DIY LEDs - The write-up

[Names Brucey]

I can get them cheaper than this dealer. There is a seller on flee bay that has them for about $3 shipped to your door. I already run XR-E on one of my tanks but I am looking to do a DIY LED for my 180 and cost goes way up so thought these might be an economical alternative.

 

[Names Brucey]

The datasheet shows the white color to have a flux of 128 lumens "typical" at 700mA and a Vf of 3.8. That's 2.66w, so 48 lumens/watt, or about half what the typical LEDs we use will do. So they're pretty inefficient.

I've never heard of "prolight" so I wouldn't be surprised if they were just repackaging someone else's emitters - it kind of looks like a Luxeon III and has similar specs.

Solves that one. Thanks. Will just have to wait out the next group buy on the smaller reef site.

 

[sammy113]

I have my set up almost finished but want to make sure of one thing. With the fan, do most people blow air on to the heatsink or have it pulling the hot air out?

Thanks in advance.

-Dave

Fans should be blowing to the heatsink so the air runs through the length of the fins

 

[ReefEnabler]

You must measure the current directly or put a fixed resistor in series and measure the voltage drop across it - which will then give the actual current in the string.

I will grab some resistors on the way home to test it correctly.

Can you comment on whether youd push a 24v supply to 26v if it helps get the buckpucks to 700ma? I still suspect theyre bein under driven.

Thanks again

 

[der_wille_zur_macht]

If you're talking about one of the common mpja power supplies (from POTRANS or other manufacturers) with the voltage trim pot, then yes - by all means, adjust the voltage up until you see the current on the output side that you think you should be getting. Set up your multimeter to measure current on the output side and turn the voltage pot up. If current increases, the buckpuck is being under-driven, so keep turning it up until current stops increasing (in theory, this should be at the point when the current is equal to the buckpuck's rated current). There's really no point to going beyond that point, since the buckpuck will just get slightly less efficient if it's fed with more voltage than it needs.

 

[der_wille_zur_macht]

Edit - weird 11 minute delay double post. . .

 

[TheFishMan65]

You can't fool us. You are just trying to reach 10,000.:lol2: I double posted the other day, because either my server had issue or theirs (since I had all kinds of problems with them that day - I am thinking them). It happens and you know somehow we manage to survive.

 

[kcress]

RE: If you've turned the current all the way up on the BPs and it just tops out you are under driving as Der said. You don't want anymore voltage than you need delivered to the BPs.

 

[ReefEnabler]

You must measure the current directly or put a fixed resistor in series and measure the voltage drop across it - which will then give the actual current in the string.

I don't have an ammeter that can handle the current for a direct read.

The Buckpuck datasheet says to use a 0.1ohm, 1% resistor and take the mV reading and multiply by 10 to get the current in mA.


I went to radioshack looking for 0.1ohm, 1% resistors to test as the datasheet describes, but of course the closest they had was 1ohm 10%

Well they did have a 5% version but you have to buy a $13 assortment pack to get it. So I got a pair of the 1ohm 10% resistors.

I assume that if I just take a mV reading across it, that should be the current in mA? Will that be fine as long as I don't adjust it over ~630mA to cover the 10% innacurracy of the resistor?

Or do you think a multimeter can tell me exactly what the resistance is of the 1ohm 10% resistor?

 

[Obi-dad]

I did the same thing yesterday, bought the shack pack of 1 ohm 10% resistors and measured them, one was 1.0 ohms on my DMM, the other was 0.9 - but who knows how accurate my DMM is anyway. For my purposes it puts me in the ballpark. Even if I am off by 10% low, that would be 770ma instead of 700, which is fine since I won't be driving them near the current limit anyway.

 

[ReefEnabler]

Are you using a 1000ma buckpuck??

I just realized that I shouldn't really be able to go above 700ma by turning the voltage up over 25v with my 700ma pucks; like kcress said I think I just need to make sure not to keep turning the voltage up after I see the buckpuck current level off.

 

[stugray]

ReefEnabler,

All standard meters can handle 1-10A.

you should be able to measure current thru the string directly through the DMM using an "in-line harness".

That's one reason I use those TYCO connectors on all of my strings & drivers.
I just un-hook the driver to string connection, then plug in my harness + DMM.

My harness has two banana plugs on it and they plug into my DMM.

I turn on the string, adjust the current, power down, remove harness, plug back in & fire up.

Stu

 

[ReefEnabler]

Odd, mine specifically says 200ma max. It's a rather small one from radioshack where the leads are attached, not plugin. Has a little lid that closes with the test leads all inside.

I had a bigger one at one point that would cut it, but the onboard connectors for the testing leads got all jacked up beyond repair. Maybe not all repair, but beyond the repair I feel like doing anytime soon...

I'll have to do the resistor test for now, I'll report back a bit later. Right now the wife is saying "spend time with meeeeeee!!"

 

[kcress]

RE: It matters NOT at all what tolerance (10%) resistor you have. It could be 38.7%! Just measure it to at least one decimal point and then use that number for the resistance.

I = V/R

 

[The Grim Reefer]

Has anyone tried attaching the LED's using plastic rivets?

 

[widmer]

Not that I'm aware of. Someone (Stu?) bought a bunch of nylon screws to attach his but I would be really wary of the plastic allowing the LED to loosen up over time from all the heat/cool cycles, and a tiny gap being created between the star and the heatsink, allowing it to overheat and fry...

 

[The Grim Reefer]

Not that I'm aware of. Someone (Stu?) bought a bunch of nylon screws to attach his but I would be really wary of the plastic allowing the LED to loosen up over time from all the heat/cool cycles, and a tiny gap being created between the star and the heatsink, allowing it to overheat and fry...

Yeah but I saw some that were made to withstand heat. Might just try the thermal adhesive but was thinking that might be a bear if you even needed to replace them.

 

[ReefEnabler]

I tested both 1ohm resistors and the readings drifted between 0.9 and 0.11 ohm.

I measured the LED string with the 1ohm resistor in series with the LED+ line.

Off the bat it was reading ~620mv with the power supply at 25v. Nuding the supply a little higher dropped the reading. By decreasing the supply voltage to 23.7v I saw the reading go up to 643mv.

So is it possible I was already overdriving the pucks at 25v? Seems odd since they say 2v buffer and my 4 LEDs supposedly would eat 23.5v @700ma.

 

[widmer]

Yeah but I saw some that were made to withstand heat. Might just try the thermal adhesive but was thinking that might be a bear if you even needed to replace them.

Don't get me wrong, plenty of folks are drilling the heatsink and then screwing standard screws into it to clamp down the LEDs. Some are using nylon washers between the screw and LED, others aren't.

From my limited experience, I think it's a pretty easy thing to drill a hole into the aluminum heatsink which is just barely smaller diameter than the screw's threads, then simply screw a screw in and it will hold very tightly.

The only key really is to use a tiny bit to drill a pilot hole first or use some other hole-beginning mechanism so that your holes are precise.

 

[kcress]

RE; What is probably happening is that the voltage is so close to that required by the BP that it's unstable. So it weaves around a bit. The higher voltage is probably the truer operating point.