Diy led

[michael_in_nc]

I would recommend that the LED's not be driven in "parallel" like this connection. It may very well work but it's risky. You could easily burn out some of your expensive LED's. Here's why I'm concerned:

The LED is a diode. The diode's voltage is V ~ Vt*ln( I / Is ). This means the diode's forward voltage is logarithmically related to the bias current, so a small change in voltage results in a large change in current. So, if you hook a stack of 6 LED's in parallel with 5 more of them, you could easily get a diode stack that is slightly different in forward voltage (variations in Is, the diode's semiconductor properties or parasitic resistances). This can cause a runaway condition over time. If one branch pulls slightly more current, it will get hotter. This activates the other sensitivity. The forward voltage is temperature dependent. So, the hotter is gets, the lower it's forward voltage which means it pulls even more current (more power = more heat) and so on.

I'm an electrical engineer, so I deal with this stuff on a daily basis, having said that I don't normally deal with LED's or diodes operating at this power level, so there may well be secondary effects I'm not considering. Just wanted to give you a heads up that this could be risky. One way to check what I'm saying is to put an ammeter in series with each string and see if the current is really splitting the way it should. Relatively easy check.

 

[GUILLO1]

I would recommend that the LED's not be driven in "parallel" like this connection. It may very well work but it's risky. You could easily burn out some of your expensive LED's. Here's why I'm concerned:

The LED is a diode. The diode's voltage is V ~ Vt*ln( I / Is ). This means the diode's forward voltage is logarithmically related to the bias current, so a small change in voltage results in a large change in current. So, if you hook a stack of 6 LED's in parallel with 5 more of them, you could easily get a diode stack that is slightly different in forward voltage (variations in Is, the diode's semiconductor properties or parasitic resistances). This can cause a runaway condition over time. If one branch pulls slightly more current, it will get hotter. This activates the other sensitivity. The forward voltage is temperature dependent. So, the hotter is gets, the lower it's forward voltage which means it pulls even more current (more power = more heat) and so on.

I'm an electrical engineer, so I deal with this stuff on a daily basis, having said that I don't normally deal with LED's or diodes operating at this power level, so there may well be secondary effects I'm not considering. Just wanted to give you a heads up that this could be risky. One way to check what I'm saying is to put an ammeter in series with each string and see if the current is really splitting the way it should. Relatively easy check.

how do you recommend we hook up our LEDs with this 200w power supply we already got from China?

 

[brassmonkyballs]

I would recommend that the LED's not be driven in "parallel" like this connection. It may very well work but it's risky. You could easily burn out some of your expensive LED's. Here's why I'm concerned:

The LED is a diode. The diode's voltage is V ~ Vt*ln( I / Is ). This means the diode's forward voltage is logarithmically related to the bias current, so a small change in voltage results in a large change in current. So, if you hook a stack of 6 LED's in parallel with 5 more of them, you could easily get a diode stack that is slightly different in forward voltage (variations in Is, the diode's semiconductor properties or parasitic resistances). This can cause a runaway condition over time. If one branch pulls slightly more current, it will get hotter. This activates the other sensitivity. The forward voltage is temperature dependent. So, the hotter is gets, the lower it's forward voltage which means it pulls even more current (more power = more heat) and so on.

I'm an electrical engineer, so I deal with this stuff on a daily basis, having said that I don't normally deal with LED's or diodes operating at this power level, so there may well be secondary effects I'm not considering. Just wanted to give you a heads up that this could be risky. One way to check what I'm saying is to put an ammeter in series with each string and see if the current is really splitting the way it should. Relatively easy check.

Yup.... +1...which is why each string should be fused.

here's the typical curves we're operating in

http://www.cree.com/products/pdf/XLampXP-G.pdf

Scroll down about half for forward Voltage vs current and associated thermal information if interested Though that data is for a cree LED.

It works very well and with a 1A quick blow fuse in each string its pretty safe.

 

[bigt0706]

i have yet to see any of these 85w ballast push the fedy led to 700 ma when wiring them in the 6x5 setup

 

[michael_in_nc]

Ok, so the fuse does protect it, but ideally you'd set it up not to have to deal with a broken fuse if the diodes are off a little between branches. I looked at the Cree datasheet and it looks like the forward bias voltage curve is more than just the log and extrinsic resistance, so there must be some 2nd order effects at play. Maybe at this power level my concern is not as much of an issue. I design on an IC, so it is significantly different in many respects.

I tried to look back at the thread closer to see what power source was being used, but couldn't find too much info. Anyone have a link that has info on it?

Sorry to jump in the middle. I'm just starting to look into this closer, as I'm considering an LED setup and building the electronics all myself. Also, trying to integrate it with my Apex controller. Partially for the reef tank and partially for the sport of it since I'm a geek at heart.

 

[bigt0706]

for the fedey thers not a ton of info on them , the thomas research and meanwell ballst have tons of info on them
http://www.thomasresearchproducts.com/LEDDatasheets/TRC-075 Dimming Series 01-31-11.pdf
http://www.meanwell.com/search/pln-60/pln-60-spec.pdf

 

[brassmonkyballs]

The driver being used in the thread is from FEDY China and I can't find anything other than gross operating info...ie 85W, 4.2A.

I personally use Meanwell drivers. Here's the most popular one being used should give you all the relevant info you're looking for:

http://www.meanwell.com/search/eln-60/default.htm

They have an entire line though...bigger, smaller, constant current, dimmable, etc. I appreciate your 'geekiness' but dude, buy a driver, its easier and they're cheap AND with the P version you can integrate it with a controller so long as it PWM.

Plus there's no good way to drive a large quantity of LEDs other than series-parallel connections. Think about hooking 120 LEDs up. You need a lot of drivers or one big one which isn't practical. So I guess if you can make a 360v 250W driver in a small package then you're in business. That or you're typically stuck with strings of 12-14 LEDs per driver.

Glad you joined in.

Ok, so the fuse does protect it, but ideally you'd set it up not to have to deal with a broken fuse if the diodes are off a little between branches. I looked at the Cree datasheet and it looks like the forward bias voltage curve is more than just the log and extrinsic resistance, so there must be some 2nd order effects at play. Maybe at this power level my concern is not as much of an issue. I design on an IC, so it is significantly different in many respects.

I tried to look back at the thread closer to see what power source was being used, but couldn't find too much info. Anyone have a link that has info on it?

Sorry to jump in the middle. I'm just starting to look into this closer, as I'm considering an LED setup and building the electronics all myself. Also, trying to integrate it with my Apex controller. Partially for the reef tank and partially for the sport of it since I'm a geek at heart.

 

[bigt0706]

85W constant power supply with 4.2A, 15-20V . --up to 40 LEDs can do with this power supply

Our 3w cool white led, with PCB, 220lm, 3.4-3.6V, 700mA , 18000-20000K , 120degree, Price is USD1.27/pc
3w royal blue led, 450-455nm with PCB, 120degree, 40-45lm, 3.4-3.6V, 700mA, Price is USD2.09/pc
60degree lenses for single 1w or 3w high power led. Price is 0.137/pc

this is the only info that i could get from them and its not much and we all know that u cant run 40 bulbs off the 85w ballast anywere close to 700ma

 

[michael_in_nc]

hey have an entire line though...bigger, smaller, constant current, dimmable, etc. I appreciate your 'geekiness' but dude, buy a driver, its easier and they're cheap AND with the P version you can integrate it with a controller so long as it PWM.

Yes, I know I should. I usually start out with alot of energy, then realize I'm lazy and don't actually do it. We'll see how this one turns out.

Glad you joined in.

Thanks.

for the fedey thers not a ton of info on them , the thomas research and meanwell ballst have tons of info on them

I looked at the first datasheet and it doesn't have much on it's operation. Based on what I saw, it looks like a switch mode power supply that is regulating power rather than current or voltage. That is, if you double the current, the voltage halves (Power = I*V) and this is being held constant. I think a switching supply is the easiest way to do this and hit the high efficiency numbers they are quoting (~90%). It looks like you could just stack them all in series. Instead of 5 stacks of 6 in parallel, you could have 30 in series. That would be 30 x 3.5v = 105v @ 0.7A = 83W.

In the specs it says the max power is 75W, so you'd probably be short on max current, but I also estimated the worst case of 3.5v which is high based on the Cree datasheet. The downside is higher applied voltage on the fixture, but if the grounding is handled correctly and the whole thing is on a GFI, it shouldn't be an issue.

Anyone tried this? is there another problem I'm missing?

 

[michael_in_nc]

Meanwell link has more info. It is a switcher as clearly stated in the datasheet diagram. I will look at this one closer as it has more info.

 

[bigt0706]

those are the thomas resaerch ballast and there dimable made to run led in series the fedy ballast dont push out as much and have to run them in parallel , series /if u try to run 30 led on the fedy ballast in series they will not turn on

 

[laud]

For the people who have done the DIY led's, I have a couple questions...

1) Has there been any irregular algae outbreaks since up and running?

2) Are your corals growing faster, slower or the same as your previous lighting?

3) Do certain corals react positively to the lights while others react negatively?


Thanks,
Michael

 

[bigt0706]

For the people who have done the DIY led's, I have a couple questions...

1) Has there been any irregular algae outbreaks since up and running?

2) Are your corals growing faster, slower or the same as your previous lighting?

3) Do certain corals react positively to the lights while others react negatively?


Thanks,
Michael

dont have long enough use under my belt to answer there are quite a few tanks with over year use on rc with crees but none with the fedy brand as most only had a few months max

 

[FLGirl]

go figure......the wiring i was going to go with was pretty much the way you had it BMB....but just when i thought I had my head wrapped around this project and it was just a matter of checking what the lights actually LOOK LIKE, ya'll throw another curve ball at me....not fair

It's fat tuesday and I can't burry myself in beer tomorrow when I read this post again

 

[bigt0706]

Fl girl depending on were u are in Fl if u Wana ever stop by and check out my fixture your more than welcome

 

[FLGirl]

I live in tampa. I am (hopefully) going to joseph's before the meeting on saturday to see what the LED's actually look like. I just don't want to throw away $500 to learn that I wired it wrong and have to do redo it all. Jae and Joseph have been running these LEDs on these drivers since around christmas. From what I can tell, neither of them has any problems with burn outs and current/voltage issues. We all know that Lightbulbs and/or LEDs don't last forever...I can tell you from running christmas lights that I had a problem this year with lights that were 5 years old.....There was a bad light, and a string of lights was shining brighter than the rest of the lights on that string. I knew it was wrong, and if i had cared, i would have taken the time to fix the problem. Since I didn't, the lights that were shining brighter lasted about 3 evenings before they were done...AND i LEFT THEM DONE... I consider myself fortunate that I am ordering extra lights, just in case of this problem....about 30.... and that I am a stay at home mom and will be able to tell if something like this goes wrong...cept for saturdays - those belong to mickey mouse-

I am not an electrician. I don't think this project should be so difficult it requires an electrical egineering degree to wrap your head around it. BrassMonkey Balls did a really good job of describing the current/wattage/volts/strings/series for us to understand what he was trying to get across.

I don't understand what it is u r trying to say. Please be a little more expletive in your position. I am really good with examples



PS... thx BMB, I was going to as you what quick blow fuse to use

PSS...I thought it was a short wire going from positive to positive then the last positive going to the Driver. Does each positive get a long wire and all positive wires go to the drivers?

THX

 

[bigt0706]

Um not getting were your going with that statement but was just offering to help sence some people learn better from seeing and doing then reading by no means am I trying to hit on u

 

[bigt0706]

And only one wire will go to the positive side of the ballast the rest u will jump between parallel series

This is basic wiring u can find many examples on the world wide web

 

[brassmonkyballs]

go figure......the wiring i was going to go with was pretty much the way you had it BMB....but just when i thought I had my head wrapped around this project and it was just a matter of checking what the lights actually LOOK LIKE, ya'll throw another curve ball at me....not fair

It's fat tuesday and I can't burry myself in beer tomorrow when I read this post again

What curveball? I'm not clear on your other posts either. Everything we have discussed is valid. I'm sure some of the guys in your area can give you a hand. Its not hard and you seem to have your head wrapped around it. Keep the faith!

 

[brassmonkyballs]

Yeah on that one driver you can do 30 in string @ 700ma but I think those driver are going to be about $80 so times 4 it starts adding up. I do like the looks of the TRCs though...well done. Right now they're are not available anywhere. nanotuners willl be carrying them sometime this month I think.

Yes, I know I should. I usually start out with alot of energy, then realize I'm lazy and don't actually do it. We'll see how this one turns out.



Thanks.



I looked at the first datasheet and it doesn't have much on it's operation. Based on what I saw, it looks like a switch mode power supply that is regulating power rather than current or voltage. That is, if you double the current, the voltage halves (Power = I*V) and this is being held constant. I think a switching supply is the easiest way to do this and hit the high efficiency numbers they are quoting (~90%). It looks like you could just stack them all in series. Instead of 5 stacks of 6 in parallel, you could have 30 in series. That would be 30 x 3.5v = 105v @ 0.7A = 83W.

In the specs it says the max power is 75W, so you'd probably be short on max current, but I also estimated the worst case of 3.5v which is high based on the Cree datasheet. The downside is higher applied voltage on the fixture, but if the grounding is handled correctly and the whole thing is on a GFI, it shouldn't be an issue.

Anyone tried this? is there another problem I'm missing?