GBA Logo horizontal Facebook LinkedIn Email Pinterest Twitter X Instagram YouTube Icon Navigation Search Icon Main Search Icon Video Play Icon Plus Icon Minus Icon Picture icon Hamburger Icon Close Icon Sorted

Community and Q&A

Retrofitting loose-fit batts in cathedral ceilings

rawlinson | Posted in Energy Efficiency and Durability on

What is the best method to retrofit loose batts that don’t fill rafter cavities? Should we slice pieces of unfaced fiberglass, or blow in loose fiberglass fill? I am concerned that loose fill might plug the ventilation space, and also that the vapor barrier would continue to have substantial gaps.

The kraft-faced 10″ batt fiberglass insulation between 2×12 7:12 pitch 24″ o.c. rafters is poorly fit, due to variation of the rafter spacing; some of the batts have up to 2″ gaps for the entire length. I guess it was deemed good enough in the 80’s.

I don’t see that any moisture problems have developed, probably because this is a mild Zone 3 in California’s Bay Area; but I would like to improve the insulating performance while we are putting on a new roof (replacing wood shakes with comp shingles).

The upper ends of rafters have collar tie rafters and a flat ceiling, and we are adding a ridge vent in addition to existing gable end vents. There is no existing eave venting but the sloped ceiling bays are only about 6′ deep dead-end bays, so I feel that the air is finding its way into each bay.

I could consider some form of spray foam, although the contractor has an aversion to the product.

GBA Prime

Join the leading community of building science experts

Become a GBA Prime member and get instant access to the latest developments in green building, research, and reports from the field.


  1. GBA Editor
    Martin Holladay | | #1

    There are several things wrong with your existing insulation, as you know:

    1. The batts are narrower than the rafter bays.

    2. The fiberglass-insulated cathedral ceiling apparently has no soffit vents (if I understand correctly).

    3. There is a ventilation gap above the batts, but there is no baffle to separate the top of the insulation from the air space.

    Here's my advice:

    1. Fiberglass batts are the worst-performing type of insulation for a cathedral ceiling. If you have opened up your ceiling, and you have a chance to fix it, remove all of the existing fiberglass batts and put them in a dumpster. Choose a different product when you re-insulate.

    2. It makes no sense to design a vented ceiling assembly unless you have soffit vents. If you can't install soffit vents, you have to design an unvented ceiling assembly.

    3. Read this article, which lists all of your options when it comes time to do it right: How to Build an Insulated Cathedral Ceiling.

  2. Expert Member
    Dana Dorsett | | #2

    Since you're replacing the shingles, NOW is once in 20-30 year opportunity moment to boost performance by putting some of the insulation on the exterior of the roof sheathing, which puts a thermal break over the rafters, and eliminates the need for interior side vapor barriers (which reduce drying rates), or the need for ridge venting. In your climate zone you could get away with as little as R5 from a moisture control point of view:

    But R10 would be a more useful thermal break. At that point blowing cellulose (not fiberglass or stone wool) into the 2x10 or 2x 12(?) rafter bays becomes the better solution, providing both thermal mass and moisture buffering, bringing the whole-assembly R (thermal bridging included) into the R50 range. Thermally breaking the rafters with R10 on the exterior will make a comfort difference you can actually feel in mid-summer blazing sun conditions, but even R5 over the roof deck and a full cavity fill of cellulose would make a noticeable difference.

    For keeping the global warming potential of the insulation itself and additional thickness well bounded use 1.5" of polyiso for the above deck insulation rather than 2" XPS. You could also get there with 2.5" of EPS, but that adds a lot of thickness to the assembly. A half-inch OSB nailer through-screwed to the rafters with pancake head timber-screws 24" o.c. is usually sufficient in non-hurricane zones, and has the least thermal bridging from the fasteners. The worst solution from a thermal bridging and shingle retention point of view is to skip the nailer deck, use 1" foam, long-nailing the shingles to the structural deck, since 5,000 nails adds up to one huge highly conductive thermal bridge, and provides a moment-arm for wind-flapping shingles to work them loose.

  3. rawlinson | | #3

    To Martin, I am reluctant to discard the batts, which appear to be in good condition, and it doesn't seem right to fill up the landfill if they can be made to work; and we are not removing drywall at ceilings, just the roofing shingles, so it may be difficult to yank the batts out of the bays. Most of the batts are a pretty snug fit, but there are the exceptions. It is possible to add the soffit vents, as we are adding overhangs where none exist; because this is located in an "urban-woodland interface" zone, the idea of poking holes at the eaves to let in sparks doesn't seem ideal, but is apparently code-compliant here if they have wire screens, or I could use a soffit board with 1/8" ventilation perforations. Because I can clearly see above the batts to the end of the rafter bays through a 1 1/2" vent space, I can't see why adding a baffle is necessary if the existing batts could remain.
    To Dana, I agree it is a good opportunity to add topside insulation over the deck, but I would then probably have to spray seal and insulate the end-gables of the attic in the center (above the collar ties), creating a dead air space about 6' high.
    Thanks to you both. Brian

  4. GBA Editor
    Martin Holladay | | #4

    You wrote, "Because I can clearly see above the batts to the end of the rafter bays through a 1 1/2" vent space, I can't see why adding a baffle is necessary."

    Here's the reason: fiberglass batts are air-permeable, and most ceilings are leaky. Your fiberglass batts will be almost worthless without a top-side air barrier. Without the top-side air barrier, leaks at your ridge will tend to pull lots of air through the fiberglass, which will act as an air filter instead of an insulation blanket.

    Almost any other type of insulation will work better in this location than fiberglass batts.

  5. Expert Member
    Dana Dorsett | | #5

    So what if there's a dead-air space, so long as it isn't a vapor trap, and dew-point protected by the exterior foam?

    Or, you can leave the gable vents open, so long as there is sufficient R on the floor of that that space.

    The benefits of the exterior insulation are many, and as configured, the benefits of venting very few, in your case. It'll meet code without venting on the sloped section with as little as R5+ on the exterior, and the roof deck in the micro-attic above is inherently moisture protected by the gable vents, though an insulated-sealed version of the gable ends would be better from a thermal point of view, it's not absolutely necessary, with or without the foam-over on the roof deck.

    If you foam over you can blow cellulose in from the exterior without taking down the ceiling by drilling & blowing through the roof deck before installing the stackup above, leaving the batts in place. The batts may compress somewhat, but that only increases their R/inch, and the net total performance will be virtually identical to what it would be if you'd pulled the batts and did an all-cellulose fill. This isn't rocket science- most cellulose installers can handle blowing (even dense-packing) over batts that have ~2" of space between the fiber and sheathing.

  6. rawlinson | | #6

    Martin and Dana, I like the idea of adding the 1 1/2" polyiso panels and nail surface above the deck and omitting the ridge vent; then stuffing the existing rafter bays with cellulose and insulating the end gables to meet requirements for an unvented attic. I can leave the batts in place so that their existing kraft face answers the inspector when he wants a vapor barrier (nominal though it may be due to gaps), and maybe add a film-forming paint in the steamy bathrooms. All this will reduce the heating load, and if I understand you both, will not create a problem for moisture and mold. Thanks, Brian

  7. rawlinson | | #7

    If all the points in Post #6 are okay, there are two issues remaining: comp shingles over an unvented deck (I understand some mfrs still exclude or reduce the warranty); and venting the exterior overhang soffits at eaves and gables. I am adding a cement fiber soffit (over gyp sheathing for fire protection because it is less than 5' from the property line along one side, so the inspector wants the rafters protected) and the inspector wants the soffit vented. The soffits would be sealed off from the attic and outside the perimeter of insulation. I could either extend the above-deck foam to the edge of roof to reduce condensation when it is foggy, or just use some 2x nailers supporting the nailer panel in place of foam. Brian

  8. GBA Editor
    Martin Holladay | | #8

    An uninsulated soffit does not have to be vented. I would ask the inspector to cite the section of the building code that he thinks he is enforcing. I don't think there is any such requirement.

Log in or create an account to post an answer.


Recent Questions and Replies

  • |
  • |
  • |
  • |