DIY LEDs - The write-up

[stugray]

"I think you need a minimum of 3 LEDs."

I have two systems up & running for months with just 2 XR-Es on a ELN-60-48

Never tried just one though.

Stu

 

[dnahas]

... I'm much more worried that my heater will be running nostop than I am about needing a chiller.

The best part of my install is the fact that I have changed my system approach from doing what I can to reduce heat and temp swings to keeping the heat without running the heater 24/7. After the swap the temp swings are 0.5 deg, with the halides it was 2-2.5. I am sure the .5 deg swing I have now is due to room air temp, poor AC in the house. FYI 10" above water

 

[der_wille_zur_macht]

Thanks, that's clearing it up. I did see that 24-48V range on the spec sheet, but I wasnt certain what it meant. I think I'll play it safe and run more than just a couple. Those LEDs are too expensive to burn them up!

Above, you mentioned running 8 LEDs on the ELN-60-48. That would be just fine, but isn't the "most efficient" use of the ELN since you're running well below capacity - if you really only need 8 LEDs, you might want to look into one of the smaller-capacity meanwells.

 

[cyrusthevirus]

I'm no expert in the field, but I would be really paranoid that welding would result in warping/distortion of the surface - and since surface/surface contact is so critical for heat transfer, that would be a bad thing.

That is a good point. We want to make sure the star has full contact with the heatsink.

 

[dnahas]

what about T SLOTS instead of u channel? Looks like more surface area and mounting options. I believe someone here picked up some free, lucky. google T SLOTS - Extruded aluminum structural building material

 

[der_wille_zur_macht]

The best part of my install is the fact that I have changed my system approach from doing what I can to reduce heat and temp swings to keeping the heat without running the heater 24/7. After the swap the temp swings are 0.5 deg, with the halides it was 2-2.5. I am sure the .5 deg swing I have now is due to room air temp, poor AC in the house. FYI 10" above water

I'm probably going to end up covering the sump and building some sort of (probably manual) control into the duct venting from above the tank such that I can direct the hot air from the LEDs over the water in the sump to help with heating when needed.

 

[dnahas]

duct venting from above the tank such that I can direct the hot air from the LEDs over the water in the sump to help with heating when needed.

You know IMO it could be beneficial to use tank water to liquid cool the heat sinks. Welding some aluminum tubing between the heat sink fins or other side of the star would cool it. It also sounds cooler, pun intended. That's also ignoring any corrosion or other bad factors of piping tank water through aluminum.

I would like an automated lid to close at night to keep the heat in. This system has no sump, and I don't want to add power consumption and noise via cooling fans. I am sure when summer hits I'll have no choice.

 

[der_wille_zur_macht]

You'd have to put a titanium heat exchanger in the sump/tank and use a closed loop between the LEDs and that exchanger. And, you most certainly wouldn't want all of the LED's heat in the sump all of the time, so you'd need a second exchanger outside the tank and some sort of control system to decide when to put heat in the sump vs. exhaust it through the other exchanger.

And at other times the LEDs wouldn't be providing enough heat (i.e. at night when they're off) so you'd have to have a conventional heater on the tank, too.

IMHO way more cost/effort/risk than it's probably worth, but it WOULD be really cool.

 

[crvz]

Above, you mentioned running 8 LEDs on the ELN-60-48. That would be just fine, but isn't the "most efficient" use of the ELN since you're running well below capacity - if you really only need 8 LEDs, you might want to look into one of the smaller-capacity meanwells.

That's a good point, I hadnt really thought about it. Right now I'm just goofing around with them, so I don't mind having more capacity for my "development" build.

To make my head hurt less, let me ask a question. Yesterday I wired up the meanwell, and at any dimming input voltage over 0V (had a variable wall wart) it gave a 48V output with no load. But what's been pointed out is that, with a load put across the output, the voltage will adjust so that the LEDs won't be in risk (considering some minimal output voltage over ~7V, or 24V if you believe the spec). Is that accurate, or am I missing something?

 

[TheFishMan65]

I believe that is correct. Even with the dimming voltage at minimum is was trying to drive the LEDs with some low current (even .01 ma), but because there is not load it tried to increase the voltage to force the current across the load. It increased the voltage to the maximum allowed and could not do it and stopped at 48 volts.

 

[der_wille_zur_macht]

Fishman is correct. With no load, the current feedback circuitry never comes into play, so the voltage spikes to the point that the voltage limiting circuit is controlling the output - and in the case of the ELNs, that means 48v (unless you've messed with the voltage trimpot).

This can lead to dangerous situations - if you were to supply power to a meanwell with no load on it and suddenly attach a load (say, a single LED) the load would most certainly fry because it would see the full 48v for a fraction of a second before the driver had a chance to respond and cut the voltage down.

 

[fstar25]

Hey DWZM, (or anybody that feels like commenting)

I should be getting all of my equipment for my LED setup this week. I am going to be using 2 of the meanwell 60-48d and I was going to use the board diagram you showed in another thread for a manual dimmer control box. But I have decided I want to go the easy route. Would the following components be sufficient and did I leave anything out?

Thank you

http://www.dimensionengineering.com/VregBreakout.htm
http://www.dimensionengineering.com/DE-SWADJ.htm
http://www.radioshack.com/product/index.jsp?productId=2062281&clickid=wish_list
http://www.radioshack.com/product/index.jsp?productId=3807944

 

[der_wille_zur_macht]

Nothing missing, but you could probably find some of that stuff cheaper - the wall wart, for sure. $25 is highway robbery for a 12v wall wart.

Also, it might make sense to add a potentiometer on the output of that regulator. Then, you could adjust the regulator to provide 10v, and use the pot to do the actual dimming. If you're using a 12v wall wart and that switching regulator is the only adjustment point you have, it would be easy to turn the trimpot on it "too far" and send more than 10v to the meanwell(s).

 

[fstar25]

DWZM, holy crap that is a great idea. And I will get the parts cheaper that was just the easiest reference I had.

That way I can put it all in a project box with the pots controlling the actual dimming effect. How exactly would I wire the regulator to the pot? Vout from the reg to the pot. Then ground from the reg to the meanwell and Vout from the pot to the meanwell?

Thanks again for your time.

 

[L98-Z]

What was determined to be the easiest way to attach some computer fans to the heatsinkusa heat sinks?

 

[der_wille_zur_macht]

If it were me, I'd either use very long screws with standoffs, or poke some holes in a few fins and ziptie the fans on.

 

[L98-Z]

If it were me, I'd either use very long screws with standoffs, or poke some holes in a few fins and ziptie the fans on.

Sounds good to me. I just figured someone may have already determined an "ideal" route. Seems like I always think of one way to do it... do it... then come back to find a better way of going about it. I think I'll go the route of zipties, that seems cheap and easy.

 

[der_wille_zur_macht]

Some would argue that the zipties will deteriorate over time and eventually get brittle enough to break, but IME they fail at about the same rate as the fans.

 

[TheFishMan65]

I think zip ties would be fine, but one other thing to think about. The holes will be sharp. Will fan vibration cause them to cut through. If so would tight or loose be better. Question that I don't have answers to, but I like to make people think. I think tight is better (no vibrational noise), but each slight vibration will wear. So in 3 years will you get back to us and tell us did the fans die or did the zip ties fail.

 

[der_wille_zur_macht]

Any time a drill bit touches metal under my guidance, the very next thing that happens is deburring via file or countersink. I don't think the fans we're using are consuming enough power to create enough vibration to cause a problem, but if someone were worried, just drill the hole oversize and put a gasket in it.

While we're on the subject of attaching things to the heatsink, something that I rarely see when people post build photos is stress relief for the wiring between the driver and the first/last LED. The tiniest little jiggle or bump against a long wire coming off a star MCPCB can cause the pad to rip off the star. IMHO the wiring should be fixed to the heatsink itself to relieve stress, and anywhere it passes through a small hole in wall or other surface (i.e. if it's entering a project box that your drivers are in).