DIY LEDs - The write-up

[chris023]

I remember ready a while back, don't know how long or how many pages ago, about the LED's run on the ELN-60-48D would flash after they were dimmed all the way down and the power turned off. I've looked through probably 100 pages and can't find it. Does anyone know if why this happens and was there a solution? It's happening to me.

On my first rig i did a rather poor job soldering some of the LEDs which was causing some of the LEDs to not light up and the "flash" on power down. I noticed one of the frays on one of the wires was touching the heat sink. Once fixed all issues resolved. You might have a short or something somewhere in your soldering job. I would inspect that first, but I am by no means an expert, just an guess.

You mixed the characteristics of several models.
To obtain 700mA per string you need this model: HLG-150-54
The output is 2.8A and 54V max. So each string can contain 10 (min) -17 max LEDs.
2.8Amps per string will toast your LEDs.

2.8 amps spread over 4 strings of 12 LEDs being run in parallel is 700ma per string. 4 string of 12 at 48 volts: 48V (driver voltage) / 3.3Vf (froward voltage per LED) = 14.5 LEDs per string (I am going with 12).
Sound right.

 

[TheFishMan65]

That sound right to me which is why I did not understand MosMike.

 

[CJO]

I remember ready a while back, don't know how long or how many pages ago, about the LED's run on the ELN-60-48D would flash after they were dimmed all the way down and the power turned off. I've looked through probably 100 pages and can't find it. Does anyone know if why this happens and was there a solution? It's happening to me.

Turn off the power to the LEDs before dimming them down all of the way.

CJ

 

[mgranato]

Would someone be so kind as to check my calculations to make sure I'm on the right track?

I'll be using 72 LEDs (48 XPE RB, 16 XPG CW, 8 XPG NW). I plan on putting all 72 on 1 driver (6 strings of 12). My calculations come out to a required 190 watts at .8amps. Looks like the driver I'll need is the HLG 240-48B.

I'm only interested in dimming for the acclimation period, after that, they will stay at what ever setting makes the corals happy. Would this driver afford me the opportunity to manually dim during the acclimation period?

Assuming I use the necessary fuses, resistors, AND match all strings by Vf as closely as possible, will this be a wise way to go about operating this number of emitters? My other thought would be to use one driver (HLG 150-48B) to run the 48 RBs, and one driver to run the 24 whites (ELN 60-48D). The advantage I see with this is that I could have the blues come on a little earlier than the whites. Could I efficiently run 24 XPGs on a single ELN 60-48D in 2 strings of 12?

Am I on the right track, or am I missing something critical?

 

[kcress]

Could I efficiently run 24 XPGs on a single ELN 60-48D in 2 strings of 12?

Yes. That's what I'm going to do. Do include the resistors so you can check the balance while they're running.

That will get you about 750mA per string.

 

[GEORGEDOPE]

Quote:
Originally Posted by GEORGEDOPE
I remember ready a while back, don't know how long or how many pages ago, about the LED's run on the ELN-60-48D would flash after they were dimmed all the way down and the power turned off. I've looked through probably 100 pages and can't find it. Does anyone know if why this happens and was there a solution? It's happening to me.

chris023

On my first rig i did a rather poor job soldering some of the LEDs which was causing some of the LEDs to not light up and the "flash" on power down. I noticed one of the frays on one of the wires was touching the heat sink. Once fixed all issues resolved. You might have a short or something somewhere in your soldering job. I would inspect that first, but I am by no means an expert, just an


I used solderless connectors, which was very easy and quick. All of the LEDs light up and dim evenly with the pot's I'm using. So if there was a short wouldn't some of the LED's not light up?

CJO
Turn off the power to the LEDs before dimming them down all of the way.


This will finally be connected to a Profilx and be dimmed at night and then the power turned off to the drivers. I just don't understand the flash of light about 10 seconds after the power is cut.

Here's a couple of pictures of my prototype.

 

[katchupoy]

georgedope, i believe this is normal???? i have the same case on mine, both on blue and white. only happens when dimming all the way down. Something like power being stored on capacitors. Nice rig btw.

 

[The Grim Reefer]

Thinking about throwing optics on my XRE blues, Who has the best deal on them?

 

[wfournier]

Thinking about throwing optics on my XRE blues, Who has the best deal on them?

Satistronics, not sure how the quality compares, but they look the same as many of those sold by retailers in the US.

Thought I would throw this up here, project I have been working on that is a little different than the general school of thought.


IMAG0312 by wfournier, on Flickr

 

[widmer]

It's fun to do some of these things differently, even if there's a way to say that something about them is not the very best ie efficiency. I'll be interested in how thermal management goes for you.

 

[TheFishMan65]

XR-E optics are $1 from ETG they just snap on which is nice 40, 60 and 80s.

 

[The Grim Reefer]

Thanks.

 

[TheFishMan65]

I am thinking of getting a par meter. But I found this thread
Has anyone purchased just the sensor and hooked it up to a multimeter (I think a few of us have those). Rather than $335 maybe only $139 here, but does anyone know the conversion formula? I read on the above thread multiply by 5.

Thanks

 

[MosMike]

I have a question regarding parallel strings:
Did anyone tried to simulate a string failure? Are the fuses fast enough to prevent LED degradation?

Here is an article which I have read long ago and since then I don't like parallel strings: http://www.ledsmagazine.com/features/6/2/2

 

[TheFishMan65]

I talked with Jeff Bingham (Apogee technical support). So yes you too can have a PAR meter for $139 dollars. The output is in millivolts and for every millivolt smultiply by 5 to get umol m^2 /s. If you order 20 they give you a 20% discount.

 

[The Grim Reefer]

I talked with Jeff Bingham (Apogee technical support). So yes you too can have a PAR meter for $139 dollars. The output is in millivolts and for every millivolt smultiply by 5 to get umol m^2 /s. If you order 20 they give you a 20% discount.

Just make sure you aren't using a cheap meter. Any error by the meter is multiplied by a factor of 5 so a 5 percent error in your 8 dollar Walmart meter will reflect a 25% error in actual PAR. The Yellow Box meter you get with their 350.00 deal isn't a real expensive unit but it is calibrated by the factory to have accurate readings. I just checked mine by the sun calibration on their website this morning. After at least 5 years since a calibration it's off by just over 10%, Pretty impressive.

 

[TheFishMan65]

I have a reasonable meter at home that I think will work. If not I will borrow a yellow box meter from the office. I will probably try several just to see. Any idea if it is the sensor or the yellow box that is off. Of do you have a full unit?

 

[The Grim Reefer]

So my light was originally set up with 52 XRE Royal Blue, 8 Warm Whites, 6 Cool Whites and 18 XPG cool whites, all in series.

If there was ever any doubt about how much more efficient the XPG is, I removed 9 of them and added 4 XRE Neutral Whites and 5 420 Violets. PAR dropped from 140 to 91. YING!!!

Also took a reading after they had been on about 5 minutes and checked again after about an hour. With the heatsink at full temperature PAR increased from 88~89 to 90~92

 

[The Grim Reefer]

I have a reasonable meter at home that I think will work. If not I will borrow a yellow box meter from the office. I will probably try several just to see. Any idea if it is the sensor or the yellow box that is off. Of do you have a full unit?

Mine is a complete unit they made. The meter isn't anything special but the difference is they calibrate it at the factory. Once you get yours hooked up they have a procedure on their website for checking the calibration using the sun. You could take readings at a few different times during the day to see how accurate your setup is, say 10 2 and 4. The site takes into account time of day, day of year, temp, humidity, elevation and location to come up with a theoretical reading you should get.

 

[TheFishMan65]

I have seen that sight, but thanks. It sound like the probes are calibrated somehow also since it is a fixed value per millievolt.