disastrous error) is magnified many times.

A .1v deviation for a plain 20mA LED isn't a big deal. For an HP LED, it can mean the difference between living a normal life and failing prematurely - or even instantly.

Certainly drivers are not "required" but for 99% of people, they are a good idea. What IS required is some way of controlling the current through the LED. and using a typical off the shelf voltage supply to control the current through the LED is like driving in an F1 car without a seatbelt. There's NO WAY I'd ever trust a few hundred bucks worth of HP LEDs to a common off the shelf power supply with no current limit (i.e. a purpose-built driver or power resistor).

IMHO the expense of a driver is a small price to pay for the benefits of HP LEDs. If you're really cash strapped, then either DIY the drivers, or use power resistors. User kcress had a really good thread about power resistor use a few months ago, might be worth searching for.

Typical off the shelf drivers (buckpucks, meanwells, etc.) will cost around $2.50 - $3.50 per driven LED when all costs are accounted for. A DIY driver or resistor approach might be as cheap as $1 per LED, so the cost savings potential is out there.

I applaud your cost-saving mindset but don't agree 100% with your approach. At any rate, it's always good to see people thinking outside the box as long as it's tempered with real world results.

[mtcs]

Originally Posted by der_wille_zur_macht

Please note that the tutorial you are linking to is aimed at plain old low-current LEDs. While the principles are the same as HP LEDs, everything (including the margin of disastrous error) is magnified many times.

A .1v deviation for a plain 20mA LED isn't a big deal. For an HP LED, it can mean the difference between living a normal life and failing prematurely - or even instantly.

High power LEDs work the exactly same way as the conventional ones. Exponential resistance. The error, if there was one somewhere, would be magnified tha same way as would in conventional LEDs, exponential. Therefor I recommend careful to design and implement the system. Once implemented, one would not have any problems.

Originally Posted by der_wille_zur_macht

Certainly drivers are not "required" but for 99% of people, they are a good idea. What IS required is some way of controlling the current through the LED. and using a typical off the shelf voltage supply to control the current through the LED is like driving in an F1 car without a seatbelt. There's NO WAY I'd ever trust a few hundred bucks worth of HP LEDs to a common off the shelf power supply with no current limit (i.e. a purpose-built driver or power resistor).

The analogy was is not very good here. What I was saying before, the same current limiting protection in the LED drive is also present in a good switching power supply. Thats why I said that you HAVE to buy a power supply matching your system.

If you buy a 12v 30A power supply and at half the implementing course you decide to use only one module consuming 1 A, the power supply overload protection wont work if there was somehow a extraordinary voltage variation. In that case the power source will fry your leds with 13A before it "knows" something gone wrong. Thats also why no one should upgrade the main power switch of a house from 60 to 240 A just because 240A is much more cooler than only 60 A...

The great difference between the constant voltage power supply and the constant current ones is the effort to maintaining one of the metrics stable . And they use the same things to do so. Both have a micro-controller watching the circuit so it can correct the voltage/current in real time if some input outage happens. And this correction happens in milliseconds (or it would be microseconds?). Good PC power supplys tha correction happens in nanoseconds. If you look closely to the datasheet you see that the voltage variation tolerance is about 1% of 12v (or 1% of 24V), that can't burn any LED.

But, nothing can protect any electronic circuit from everything, I sense that this protection that the LED drivers crave so much for are not so different from the protections from the other power sources so widely use in the market.

Originally Posted by der_wille_zur_macht

IMHO the expense of a driver is a small price to pay for the benefits of HP LEDs. If you're really cash strapped, then either DIY the drivers, or use power resistors. User kcress had a really good thread about power resistor use a few months ago, might be worth searching for.

I wont use resistors in my setup. No need for them I can get the exact voltage I need to drive my leds. Resistors only dissipate unnecessary heat and make the system more inefficient.

Originally Posted by der_wille_zur_macht

Typical off the shelf drivers (buckpucks, meanwells, etc.) will cost around $2.50 - $3.50 per driven LED when all costs are accounted for. A DIY driver or resistor approach might be as cheap as $1 per LED, so the cost savings potential is out there.

You cant do the math like this. Only $2.5 -$3.5 per $7 LED is the same as 50% more cost! $1 dollar per $7 LED is 14%. So the difference is not so little.

Originally Posted by der_wille_zur_macht

I applaud your cost-saving mindset but don't agree 100% with your approach. At any rate, it's always good to see people thinking outside the box as long as it's tempered with real world results.

Look, if I had the $ to spend that way I would probably use drivers. They a re easy to use in a system and have a great flexibility. But considering that, I think I would buy a off-the shelf LED system... with controllers etc...

And thats me... crazy thoughts

[mtcs]

Actually, I noted that meanwell drivers are not sooo expensive... maybe this discussion will soon be outdated. I think due to the mass production of LED driver they will eventually be even cheaper and handle more power.

Maybe its cheap already and I'm looking in the wrong places... hehe

[redfishsc]

The Meanwell ELN 60-48D is $33 from nanotuners.com now. That's $7 cheaper than I paid 8 months ago.

That will drive and dim 13 XRE or XPG. You still have a few nuts and bolts to buy (dimmer voltage input, $6, 10K pot, $2).

[westmoorenerd]

You can actually run up to 26 Crees as long as you run 2 parallel strings with an ELN-60-48D. So 33/26= $1.26 per LED to be driven.

[uncleof6]

Yeah, but you need a couple transistors, and resistors, added to the circuit, to avoid an LED melt down, if one goes out, or the forward voltage varies .1 volt between leds......... normal production tolerances can cause that to happen.

Incidently, how is the Agana Mickey D's doing these days?

Jim

[kcress]

You cannot run HBLEDs with only a voltage controlled power supply. C o m p l e t e folly.

That is the same as controlling a car's speed by keeping the fuel rate fixed. Up a grade, down a grade, into a head wind, with a tail wind; they will result in radical car speed differences - some possibly fatal...

[TheFishMan65]

uncle, I think if you set the internal voltage control also then it is safe. SInce with the voltage set the current would not be able to go up and fry the LEDs. I don't have one to try, but the logic seem good.

kcress, Any evidence that it is folly? Not disagreeing, but just wondering.

[der_wille_zur_macht]

Originally Posted by mtcs

High power LEDs work the exactly same way as the conventional ones. Exponential resistance.

Arguably, yes. The difference is - if the guy in that article has a problem, he's fried ten cents of LEDs. If I have a problem with my build (and, IMHO, you WILL have problems) I've fried $500 worth of LEDs.

The analogy was is not very good here. What I was saying before, the same current limiting protection in the LED drive is also present in a good switching power supply. Thats why I said that you HAVE to buy a power supply matching your system.

It is NOT the same current limiting protection. A constant current LED driver operates entirely off the current monitoring portion of the circuit. It's designed to work on the current feedback loop. An off the shelf power supply's current limit is a last-ditch failsafe. It's not accurate, it's not quick enough, and it's not designed to be part of the normal operation of the device.

If you look closely to the datasheet you see that the voltage variation tolerance is about 1% of 12v (or 1% of 24V), that can't burn any LED.

Heck yes it can. Or, at least, it'll take an LED designed to last 20 years and ruin it in 5 years.

But, nothing can protect any electronic circuit from everything

Agreed - BUT, on the other hand, electronics are designed with specific operating principles in mind. Off the shelf DC power supplies are designed to supply a roughly constant DC power source. LED drivers are designed to drive HP LEDs. The cost added to a build by using the correct parts designed for the correct function isn't really that big - maybe 10 - 20% of the light rig build, which for many people probably translates into 3-5% of the tank's set up cost.

I sense that this protection that the LED drivers crave so much for are not so different from the protections from the other power sources so widely use in the market.

Sorry, but you sense wrong. Again, an LED driver is designed to operate based on a constant current feedback loop. The concept of "over current protection" doesn't really even apply - it results naturally from the driver's constant current operation. An off the shelf DC power supply is NOT designed to provide a constant current and has NO MECHANISM to ensure that current is kept constant. You keep latching on to the idea that you can determine a magic voltage number and just run the array at that number and the current will work itself out. In theory, on paper, in an ideal world, that's essentially true. BUT, in the real world, it's extremely risky at best. As kcress pointed out in his automotive throttle analogy, there will be variances. Things will heat up, cool down, tolerances will change. Your power supply will drift, and the LED's I/R curve will drift. These drifts WILL be enough to throw your math WAY off. So, the only real option if you want to use your method reliably in the long term is to build some automatic current monitoring and control scheme. But, then you've effectively designed an HP LED driver, so you might as well just admit it from the beginning.

Look, if I had the $ to spend that way I would probably use drivers. They a re easy to use in a system and have a great flexibility. But considering that, I think I would buy a off-the shelf LED system... with controllers etc...

And thats me... crazy thoughts

You're misinterpreting my math. You CAN DIY this stuff for dirt cheap - ALMOST as cheap as your power supply setup, but WITH the correct constant current functionality, and with controllability built in. Have a gander at this thread:

http://www.reefcentral.com/forums/sh....php?t=1759758

Near the end, we discuss a design based on the CAT4101 LED driver IC. This chip costs lerss than $2, and needs about $2-$3 of other components. It can drive 6 or 7 HP LEDs on a 24v DC supply. By the time you factor in DC power supply cost, an enclosure, etc. - the cost per driven LED is around $1. That's chump change. And you get a purpose-built constant current LED driver with good performance, good stability, good efficiency, and controllability via PWM or analog signal if you wish to dim your LEDs. It's a fraction of the cost of an off the shelf LED controller, and in a vastly different price range than a complete off the shelf LED unit built for reef tanks.

You seem bent on the idea that "LED drivers only exist so manufacturers can rip us off" which is not true. LED drivers exist to provide specific functionality you don't get in an off the shelf DC power supply - namely, constant current control. Also, typically, some form of controllability.

And again, I apologize if I'm coming off as argumentative, but I've put a few hundred hours of effort into designing and building LED drivers and LED arrays over the last year or so, specifically with goals that seem similar to what you're looking for: cheap cost, high reliability, long life, good efficiency, and controllability. I've pondered many of the things you're pondering, tried a few of them, and am trying to explain the conclusions I arrived at. You're free to take my advice or leave it.

[mtcs]

Originally Posted by der_wille_zur_macht

Arguably, yes. The difference is - if the guy in that article has a problem, he's fried ten cents of LEDs. If I have a problem with my build (and, IMHO, you WILL have problems) I've fried $500 worth of LEDs.



It is NOT the same current limiting protection. A constant current LED driver operates entirely off the current monitoring portion of the circuit. It's designed to work on the current feedback loop. An off the shelf power supply's current limit is a last-ditch failsafe. It's not accurate, it's not quick enough, and it's not designed to be part of the normal operation of the device.



Heck yes it can. Or, at least, it'll take an LED designed to last 20 years and ruin it in 5 years.



Agreed - BUT, on the other hand, electronics are designed with specific operating principles in mind. Off the shelf DC power supplies are designed to supply a roughly constant DC power source. LED drivers are designed to drive HP LEDs. The cost added to a build by using the correct parts designed for the correct function isn't really that big - maybe 10 - 20% of the light rig build, which for many people probably translates into 3-5% of the tank's set up cost.



Sorry, but you sense wrong. Again, an LED driver is designed to operate based on a constant current feedback loop. The concept of "over current protection" doesn't really even apply - it results naturally from the driver's constant current operation. An off the shelf DC power supply is NOT designed to provide a constant current and has NO MECHANISM to ensure that current is kept constant. You keep latching on to the idea that you can determine a magic voltage number and just run the array at that number and the current will work itself out. In theory, on paper, in an ideal world, that's essentially true. BUT, in the real world, it's extremely risky at best. As kcress pointed out in his automotive throttle analogy, there will be variances. Things will heat up, cool down, tolerances will change. Your power supply will drift, and the LED's I/R curve will drift. These drifts WILL be enough to throw your math WAY off. So, the only real option if you want to use your method reliably in the long term is to build some automatic current monitoring and control scheme. But, then you've effectively designed an HP LED driver, so you might as well just admit it from the beginning.



You're misinterpreting my math. You CAN DIY this stuff for dirt cheap - ALMOST as cheap as your power supply setup, but WITH the correct constant current functionality, and with controllability built in. Have a gander at this thread:

http://www.reefcentral.com/forums/sh....php?t=1759758

Near the end, we discuss a design based on the CAT4101 LED driver IC. This chip costs lerss than $2, and needs about $2-$3 of other components. It can drive 6 or 7 HP LEDs on a 24v DC supply. By the time you factor in DC power supply cost, an enclosure, etc. - the cost per driven LED is around $1. That's chump change. And you get a purpose-built constant current LED driver with good performance, good stability, good efficiency, and controllability via PWM or analog signal if you wish to dim your LEDs. It's a fraction of the cost of an off the shelf LED controller, and in a vastly different price range than a complete off the shelf LED unit built for reef tanks.

You seem bent on the idea that "LED drivers only exist so manufacturers can rip us off" which is not true. LED drivers exist to provide specific functionality you don't get in an off the shelf DC power supply - namely, constant current control. Also, typically, some form of controllability.

And again, I apologize if I'm coming off as argumentative, but I've put a few hundred hours of effort into designing and building LED drivers and LED arrays over the last year or so, specifically with goals that seem similar to what you're looking for: cheap cost, high reliability, long life, good efficiency, and controllability. I've pondered many of the things you're pondering, tried a few of them, and am trying to explain the conclusions I arrived at. You're free to take my advice or leave it.

yes, you are right, both power sources work differently and definitely serves to different purposes. I just don't believe that the voltage variation from a switching power supply would impact that much when you put the LEDs in series. As I said, those power supplys are mato to mantain V contant with 1% of voltage variation. I also believe that 1 ms is not fast enough to heat the LEDs to decrease lifetime.

And yes, you is also right as I said earlier. The LED drivers are getting cheaper. The price now is not the same when I started researching. Back then I found led drivers costing $70 that could be replaced by a $20 DC power source. But they are not cheap enough thou. If you wanted like 300W of LED light you would pay a fortune in small drivers, because the 350W DC power supply costs only $ 40.

I just dont think LEDs will be default to aquarium if they dont increase the sizing of the LED solutions, I know a few guys that just dont even want to try LEDs because of the thousand stars they would have to solder and the dozens of drivers they would have to use.

I apologize for my argumentative behavior too.

[der_wille_zur_macht]

Originally Posted by mtcs

I just dont think LEDs will be default to aquarium if they dont increase the sizing of the LED solutions, I know a few guys that just dont even want to try LEDs because of the thousand stars they would have to solder and the dozens of drivers they would have to use.

I apologize for my argumentative behavior too.

I think an important thing to remember is that HP LEDs are relatively new products in the electronics world, and development is still happening very rapidly (If you look at lumens/watt for white LEDs, it's following Moore's Law pretty well!)

In two or three years, your friends might find that the LEDs and components available present a more attractive project!

[mtcs]

Originally Posted by der_wille_zur_macht

I think an important thing to remember is that HP LEDs are relatively new products in the electronics world, and development is still happening very rapidly (If you look at lumens/watt for white LEDs, it's following Moore's Law pretty well!)

In two or three years, your friends might find that the LEDs and components available present a more attractive project!

thats exactly what I am waiting for.

do you guys know of any LED driver that can handle more power for a reasonable price?

[der_wille_zur_macht]

Originally Posted by mtcs

thats exactly what I am waiting for.

do you guys know of any LED driver that can handle more power for a reasonable price?

Commercial drivers scale pretty linearly with respect to cost per unit power or cost per LED. Meanwell has some products that can drive many more LEDs per driver but you pay more. And in some cases you miss out on features (i.e. it'll be the wrong current range, or won't easily be dimmable, etc.) You can browse the LED section of their website for ideas:

http://www.meanwell.com/webnet/searc...earch.html#led

Their drivers (or any, really) can also be used with several parallel strings of series LEDs, but that introduces some complications (balancing strings, protecting them in case of failure, etc.)

[sfsuphysics]

While I use a regular dc transformer to test my HP LEDs, I do so first through a decade box (fancy name for resistor that I can change the value to a specific amount), and do so regardless of the output of transformer, even if it's within "safe" parameters.

Now hooking them up in the "ok this is going over a tank" setup, I absolutely use the current sources (Meanwells, Buckpucks, etc) as the DC transformers are voltage sources, I figure it's $20 for a non-adjustable power supply when I'm using $7-8 per LED is not that much of an expenditure considering the safety, it supposedly provides.

However I will admit that I never learned about current sources in electronics class, and it does confuse me that something that can supposedly only put out say 700mA can fry an LED if only one is hooked up to it.

[der_wille_zur_macht]

Originally Posted by sfsuphysics

However I will admit that I never learned about current sources in electronics class, and it does confuse me that something that can supposedly only put out say 700mA can fry an LED if only one is hooked up to it.

I never even took electronics class, and those who have will probably be along shortly to correct me, but generally from the perspective of current control, DC power supplies fall into three categories:

1) Cheap, more or less unregulated supplies. They'll be rated with a max current, based on what it'll take to burn them up. No actual current regulation. The current rating is an arbitrary number based on component failure. If you pull too much current, you fry them. This type of supply will often provide more than the rated current if you short it or connect an LED, but it'll soon burn up. Hence they're totally inappropriate for LEDs without some sort of current control.

2) Voltage regulated supplies. Such as the DC supplies we're typically using when driving a buckpuck, etc. These typically have active over-current protection, but are designed to be constant voltage supplies, i.e. they'll let as much current through as your circuit wants, until the point at which the over-current feature trips. This is indeed designed as a failure prevention feature (i.e. it keeps the supply from catching fire), not as a normal operating feature. If you are grossly under-driving an LED, or like to live dangerously, this sort of supply CAN power an LED, but it's unstable and far from optimal, especially if you even think about driving an HP LED at a typical drive current.

3) Constant current supplies, such as a Meanwell ELN or the output side of a buckpuck, etc. These typically have an active current feedback loop. There's usually a sense resistor between the load and ground, connected to a comparator in the device. The other side of the comparator is connected to a low, extremely well regulated voltage source. This way, the device "knows" the current in the circuit, since it's comparing the voltage drop across the resistor to the known voltage source. If the reference voltage is above the resistor's drop, the current is below the target and the device will increase current. If the reference voltage is below the resistor's drop, there's too much current, and the device will decrease current. Typically, a specific operating voltage isn't spec'd, rather a range of acceptable voltages is. As long as your load is within that range, the device will regulate current and let the voltage "fall where it may" as a result of the current regulation. Off the shelf LED drivers fall in to this category.

[kcress]

Originally Posted by mtcs

I just don't believe that the voltage variation from a switching power supply would impact that much when you put the LEDs in series. As I said, those power supplys are mato to mantain V contant with 1% of voltage variation.

Your belief here is wrong. Very wrong. You should stop repeating what is blatantly, patently wrong. It's not helping you.

You can believe that you could walk to the Moon by this evening. It doesn't mean in reality that you can.

You clearly do not understand "negative temperature coefficients" and what they mean to a bunch of expensive HPLEDs driven by a nonsensical voltage regulator.

[stugray]

cmon guys!

what mtcs has to say is perfectly valid if you use something like the Agilent (HP) 6674A 2000 Watt System Power Supply, 60V, 35A.

You could drive 35 strings of 18 LEDs with that in a constant current mode.
I use them all the time at work.

A simple Labview VI could drop the ConstCurrent value if there was some way to detect a string failure ;-)

New -> $5000

ebay -> $2000

Stu

[kcress]

Yes but...
You are cheating with the "constant current mode".

BTW; I have that same supply!

[stugray]

"I have that same supply! "

Used it to charge the batteries on Mars Global Surveyor, Stardust, Genesis & Deep Impact for launch.

There's a trick to flip it into CC in those cases... well worth the $5k but you have to be careful with those pesky remote sense lines ;-)

Stu

[TheFishMan65]

How is his plan different than using the voltage control on an ELN. I understand all the theories, but I lack experience since I do software. I have been suggesting (need to know if I am wrong) that you can do multiple strings on a meanwell if you adjust the voltage after the current limit has been set. Theory if you can't increase the voltage to the string then you can't increase the current enough to damage the LEDs. Now maybe it has been missed (there are a lot of LED threads), but no one has contradicted my. I can't imagine the meanwell is that much better at regulating voltage than a true power supply.

thanks in advance

[BeanAnimal]

Originally Posted by kcress

Yes but...
You are cheating with the "constant current mode".

BTW; I have that same supply!

I keep bidding on them... (2 years now) and will get one at the price I am willing to pay eventually. I just got a new bench supply (tripple output) with constant current, but it is not programmable. Not bad for $200

[der_wille_zur_macht]

Originally Posted by TheFishMan65

How is his plan different than using the voltage control on an ELN. I understand all the theories, but I lack experience since I do software. I have been suggesting (need to know if I am wrong) that you can do multiple strings on a meanwell if you adjust the voltage after the current limit has been set. Theory if you can't increase the voltage to the string then you can't increase the current enough to damage the LEDs. Now maybe it has been missed (there are a lot of LED threads), but no one has contradicted my. I can't imagine the meanwell is that much better at regulating voltage than a true power supply.

thanks in advance

It's the difference between a normal mode of operation, and a failsafe. Driving HP LEDs with a constant current driver is nearly 100% foolproof, and will work under a vast range of environmental and system configurations, and will automatically account for things like temperature coefficients and variations from LED to LED. Using a voltage limit as a failsafe (as you're suggesting with the ELN) is totally fine, BUT if you were to crank the current limit on the ELN way above spec, and then try to use the voltage limiter to control current in the string, you'd have a hard (if not impossible) time of it, because you'd have to take all those variations and dynamics into account when you set the voltage limit. That's essentially what the OP was suggesting.