I have an under-insulated roof above a "cathedral ceiling".  The shingles need to be replaced this year.  It occurred to me that this could be an opportunity to insulate from above.

Can it make sense to insulate above the plywood roof deck? Or to remove the plywood roof deck, add insulation, and re-install the plywood roof deck on top of that? 

Accessing the insulation from inside would require ripping down and re-installing a perfectly good ceiling.  I believe there's no more than 6-8 inches of fiberglass insulation in that ceiling, and in some places where there's an empty cavity, it appears the fiberglass has fallen out of position. 

The home is in Massachusetts.  I would love to hear from anyone on the theory and practice of doing this, and any real-world experience. 

Tags: Insulation, Retrofit, Roof

Views: 22419

Reply to This

Replies to This Discussion

Sorry I can't access where it is but your best bet is to remove the plywood and add Closed cell SPF leave a gap below the new applied plywood sheathing and ventilate from the soffit to the ridge vent, this will prevent the ice dams.

A project was done on Net Zero Energy Efficient home and this is what they did along with other upgrades.

I believe your best option is to deal with this when you have the roof off. I would suggest, if possible is to remove all of the existing insulation and replace with dense pack cellulose. If you achieve the proper density of 3.5 to 4.0 #'s per cubic foot ventilation will not be an issue. I live in upstate NY and have done this several times with great results

we have many successful jobs done where a roof was removed, we sprayed closed cell foam down on the interior sheetrock and a new roof was applied.  it is very expensive to do this, but it works well.  you need a minimum of 3 inches of foam, and if you have space for 5 go ahead, but over 5 is frivolous. In this situation you can end up with a vented roof becasue the foam does not completely fill the cavity. 

If you are a true energy geek, put a continuous layer of rigid foam between the rafters and the new deck before you roof it.  that rigid will far outperform another 4 inches between rafters.  I don't know what the snow load requirement in your area is, but most rigid foam is rated around 25-27 psi compressive strength.    in this case, don't vent it,a nd seal the rigid to the underlying spray foam along all edges.  Skipping the edge seal is criminal to me. 

adding a roof over a roof is OK if you remove the old roof at the intersection of roof and walls and seal between the original ceiling and the new roof insulation system.  if you miss that step you are wasting your time as the new insulation system is suspended over a failed system with infiltration from all sides. 

recessed lights and side attics like the one you describe over the closet are tricky.  you can suspend  piece of rigid foam under the rafters in the 'attic' space above the closet, and spray down on it.  the spray foam will 'glue' the rigid foam to the underside of the rafters and your roof will work. 

for recessed lights you need thick roof cavities.  otherwise, have them removed.  at minimum replace them with AT/IC units and create an insulated dome over them.  TenMats are good for this, but your rafters need to be 12 inches or more for it to work. 



A couple of replies (Robert and Sean) are on it -- fill the cavities with cellulose, add enough foam above the deck, then re-roof without venting. But, you wanted some theory to back this up...

Note that cellulose in unvented ceilings is a bad idea UNLESS you get enough foam on the deck above to keep the roof sheathing above the dew point. See Building Science Corp. -- http://www.buildingscience.com/documents/insights/bsi-043-dont-be-d... for the science .

Why install the cellulose at all? Without some "filler" in the air space in the rafter cavities, the existing insulation is essentially non-functional. 

I am always a fan of leaving intact drywall intact -- I just hate mudding joints and re-painting! So I'd work from above.

That means the simplest approach is to strip the existing roof, blow the cavities (with celloluse OR dense-blow fiberglass) then apply foam on top in accordance with the tables in Chapter 8 of the IRC. Then lay another deck surface and re-roof.

Then, if you really want to do this so it'll never need attention in your lifetime or mine, lay down a metal roof. And, that way, no one will have to deal with the usual shingle warranty hassles -- the shingle manufacturer arguing that their product failed because you didn't vent under the decking.

Thanks to everyone for your replies.  The response was way beyond my expectations!   I really appreciate it.  I'll be glad to see the snow lasting longer on the roof next winter, and the room not getting so hot in the summertime.

Lots of great information. This string of information will be useful for many others besides me, I'm sure. 


Article by Joe Lstiburek. I don't think its a project that I would undertake, but it makes for a good read.


I have a similar cathedral ceiling. Unfortunately, it has pretty new shingles, as we re-roofed the house a couple years ago without my realizing that this presented an opportunity to improve the cathedral ceiling.

I got a quote from a local (I'm also in MA), experienced insulation contractor to blow in dense-pack cellulose without having to tear (much of) the roof off. This certainly sounds good to me, as we love this room, but I'd bet that this room is responsible for a disproportionate amount of our heating bill, in large part due to the existing "insulation."

The room also has a few recessed cans, which are easily located from the outside during the winter via melt patterns. I am going to replace these with low-profile LEDs (again, we like the lighting the recessed lights provide, but hate the energy performance). This seems like a no-brainer.

I'm interested if anyone else has done a similar job, and would like to hear how it turned out.

Two reasons not to attack a new waterproof roof:

1.  100% of the patches have to be perfectly waterproof for near 20 yrs, which for me means a double patch, which...

2.   Leaves you with an irregular roof, assuming you can get shingles that match to begin with (heh, looks count too)

When retrofitting insulation, adding to existing insulation, you need to keep it simple, keep costs low, or the payback time is just too long for anyone not trying to pass the home on to the next generation.  Doubling up with other work, like roof replacements is GREAT.  Ripping off perfectly good sheathing and siding, not so great.  Remember, double the insul does NOT mean double the savings, you need to run the numbers and just do what makes sense.

I hear that you like the room under the roof alot, but I'm going to suggest you work from below for now.  It'll help with the hi hats too, and before you go all LED, run the numbers and see WHEN you'll start saving money.  Compact Fls probably have a better payback time, and better color quality is available. 

Get some sheets of HIGH density rock wool, and a saw to cut it.  The price is comparative to foam, no smoke problem in fires, no chemical outgassing, can take water and dry out and still work, it allows vapor back out to dry the system if needed, and the R value you buy, is the one you'll still have in 20 yrs and will match the aged R of foam. 

Put up 2x's at right angles to the rafters - if you want 4 inches, nail a piece of 2x over each rafter.  Space 'em a bit less than the width of the glass wool to get a tight friction fit, butt sheets tight against each other.  Gypsum board over and paint, oh, might as well caulk the GB to the 2x's for an airtight drywall installation.  Overall, iIt's not going to be that much messier than painting the room, just take a couple days more.

You'll then have broken thermal bridging through your rafters, and gained 3 - 10% right there, depending on how the roof is framed.  You'll get rid of the thermal holes at the lights, and still be able to use recesssed lights probably.  Thermally, you'll go from probably a badly installed batt (alas, batt is usually badly installed) in the 6 inch rafter space, with an assembly R of maybe 15, to an assembly R of 27 for 2 inch rockwool.  Which is just about double, which is just about the MINIMUM insulation you need to add to an assembly to have any hope of a good pay back time.  Triple may or may not be worth it, you'd need to run the numbers.  More than happy to run them for you, gratis, go to roughdesigns.com, hit the mail link, and I'll fill you in on all I'd need - it's lots more than just what the insulation is.

Russell Higgins



We have been using the technique of removal of the roof deck along with the shingles for a while now.  Results are quite amazing.  However the only way to get amazing results is to use closed cell foam in the rafter cavity.  We have been doing it this way leaving a minimum of 3/4" air space above the finish foam for roof deck ventilation when complete. Be cautious of Can lights spray foam is not supposed to be used even on IC rated cans.  Our solution is to retrofit the cans with an LED kit that is tagged as a modified fixture.  The added benefit is that you can seal the wall to ceiling connection from the top side.  Keep the soffit ventilation (or add new) and add a new ridge vent when done. 



George, you got many great comments and suggestions.  All the issues were considered by the combination of the comments.  If you go to the Building Sciences web site and access Research Report - 0404 you will get a detailed explanation of the issues and diagrams of the appropriate roof design.


No way would I start from the bottom. It is so much easier to re-due the roof then remove drywall and trim and whatever else is on the ceiling. Other thing not considered is the control of debris, fiberglass, drywall dust, possible mold exposure, mice, bird or vermin droppings. However I do agree with Roxul over closed cell foam as it must be applied correctly or you may encounter some very fishy problems.

Definitely anytime the roof is on its last legs work rip it off and work from the top. 

Otherwise, Instead of removing the exisitng GB below the ceilinig you just leave it and work on top of it - (working on the assumption there's insulation in between the rafters already, which seems to be the case in these postings.).

The messiest part of the furring out a cathedral clg is taping and sanding the GB, you can "sponge" the joint compound, I do, it works.  Though, you don't want to lard on the compound, as I've seen crews do and rely on snow storm scale sanding to get rid of it.

Adding to existing insulation is usually a marginal energy savings. 

To get a justifiable pay back for most folk demolishing GB from below, or roof deck from above, or ripping off or covering up an good roof with years left in it will push the payback to 10 yrs or beyond.  I'm all for Deep Energy Retrofits, but ONLY if you have "spare' money you definitely won't need for a decade or more. 

I don't, lots of folks don't.

Also, furring out below and uisng high density rock wool is feasible for a DIY'er over a few weekends, for those who can't afford contractors - I keep telling myself it's a good thing working all those weekends as it cuts the payback time in half!  I prefer working a bit more, and using the money for a GC, but I can't find enough work right now..



Latest Activity

Sean Lintow Sr replied to David Meiland's discussion Air gap above floor insulation--how big is the impact?
"Yes & no - newer codes added in "Exception: As an alternative, the floor framing-cavity…"
13 hours ago
Dan Phillips replied to tedkidd's discussion It is time for government programs to end. They become monopolies that block market solutions.
"Not all states used deemed savings or modeled savings.  Don't throw the baby out with the…"
15 hours ago
Dan Phillips replied to David Meiland's discussion Air gap above floor insulation--how big is the impact?
"I wonder if you could do an experiment.  Leave a section of the floor insulation as is. …"
15 hours ago
Dennis Heidner replied to Tim Kendzia's discussion Cooling effect from solar panels
"I had missed your query originally made a couple of years ago.   Yes it does make a…"
Dannie Jackson replied to Tom Smith's discussion Are Energy Efficient Windows worth the investment?
"If you look at the R factor of the window and compare it to old school single pain glass windows…"
Dannie Jackson replied to tedkidd's discussion It is time for government programs to end. They become monopolies that block market solutions.
"Something that bothers me which is very similar to the concerns about agencies and programs…"
Dannie Jackson replied to Tim Kendzia's discussion Cooling effect from solar panels
"I would assume two things about this idea, first of all you would definitely shade the roof and…"
Charles Buell replied to David Meiland's discussion Air gap above floor insulation--how big is the impact?
"You want to install enough insulation that the room temp line is within the insulation---not…"

© 2018   Created by Home Performance Coalition (HPC)   Powered by

Badges  |  Report an Issue  |  Terms of Service