GBA Logo horizontal Facebook LinkedIn Email Pinterest Twitter 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

Recessed Lighting / heat loss

Justin Brown | Posted in Energy Efficiency and Durability on


New homeowner. I have 10 6″ prescolite IC non-airtight cans atop my 2nd floor going to my attic and 11 of the same IC non-airtight cans in cathedral ceilings going to the roof. They are square housings.

I have explored all the options. The attic is covered with I believe r-30 batts and about 12″ of blown-in cellulose on top of that. They put additional bats all around the fixtures like a moat but not on top of them — not sure why, since they are IC rated.

Access to the attic is very bad. Most of the fixtures abut joists and digging around the cellulose to build air-tight boxes is my last resort since I’d have to do a lot of custom cuts etc.

Commercial electric makes a gasketed led retrofit which they claim/certified airtight. I assume that will greatly cut down on air-flow to the attic? If I install those and add some insulation directly over the cans in the attic, is that a good solution for cutting air passage and bolstering insulation in those areas?

If I use those retrofits in the cathedral areas, since the retrofits are IC rated too, can I stuff some fiberglass into the old IC housings themselves before installing the retrokit or is that a no-no?

Lastly, I know there are a lot of factors at play here, but can anyone help ballpark for me annual heat loss costs per can in a situation like this? I’ve seen $5-30 per can per year. I’m in New England so winters are cold.


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

    You have my sympathy. It's outrageous that there is no law against building houses with massive holes in the ceiling air barrier. I'm sorry that you bought a house with this problem.

    I'm not going to try to estimate the annual cost of the heat loss associated with these recessed can lights, because there are too many variables. Suffice it to say that the air leakage associated with these fixtures is a major problem, and you are certainly justified in pursuing a solution.

    I think that your plan to install LED retrofit units is a good one. I'm fairly sure that you won't violate the electrical code if you install a handful of fiberglass in the empty cavity of each recessed fixture, but I'd be interested to hear from any electricians on that issue.

  2. Justin Brown | | #2

    Thank you Martin for your input.

    Okay, I will proceed with the LED retrofits.

    Regarding insulating inside the existing cans (recall both the current can housings and LED retrofits are IC-rated), would 2" XPS pushed tight inside the top of each recessed can cavity be an effective strategy? There would be about a 5" gap between the LED housing and the foam board. The LED retrofits would solve 99% of any air leaks...the insulation would be to limit heat transfer where I can.

    Alternatively, I could stuff in loose fiberglass...or neither. Any input would be great.

    Thank you!

  3. Expert Member
    Dana Dorsett | | #3

    If you're going to take this chance, don't use XPS inside the can. XPS melts while burning, whereas polyiso chars in-place, and has a higher kindling temperature than polystyrene.

    If doing it with fiber insulation, use rock wool rather than fiberglass. Fiberglass can melt in the presence of a fire too, rock wool cannot. Fiberglass is also prone to creating suspended glass particles in the conditioned space air if there is any air leakage from the attic to the conditioned space, and putting it right in the potential leak path probably makes that even more likely.

  4. Justin Brown | | #4

    I assume you are talking about worst-case scenarios if there is a fire, not the day-to-day heat generated by the LEDS? Okay so some rock wool like this:

    stuffed into each cavity?

    To be clear, I would keep a 4-6" gap between the LED and the insulation.

    Would the insulation even help much? I guess I am thinking the primary benefit is reducing the air flow which the LED insert affords. That is the biggest source of heat loss here, correct?

    Any other suggestions in scenarios where sealing the can from above is not feasible? I could spend the time to tape all the cracks in the cans with foil tape for instance but it would take a looong time and I'm not sure if that's redundant given the LED insert provides a pretty tight air seal to begin with.

  5. Robert Hronek | | #5

    Are you talking about changing just the bulbs to led's. I would think a guy could rent a machine and blow more insulation in the attic and fill up around the cans. You wouldnt have to get close to the cans as in crawling on your belly and building boxes. I even think you could spray foam right against the can if you were using leds. The led's put out very little heat compared to bulbs.

  6. Justin Brown | | #6

    I am talking about using an LED insert that is gasketed and seals tight and covers the entire ceiling opening to limit air passage into the can at all. This specifically:

    I suppose I could blow more insulation over the cans I can access from above. However, the cans in cathedral ceilings are a different story, hence my question about putting some insulation inside the cans themselves.

  7. Expert Member
    Dana Dorsett | | #7

    Justin- your insulation link was to some cotton batts (denim), not rock wool. For anything inside the fixture, is what you're looking for:

    Say the power supply in a cheap LED fixture blows a capacitor and starts a small electrical fire- you DON'T want flaming melted polystyrene dripping out of of the fixture. That can't happen with rigid polyiso even if it persists long enough to light off the foam, but it TOTALLY can't happen with rock wool, since even the steel of the fixture would begin burn before the rock wool.

    Most LED recessed can retrofit assemblies rely on air-convection to keep them cool. Don't expect them to last as long in a sealed & insulated recessed fixture. The surface-mountable Philips units designed to be mounted to regular electrical boxes will probably last longer:

    The 5.4" outside diameter round ones are on the order of $50/pop at internet pricing for the 90 CRI (=better color rendering) flavor, the 7.9 inchers run ~$55. There may be a way to cobble these (and an electrical box) inside your existing fixture stuffed full of rock wool, tossing out your trims.

  8. Michael Geoghegan | | #8

    To piggy back off of this question, if you are using rigid foam to build airtight boxes around an IC rated can light, is there a minimum distance the foam should be kept back from the can? Or does IC include rigid foam insulation as well?

  9. GBA Editor
    Martin Holladay | | #9

    To the best of my knowledge, IC-rated fixtures do not require an air space between the fixtures and the insulation. When it comes to spray foam, however, you don't want to gum up the works of the fixture -- so some type of barrier around the fixture probably makes sense.

    There are several issues here. One has to do with heat build up, which is a problem that is obviously most severe with incandescent bulbs. For more on this issue, see A Recessed Can of Worms.

    Another issue has to do with the mechanics of insulation installation and the field practices of installers. If any spray foam installers are reading this, I'd be interested to hear whether they wrap IC-rated fixtures in some type of paper or cardboard before they spray around the fixtures.

    The third issue is the electrical code requirement that electrical boxes include a certain volume (expressed in cubic inches); this volume can't be filled with insulation, because the air in the electrical box helps dissipate heat if there is a wiggly wire nut in the box. I'm not sure whether the prohibition against stuffing insulation into electrical boxes might apply in this case, but I suspect that it doesn't.

  10. Justin Brown | | #10


    Appreciate the feedback. Okay I will stuff rock wool inside the fixtures in cathedral ceilings, leaving an air gap between the rock wool and LEDs. I will place rock wool either inside the fixtures or over the fixtures venting to the attic, where I can access from above. Because rock wool doesn't air seal I would imagine they will all still "breathe" a bit via air transfer from above, correct?

    I will put no insulation inside the cans between 1st and 2nd floors.

    With 36 total fixtures, I am not eager to jump to $50/ LED unit you linked to from $13/LED unit I linked to but I appreciate the benefits of doing so. The units I linked to carry a 5 year warranty. Frankly, and frustratingly, I wont even be using many of these fixtures very often...of the 36 recessed lights in the house, if we use 10 of them daily for 3 hours total per day I'd be surprised. This is mostly about limiting air / heat loss.

    Back to that, am I correct that in this situation, from an energy savings perspective, the primary benefit comes from air sealing / limiting air loss by sealing the hole to the can from the house, with insulation in/above the can providing the secondary benefit?

    Thanks for everyone's help!

  11. Expert Member
    Dana Dorsett | | #11

    Justin- while rock wool is not an air barrier, it's fairly air retardent (far more air retardent than low density batts such as R11s. & R19s.) You won't be convection-cooling the LED assemblies if installed tight to the rock wool.

    Air sealing the cans would clearly be better.

  12. Dan Kolbert | | #12

    Or get rid of them, but that's obviously more labor- and capital-intensive.

  13. Justin Brown | | #13

    Yes I thought about getting rid of them but it's just more work and cost in the end.

    Okay so primary benefit comes from air sealing...rock wool helps but less critical. Let me know if I've got this wrong. If not, I will plan to use rock wool more significantly in the lights we use infrequently, which in this case happen to be the units venting direct to the roof. Win-win.

Log in or create an account to post an answer.


Recent Questions and Replies

  • |
  • |
  • |
  • |