PDA

View Full Version : LED build for my 210


Tbduval
11/16/2010, 11:31 AM
Hi LED guyru's!

So I am in the planning stages of an LED build for my 210. Tank size 72"*24"*29". I am replacing my 250 MH and 4-5ft 80w t-5's with leds. I am think somewhere around 100-120 leds unless you don't think that we will be enough.
I have study almost every LED unit I can get my hands on. Went to the LFS and study the fixture. They even let me look into the guts on the AI. I must shop there too much. I am after nice 20k look but need to keep my PAR pretty high, I have SPS growning near the sandbed. So off to my thinking:rolleye1:

2RB:1CW- All Cree's the CW will be the XP-G's. Since I have a 2:1 ratio, I was think that all the RB will have a 40 degree optic and the CW will have a 65 optic. Do you think this will work without spot lightning affect and retain the PAR I am after? FYI, I will be making the unit dimmable.

Thanks for the help!

Bamm Bamm
11/16/2010, 02:48 PM
Tagging along I'm thinking of doing a led setup on my 210 also

Larry Grenier
11/16/2010, 03:25 PM
I just build a 12 3W Blue Cree LED strip from a kit frm Rapid LEDS for just the actinic portion of my 96gal and I was amazed at how much light the 3W Crees put out. It seems to me that you should look into 3W Crees with all 60degree optics. Maybe a ratio of 4:3 blue:white to get a bluer look than 1:1.

http://www.rapidled.com/servlet/StoreFront Mike @ RapidLEDS was very helpful.

der_wille_zur_macht
11/16/2010, 03:32 PM
2RB:1CW- All Cree's the CW will be the XP-G's. Since I have a 2:1 ratio, I was think that all the RB will have a 40 degree optic and the CW will have a 65 optic. Do you think this will work without spot lightning affect and retain the PAR I am after? FYI, I will be making the unit dimmable.

Thanks for the help!

Don't pick optics based on ratios. Pick optics based on desired spread, mounting height, depth of the aquarium, and desired intensity. I would not suggest using different optics on different colors, as this may result in differences in coverage that you find undesirable.

I think you're pretty close with the plans you have described, assuming your build follows typical practices. Maybe a hair on the low side (i.e. I'd stick with 120 LEDs vs 100) if you're planning on high-light corals down at the sandbed.

Tbduval
11/17/2010, 01:28 AM
Bamm Bamm- tag right along. I plan on getting the some parts within the next week or so. I am going to build a test setup for frag tank in my sump. Then I will build one at a time for my 210. Total of 3. 40 per heatsink.

Larry Grenier- I plan on using 3w Cree's. Its the only LED i know of! j/k. I own stock in the company because of the reefers....!!!:lmao:

der_wille_zur_macht- Ok. 40 per bank seems doable. I figures having more blue's with tighter optics would help colors issues. What optics do you recommend? Also how high should I mount them in canopy? I have around 14" or so to play with.

Thanks!

der_wille_zur_macht
11/17/2010, 08:41 AM
If you have 14" of height I'd mount them right up at the top and use something in the 40 - 50 degree range.

Tbduval
11/17/2010, 11:53 AM
Thanks. Now I have some sort of baseline. Let me ask one more question, ya right, do I still want a 2RB:1CW ratio or 50:50? I like the look of my MH 250 XM 20k's but hate the power it uses and light spill. Will it give me the coverage a need with 40 degree optics. If it were you, would you build with dimmable Meanwell drivers or standard Meanwells? Do you really feel the XP-G's are worth the money difference over the XR-E's for my build?

Thanks!

der_wille_zur_macht
11/17/2010, 12:02 PM
If you're doing XP-G for the whites, I would definitely tend towards a 2:1 ratio (blue:white) if you like a blue look. The 1:1 ratio was based on XR-E for both colors, and XP-G is significantly more intense - so you need more blues to balance them out.

jefathome
11/17/2010, 04:12 PM
I just build a 12 3W Blue Cree LED strip from a kit frm Rapid LEDS for just the actinic portion of my 96gal and I was amazed at how much light the 3W Crees put out. It seems to me that you should look into 3W Crees with all 60degree optics. Maybe a ratio of 4:3 blue:white to get a bluer look than 1:1.

http://www.rapidled.com/servlet/StoreFront Mike @ RapidLEDS was very helpful.

Tagging along....


So what are the dimensions of your 96g? What are you using for your normal lighting and how much difference does the new unit make in turning the tank more "blue"?

I have a 150g (48x24x30tall) with two 14in Solar-tubes on it. I love the shimmer I get, I love the FREE light I get from the Mrs (Mother Nature). It is just a tad too yellow for me (since I guess these are in the 5,000k range).

Right now I have 3 T5's (1 Actinicplus and 2 AquaBlues) and that still is a little too yellow for me. Was thinking that one of these Rapid LED's would be what I needed instead of the T5s. Wanted to talk to someone with actual experience though....

so do you think I would want a 24 LED unit?

Tbduval
11/23/2010, 12:12 PM
der_wille_zur_macht- Question for you. What should I use to power the LED's? Running a million or 6-8 Meanwell drivers seem very strange to me? Is there something better I can use to use, considering I will be running 120-148 LED's? Thanks for the help!

TheFishMan65
11/23/2010, 12:14 PM
Look into the HLG series of Meanwell drivers. Or build you own (http://www.reefcentral.com/forums/showthread.php?t=1759758)

Tbduval
11/23/2010, 12:38 PM
Thanks. I thought DIY but I not sure I want to go that route just yet. How many LED's can I drive off HLG drivers?

der_wille_zur_macht
11/23/2010, 12:42 PM
Depends on which one you get. They range from 80 - 320w, so depending on your drive current choice, you're looking at a range of maybe 30 - 120 LEDs per driver. They're really the best solution for very large builds, if you don't mind potentially having to deal with parallel strings (which isn't a big deal as long as you understand and prepare yourself for the consequences) and you're OK with having low granularity in terms of control (which most people end up with anyways, via ganging multiple drivers onto a single channel of control).

Tbduval
11/23/2010, 01:36 PM
Thanks. Do you think it would be best to get 2 of the drivers? Which size would you recommend? One for the blue's, one for the CW. I don't care about control that much. I'd have less drivers laying around.

TheFishMan65
11/23/2010, 01:39 PM
Someone just got some. He even posted in one of these threads. Not much help. I had suggested CDIWeb.com so you might want to try them.

Chris27
11/23/2010, 01:56 PM
Are those drivers available yet? I was checking Mouser the other day, and they listed that they are on order and won't be available until March. If someone knows a distributer that has them in stock I would appreciate the heads up.

Tbduval
11/23/2010, 02:10 PM
When trying to determine how many LED's can run off of one driver, what should I be looking at exactly? Wattage and Volts? Thanks!

TheFishMan65
11/23/2010, 02:12 PM
Here is the other thread I mentioned. I think he used 185 so maybe they are more readily available.
Need some LED guys to check out my numbers before i order please (http://www.reefcentral.com/forums/showthread.php?p=17943398#post17943398)

der_wille_zur_macht
11/23/2010, 02:18 PM
When trying to determine how many LED's can run off of one driver, what should I be looking at exactly? Wattage and Volts? Thanks!

Wattage will give you a very rough idea of how many LEDs a driver can handle. You also need to consider voltage and current to determine the layout for the LED array.

The typical approach is to wire a bunch of LEDs into a single series string. You basically add LEDs to the string until the string's Vf is just below the driver's maximum output voltage. This works fine assuming the driver is providing the current you want in the LED string (i.e. 1000mA, for instance). However, many of these larger drivers operate at currents much too high for use with a single series string. For these drivers, you would basically size a series string according to the output voltage, then add parallel strings (of that length) according to the output current.

Think of the output voltage as dictating the "length" of the array, and output current dictating the "width" of the array.

Imagine you have a theoretical LED that wants 1A current and runs at 3.5v.

Now, imagine the following drivers:

1) 48v max, 1A. This one's easy. It's essentially a Meanwell ELN60-48 turned down a bit. You can fit 12 LEDs in a series string, and you want only one such string (since the driver is providing your target current.) So, this driver can handle 12 LEDs total.

2) 48v max, 6A. Same voltage, so you're using strings of 12 LEDs. However, since the driver provides 6A current, you'd want 6 strings, not just one. So, this driver can handle 72 LEDs total.

Tbduval
11/23/2010, 02:31 PM
Gotcha.....mostly anyways. In your example #2, is that in series or parallel. What happens if one of the LED's blow? I can live with 3 drivers, since I am building 3 seperate fixtures of 48.

der_wille_zur_macht
11/23/2010, 02:51 PM
In example 2, you're have six strings in parallel. Each string would consist of 12 LEDs in series.

LEDs can fail open (not conducting) or closed (conducting). If one LED in a series string fails open, that string is effectively removed from the circuit. When that happens, the driver is still configured to supply the same current, so that current is now distributed across a smaller number of strings.

Imagine you have a driver providing current I. You have n strings attached to it. By default, the current in each string will be I/n. Reference example 2 above. With 6A and 6 strings, each string gets 1A current. If one string dies, you have 6A and 5 strings, so each string now gets 1.2A.

So there are basically a few things to think about. First, it's "safer" to have a large number of parallel strings - because, when the number of strings is larger, a failure causes a smaller rise in current in the still-functional strings, meaning it's less likely to cause damage to them as well. Second, it's wise to plan your string count and driver current such that if a single failure occurs, the other LEDs will still be operating at safe currents. Back to that same example - assume the LEDs in question were XP-G, which can run at 1.5A. You'd be all set - if a single string failed and the other strings jumped from 1A to 1.2A, they would still be safe. However, if you were using LEDs designed with a 1A max, you'd be in trouble - a single failure would cause other strings to go above their max, which would likely cause them to fail as well.

The second consideration for safety is to come up with a method to limit current in the event of failure. There are a few ways to do this. First, you can put a current limit resistor in each string of LEDs. This is pretty safe and foolproof but now you're burning off power and operating less efficiently. Secondly, you can use rapid-blow fuses such that the fuse blows if an unsafe number of strings fail - you'd put a fuse in each series string. The third method is to utilize the drivers' own controls. Most of these drivers have an internal voltage limit adjustment. You can turn that down such that in normal operation (all strings live) the limit is set JUST above operational voltage. Then, if a string fails, the driver will try to bump voltage up to get higher current through the remaining strings, and it'll hit the voltage limit - preventing it from actually reaching it's current target and damaging the LEDs.

The second major consideration is Vf differences from LED to LED. This can be significant. If you end up with strings in parallel that have different total Vfs, then they won't be balanced in terms of current. A very small difference in Vf could result in hundreds of mA of difference in current between the strings - at best this will create a very uneven distribution of light over the tank, and at worst it will create unsafe conditions for the LEDs in one or more strings. To combat this, you can test the Vfs of the strings and move LEDs from string to string until things balance out.

This is all relatively uncharted territory in terms of typical builds on this forum, but IMHO this is the direction that makes the most sense if you have a very large tank.

kcress
11/23/2010, 02:55 PM
{cross post}{hopefully willie and I didn't contradict each other greatly}
If one blows OPEN then that string drops out and the current will be shared between the rest of the strings. If you aren't running the string currents right up against the wall anyway, there is no problem. The rest of the strings get a little brighter and a little warmer.

The more strings you parallel the less loosing one will increase the current to the others. On these big HGL setups it would be pretty minor.

You should notice one missing string and fix things sooner than later.

Now if one LED blows SHORTED. You can have a bit more of a problem. That one string will start hogging current. Probably to its great detriment. However the likely result will be either the already damaged LED switching to being blown OPEN or another LED in that chain blowing OPEN which will revert you back to the original case. Worst case would probably be, (and I'm guessing), two to three LEDs being lost in the case of one LED originally blowing SHORTED. Which is the least likely failure mode.

If you're going to parallel two strings off of a ELN60-48 then you better dang well use fuses and correctly set the voltage limit because the second string can not carry the resulting current if one string opens.

If you spread an ELN60-48 across three strings then two strings can carry the resulting current. Fuses would still be suggested.


If you are running an HLG120, or 240, or 3??, then you don't likely need fuses as long as you are running the string currents below the LED maximums, as you should be, because you'll get substantially longer lifetimes and greater efficiency.

In ALL parallel cases you should be running 1 OHM power resistors - one to a string - as that is the ONLY rational way to measure the individual string currents to see if they're balanced reasonably well. Breaking a bunch of strings to insert an ammeter is not the smart way to do it.

der_wille_zur_macht
11/23/2010, 02:59 PM
Now if one LED blows SHORTED. You can have a bit more of a problem. That one string will start hogging current. Probably to its great detriment. However the likely result will be either the already damaged LED switching to being blown OPEN or another LED in that chain blowing OPEN which will revert you back to the original case. Worst case would probably be, (and I'm guessing), two to three LEDs being lost in the case of one LED originally blowing SHORTED. Which is the least likely failure mode.



I think it's fair to point out that this scenario (a failure to a shorted state) would have the same impact on a "traditional" single string configuration.


In ALL parallel cases you should be running 1 OHM power resistors - one to a string - as that is the ONLY rational way to measure the individual string currents to see if they're balanced reasonably well. Breaking a bunch of strings to insert an ammeter is not the smart way to do it.

Can I pick your brain on this? To date, when I've experimented with parallel strings, I have not included resistors as a current measurement tool. Instead, I have simply measured voltage drop across strings (and then across individual LEDs if warranted). Is this not a valid method?

TheFishMan65
11/23/2010, 03:17 PM
It will tell you the Vf, but not the current. You can guess the current from the charts printed ie an XP-G with a voltage of 3.5 is dropping 1 amp. Number may not be right , but you get the idea. The problem is the chart is just an estimate.

der_wille_zur_macht
11/23/2010, 03:19 PM
Regardless, the goal is balance across strings. If you know Vfs are the same across all strings, I'm assuming the V/I curve is the same for identical LEDs, meaning current would be the same.

TheFishMan65
11/23/2010, 03:23 PM
All string will have the same Vf if they are hooked up to the same power supply. They will measure the output of the power supply. Are you measuring each string before you hook it up at a known current?

TheFishMan65
11/23/2010, 03:25 PM
Try it this way. Take a 6 string of XP-G and a 6 string of XR-E. Hook them up to an 18 volt power supply. If each LED of a type are the exactly same (Vf / i curve) then each LED has a Vf of 3 volts. Now will they have the same current. No, the XR-E will have way lower current. But the Vf are all balanced.

der_wille_zur_macht
11/23/2010, 03:26 PM
I'm measuring Vfs of individual LEDs and/or entire series strings of LEDs, arranged in a single series string on a driver operating at a known current.

Edit because of your second post - I'm assuming we're talking same model/bin of LED...

TheFishMan65
11/23/2010, 03:41 PM
If the Vf/I curve is the same then yes, but you are the one who convinced they aren't. If the curves are the same then why bin (sort I based on fixed Vf)?

S2minute
11/23/2010, 04:47 PM
Try it this way. Take a 6 string of XP-G and a 6 string of XR-E. Hook them up to an 18 volt power supply. If each LED of a type are the exactly same (Vf / i curve) then each LED has a Vf of 3 volts. Now will they have the same current. No, the XR-E will have way lower current. But the Vf are all balanced.

So Fishman, your saying that even though the Vf is balanced, the limiting factor will be the lowest current of a pictular LED string? Like, say the string of XR-Es max is 1000MA and the XP-Gs string is 1500MA, even though they are hooked up to the same power supply, you`ll still only be able to drive the 2 strings a max of the XR-Es 1000MA? Dispite the Vf being the same?

If that`s true, then while you can mix different colors on the same string, you`ll still be limited by the lowest colors MA max? I guess that would be a more inefficent way to run it, but it can be done then??

I figured, that`s why i`m isolating colors to their own HLGs because one group,(violet @max 500MA), XP-G(NW,CW @ 1000MA) and RB(700MA). The only thing the Vf of each let me do is figure how many LEDs in how many strings.

So, if i`m running HLGs and x amount of MIXED string colors, the power supply will still even out the VF to a standard BUT the current will be limited to the color LED that has the lowest max MA.?? I`m my case the VIOLET at 500MA max. Then i`m guessing if one blows, the rest of the LEDs that can run at a higher MA will absorb the extra current??

Boy if i`m on track with all this i`ll shi* myself :beer:

TheFishMan65
11/23/2010, 05:12 PM
It all leads back to this statement by kcress made and I made the mistake of trying to answer.
"In ALL parallel cases you should be running 1 OHM power resistors - one to a string - as that is the ONLY rational way to measure the individual string currents to see if they're balanced reasonably well. Breaking a bunch of strings to insert an ammeter is not the smart way to do it."

DWZM questioned the need for the resistor. He is "measuring Vfs of individual LEDs and/or entire series strings of LEDs, arranged in a single series string on a driver operating at a known current." So his known factor is the current. He balancing before everything is hooked up. If a string is high or low voltage he checks each LED and switches some low for highs in different strings.

In kcress's case he is trying to measure a fully implemented array. He has a known voltage and wants know what the current is in each string. He has say ten string hooked up and want to know if they are balanced. He can either open each of the ten string and measure the current (which is a minor pain). But by placing a 1 ohm resistor in each series he can measure the current in directly. This is done by ohm's law (I think that is the right name) which stages Voltage equals current time resistance. So if he measure .7 volts across a 1 ohm resistor current = voltage / resistance so that string has .7 amps going through it. These need to be relatively hefty resistor in the 1-2 watt range.

I hope this clears things up.

Tbduval
11/23/2010, 06:49 PM
You guys are way past me. It might help if I knew what Vf actually stood for!!!!! My school days have failed me..... its been a while also! I guess as much as I have learned I am still stuck. I realize for large builds, going with a parallel layout. First, what is considered a large build? 100 LED's? Second. Do you think it is better to just try and run 2 drivers, one for whites, one for blue? Thanks again!

TheFishMan65
11/23/2010, 07:01 PM
Vf (forward voltage) is the voltage drop across the LED. 100 LEDs is probably a reasonable point for saying large build.

Two driver should let you adjust the white or blue to get the color wanted. I think that 2 is probably the minimum number of drivers someone would want.

S2minute
11/23/2010, 07:13 PM
Well,..i`m just a tailgater :) literally, i`m hanging on fer dear life,dragging muh feet in the dirt but, i`m smiling the whole way! Vf : is Voltage Forward.

Tbduval
11/23/2010, 07:22 PM
I had a feeling it was but was not 100% sure. I am off to research which driver I should use unless someone already know off the top of their head. You guys a huge help! Thank again!

S2minute
11/23/2010, 07:46 PM
ok wille,fishman i`m getting a much better handle on this...i think :) Before i make a further decision. The setup your going to do for your large tank wille. I know you like the real sunrise/sunset/cloudy/stormy day effect, and so do i,(and that`s what i want over anything). Now from what i`m assuming, your setup will have all mixed color strings reguardless of the different Vf of the LEDs, so you can get that gradual power up from one side of the tank to the other? Just like the sun progressing from east to west? Or are your colors going to be seperated on their own strings and do it by power up on blue, then white to do the sunrise/set sim?
Now i figured out what and how many HLG drivers i need with the amount and type of LEDs. I`d not just Dim to get the color i want, that`s a waste to me. I`m set with what i have figured. If i could do it, i`d trade off the color seperation for the more realistic simulation. I just don`t know how to lay it out ..yet.
1x HLG-240-42B 60 white XP-G(20NW,40CW)
1x HLG-100-20v(adj to 17)4.8A(adj to 3A) for 30 Violet
2x HLG-185-42v for 120 RB XR-E
Total: 210 LEDS over a 175 Bow
3x 10"x 22" HS from heatsinkUSA which will fit 70 leds ea.

TheFishMan65
11/23/2010, 08:36 PM
If you get the B version you should have the ability to dim. This will let you bring up each driver slowly. But everything on that driver sill come on at the same time.

Tbduval
11/23/2010, 09:57 PM
So let me see if I have figured this out.

HLG-185-42

42 Volts 4.4 Amps 185 watts

Running XR-E RB, volts at 700 mA is 3.5, so 42/3.5=12 LED's per series, then to find out how many strings in parellel take 4.4/.7=6.12, 12*6=72. 700 mA @ 3.5V equals 2.45 a watt LED. So I should be able to run 72 LED's on this driver.

Am I even close or do I need to "Electronics for Dummies" again!

TheFishMan65
11/23/2010, 10:02 PM
Grade A+

kraut3253
11/23/2010, 10:32 PM
Why not use the new Cree XM-L bulb instead of the XP-G for the CW light? It gives you twice the light output of the XP-G with roughly the same cost per bulb. The XM-L is also supposed to be more efficient which will cut down on power consumption and heat production.
One XM-L bulb give you 910 lumens @ 3.35V,3A for a cost of $7 per bulb.

Tbduval
11/23/2010, 11:24 PM
I am not the one to ask that to. I am still young in LED's. Hopefully someone will explain one way or the other at why or not. Where are you getting you pricing from on the LED's from?

Tbduval
11/23/2010, 11:25 PM
What about this bulb Cool White MC-E CREE LED? Looks like it maybe good but still kind of costly!!!!

S2minute
11/24/2010, 12:03 AM
What about this bulb Cool White MC-E CREE LED? Looks like it maybe good but still kind of costly!!!!

MC-E is a 3-up LED at what somewhere between 4-500 lumens? if i`m not mistaken for about $19-20 US. I was told it might it may put out to narrow a beam. Alittle painful in the wallet if it shorts.

S2minute
11/24/2010, 12:06 AM
So let me see if I have figured this out.

HLG-185-42

42 Volts 4.4 Amps 185 watts

Running XR-E RB, volts at 700 mA is 3.5, so 42/3.5=12 LED's per series, then to find out how many strings in parellel take 4.4/.7=6.12, 12*6=72. 700 mA @ 3.5V equals 2.45 a watt LED. So I should be able to run 72 LED's on this driver.

Am I even close or do I need to "Electronics for Dummies" again!

Yeah man :) right on. Let the fun begin!

Tbduval
11/24/2010, 12:55 AM
The fun has already started. I have learned so much by reefing. Just think....I am now a plumber, electrician, mechanic, and now electronics!!!! I now fix everything my house now because I have the motto " I can do that myself, it cant be that hard!" Now on to where to buy everything from. Where are you guys find the best prices at? One more thing. What about the XP-E RB? Do you think it worth the cost difference? Thanks again!

kcress
11/24/2010, 01:41 AM
Shesh, I look away for a few hours and you guys run a muck!! LOL

Let me see if I remember all the points..

Willie: No you can not measure the string voltages and get a meaningful current. As FishMan pointed out, if you measure across the strings you are only measuring the same voltage on all the strings. They will all measure the same even though the currents can vary greatly. You need something with a linear, fixed, temperature invariant, resistance. A LED is the antithesis of that.

The 1 Ohm resistor provides that without adding a lot of wasted power. If the measly 0.7W you'd lose in a 700mA application is too much for one, it could be dropped further if needed.. By using a 0.1 Ohm resistor. That would drop any loses to 70mW. You'd have to multiply your voltage reading by 10 to convert the reading to amps.


S2; You should NOT run parallel strings of different types of LEDs. That could work but...damn!... it would be risky. And, trouble prone. Just say no.

That said, you need to do a "one for you", "one for you", "one for you", etc., etc., for each string. Each string must have the same number of each type of LED. (This is why I shuddered at all your various LEDs. It's certainly do-able but it adds a little complexity to the build.)

Once you have all the strings with exactly the same numbers of the various LEDs you can measure the currents using the trusty resistors and directly compare how they're sharing the current.
(Note: You can order the LEDs in any string any way you want. No specific order is required.)

Somewhere I described the process of balancing but I'll paraphrase it here:

1) Find the string with the highest current.

2) Find the string with the lowest current.

3) Measure all the individual LED Vfs in the highest string and write them down.

4) Measure all the individual LEDs Vfs in the lowest string and write them down.

5) Pull the LED with the highest Vf in the lowest current string

and swap it with

the LED with the lowest Vf in the highest current string.


Repeating. The highest drawing string's lowest Vf LED is contributing most to that high current.

The lowest drawing string's highest Vf is contributing the most to the low string current.

Swap them.

Remeasure all the strings again.

Step and repeat as required to bring all the strings within about 5% if you can. That's about 35mA in a 700mA / string setup.

der_wille_zur_macht
11/24/2010, 07:31 AM
I understand what you guys are saying but I think you're misinterpreting my test procedure. I'm "binning" LEDs in individual strings or one LED at a time, not trying to measure voltage across multiple parallel strings on the same driver.

At any rate I like the method kcress is describing better anyways. :D

TheFishMan65
11/24/2010, 07:35 AM
I eventually figured that out, and thought I explained it a few post back. But I think we are all clear now.

Tbduval
11/24/2010, 09:23 AM
Hey guys! Do we ever sleep? What RB's should I use? XR-E or XP-E? This would be mixed with the XP-G CW? Thanks!

der_wille_zur_macht
11/24/2010, 09:31 AM
Personal preference. Efficiency-wise they are (likely) the same. The only significant difference is that the XP-E has a wider viewing angle, meaning it's light is spread over a slightly wider area. If you're using optics, this becomes moot.

My approach would be to get in touch with the vendor(s) you plan on using, and see what bins of the two LEDs they have available, then pick based on which model comes in the most attractive bin. Most would agree that you want the highest brightness bin (16). Color bin is open for debate, but I'd bet that most would want the lowest wavelength they could get (D3).

Tbduval
11/24/2010, 09:34 AM
Where are you finding the best place to buy LEd's right now? If they are the same or close to it, I will go with the XR-e's because they are cheaper and plus I will be running 40 degree optics.

der_wille_zur_macht
11/24/2010, 09:36 AM
I've seen XP-E cheaper depending on the source, so don't assume XR-E will always be cheaper.

Cutter and ETG are my favorite sources. They both have reasonable prices, and both are happy to talk bins with you - letting you pick from a good selection, or at the very least telling you what they actually have. Compare to most hobby-targeted vendors, who, half the time, can't even tell you which bin they have, much less offer a choice.

For optics, I'd check with ledsupply, as they seem to have good selection and good prices.

TheFishMan65
11/24/2010, 10:11 AM
I am finishing up a group buy I did with ETG. There is a minimum quantity of 10 and I think a minimum order of $75.
XR-E RB $3
XR-E CW $4 if you order over 100 (we did) I think $4.50 for less but not positive
XP-E RB $4 for 100 or less less $ for more
XP-G $75 for 100-500
Lenses are $1 or less for XR-E (40, 60, 80) and $1.20 or less for XPGs (40,60)
This assumes prices have not changed. They changed twice as I got the order together.

Tbduval
11/24/2010, 11:25 AM
I am finishing up a group buy I did with ETG. There is a minimum quantity of 10 and I think a minimum order of $75.
XR-E RB $3
XR-E CW $4 if you order over 100 (we did) I think $4.50 for less but not positive
XP-E RB $4 for 100 or less less $ for more
XP-G $75 for 100-500
Lenses are $1 or less for XR-E (40, 60, 80) and $1.20 or less for XPGs (40,60)
This assumes prices have not changed. They changed twice as I got the order together.

Are these mounted to stars? If so, that not a bad price at all!

TheFishMan65
11/24/2010, 11:33 AM
These were on a 20mm start - oh and add shipping. I just checked DealExtreme adn you can get 10 XR-E Q% for 44.11 if you need small quantities of those.

Tbduval
11/24/2010, 12:07 PM
What's the consensus of the mA running of the XR-E's RB? What about the XP-G's?

TheFishMan65
11/24/2010, 12:13 PM
I think most are in the range of 500-700 for both.

der_wille_zur_macht
11/24/2010, 12:51 PM
I agree. I bet 700mA is the single most popular drive current for the LEDs commonly used.

Keep in mind that as long as you are under the max current you are technically OK (given proper thermal design). However, the lower you run an LED, the more efficient it will be, and the longer it will last. If you can afford the upfront investment, and plan on keeping your rig for a long time, it's almost always better to use more LEDs at a lower drive current.

Tbduval
11/24/2010, 01:17 PM
So on my build, does 144 LED's sound good? 48 XP-G's CW & 96 XR-E RB with 40-50 degree optics mounted 14" above my tank. Will that be between 15k-20k? Still trying to iron out exactly what power supplies to buy. Thanks!

TheFishMan65
11/27/2010, 08:10 AM
Tbduval, I can't seem to find your tank size. But 144 would be good for a surface are of between 1440 and 2880 according to common numbers.

SkorFred
11/28/2010, 07:15 PM
From TBDuval
So on my build, does 144 LED's sound good? 48 XP-G's CW & 96 XR-E RB with 40-50 degree optics mounted 14" above my tank

I have the same tank and make a similar led setup like you 'rs
Ive order from Cutter.com 48 XP-G CW and 63 XP-E RB + 9 XP-E Red all no optic. 120 Led on a 10'' x 64'' heatsink. But as you, I'm not sure about the drivers. I want the less driver possible and i dont really care about dimming for my first setup. I Thought doing like that: plug the XP-G on a Meanwell HLG 240-48 like this: 4 parallel row of 12 led in serie and the XP-E on 6 row of 12 led on a other HLG 240-48. Is it a good setting? I've probably choose the HLG 185 but no one have it in stock unlike the HLG 240 who are the most '' in stock'' driver High Power and less expensive.

Tbduval
11/28/2010, 08:01 PM
Tbduval, I can't seem to find your tank size. But 144 would be good for a surface are of between 1440 and 2880 according to common numbers.

210-Gallons- 6ft * 2ft * 29" inches

I am trying for 2 drivers but I don't see how I can do that with number of LED's for the blue's vs. the white. I can live with 4-6 drivers but would prefer less. Thanks for the help!

TheFishMan65
11/29/2010, 07:49 AM
144 sounds like a good number. IIRC te HLG 240 is 240 watts so it should be able to power close to 80 LEDs.

Tbduval
11/29/2010, 08:55 AM
I looking at 96 RB and 48 XPG cool whites. I think I have decided to build one array of 48 first. See how I like the color and get some par readings. Once that is completed, I will go from there.

kcress
11/29/2010, 04:49 PM
Yeah FishMan, that sounds pretty good. You can also dial those down a bit, so that would work for a few less LEDs.

Tbduval
11/29/2010, 06:51 PM
Since this my first LED build, I want to build an array of 48. Play with it, test it out, make sure everything works the way I want it to. Do I get 2 drivers, turn them down and run the 48 array. When I decided to build the other 2 array's of 48 intergrate them with the other drivers? Does that make sense? :hmm3:

TheFishMan65
11/29/2010, 08:31 PM
I don't remember enought about the HLD drivers, but there should be a way to limit it for a small build and add later.

Tbduval
11/29/2010, 08:55 PM
TheFishMan65----Thanks! Now off to find my drivers but I will need some help with wiring in parallel. What size fuses do I need and where exactly do I place them in the line?

TheFishMan65
11/29/2010, 10:12 PM
You want fast blow probably about 750 ma, but would be ok upto 1 amp. One in each string.

kcress
11/30/2010, 01:02 AM
Since this my first LED build, I want to build an array of 48. Play with it, test it out, make sure everything works the way I want it to. Do I get 2 drivers, turn them down and run the 48 array. When I decided to build the other 2 array's of 48 intergrate them with the other drivers? Does that make sense? :hmm3:

In your case I'd probably go with the ELN60-48. Run dual strings with resistors and fuses. That way each string can run at 650mA. So you'd have two strings of white(24) and two strings of blue(24). And 2 ELNs.

The problem comes if you want something other than 1:1 of each color.

That would result in using 3 ELNs and 48 blue and 24 white for 2:1. But now you'd have 9 ELNs with your three fixtures. That's right on the edge of a poor idea.

If you are not going to try to control the individual colors there are other ways to go.

der_wille_zur_macht
11/30/2010, 08:35 AM
If you are not going to try to control the individual colors there are other ways to go.

Here here! Dimming of individual colors is overrated IMHO - we are best off designing the correct color with ratios and currents.

Still, even with two HLGs, if you had to build "hybrid" strings containing more than one color to hit your overall ratio target, you could bias one of the HLGs towards one color and the other towards a different color - to give you at least a little bit of color-changing dimming, without having to resort to a huge pile of less desirable drivers.

TheFishMan65
11/30/2010, 10:48 AM
If you do XP-G for whites and XP-E or XR-E for royal blue then you could replace 2 or 3 of the white with RB to get the 40/60 mix. Most turn the blues on first and off last so I am not sure that you would ever notice the blues in with the whites.

Tbduval
11/30/2010, 12:15 PM
I was thinking a a total of 3 drivers. 2 set on the blue and one on the whites. I had the driver picked out but cant remember which model. Let them throw this at you. If I can a driver thats more powerful for my big build but want to use it on my test rig, can I throw resistors in place of LED to even out the current/amps? Because the driver I was looking at would over power the RB's by a string in parallel. Mixing XP-G and XR-E on the same string is a bad idea right? Is that because of the Vf?

Tbduval
11/30/2010, 12:21 PM
I am thinking 3 of CLG-150-48 for LED's. Which one do I want to get blank, A,B,or C driver? What you guys think?

TheFishMan65
11/30/2010, 12:42 PM
None of those would work well for dawn to dusk. The difference between A and C are the terminal it looks likes C has screw terminal you connect to and A has wires. Peronsal choice, but I prefer screw terminal (verify that is what they are) over wire nuts or the like. B would be nice if you will keep adding corals and need to acclimate. It allows for a potentiometer for dimming. A/C also have this but you need to remove a rubber stop to adjust. I would go with C personally unless I read the data sheet wrong.

Tbduval
11/30/2010, 12:55 PM
Not looking for a dawn to dusk setting. Just the ability to turn on the whites and blue independently. Do you think this driver would be better HLG-150H-48?

der_wille_zur_macht
11/30/2010, 01:00 PM
I really think you want the HLGs, not the CLGs - unless you don't care about dimming. The HLG series has much more convenient dimming functionality.

der_wille_zur_macht
11/30/2010, 01:04 PM
Ok, just saw your last post. If you really don't care about dimming, and don't think you ever will, then fishman's recommendation is fine. If you think you'll ever want to automate dimming, I'd spend the extra few bucks on HLGs instead.

TheFishMan65
11/30/2010, 01:06 PM
The HLG is much more convenient. Anyone have any thoughts on why they took out the voltage set point? I kind of feel this is required for multiple strings.

Tbduval
11/30/2010, 01:10 PM
I say I don't about dimming now, but that could change. I don't want to be 1-2 years down the road say, I wish I had the dimming ability. Can you explain what this means, Constant current level adjustable through output cable with 1~10Vdc? Also, will check my math, 3 of HLG-150H-48 for a total of 148 LED's.

TheFishMan65
11/30/2010, 01:17 PM
The rated current is 3.2 amps so at 10 volts (I think it maybe 0) or 0 ohms, or 100 PWM you will get 3.2 amps out at whatever voltage (upto 48V) is is required. At 5 volts, 50K (IIRC), or 50% PWM you will get 1.6 amps. At the 1 volt .32 amps.

I thik you mean 144 LEDs (3 * 48). That would be 4 strings of 12 on each driver. 3.2 amps / 4 strings is 0.8 amps or 800 ma. So as long as you don't want to driver over 800ma it will work.

Tbduval
11/30/2010, 01:21 PM
You are right. My math and typing is exactly bad when running off 3 hours of sleep. Now that I have found the driver, do you know of some sites that sell them. I looked at cdiweb.com but they do not have any in stock. One step down, 500 hundred more to go!!!! Thanks again for the help! I would be lost without the advice from all of you!

TheFishMan65
11/30/2010, 01:25 PM
I have heard they are not out yet (but they be some other model) and I have heard someone ordered some (once again maybe some other model). But I am afraid I don't know who ordered or what from where.

der_wille_zur_macht
11/30/2010, 02:09 PM
Neither of these drivers are in any of the hobby vendors (which is too bad really) but at least a few of the industrially-oriented vendors out there are showing them in stock, at least the last time I checked. Pricing is a bit of a sticker-shock until you add up the cost for all the ELNs you're replacing with a single HLG.

Tbduval
11/30/2010, 02:24 PM
I really like the idea of a DIY driver but I get in that thread and just get lost.....:confused: It seems like the way to go for a large build. I am studying on them but at this point it might be my only option.

saf1
11/30/2010, 02:25 PM
Hi all. I'm a bit rusty here but in the 3 setups that I have (1 nano with 4 LED's, and 2 11 LED rigs over 20 gallon tanks) I'm using the ELN's which I believe are the 35 750's. I like them and they are affordable.

I wanted to ask though since I've been reading this thread. On larger projects there are a couple drivers out now that can actually run more led's? Do you use them the same way as it relates to setup or does it require a different wire method?

I wasn't sure so wanted to ask. I was also wondering if you could define sticker shock :)

TheFishMan65
11/30/2010, 02:30 PM
Pretty much the same wiring - take string and tye the ends together with a fuse and/or resistor.

ELN are around $30 HLG I think I saw running $90.

Tbduval - there is a summary on the driver thread within the lst few pages IIRC

Tbduval
11/30/2010, 02:33 PM
There are bigger drivers that can run more LED's than what you are using. THe do require that you run them in parellel. It is do able but I can not find the one's I want in stock anywhere. When he says sticker shock anywhere from 75-150 dollars per driver.

der_wille_zur_macht
11/30/2010, 02:35 PM
I really like the idea of a DIY driver but I get in that thread and just get lost.....:confused: It seems like the way to go for a large build. I am studying on them but at this point it might be my only option.

DIY drivers are definitely "for DIY'ers." If you're in to that kind of thing it'll be a great project, and cheaper than most commercial drivers with the same functions. If you're not in to that sort of thing, don't fret. There's no magic bullet that's only available on DIY drivers - you can pretty much get commercial drivers with the same functionality. And, right now at least, the commonly-used DIY designs are all powering very small arrays at low voltage, which is a little bit of a pain on a very large tank.

I wanted to ask though since I've been reading this thread. On larger projects there are a couple drivers out now that can actually run more led's? Do you use them the same way as it relates to setup or does it require a different wire method?

I wasn't sure so wanted to ask. I was also wondering if you could define sticker shock :)

The driver's we're discussing here - CLG and HLG - are Meanwell-brand products meant to power larger arrays. They are basically the bigger brothers to the ELN line. Similar overall use but more power. Besides that, the major differences are better response to dimming (no cutout low down in the range) and they've got active PFC front ends, so you don't have to worry about poor power factor when running a ton of ELNs.

These drivers do indeed require a bit of a different technique, since they all operate at much higher currents than we're used to. What this means is that you're basically forced into doing parallel strings, instead of a single series string. This isn't inherently bad, but it IS something you need to understand before implementing, since there are some potential pitfalls to this arrangement that aren't present in a single series string.

Sticker shock - when I checked, the more powerful HLGs were all around $100 - $150 or so, iirc. It sounds like a lot when you're used to paying $35 for an ELN, but one of these big boys can replace half a dozen ELNs so it ends up being a bit more affordable.

saf1
11/30/2010, 02:42 PM
Thanks all. It is nothing I would use today but as I'm slowing starting to plan another 100 to 125 gallon build it is good to know. So far I have been very happy with the LED's that I am using and the growth per my coral type is really good.

Then again green star polyps, hammers, mushrooms, and kenya trees can probably grow in the dark :)

TheFishMan65
11/30/2010, 02:46 PM
DWZM always words things so much better. I guess what I was trying to say and thought I was answering was that you do not need to rewire all your strings. String theory is still good to pint where all the strings collide.

CDIWEB (they are out of stock) list the HLG-150-48 (blank, A, or B) at $71.63.

Salt_in_MyBlood
11/30/2010, 10:35 PM
Just read the thread, great info. I am planning on setting up a 125 Gal tank, and looking into LEDs. Tbduval, What are you looking at for total price? i know you still have a few more things but if you can get the driver you want, what is an approximate price?

Again, thanks for the info

Tbduval
12/01/2010, 08:03 AM
Give a take a thousand or so. To lighten the load of cash at once, I am building 3 array's. Hopefully I can sell my MH & T-5's to recoupe some of the cost. My goal is by this time next year to be on all LED's.

Tbduval
12/01/2010, 08:04 AM
Looking at the HLG drivers, I notice there is a model 150H and just a 150? Is there any difference between the two?

TheFishMan65
12/01/2010, 08:09 AM
A quick look says it is only the input. The H version will go up to around 300 AC for the input and the non H only goes to 264. If that really is the only difference then either will work but I expect the non H will be a lot cheaper and easier to find.

lokian
12/01/2010, 08:10 AM
The HLG series is mostly a back order item at this point... Last i checked i was 14 weeks out on HLG-185-42. I called meanwell to get a better understanding on why most of these are not in stock and was told that it was a newer product that they built on request... kinda strange but o well.. got my order in and waiting patiently.

Tbduval
12/01/2010, 08:17 AM
The HLG series is mostly a back order item at this point... Last i checked i was 14 weeks out on HLG-185-42. I called meanwell to get a better understanding on why most of these are not in stock and was told that it was a newer product that they built on request... kinda strange but o well.. got my order in and waiting patiently.

That is kind of strange. In meantime, I am building my manifold to take out 5 MJ's in my sump. Switching my main pump to a better, yet more efficient one. Switch out the T-5's over the frag tank for LED's. Re-plumb my tank for better efficiency. You know, all the fun stuff. My goal is by next year to have my tank consume 1/2-1/3 less power than it did this year.

der_wille_zur_macht
12/01/2010, 11:26 AM
Keep in mind that "way back when" they were pretty much saying the same thing about ELNs, and they were very hard to find. Now there are all kinds of hobby-oriented vendors able to keep them in stock. They seem able to respond to demand even if it takes a few months.

Tbduval
12/01/2010, 12:21 PM
Keep in mind that "way back when" they were pretty much saying the same thing about ELNs, and they were very hard to find. Now there are all kinds of hobby-oriented vendors able to keep them in stock. They seem able to respond to demand even if it takes a few months.

That makes sense or cents! DWZM- What would it take to convince you to do a DIY driver thread for something similiar HLG's. Something for large builds to control 100 to a couple hundred LED's.

der_wille_zur_macht
12/01/2010, 12:39 PM
Someone to fund the R&D? :lol:

It shouldn't be too hard. There are a few good driver ICs out there that could give this sort of functionality. Designing the circuit would not be hard and I've kind of been wanting to anyways. I'd probably use the LM3409HV from National Semiconductor, which can handle 75v input and 5A drive current. You could run 100+ LEDs off one of these with the right design. The tricky part would be the PFC front end. Most DIY drivers are powered by low-cost 24v DC power supplies that have poor power factor. Best way to handle that would probably be to use an off the shelf PFC-corrected DC power supply. But especially in higher voltages (which you'd need/want for a large array) they're expensive.

Tbduval
12/01/2010, 12:54 PM
When you say expensive, could you please define?

der_wille_zur_macht
12/01/2010, 01:05 PM
Well I've snagged three nice PFC supplies on eBay - I think I paid ~$40 for them. They're 24v supplies, 9A or something like that. I've seen 48v supplies on eBay for a little more but you'd have to get lucky. Buying these off the shelf, they're probably $100 - $200. At which point it's not much of a savings over just going with an HLG, by the time you've built the driver part. Unless you just happen to really like DIY'ing.

Tbduval
12/01/2010, 11:27 PM
Ok...I have finally narrowed down my driver selection. It will be either 1 HLG 320-48-B and one HLG-185-48-B or 3 HLG 185-18-B. It will really depend on on how long it will take to get them and the price of the HLG-320. I just want to see what will be, since I will have the same functions no matter which way I go.

On another note, which does not apply to my big build. I got 12 LED rig going today for my frag tank. Can I say these little things are freakin awesome. It amazes me how much light these put out. I will post some pictures in next day or two, hopefully. I will be going out of town this weekend!

Thanks again guys for help and understanding of this gibberish.:bounce3:

Tbduval
12/06/2010, 01:08 PM
Just to add on. I am playing with my big build. I was going to build 3 banks of 48 LED arrays. As I been thinking lately of going with only 24 on each bank. Placing each LED pretty tight, over the coral, on the bank. I am going to play with optics between 40-70 degrees. Then placing some LED's on some u-channel to to give some spread to help light up the tank a little more.

Now with saying this, I am going to to play with all sort of optics. I want to try and get the max spread and par with the least amount of LED's possible. I know this will work in progress. Light placing and and optics will need to be placed correctly to have the colors blend correctly.

I plan on running, hopefully, no more than 90 LED's. Each LED will run at roughly 2.5 watts. Total power 225 watts. That's less than 1 of my 250w MH. I would say that's really good for a large tank.

Now tell me where I screwed up and what problems I will run into. Feedback is appreciated.

der_wille_zur_macht
12/06/2010, 01:43 PM
Sounds perfect.

I think the biggest potential savings is in careful matching of optics and placement, with respect to your corals. Not all people will want to go through this effort, and not all people will want to "commit" to coral placement when building a light rig, but the savings can be astronomical.

Tbduval
12/06/2010, 02:48 PM
I like the idea of less being more! I also just read your thread on being creative with LED's..

I am going to experiment with 3 LED's on each end of the tank, pointed at an angle, to make it like a sun rise/ sun set. I will also have one set in the middle. They will no optics. 2 RB 1 CW, all XR-E's. Turn on it order from left to right. I have many idea's in my head now because of that thread.

I have some mounting idea's but what do you think the best angle to simulate a sunrise/sunset would be? I am thinking around 45, give or take by looks.

Looks like some automation will be in works now to! O Boy! BTW DWZM, I want to see that build over that 60 with very LED's.

der_wille_zur_macht
12/06/2010, 02:56 PM
I think your sunrise/sunset bit sounds good, but I'd slap optics on. Half the "beauty" from having the angled LEDs comes from nice sharp shadows and the "beam of light" look, which optics enhance. You'll have to play with angle/position to get the best results, but 45 degrees is probably a good starting point. It really depends on what you want to achieve and what your rock structure looks like - do you want to light the tank up with angled light, or just create the look of sunrise "peeking" over the top of a particular live rock structure?

Check out the Typhon thread if you haven't yet, to get some ideas on automation:

http://reefcentral.com/forums/showthread.php?t=1847680

Tbduval
12/06/2010, 03:12 PM
I think the peeking over the top of a live rock structure would look pretty neat.

I will read that thread.

It looks like I will have a few more drivers than originally planned but if the new plan comes through, it should be worth it!

patmack
05/31/2011, 08:38 PM
Sooooo what ever became of this??