Screened porch build, LED waterproofing and advice?

[Illum]

Since dad is planning sometime next year to modify the gutters, I designed everything to fit together with breakouts as well as mounting them to the beams the way one would hang a picture. All I need is a Philips head and I can take apart everything and put them in rudimentary piles

Continued from yesterday
the last of the soldering


As mentioned yesterday, I was pondering the use of cellophane wrap.





Its remarkable ability to cling to the aluminum and acting as one single layer of plastic when overlapped had me optimistic...until

failure!


The LED still worked, but that made me change my mind about using it as a moisture barrier, and I went with the cheapest and readily available supply in my inventory room: tape

I used boxing tape to create a weather guard awhile allowing some air convection to access the LED. To prevent the tape from yanking the dome off, here's what I did

cut a 1" thick piece of tape and lay it sticky side up


Cut a ~3" thick piece of tape and adhere it with the 1" at the middle


Place the "assembly" over the LED, and only allow the sticky parts to contact aluminum


the heatsink design will not reach the temperature sufficient to melt the tape into a sticky mess, if it does, at minimum none of it will touch the LED





For the sheared dome LED, I decided to tape it down and seal it:shrug:




More pics will continue on the next post, for now my computer is still resizing them

 

[Illum]

The central junction box had me working for a couple hours. In design I specified 18 gauge and had simulated the wiring using spare wires..but when dad came in and said we have weather resistant wires in the garage I went with it, but as I started splicing I realized it wasn't such a good decision

change in specs
Stranded copper -> Solid copper
18 Gauge [AWG] -> 15 Gauge [AWG]

So I had to open up the holes a little bigger, and force the wires around to get it to fit

What I began with


the new wire couldn't fit the silicone tubes, so that plan was scrubbed


Spliced, I bent the ground wire down and crimped it backwards into the wire. this made a bump on the wire surface that, when forced into the enclosure, stays in the enclosure. Should one assembly fall from the mounting place, it'll act as a strain relief from directly pulling on the wires in the breakout


The jumpers on the terminal block is 18 AWG, the copper alone on this new wire is the same thickness as the 18AWG with the insulator on




Not wanting to put new holes where I expect water to hit, I taped the terminal block to the enclosure


I couldn't wedge the wire beside the screw so I had to take the screw out and stick it in. It cross threaded a couple times but it went in.




Solder the switch on, my 25W iron took awhile to heat the new wire up, I think I need a new tip or a retinning:shrug:

Soldered joint

Insulate joint

As mentioned before I want the parts to be removable, so I sticky taped the enclosure to an angle, taped it down for extra hold. this tape is not the 3M variety and you can't tear it apart with giving it a little cut to start



After the connections were tested and function, the unit is sealed and hung. Here we are, the wire leading off to the right is to the xitanium, the one leading to the overhang is to the receptacle, I decided not to fuse the circuit, the xitanium draws 200ma max, I don't know if 125V .25A fuses exist


Wiring the xitanium box is in the next post:thumbsup:

 

[moonfish]

The downside of that approach is that when it gets moisture in there, it can't get out.

I plan to give some a coat of epoxy. There's a big thread on here speculating about it. It didn't look like there was any proof about it working or not.

 

[Illum]

that's true that's something I'm still trying to figure out. I wonder if I leave it on 24/7 would it retain moisture?:shrug:

The enclosure with the xitanium, what I like about the enclosure is not that its big enough to stuff in a couple xitaniums...but that's certainly a good thing to have.


Note how close the length is of the xitanium with the length between the standoffs, its a snug fit




I've added a 4 block terminal bar here that will receive incoming AC and departing DC, this allows me to isolate the AC input or isolate the DC end [good for breakdowns]. I changed the configuration after the picture, the block diagram from top to bottom is Line, Neutral, Positive, Negative


This also allows me to swap out xitaniums [if 700ma was too strong and 350ma is ideal] without having to manipulate the connections, just unscrew the AC end and DC end then do anything on the floor instead of standing up and resting my head on the gutter

Hanging grapes...

 

[Illum]

Results, it was better than I expected, I did not want the wall to be screaming bright but I was afraid it might be too dim. It came out balanced pretty well, but I might have to move some of the hanging plants for the added effect

Evening


Dusk, Since the screened porch is flanked by a wall of plants, there's little to no need to zip up the curtains, the daylight that goes in the windows saved us alot of money for lighting


Result


The picture is dimmer than it appears, NW had the best color rendition over WW and CW. But it illuminated the ground, the wall, and vegetation with adequate intensity. Since all fixtures are movable, I might adjust the placements in the future, for now I'm done

 

[blasterman]

The problem with clear tape over the emitters is you'll eventually get moisture / condensation in there, and it will have a hard time going anywhere then. This might actually cause the emitters to fall apart faster.

For general protection, a cheap optic over the emitter along with a generous spread of expoxy around the base works good. Acrylic and hardened epoxy are among the toughest materials in a harsh environment.

I hate to give away ideas, but one thing I'm working on you might want to try for outside duties is this: get a couple pieces of 3/4 inch wood strips from a hobby or home supply store. Cut the strips the length of your alu bar, and glue them to either side of the emitters along the alu bar length. Now, get a cut piece of clear plexi, or, acrylic diffuser from a home supply store meant for fluorescent ceiling lights and attach it to the wood strips. This way your LEDs are inside a nice housing. I use this method for my reef tank light because bare LEDs don't last long under the salt splash. I guarantee it will take plenty of abuse outdoors.

 

[Ken_McE]

Illlum, well done.

 

[Illum]

Given the success of this project came the failure of meeting to budget and deadlines. I overshot this by two weekends so Part B has been placed on hold indefinitely. I'll be away on a family vacation by the 16th.

The problem with clear tape over the emitters is you'll eventually get moisture / condensation in there, and it will have a hard time going anywhere then. This might actually cause the emitters to fall apart faster.

For a cheap fix would cramming the area below the tape with silica dessicant help at all?

I understand epoxy might be the best answer for everything, but I failed to understand it prior to buying stars.:shakehead I could've used emitters
where potting with epoxy would have been very easy and will save me some $10.

I have been looking into optics, specifically
http://ledsupply.com/10210.php
But I'm not sure whether or not that has the effect I'm looking for

I have a couple of these: http://ledsupply.com/247-10-lens.php
Bought for the bike light project, which works well for a headlight but not exactly what I'm looking for in this build

For general protection, a cheap optic over the emitter along with a generous spread of epoxy around the base works good. Acrylic and hardened epoxy are among the toughest materials in a harsh environment.

Will this work?
http://tinyurl.com/y923n6g

Whenever people speak of epoxy in relation with LEDs I think of stuff like Norland 61...and that stuff isn't cheap:candle:

I hate to give away ideas, but one thing I'm working on you might want to try for outside duties is this: get a couple pieces of 3/4 inch wood strips from a hobby or home supply store. Cut the strips the length of your alu bar, and glue them to either side of the emitters along the alu bar length. Now, get a cut piece of clear plexi, or, acrylic diffuser from a home supply store meant for fluorescent ceiling lights and attach it to the wood strips. This way your LEDs are inside a nice housing. I use this method for my reef tank light because bare LEDs don't last long under the salt splash. I guarantee it will take plenty of abuse outdoors.

I'm sure it would, but then the light will be directed downwards, perfect for a walkway or a fishtank, but that would post a conflict with what the purpose of this design

This porch is like a black hole at night, with dark walls lined with dark vegetation. I mounted the bars on the center of the 2x6 beam to faciliate light downwards but not outwards where it would be distract anyone sitting on the deck. While lighting up the vegetation I also wanted light to light the wall as far up as possible, which was one of the main reasons I mounted the LEDs on the bottom of an aluminum angle rather than on the inside of an aluminum channel.

If I wanted to make a fish tank light I'd use a 3/4"-1/2" wide 1/16" thick channel where LED stars can easy "drop" into the channel with a window and end caps fitted over it. Depending on mounting methods one could buy some long finned heatsinks made for passive cooling and epoxy them to the top or side of the channel

I hear what your saying though. And something similar was designed in the past as well. It concerns the use of an aluminum gutter and very amateurish metal fabrication. I apologize for the messy and ill-proportional picture drawn in mspaint. If my drafting license didn't expire I'd come up with a technical drawing


What I came down with is:

1. If I was to seal it up using some medium or method I have not yet devised, moisture will find its way into it and every time the light is turns off it'll condense to water and stay in the medium and collect on the lowest point, where happens to be the LED dome in the current setup.
2. If I left it open for air convection flow, then the LED will contact moisture very easily and may perish through electrolytic corrosion and sequential oxidation. But it is also self drying as each time the fixture is turned on the heat produced from the LED will heat the metal to 85-90F and it will slowly "cook away" the water condensate.
3. If I did nothing about it by spring the monsoon season will trip a couple breakers, fry a couple LEDs, and laugh at me in the face for building such an ill-protected contraption in the first place
4. If I built a sophisticated air cooled/humidity controlled environment for them where I can store my stamps, cameras, and photos in its air reservoir cell I will have overshot my budget so many times over that it'll be cost effective for me to buy low dome outdoor HIDs and toss the whole LED assembly in the trash

So, my proposal for now, until I can achieve the 5th alternative, I've wrapped all the LEDs in tape [not sticking to the dome except the failed one] but left the sides open and leave them on running 24/7, keeping the metal warm and the LEDs dry.

Its going to be the final exam season for the next two weeks so I don't have time this year to go back and do any substantial modification. However, I can still sketch up new ideas for it.

 

[lonesouth]

Any updates?

 

[usLEDsupply]

you might try something like "Electro-Seal" it comes in a spray can and goes on like a thick contact cement but dries quickly and is quite rubbery so can handle expansion, it says it will work even if the item is going to be used under water and while it looks like it would work well i haven't tried dunking my cell phone yet to see

i have used it on some of my DMX Decoder boards that were to be outside somewhat unprotected and it seems like really good stuff, it ranges from $8-$15/can so it is not expensive. i have even sprayed the circuit boards including the front screen "not the main one tho" on my cell phone to help make it more water/shock proof.

i would still tape off the center of the dome as it is clear but not as clear as glass and you don't want to loose any light output.

it would be a lot easier than making a box or cover for them and you wouldn't have to worry about condensation or moisture buildup

they also make a high temp version but i have not tried it nor do i think you would need it for this

 

[Illum]

well, its still going strong.
We had a couple torrential downpours so far with no issues. Instead of turning it on and off I decided I'd leave it running 24/7. Even when its 40 something out the heatsinks are quite warm throughout the night.

No condensation has yet to be observed to accumulate under the tape seal.


However, I have one emerging issue....one of the CFLs from the wall fixture burn out prematurely, upon inspection the globe is half filled with water possibly resulting from the rainwater trickling down the power line I installed adjacent to it

By dad's request I might go about modifying the wiring and possibly create a channel to duct the rainwater away. For now I'm stuck on another project...this time for an astronomer friend who needed to illuminate a remote 10ft by 10ft work deck in red LEDs...preferrably dimmable and 12V input. I'm currently tapping out the bugs of running close to 300 Piranha red LEDs on a single Luxdrive boostpuck.

 

[purduephotog]

There is some blue RTV compound you can get at auto parts stores- I used to use it to pot CPUs back when the 'peltier' craze was going on- you could coat the back of a Slocket adapter and put neoprene around it to help insulate. No issues booting, which meant no conductivity- and no water intrusion over several thermal cycles.

Anyway- what I was getting at was I really like the ideas. I wonder if theres an easy way to bore out the inside of the copper caps (or just go to 1" copper) to try and smooth out the surface. Wouldn't be but two seconds for someone with a functional lathe or end mill (do they make 3/4" end mill bits? Hrmm...)

You could easily pot a piece of lexan or glass (no plexi!) over the end of the cap, and attached to the aluminum bracket you'd be in pretty good shape for conducting the heat away.

Very interesting project.

 

[Illum]

Since its completion back in late November of 2009 the assembly has been on every night from sunset to sunrise...on occiasion left on to dry out the metal if a strorm passed at night.

To date it has ran a total of 220days, assuming 12 hours/day the assembly has operated without an issue from any part for over 2600 hours
The house had new windows put in as the old ones were some 30 years old and the insulation is essentially non-insulating. The new ones has dual reflection coatings [two layers of glass, argon filled] and these LEDs provided just enough light to create a one way privacy window to the master bedroom at night.

Mom noticed that the plants under the LEDs seem to grow better, despite the lack of evidence that such LEDs has what it takes to stimulate chlorophyll the plants do see to like it, it changed mom's perspective on residential LEDs and so it did I.


There were issues though regarding the LEDs attracting bugs, being lukewarm all the time it provided a haven for spiders and flies who likes to congregate between the tape shield I placed on the emitters...it takes one bug to dehydrate and die at the entrance to cause the entire assembly to die under it. Out of 6 LEDs 2 had this fate, I had to cut it open and rub down the LED to allow light to come out of them.

Since the intensity hasn't appeared to dim after over 2600 hours, and that the heatsink is hardly warm, I'm thinking of going up one tier higher and use a Xitanium off-line 1000ma driver. Last I recall CREE 7090 XRE Datasheet indicated the absolute maximum foward current as 1000ma, but the 25W Xitanium's 1050ma...I'm not sure if I'm willing to lose $50 worth of LEDs for kicks.

 

[Illum]

Well, a year and a half down the road running 24/7, over half of the LEDs in one string failed [close circuit] and the other half the xitanium went bottoms up.

The XREs were so fogged up by the humidity and the resultant corrosion that they were actually becoming discernably dimmer.

the dead Xitanium 700 on top, notice two resistors are missing, a transistor blew up, and several diodes have failed. The cause was traced to a water leak and a spider egg sack that was nested [of all places] under the output connectors.

Given the condition of the xitanium [melted plastic, a gut wrenching smell of burnt electronics] I was blessed to know the house didn't catch fire.

The existing system has been decomissioned, a future system shall employ a dusk to dawn switch and a modular system operating on a 12V rail. This way failed modules can be removed and replaced without shutting down the entire system. LEDs will be sealed with dielectric spray and be configured to increase the heatsink temperature as high as safely possible [to ward off the critters that likes to sleep on it, such as frogs]. 12V as opposed to other voltages would allow me to use different LEDs up and down the stream, plus the ease of using tri-XPG stars vis current limiting resistor.
I wanted to create waterproof modules that have integrated buckpucks, but thermal dissipation has long been the main obstacle. I could use die-cast aluminum enclosures, with a hole opened for the LEDs, but the material cost of that idea easily overrules its benefits.

 

[Steve K]

I wanted to create waterproof modules that have integrated buckpucks, but thermal dissipation has long been the main obstacle. I could use die-cast aluminum enclosures, with a hole opened for the LEDs, but the material cost of that idea easily overrules its benefits.

thanks for the follow-up report! Always good to see how things survive the environment.

I'm not familiar with the moisture resistance of the buckpucks. Would it be feasible to epoxy it to some aluminum extrusion and then spray it with conformal coat to provide the needed moisture resistance? I'm assuming that the proper aluminum extrusion, such as C channel, could keep rain off the buckpuck, and the conformal coat would resist the humidity and condensation.

regards,
Steve K.

edit: conformal coat is the standard material used in the electronics industry, but I've had good results with Plasti-dip on my bicycle electronics. In either case, use multiple light coats to get better coverage and less trouble with tiny pinholes letting moisture in.

 

[mds82]

for sealing the LED's, LEDSupply sells a product called LEDSeal - http://www.ledsupply.com/9013-07.php Its meant for outdoor products liek this. Also their PuckPucks are fully waterproof since they are fully encapsulated

 

[Illum]

for sealing the LED's, LEDSupply sells a product called LEDSeal - http://www.ledsupply.com/9013-07.php Its meant for outdoor products liek this. Also their PuckPucks are fully waterproof since they are fully encapsulated

Hello,
I've been testing Ducky's Electro Seal spray-on on a number of outdoor mounted LEDs. But since then it has yet to go through a downpour while its on. :thinking:
this stuff: http://www.carolinaboatoutfitters.c...-seal-rubber-film-coating-ducky-d1005-p-76490

you are correct, the wired buckpucks are confomally coated which makes then essentially "dunkable"
I have bought several ABS enclosures with clear tops to test the feasibility of outdoor fixtures.

I used one of make a fishing light runnig 8 green Lux1s with optics and a boostpuck off 12V:twothumbs
http://i43.tinypic.com/6frz1g.jpg
It ran well until the power outlet joint got corroded, water flooded in, and ruined everything on the inside.

I got the idea to drill 1" holes on the side, snake a PVC conduit in and epoxy around the joint to seal the plastic. The problem now is the air is completely still inside it, which insulates the heatsink:sick2:, and it gets to a point where its so hot that the ABS starts to melt.:green: Only way I have found to remedy the problem would be to force air through the conduit from one end using a blow-fan and have the other end covered with mesh as a exhaust. But for every piece of hardware that blocks the line of sight airflow is reduced and every bend it enounters the air flow is effectively eliminated.

I have converted from force air heatsinks to passive heatsinks and currently found that a tri-rebel star driven at 700ma works well with this TO-3 heatsink. Its hot but not so hot I cannot comfortably touch it.
http://www.jameco.com/Jameco/Products/ProdDS/16513.pdf

Dale 2 ohm 1% 5w aluminum clad wirewounds level the rebels at around 680ma on 12V

My plan now is to see whether or not I can find six die-cast aluminum enclosures without costing me an arm and a leg or a size so large I cannot effectively use it and house the heatsink in it [or the LED itself] and cutting a 20mm hold on the lid for optics.

Current tests tell me for the current I'm working with JST connectors are sufficient. Installation of modules will use heatshrink over the connectors, which can be peeled away to expose them if modules needs to be swapped out.

The dusk to dawn sensor is currently wired to a receptacle [actually, two receptacles] which the 12V SMPS will plug into. The SMPS will be rated for 10A, but I speculate only 5.5A will be used.
4x Rebel tri-stars at 700ma, ~3A
2x XPG tri-stars at 1000ma, ~2A

the XPG tri-stars will be connected on one of these: :thumbsup:
http://www.allelectronics.com/make-a-store/item/RC-10/KEYCHAIN-REMOTE-CONTROL-12VDC-6-AMP/1.html

Since the patio is facing the back yard, these will act as on demand walkway lights that will illuminate the rear of our property between the garage and utility shed. Field tests gave me excellent cover and sufficient to identify trespassers without them being able to look into the windows. I can activate the reciever anywhere in the house using the keychain. My SMPS allows me to put at least 5A of LEDs on that remote control. The poor fellow that stumbled into our yard and accidentally kicked a bucket or a pot will get one hell of a surprise when I put those XMLs in...

 

[HotWire]

What a wonderful project! I like it. The last time I was in a big-box hardware store I saw some fairly inexpensive waterproof LED fixtures designed for outdoor use. They would have the advantage of being a turnkey installation. Just a thought.....

 

[Illum]

Here much of the waterproof LED fixtures are based on clustering 5mm LEDs, which if driven correctly works well for general illumination, but a new work bench has been added that was sized to accomodate as a useable surface to lay out outdoor barbeque gear/meats on. Illumination on these areas must satisfy at least two criterias: on par color rendition, and bright enough to comfortably work but not so bright one would have to squint.

Heres one of the original strings that hasn't been disassembled
This one was closest to the house, pristine condition, still works

The one is in between, minor corrosion on the ring, one LED is closed circuit

This one is parallel to the leading edge of the house, about 20ft from the center of the house. It has been repaired twice, the second time new CREE XRE emitters were laid on the heatsink, one LED is closed circuit.

First installation in November 2009 they have worked 24/7, considering the fact that these LEDs were not coated with dielectric spray on and were shielded only by a gutter mounted above it, I have no complaints. Having the freedom to mix and match LEDs on a common voltage rail makes me want to use the Nichia 219 Tri-stars over the work bench side. But I have little familiarity with these new LEDs. I do know the Rebel Warm whites used in the evaluation of a feasible budget CRI LED was a :fail:. It made the place feel like I was lighting it with low pressure sodiums. :green:

 

[idleprocess]

My plan now is to see whether or not I can find six die-cast aluminum enclosures without costing me an arm and a leg or a size so large I cannot effectively use it and house the heatsink in it [or the LED itself] and cutting a 20mm hold on the lid for optics.

I used these (the bare aluminum 1-1/4" variety) on my outdoor lighting project and they're still running strong. I just AA'ed the star to the canister and the canister to the (admittedly a bit too small) quasi-heatsink. Been running a ~75% duty cycle for about 18 months now including one of the hottest summers on record. Inside are some Rebel tri-stars running at 350mA. Getting the wires routed is a bit of a challenge thanks to my awful soldering skills and resultant cold solder joints (not a fan of RoHS solders).

A bead of some sort of adhesive will keep the lid in place and the windows on the lid are actual glass. Appears to be sufficient vertical clearance inside for an optic, but I've admittedly never experimented with such.

I'm using the same canisters on a different project that's about two-thirds complete (need only spend a few hours in the attic ... yuck), albeit mostly as additional radiating surface area and as a means of keeping dust off the LED's.