LED buckpuck vs drivers

[AngelAddict]

Fish, I saw your review, very good stuff there, but I have lots more questions/concerns then what that addresses. Don't get me wrong, that's a great idea and your posts there should be stickied.

Henery, makes sense. Take some risk to get more bang for your buck, wire parallel. Play it safe, spend a little more money, wire series.

If one LED goes out in parallel, wont the whole string go out? I've heard bad things about parallel, so I think I'll stick with series.

 

[HeneryH]

If one LED goes out in parallel, wont the whole string go out? I've heard bad things about parallel, so I think I'll stick with series.

I think the bad things that people hear are related to the risk of current imbalance during normal operations. While this discussion has merit in theory I have never seen anyone actually complain about it in practice.

I have one data point of success in my system. Maybe there have been data points of failure (actual failure, not a mathematical calculation) but I have not seen them.

Failure Conditions:
I run three strings so a failure in one string will push the load onto the other two strings. For my XPGs, I run the 2.5A driver at full power providing 833mA (2.5A/3) to each string in normal ops. A failure of one 1/3 would put 1.25A (2.5A/2) into each which is still within spec for the XPGs.

For the XPEs I throttle it back a touch to keep the failure condition such that it doesn't burn out the remaining XPEs.

 

[AngelAddict]

Its good that you have a back up plan. I think I'll still play it safe and run extra drivers in series.

 

[TheFishMan65]

HeneryH do you also have fuses? Or are you just hoping you will catch the problem before a second string can go since that would overload the third one? And just as point of interest how long have you been running and have you ever lost a string?

thanks

 

[Zimster]

Quote:
Originally Posted by AngelAddict
If one LED goes out in parallel, wont the whole string go out? I've heard bad things about parallel, so I think I'll stick with series.


If one LED goes out in serial or parallel, all LEDs connected to it in series will go out. Say you have 3 strings of LEDs running parallel and one in string 2 fails, the entire string 2 will go out. However, when this happens if you have the same power entering the system, it has to be dissipated through out the rest of the circuit thus making the rest of the LEDs become much brighter. This in theory would end up burning your LEDs (in strings 1 & 3) to the point where they all fail.

When you have just one string (series) and one LED fails, the circuit is broken thus no current can flow. This shuts off all the LEDs in the build. This is much safer, imo, in that you can save the rest of your LEDs when one fails.

I would stick to series as well. Not just for that reason but because of the amount of current that you have to provide to a parallel circuit versus a serial. (2.1Amps vs 0.7Amps)

 

[HeneryH]

..I would stick to series as well. Not just for that reason but because of the amount of current that you have to provide to a parallel circuit versus a serial. (2.1Amps vs 0.7Amps)

Why would more current total output matter? The power is the same. 24v@2.5A vs 48V@1.25A. While using serial may have some advantage over parallel (and vice-versa) the current usage shouldn't matter either way.

 

[Zimster]

Why would more current total output matter? The power is the same. 24v@2.5A vs 48V@1.25A. While using serial may have some advantage over parallel (and vice-versa) the current usage shouldn't matter either way.

The current matters since some drivers cannot drive enough current. For example the 48V@1.25A would only be able to support 1 string since the LEDs need 0.7A to operate at maximum. However, it can support more LEDs in series than a 24V@2.5A (13 vs. 6 @0.7A). The 24V@2.5A can support 3 strings of LEDs in parallel given that the LEDs run at 0.7A. You can have as much power as you want but if there isn't enough current, the system won't work efficiently or the LEDs won't be driven at maximum.

If you don't want to drive them at maximum then i guess current won't matter as much, just voltage to forward bias the diode will matter at that point.

 

[Zimster]

**duplicate posted for some odd reason** edited

 

[HeneryH]

Zim,
My point was that the 60-24 was well suited for parallel and the 60-48 was well suited for serial.

If you go parallel you can get 21 LEDs /60-24 driver. If you want to go serial you can get 14 LEDs / 60-48.

Current draw would not factor into the choice.

Clearly other issues would factor in, just not current draw.

 

[sfsuphysics]

Stu, out of curiosity why can the LPC be run in parallel and the ELN not? Do the dimming capabilities really toss that big of a monkey wrench into things? I would think that if you had (ideally) matched LEDs and ran them parallel you'd simply split the current to each side.

 

[TheFishMan65]

I had not studied the drivers with as much consideration as HeneryH, and just to the initial assumption the parallel was bad (and I thought supported by the driver manufacturer) since most people picked drivers that were more suited for series.

What I have yet to decide on is the best configuration for parallel. Is the loss of one string as HeneryH designed enough or do we need current limiting resistor, or current stopping fuses. HeneryH you never did answer me about the fuses Another option (is it enough ??) would be to not only set the current, but also set the voltage when they are run in parallel. Then if you loose a string the voltage is limited and if can not drive high enough to put out the current that was set earlier. I almost think that this is the way the Meanwell drivers are meant to be operated.

 

[Zimster]

Zim,
My point was that the 60-24 was well suited for parallel and the 60-48 was well suited for serial.

If you go parallel you can get 21 LEDs /60-24 driver. If you want to go serial you can get 14 LEDs / 60-48.

Current draw would not factor into the choice.

Clearly other issues would factor in, just not current draw.

I was just trying to explain why one should probably go with a serial circuit. It's a little less dangerous and you can use a lower grade wire because of the lower current. One wastes a little more power with serial and there are more parts to buy, though.

However like you said, you can get more LEDs with the parallel version (18 LEDs @3.5V & 0.7A with the parallel circuit and only 13 with the serial ) while having the same power usage capability. If you know what you are doing, you can probably go with parallel but one risks all the the LEDs on one circuit if one fails.

There are so many factors to a build like this. Its all up to the builder. There are ups and downs to both.

 

[jerryz]

This has actually been covered ad naseum in the main thread. But a current mirror will be of help if want to wire in parallel. The Meanwell data sheet has a decent example. Although for simplicities sake the serial versions have Ben wired up by most people. Fewer parts and a simpler wire up.

 

[HeneryH]

Fishman, I didn't use fuses but probably should now that you mention it.

 

[laverda]

"Wow! There is a whole lot of incorrect information in this thread."

"The ENL-60-48 is designed to run multiple strings in parallel"

They are ELN-60-48s, and they are not designed to "to run multiple strings in parallel" they are designed to run ONE string in series.

The LPC models can be used in parallel configurations.

ALL of the ELN series have internal adjustments for current that sets the MAX that the driver will put out.

The ELN-60-48P & ELN-60-48D have ADDITIONAL dimming functions that let an external device control the brightness ( UP TO the internal setting ).

The P version requires a "Pulse Width Modulation" signal whichis most easily done with a micro-controller.

The D version requires a 0-10 Volt DC signal which can be done by various reefcontrollers or the radio shack DIY flavor where you have to turn the knob.

Stu

I have run two strings of 14 in parallel on the ELN-60-48D with out any problem. Granted I have not run them for weeks, months, or years yet.

 

[laverda]

If you use properly sized fast blow fuses, there should be vertually no risk of losing any strings of LED due to over current when wired in parallel.

 

[AngelAddict]

I'm Just going to play it safe and run series, at about What $35 each, the extra drivers are worth the piece of mind. I just need to read up on and figure out How to work the dimmer without a controller.

 

[AngelAddict]

A little off topic, but since I'm running the drivers in series, in order for the right color scheme, I should get 13 XP-Gs and 24 XR-Es? What's a good price for those? I can get them shipped for around $220, I'm not sure if the XP-G come with 60 degree optics or not. I just want to make sure that's a good price before buying. This is going over a 34g solana 20" cube. Will that be enough light? I'd like to keep clams on the sandbed and SPS pretty much anywhere.

 

[TheFishMan65]

I think that is a reasonable price. I just paid $6 for XR-E royal blues, not sure of the price for XP-Gs. You might check with ETG Tech I think they misshipped me 80 degree XPG rather than XR-E. Should be plenty of light based on the current number per square inch you are near the upper end 37 vs 40, but you also have XP-Gs. I don't know about the clam haven't read how they do with LEDs.

 

[AngelAddict]

Well I know the XP-Gs are What 30-40% more efficient then the XR-Es? I don't really count the 24 XR-E royals, there only for color. Its the 13 XP-G whites That will be giving me the par/light I need for corals. I will gave 60 degree optics on the XP-Gs as well. Thus will total me $230 shipped, $5.25 for XR-E, $6.50 for XP-G and $1.00 per optic. I'm going to put in the order now.