DIY LEDs - The write-up

[marcer]

I am not a physicist or mathematician, but this formula is not correct - there is nothing in there to account for the number of LEDs. You get the same number with this formula for 1 LED as you do 10,000 LEDs.

It looks impressive to throw out physics terms like Planck's constant, but is not a valid formula.

I made mistake by writing it into forum, I have the formula in Excel table. I did correct the formula in my next post:

PAR/W = lambda / (c . h . NA) in mol of photons.m-2.s-1

h = 6.6626E-34 (Planck's constatnt)
c = 299792458 (Speed of Light)
NA = 6.02214E23 (Avogadro's number)
lamdba = 450nm = 4.5E-7 m (wavelength of Royal Blue)

Look at http://en.wikipedia.org/wiki/Photon
you can see there that energy of Photon is E = c. h / labmda in J (W.s-1)

Then look at Example problem #2 and #3: where they calculate inverse KJ/mol as E multiplied by Avogadro number. 241 KJ/mol inverted is 4.14E-6 mol/J (same result as you get from 496.36nm / (299792458 . 6.6626E-34 . 6.02212E23 ) = 4.13E-6 mol / J = 4.13 umol/J
http://www.chemteam.info/Electrons/LightEquations2.html

If we invert the equation we get umol/J and J per one second is Watt. So in the final we get umol/W formula:

umol/W = 1/ (E . Na)= lambda / (c . h . NA)i

The result you get is how many PAR you get from 1W of Radiant flux of Royal Blue per square meter. Cree XP-E Royal Blue gives you 500mW at 350ma, at 700ma around 900mW. If you get 3.76 umol for 450nm light then from 0.9W RB radiant flux of 2.4W (700mA*3.4V) of RB led you should get 3.4 umol.

If you put 41 Leds (~100W consumption) in one square meter you get average 41x3.4 = 140 umol per square meter. Focused in smaller area e.g. into 50x25 tank (0.75 sq. m) you get 140 / 0.75 = 187 umol/s into water surface of your tank.

To me it perfectly correlates with other calculations such as umol/lux, etc.

 

[S2minute]

Ok Marcer...that`s cool but ummm holly crap. I`d love to sit and think on that for awhile but i did not bring my life jacket So, ok, let me throw this at you. Say you had a 2`x2` sq tank and you placed 56 3w leds over it, 48 were XP-Gs and XP-E RB run all at 700MA. The other 7 are XM-Ls run at 1.5 A but only on 4 hours of the day. What then, would be my PAR at the water surface? According to that formula.

 

[S2minute]

LEDs

LEDs

bump

 

[marcer]

Ok Marcer...that`s cool but ummm holly crap. I`d love to sit and think on that for awhile but i did not bring my life jacket So, ok, let me throw this at you. Say you had a 2`x2` sq tank and you placed 56 3w leds over it, 48 were XP-Gs and XP-E RB run all at 700MA. The other 7 are XM-Ls run at 1.5 A but only on 4 hours of the day. What then, would be my PAR at the water surface? According to that formula.

OK, I will suppose that you had 24 CW XP-G and 24 XP-E RB at 700ma,
a) from 24 RB - 24x0.9W= 21.6W radiant flux = 81 PAR (umol.m-2.s-1) from RB
b) from 24 XP-G CW - assumption same efficiency = same radiant flux, but 50% from RB and 50% from green/yellow/red
- 10.8W RB Radiant flux = 40 PAR (RB kind)
- 10.8W RB Radiant flux = 50 PAR (G/Y/R kind) (4.6 PAR/W at 550nm)
c) 7 XM-L 1.5A - XM-L T6 has around 540lm ay 1.5A comparing to 250lm by XP-G R5 by 0.7A (2.16 ratio)
- for RB phosporus - 50%x0.9Wx2.16 x 7 leds = 7W radiant flux = 26 PAR (RB kind)
- for yellow phosporus - 7W = 32 PAR (G/Y/R kind)

If we divide the PAR by 4 sq. water surface (converted to) and calculate 90% optics efficiency we get:
293 PAR of RB and 120 PAR of G/Y/R
plus additional 63 PAR RB and 77 PAR of G/Y/R.

Of course these value you would get if you had focused evenly all your photon just onto 2'x2 water surface. In reality some of light goes out of tank (10-20%) and in middle of tank there is usually more PAR than on sides, all depends on your design.

Check of correct calculation. 24 XP-G with 250lm focused on 2'x2' tank means 16,000 lux, comparing to 250 PAR (90/2'x2'). So the ratio is 0.01525 which is quite in line with 0.013 ration of fluorescent CW and in middle of other light sources.
http://openwetware.org/images/e/e8/Conversion_lux.pdf

 

[marcer]

double post

 

[S2minute]

Thanks Marcer ..i should have been more specific. The 56 LEDS are going over just one of 3 sections of my tank. Broken down it will be 36 XP-E RB, 12 XP-G(6 CW, 6 NW) and 7 MX-Ls,(4 CW T6, 3 WW T3), 40 degree lenses will be placed on outside edges and 60s for the rest. The XM-Ls may not have any lenses but, if i do use them, they will be 80s. According to your last post, based on what your assumed i had would have been just over 500 par at the surface?

 

[bamf25]

Stupid question. When you dim the meanwells what exactly is that doing. I was dimming them when I had my meter on the sting and it showed the current in the string was being reduced. So does the dimmer in the eln-60-48D reduce the current to achieve dimming or does it do it by another method?

 

[S2minute]

Bam, are you using a Potentiometer to dim them? The meanwells have the circuitry to dim but i think the Potentiometer is inbetween your leds and the meanwell. Unless the pot is built in..so then the dimmable MW makes it possible while the potentiometer makes it happen? Am i even close here guys n gals lol.

 

[brassmonkyballs]

Bam, are you using a Potentiometer to dim them? The meanwells have the circuitry to dim but i think the Potentiometer is inbetween your leds and the meanwell. Unless the pot is built in..so then the dimmable MW makes it possible while the potentiometer makes it happen? Am i even close here guys n gals lol.

theres a separate 10vdc power supply fed back into the driver and controlled 0-10v by the pot into a feedback circuit which in essence lowers the current to dim. 0 V = open circuit, 10v=100% current....I think the claim is for every 1v you get 10% current.....something like that....but its not a pot between the driver and the LEDs. I'm sure one of the electronics gurus can explain it better and more technically....

 

[widmer]

Stupid question. When you dim the meanwells what exactly is that doing. I was dimming them when I had my meter on the sting and it showed the current in the string was being reduced. So does the dimmer in the eln-60-48D reduce the current to achieve dimming or does it do it by another method?

Correct, the output of LEDs is considered to be limited by current because they have such flat voltage-response curves relative to other electronics.

The point is that when you have installed your potentiometer correctly, it is not in line with the LEDs, but rather it is in line with a 10v source that essentially "gives orders" to the meanwell for how much current it should let out of the gate. So yes, as you reduce the current, the LEDs will dim. You can actually know exactly how much their brightness is changing if you open up one of the data sheets and see the chart which shows you the current vs response curve.

 

[TheFishMan65]

On the D series the Pot is not between the driver and the LED. It actually causes the ELN to lessen the current through the circuitry.

[EDIT]
Read the answers above they are much better

 

[S2minute]

eh ...that`s what i was in my head but it obviously came out bassackwards. Hey Wid, your back
Now if only Willie would come back from being MIA.

 

[NasotheHutt]

I'm confused. I want to build 3 arrays of 48 LEDs for a 180. I was thinking that 4 meanwells per array would be they way to go. Then I read that too many meanwells will cause electrical problems and maybe even fire. Then I read that wiring in parallel is unadvisable due to the potential to fry all your LEDs. What's the alternative?

 

[bamf25]

^^^^

Thomas Research makes high voltage drives as opposed to high current drivers (Meanwells ). Nanotuners is the only place that sometimes carries them. The issue there is high voltage is much more dangerous for you personally. 48 volts may hurt, 150 volts can kill.

From what I have read parallel can be done safely, but you must include a fastburn 1 A fuse one each individual string. Kcress has a nice writeup on it here http://www.reefcentral.com/forums/showthread.php?t=1751598. You can also use Meanwell's hln series, which are low voltage, high current drivers which work great of parallel applications. That series can run upto 4 strands of 14 depending on the specific driver and LED picked.

 

[bamf25]

Ok I am running the leds through my Apex lite. So I am not dimming (doing that though simple pots) the leds through the Apex, just on/off. The Apex gives a amps readout on your system. When I turn the system on full blast it is not pulling the amps I expected. I have 2 strings running at 1000 mA and 3 strings running at 700 mA. In theory this means this lighting should be pulling 4.1 A, plus a little for the fan ats full power correct? Instead at full blast the Apex is telling me I am drawing about 1.5 A total for my lights. Any explaination (not that i am complaining less stress on my home circut)?

 

[brassmonkyballs]

^^^^

Thomas Research makes high voltage drives as opposed to high current drivers (Meanwells ). Nanotuners is the only place that sometimes carries them. The issue there is high voltage is much more dangerous for you personally. 48 volts may hurt, 150 volts can kill.

From what I have read parallel can be done safely, but you must include a fastburn 1 A fuse one each individual string. Kcress has a nice writeup on it here http://www.reefcentral.com/forums/showthread.php?t=1751598. You can also use Meanwell's hln series, which are low voltage, high current drivers which work great of parallel applications. That series can run upto 4 strands of 14 depending on the specific driver and LED picked.

Lots of people wiring in parallel with fuses for protection. i never heard of potential for fire with too many drivers????

There is only 1 HLN and its an 80W 48V, 1700ma. There's a PLN thats 100W.

I'm going with the HLG line. Lots of options and available. Check out the meanwell site for drivers and specs http://www.meanwell.com/webnet/search/seriessearch.html#led

 

[bamf25]

With the Meanwell drivers you also need to make sure you are getting the B versions of the drivers. The A version is generally NOT dimmable.

I had not heard of fires either with the eln's, but I have heard of harmonics issues with too many of them on one circut.

 

[NasotheHutt]

With the Meanwell drivers you also need to make sure you are getting the B versions of the drivers. The A version is generally NOT dimmable.

I had not heard of fires either with the eln's, but I have heard of harmonics issues with too many of them on one circut.

I read a post that the harmonics 'could' lead to the wires overheating and catching fire. How many Meanwells per 15A circuit would be acceptable?

 

[NasotheHutt]

Lots of people wiring in parallel with fuses for protection. i never heard of potential for fire with too many drivers????

There is only 1 HLN and its an 80W 48V, 1700ma. There's a PLN thats 100W.

I'm going with the HLG line. Lots of options and available. Check out the meanwell site for drivers and specs http://www.meanwell.com/webnet/search/seriessearch.html#led

I've looked for a build using the fuses but can't find any. Is there one documented with pictures? I looked at the link bamf25 posted and it is really heavy stuff. I'm decent with electronics, but no where near that good.

 

[widmer]

There's really no problem with running them in series at all as long as you balance the strings when you're setting it up, and add a $1 fuse to each string for good measure if the combined amperage going through your strings would be enough to fry one string should the other string become open. It has been described countless times in this thread and others.