Home>Discussions>INSULATION & HVAC>Insulating a large cathedral ceiling
14 posts / 0 new
Last post
Insulating a large cathedral ceiling

I'm looking at insulating the ceiling of a large gothic barn, in the process of converting it to a house. The problem is trying to preserve the shape, whilst getting sufficient insulation.

The rafters are laminated curves (like a glue-lam, but nailed and screwed, not glued), and the cavities between are only 4" deep (real, not nominal). Beyond that, there's thin lath, then plywood sheathing, then an asphalt shingle roof.

So far, the plan is closed cell foam in the cavities, then a 2" high density foam board over that (will hopefully bend to conform smoothly), and screw drywall or wood finish with 4" screws to penetrate it back into the rafters. 6" x R6.6 = R39.6 However, it's about 7500sqft of coverage, so the closed cell foam is going to be about $30K, plus another $7.5k for the foam board plus install of that, so ~ $40K. Ouch.

Anyone have a better way? I think building anything out with wood to match the curves is going to be very difficult.

I've seen some people spray foam round the edges of rigid foam board. The board seems to be about half the price of the spray foam per volume. Does it really give good tight air-sealing, etc? I guess another possibility is to mount foam-board in the outer 2", then spray foam the next 2, then board again. There's a bunch of roofing nails sticking through, but maybe it'd just press in OK.

Re: Insulating a large cathedral ceiling

Not knowing your zone R-60 may not be required. Here in southern Ontario Canada R-24 is to code for a ceiling, however no matter where in the case of insulation the more the better. With the cathedral or vaulted ceiling closed cell isostyrene is the only way IMO to go.

The max you can get in with batt is R-12, 3.5 inches because you will need a 2" baffle behind it (DURAVENT). The same goes for rigid board. You need, I believe, venting behind the rigid insulation as it will not act as a vapour barrier.

The layering of foam, rigid, foam etc. is nothing I am familiar with so I not going to guess. Sealing the rigid with closed cell will create a thermal break but you are not going to get anywhere the R value of closed cell.

Have you on investigated closed cell roofing panels. We use these for a number of applications from barns to sunrooms. But you will not maintain the curve you describe.

I don't know if you have a had a heat-loss calculation done if not, do so, you may save yourself thousands if you are good with less closed cell.

There is no need to put anything over that closed cell other than your finished ceiling.

The closed cell would be the way to go IMO.

7500 ftsq ceiling sounds magnificent.

Hope this helps.

Re: Insulating a large cathedral ceiling

I think you need to get a new calculator, 6x6.6=39.6, not 59.5. The 6.6 is also a little optimistic, especially for the foam boards. They actually measure about R4 per inch. Closed cell foams have a higher initial R value but they usually degrade over time to about an R4 per inch as well. Realistically you are really about an R 24 between the beams. Over the beams themselves, the R value drops to about R13.

You can figure your average insulation by calculating your total area minus the area occupied by the beams. For example, if the laminated beams are 3.5" wide on 36" centers, then the beams makes up about 10% of the total area. The average R value would be [(24x9)+13]/10=22.9

The good thing about the solid insulation is that it is very consistent. The R value of batt insulation is based on the insulation being dry. A moisture content of as little as 3% can reduces its insulating value by half.

Heres an idea if it will work into your plans. Leave the cavities alone, no fill. Laminate either three layers of two inch or six layers of one inch foam boards. The board on the outside should have a foil face on the side that faces the cavity. There is a special adhesive that you can use to glue the foam boards to the beams and to each other and then glue what ever inside surface you want to the foam.

Now for the mechanical part. Make false laminate beams to go on the inside so that the foam is sandwiched between the roof beams and the false beams. The false beams do not have to be very thick, they are there just for looks. You could use a 1x4 with a long screw, if you can find an 8" screw through the first layer of the laminate, the foam and into the roof rafters, The second layer could be screwed to the first layer or the second layer could be a thin veneer just to cover the screw heads. The inside beams are just for accent mostly.

This would give you a nice even R value at less cost than the spray foam.

Re: Insulating a large cathedral ceiling

You are correct Keith, I missed that. Oxfletch, here is one of thousands of sites re: R values that might be of service.


Re: Insulating a large cathedral ceiling

I'll try to combine replies.

It's Northern NJ (zone 5), so this is my understanding of what's required: Ceiling R38, Wall R20, Floor R30. There really are no walls, it's all building out in the "hayloft", but maybe they will say the first 8ft of height can be "wall". So my original point (not explained well) was, that 6" of closed cell seems to be > R38. I'm not sure if they count the rafters when calculating R values.

My understanding is that now for new construction (and presumably conversions to residential use?) they use a software package, not just R-values. I'm starting to get very worried that this will present a problem, as it's a long thin building with huge roof space, so inherently inefficient per sqft of living space. If I can actually just go by R values, or they go by sqft of building envelope, not floor space, it's probably fine.

Re venting: it seems you need venting on any air-permeable insulation (eg batts/cellulose) but not in an impermeable one (eg closed cell foam). It may be installed directly to roof sheathing.

Sorry about the R-value ... typo. Embarrassingly, I have a math degree - sigh. Fixed. My understanding is that there are both closed and open cell rigid foam board insulations. For example Home Depot sells "Thermasheath 3 R-12.9 2 in. 4 ft. x 8 ft. Foam Insulating Sheathing" (R6.45) per inch for about $1/sqft. I also hadn't factored in the beams degrading the insulation value :-(

Re aging: my understanding is that closed cell foams initial R value is 7-8, and degrades to about 6.5. I'm not sure which value inspectors use though? R4 sounds more like open cell foam?

I like Keith's idea of the lamination a lot ... presumably I don't even need the inside beams if they don't fit the design ... I could just screw all the way through the stack from drywall to hold it? I would think with such a huge continuous surface (assuming I spray-foam round the edges and overlap the joints), it would form a very air-tight building. I wonder if I can achieve much the same thing by just stuffing the closed cell foam panels in the gaps, and having less layers on the outside.

I would like to avoid panels / SIPs, as the roof curve is the main architectural feature of the barn (it won't let me post links, but if you do a Google image search for "sears gothic barn", that's it). It also has an almost-new roof on it (I forgot to mention that too!), so replacing that would add a significant cost. If this wasn't the case, maybe one could effectively construct curved panels on site, by making the same sandwich on the outside of the existing roof with lamination and glue.

From that link, Phenolic foam looks interesting ;-)

Thanks very much for the input

Re: Insulating a large cathedral ceiling

The beams are not counted which makes those minimum R values bogus for the most part. Most so called R20 walls are no where near R20 in the real world. They fill a 5.5 cavity with 6.5" of fiberglass and call it R20, but lets look at it really. The fiberglass is over compressed so it is slightly less than R20. The wall consists of pine 2x6's which are 5.5" actual dimension, has an R1.25 per inch for a total of slightly less than R7, and that comprises almost 10% of the wall.

Too that, you add the through R value of the interior and exterior sheathing. Sheet rock has an R value of about 0.4 per inch so a 1/2' sheetrock adds R0.2 to the wall. The exterior sheathing could be anything from concrete clapboards which have an R value somewhere around 0.0625 per inch for a total of 0.0313 up to 2" foam board with siding over it for an additional R13. Most common is 5/8 or 3/4" wood siding that adds roughly R1. So in most cases we have a wall that is 90% R20 and 10% R8 as the through value

Now you have to add the effect of lateral surface conduction. The sheetrock will conduct heat laterally along the surface to the studs as well as through it. From the center of a stud bay made of 2x6 on 24" centers, the lateral R value would be about 4.8. So the heat at this point will see an R20 through the wall, but also path of R12.8 by going along the surface of the wall to the studs. Heat trying to escape at various other points along the wall sees even less R value as it gets closer tot he studs. But not all heat can take this path To calculate the actual R value would take a calculus equation that is beyond my capabilities.

What I am trying to really say here is that a solid foam layer of R38 is going to be far more effective than a typical ceiling with R38 batt insulation. The codes do not account for this and that is too bad. You can stuff the foam boards between the beams if you want. Because you still have 2" of foam over the tops of the beams, the loss due to the beams wont be much and the codes don't consider it anyway. R30 for the floor really baffles me. Energy loss through the floor is not high enough to justify that level of insulation.

Using a screw through the sheetrock and 6" of foam, I am not really that comfortable with, but if the foam layers are offset and all glued together with that special adhesive made for foam, and the foam is glued to the beams and the sheetrock, then I guess there wouldn't be much stress on the screws. I would be more comfortable with some false beams that kind of brace everything together.

BTW, the bottom 8 feet of the gothic arch is considered wall. Are you going to have two floors in it or just one huge vaulted room. It sounds like it will be an interesting and unique home when you are done.

Re: Insulating a large cathedral ceiling

I totally understand (and agree with) your points about the calculated R values being incorrect. In this case, at least the inner layer of foam board would provide some continuous insulation layer, which would help, but not nullify.

From what data I have seen, above R20 or so, the returns diminish quickly enough that it's a waste of money, at least when you have to use closed cell foam instead of fiberglass like this. So in my mind, it has become an exercise in jumping through whatever legal hoops there are for code as cheaply as possible ... keeping the look of the place is more of a concern.

I think the floor code for NJ actually says (in some later footnote I found) something like "R30, or at least the thickness of the floor". Sadly it doesn't say the same for the rafters.

Re walls/floor - thanks ... if they'll let me average the target values between wall and floor, life will get easier. I think a step out at 8ft up would look odd unless we end up running a beam laterally there, but maybe I can use open cell below 8ft or something for cost.

The "walls" of the barn will form a basement area about 8 ft high; the living section of the house is built on a single level into the roof section, so the walls will curve all the way to the bottom.

PS. Also found that you can get R9 for 1 inch of foam board with a radiant barrier on it, if you leave a 3/4" non-ventilated gap on the outside. If the roofing nails penetrate about 3/4 of an inch, that might work quite well for holding that gap ;-)

Re: Insulating a large cathedral ceiling

If you put a radiant barrier on the underside of the roof and the outermost layer of foam has a radiant barrier so that you have two radiant barriers facing each other, then you add an R3 to the wall. The radiant barriers need to have an emissivity or 0.03 or less. That extra R value only counts once, having radiant barriers between the layers of foam will not help.

Having the radiant barrier on the foam side only also helps, but I don't remember the value it adds, but its not quite as much as with the barrier on both sides.

You are right that additional insulation reaches the point of diminishing returns very quickly, but the codes are not made on return on investment calculations, they are for the most part political. You can make a case though that there is an unseen ROI. If no one increased their insulation, then there would be more demand for energy and the cost of energy would go way up. Because so many people increase their insulation and furnace efficiencies, the cost of energy is lower and that reduces the visible ROI.

Re: Insulating a large cathedral ceiling


You state,

" If you put a radiant barrier on the underside of the roof and the outermost layer of foam has a radiant barrier so that you have two radiant barriers facing each other, then you add an R3 to the wall."

Does this hold true for floor insulation as well?

Cold Canadian.

Re: Insulating a large cathedral ceiling

When they put all this into rescheck to get a building permit ... any idea if that calculates efficiency per sqft of building surface, or per sqft of livable floor area?

Re: Insulating a large cathedral ceiling

Cold Canadian, floors are different than walls and ceilings. In walls, there is a cold vertical wall and a warm vertical wall. The air against the cold wall cools and falls, the air next to the warm wall warms up and rises. The air that falls ends up next to the warm wall at the bottom and the warm air that rises to the top ends up next to the cold wall, The air is in constant circulation.

In a ceiling, the bottom surface is warm and the top surface is cold so there is circulation there as well. In both the walls and the ceiling, the added R3 for a 3/4" space is the maximum available. Increasing the air space does not increase the insulation value.

But in a floor, the top surface is warm and the bottom surface is cold. The air next to the top surface warms up and stays there, the air at the bottom cools down and stays there. For that reason, as the spacing between the boundaries increases, so does the insulation value.

So yes, it sort of works the same, only better. You just have to take in the dynamics that occur around the rim joist. The rim joist needs to be insulated and sealed if possible.


TV Listings

Find TV listings for This Old House and Ask This Old House in your area.