The light of life..mp4
More ideas from Mark J. Comerford, architect:
from mark comerford email@example.com
to [Available] Catherine Todd
date Mon, Sep 12, 2011 at 2:13 PM
subject Re: bottle lights caulk
Important mainly because of your interaction with messages in the conversation.
hide details 2:13 PM (11 hours ago)
Here's what I have thus far, and I hope to have an update for you later.
Your project and blog discourse have got me thinking about issues with the original youtube design that an improved version should try to avoid.
Creating a Leaky Roof?
One of your readers comments ("be careful about poking holes in someones roofs) got me thinking how the water integrity (no leaks) especially in the rainy season--where I suspect it may be too rainy for days on end to repair, is important, and how you might be doing them a service in terms of creating interior light but a disservice in terms of a leaky roof.
Then, thinking about the Youtube approach, chiseling the hole in the rusty corrugated roof, ( I understand how this is done working on a shet metal on the ground, where you could back it up with something firm, but how this is done in installed panels escapes me) gooping unprepared substrate with sealant and expecting it to adhere to both a rusty roof and a slick plastic bottle. I think the method they are proposing is just too crude to expect to get a "permanent" solution. While the silicone will last 30 years the entire system will fall apart rather quickly due to thermal contraction /expansion, roof movement due to wind and possibly foot traffic, and most importantly inability to prepare substrate properly resulting in adhesion gaps leading to water entrance, etc.
Skylights in modern house construction are tricky to flash/seal properly and are often avoided by builders and owners due to fears that they will inevitably leak, and this is all in the context of a much more precision oriented and better trained workforce.
Is Disposable Skylight OK?
What you will be giving them, at this level of technology, is a disposable skylight. This may not be bad as the time and effort to make a new one every year is not so great and it is consistent with the technology and unemployment of the vernacular/economy which is in itself largely recycled/disposable. If this is the approach you take then be sure to make it clear to everyone it is disposable. And it may well be the best approach
Thinking about it, in the context of the disposable option, The urethane makes better sense than silicone. While it wont last as long as silicone, it will definitely outlast the assembly. Urethene requires less skill and procedural diligence to apply effectively than silicone, and costs less than half. It is a bit softer and more forgiving and probably adheres better in the short term than silicone. I'd bring a few tubes of each and then experiment.
More Permanent But More Difficult Option:
I think a more permanent solution ( i e 10 year) would be much better, and allay the fear that in the process of providing light you are creating a hole in their roof for rain. That solution would address the deficiencies that make the Youtube model so temporary. I think it is "doable" but will add a few steps and complexity to the skylight design and construction process. And you will need to build several experimental mock-ups to fine-tune process. In terms of our Appalachia analogy, instead of doing 100 skylights you will do 20. But they should have less leak/maintenance problems. Here are some issues/suggestions whose solution could yield a more permanent skylight:
Glass or Plastic Bottle?
Replace plastic 2 litre bottle with glass bottle. The glass will provide an ideal substrate for the silicone, as opposed to the plastic, which has to be sanded/scratched, and also probably suffers from UV degradation when it bakes in the sun, and may have chemical compatibility issues. Find bottles with broad caps (like the Nagene hiking bottles for example or canning jars)and position the bottles so that the caps are on top. Thus if the water clouds the bottles the water can be replaced from the top without tearing the skylight apart.
If Glass, Support More Critical
Design a wire basket or strap assembly to support the skylight bottle from the underside. This will have to be thought out--not sure whether rivets into potentially rusty room metal is good enough, but a simple, cheap design should be easily figured out. This could be a safety issue. Do they have earthquakes?
This is the hardest, but potentially most beneficial step if done properly: Flashing. Flashing is sheet material used to redirect water. You have in your blog a link to a thru-roof penetration sleeve; what that is doing gets to the heart of flashing. What the flange at the base of that sleeve functionally does is elevate the opening of the roof above the plane of the water. But for it to work the lower portion of the flange must be sealed/integrated into the design of the roof. So really you need to know exactly what type of roof you are penetrating prior to designing the flashing
You might search the web for the SMACNA manual on sheet metal flashing. It consists of hundreds of different diagrams for sheet metal flashing addressing hundreds of different situations. For flashing to work, it is highly customized to the specific site/roofing conditions. Failure of flashing is the most common source of building links.
By the way, if you are interested in building technology issues, one of the best sites I know of, sponsored by one of the worlds leading building scientists, Lsturbek, is www.buildingscience.com
Flashing for Corrugated Roofs
The problem I see with using a sleeve of the type you have shown is that it is not made for corrugated roofs. I cannot envision how this would work with a corrugated roof. Unless you position the leading (upper) edge of that flange beneath one of the bottom edges of the corrugated roofing sheets, you will have created a condition where the gaps between the flat flange and the undulating roof panels will have to be dammed up with globs sealant. You might use a tube or more of sealant on one skylight--expensive stuff.
I'll try to look into flashing solutions for corrugated roofs. Do you only have corrugated roofs?
This is where the water penetration would likely occur, as the water works its way on the interface between the sealant and the rusty corrugated roof panel at that dam. All this is aggravated by thermal expansion/contraction, roof panel flexing in high winds, etc. The primitiveness of their technology makes it all the more challenging. The youtube video approach doesn't really seem to use flashing as I understand it, it must rely on the sealant to act as a dam against the water. I need to look at the video again
Build a wooden curb/box Idea
( or maybe modify already made plastic tupperware type box)
What would work better, I think, would be to build a small wooden box that is fashioned to mate with the undulating profile of the corrugated roof, sit on top of the roof, and then the bottle penetrates the top (base) of the box.
Wooden box customized to roofing material
Imagine taking a wooden box, say 12" x 12" x 4" high, then cutting it, probably with a jig saw, at two opposite edges in an undulating pattern that matches the profile of the roofing. You would have to determine the undulating pattern, using a simple carpenters scribing technique, and it may vary house to house.Then, turn the box upside down, and sit it on top of the roof, centered above a previously cut bottle hole, and seal all around with silicone.
Note that this sealant profile (the cross section through the sealant attaching the box to the roof panels) is now of a consistent dimension rather than having gobs of sealant here and little sealant there. The undulation is handled by the box geometry. Having the sealants profile of a consistent dimension (ideally for movement conditions in an hourglass shape in section, a 3/4" wide joint should be 3/8" deep at narrowist, and this geometry allows the sealant to be elastic, and act more like a rubber band.) Note with a rubber band that it is very elastic pulling in one direction, but less so in the other. This is because the cross sectional shape of the material (rubber band, or sealant) effects its elasticity. )
Now you essentially have an upside down wooden box glued to the top of the roof, and because of the undulating cut of the opposing edges of the box, the box matches geometrically with the roofing and the sealant width can be uniform. Now cut a hole in the top of the box and insert bottle through. Note that this is now an easily sealed joint. You are not gooping sealent around an undulating material. Plus, this joint is located above the plane of the water cascading down the roof and thus less vulnerable to the ravages of time and water volume
Then there might be two options. You might ignore the bottle and just slap a sheet of glass on top with silicone used as an adhesive. This would actually be more foolproof/simpler, a less tricky joint, as the glass protects the joint, and the joint is not weather exposed
But I suspect that the suspended bottle full of water might have an optical effect better than simply a hole in the roof, and probably look and behave more like a light bulb. Not sure. But you might experiment with ignoring the bottle altogether. Maybe even just make the base of the box out of plexiglass or clear plastic and call it a day.
Logistics: Benefit of an offsite shop
Now the more permanent method I am proposing involves more precision, translating into more equipment, training, and time. I might consider setting up a shop in town to build the boxes. Here you have electricity, controlled environment, level tables and lighting, proper tools, and security (locked door) to keep the tools from being swiped. I suspect a sawzall is equivalent to a years salary, so tool security will be an issue.
Then you go to the field, meet the occupants, determine the profile of their specific roof panels ( I bet there is a variety of roof sheeting used, varying from place to place), and make some field measurement/ sketch. Do this for each 20 different houses. Then return to the shop with the electricity and make the 20 different boxes in the shop. (or modify the found tupperware type boxes) Here you could also paint them. Return to the field to install.
Shop not required though--all required tools are available rechargable. But you will need to be recharging batteries often, probably daily. May
More permament solution will require elevated skills and organization. Electric tools will cut through metal like butter compared to hand but will present theft problems. More information is necessary about the roof systems you will be encountering, and permanent solutions will vary depending on type of roof systems encountered.
Do you have a sense of which direction you want to go? In either case I would try to set up a cottage industry where they are installing these themselves like the chap in the video. The process of teachng them some basic carpentry skills and in a sense elevating the technology a bit might have its own benefits as well. Maybe as a part of this you establish a little carpentry school where students are building these roof boxes/skylights. The community gets better skylights and also a cadre of people who now understand some basics about field measurement, drawing, working with wood, basic power tool operation, etc. This might benefit some to "move up" to become carpenters for whatever richer class or group is nearby.
The questions raised above (permanent vs temporary, on site construction versus remote/ shop construction, and hand versus elect tools, all affect your shopping list. If you can provide me input on the level of permanence you think is appropriate, and the amount of modern technology (electric drills, etc) that would be best, I can provide you with a better shopping list. Ideally, we would figure the design parameters out better before shopping, but I understand your dilemma about your flight Thursday. To figure out the design, we really need to know specifically about what type of roof panels they have--which again may vary considerably.
Let me know your reactions and I will make a recommended list. I'll also go through the videos you've attached which I haven't yet dug into to see whether the design concerns I have are addressed already.
It boils down to doing the Youtube video way, which is inherently simple but also is probably not going to work long term, and is definitely going to create maintenance headaches on a yearly (at best) cycle, or elevating the technology, to get a more permanent solution, which will also make the whole process more complex and in a more difficult. Which way to go?
All the best,