Ice dams — new construction
Hello all – My wife and I purchased a newly constructed home in the summer of 2014, and we live in SE Massachusetts. Fast forward to the winter, and we (along with everyone else in the state) ended up with giant ice dams. The next winter, we rarely had snow on the roof, so no ice dams. The winter after that when the first decent snowfall occurred (5”), boom, ice dams over night. I religiously raked the roof, and almost immediately after ice dams show up.
Basically now whenever it snows I’m waiting to hear that dreaded drip drip noise outside.
We have had multiple solutions proposed from fully sealing the attic and spray foaming, to heat tape along the edges and down the gutters, to installing power vents or passive vents, so I’m hoping someone might be able to recommend something or provide pros/cons since nobody has provided a consistent recommendation across at least 4 independent contractors (builders, roofers, Mass Saves assessors)
A little info about the house:
Built in 2014
colonial style, no dormers, etc.
soffit and ridge ventilation (daylight visible so not apparently blocked)
Blown in insulation on the deck, about 12″ deep
Energy efficiency contractors from the state’s program came out and did some air sealing in the attic attic deck
HVAC system (heat and air handler) along with ductwork (insulated) are all located in or originating in the attic
first 3’ of shingles and ice/water shield replaced along entire length of roof (front and back) to prevent water penetration. Apparently previous ice&water shield was low/poor quality.
roof pitches to the North and South. Ice dams occurred on both faces of the roof along gutter/drip edge
GBA Detail Library
A collection of one thousand construction details organized by climate and house part
Replies
A.,
First of all, can you tell us your name? (I'm Martin.)
Start by reading these two articles:
"Prevent Ice Dams With Air Sealing and Insulation"
"Ice Dam Basics"
Thanks Martin - I'm Luke. I'll have a look at those articles
Once you have done as much source control as is economically feasible, you should then add more pressure balanced ventilation (powered is an option) to remove the heat.
https://buildingscience.com/documents/building-science-insights/bsi-097-de-icing-ice-dams
From what you describe, I would bet that the HVAC system is bleeding heat into the attic. Most HVAC ductwork is incredibly leaky, despite the insulation. A duct blaster test is similar to a blower door, but just for the ductwork. Ask your Mass Saves assessors if they do that testing. Once the ductwork is sealed, take a look at burying it in insulation. A blower door test can also tell you whether the duct boots are well sealed to the ceiling. That's another big area for heat loss to the attic. Pressurizing/depressurizing with a blower door and using an IR camera from the second floor and the attic is another great way to find heat leaks.
Thanks Peter. I'll give them a call and see if they will do that test, as well a further burying the ducts further.
>"HVAC system (heat and air handler) along with ductwork (insulated) are all located in or originating in the attic.
Peter is right, that's the attic HVAC is most likely proximate cause (and and an absolutely TERRIBLE way to build in MA, even if it's code-legal!!)
Fix all the air leaks (ducts and ceiling) and bury as much of the HVAC as you can in blown insulation (whether MassSave subsidizes it or not.) The R8 (or was it only R6 in 2014?) duct insulation is just nowhere near adequate for dealing with the ice damming issue (try R48).
If it's a fossil burner there may be limitations about what you can do about insulating the cabinet & flue. Can you be more specific about what equipment is up there (down to the model name & #, if you can)??
Hi Dana - attaching pictures of the furnace labels here. It's all installed as one unit, but there are two model# / serial# labels.
Luke,
If you follow Dana's advice (to bury the attic ducts in insulation), you'll want to read this article: "Burying Ducts in Attic Insulation."
If you have a furnace up there, however, I suspect that burying the ducts won't solve your problem. You're probably going to have to move the insulation to the sloped roofline (creating an unvented conditioned attic) to have any hope of solving your ice damming problem.
Here are three more relevant articles for you to read:
"Solutions to the Attic Duct Problem"
"Should I Build a Mechanical Room in My Attic?"
"Creating a Conditioned Attic"
Thanks Martin. All good reads. I had heard about people having issues with mold, unseen leaks etc in regards to making the attic an unventilated conditioned space. Seems like this may be a concern in Massachusetts, but not quite as much as it would down south in GA or FL
A 70,000 BTU/hr furnace and 3-ton air conditioner- it's either a fairly large house (3500' +) or the equipment is (as is typical) oversized.
But at least the cabinets are small enough that it may be possible to build a super-insulated doghouse over them with a hatch for servicing the equipment.
How long is the (presumably uninsulated) flue run inside the attic?
Dana - House is under 3,500', flu is uninsulated PVC and runs about 1'-1.5' before exiting through the roof
If it's only a 1-1.5' run to the roof deck it means there's only 18" between the HVAC cabients & roof deck?
If yes, that's a very localized "hot spot" for snow melt for starting ice dams. Installing some rock wool batt insulation in those rafter bays extending three to five feet either side of the furnace & air handler may slow down the process to tolerable levels. Install chutes maintain the code-minimum 1" clearance between the batts & roof deck, and install the thickest amount that fits. Can you take some pictures of the proximity of the HVAC to the roof deck? I'm guessing that those rafter bays are where the major melt-out begins, and the major ice dams are just down-slope of the HVAC(?) which would be easy to spot by the flue pipe sticking out.
Insulating around the flue will help some- 10% of the source fuel heat is going out that pipe. Hopefully it's directed straight up on the exterior, not horizontally where the warm flue gases will accelerate melting.
Another thought:
If the exhaust flue/vent can run horizontal and run out of a gable end it would get rid of that very hot-spot where it exits the roof. According to the installation manual you'll have over 100' of allowable vent length even with 7 ells along the path. See the chart on page 17 of the installation manual:
https://www.hvac.com/media/manuals/installation/GOODMANGKS9.pdf
Even though a longer path within the attic puts more total heat into the attic space, it doesn't have the concentrated hot spot at the roof penetration melting snow to get the ice dam started.
Good thought Dana. The house doesn't have Gable end vents, but it was something we had noodled over at one point, but it seemed like the consensus was that with soffit and ridge vents already in place the benefit of Gable vents wouldn't really be there.
As an aside, we had some decent frost overnight and I wanted to share some telling pictures of the front and back of the roof before the sun hit it this morning. It was about 40° outside when I took these
Does the ice dam appear in the center of the roof edge or all along it?
Some wireless temperature probes would be useful in measuring success.
I agree that insulating the roof deck (while leaving the attic ventilated) would help significantly. I don't think code requires a 1" vent below the roof deck in this "unconditioned" already-vented attic case.
The only real frost is on the overhangs, which is pretty common in light-frost conditions.
It looks like melt-out at the (bath fan?) vents. If the frost were heavier you might be better able to spot exactly where the furnace is hovering below the roof deck.
Your photos make me think your attic venting is not working. Can you see day light thru the ridge vents or just the soffits?
If your venting is working the temperature inside your attic should be almost the same as outdoors on cloudy days, is it?
Walta
The attic venting can work just fine and still have that characteristic very strongly when there is the distribution losses of a furnace & ducts heating up the attic. No amount of venting is going to overcome the hot-spot of a thinly insulated or uninsulated cabinet on a furnace. That's the fundamental problem causing the ice damming here.
The colder it is outside, the more the furnace runs, with more parasitic heating of the attic space. The difference in interior & exterior temperature will increase as the outdoor temperatures drop.
You can see from the frost on pitch that doesn't have the vents sticking out that the attic is somewhat warmer in the middle (where the furnace is) than at the gable ends. The side with the roof penetrations is mostly melted-out, probably due to the furnace being on that side, but it could be partly due to higher amounts of sun.
The roofer I just had out that recommended the power vent went into the attic and his comments were "jeez, it's warm up here.. and there doesn't seem like there's much fresh air either"..
Outdoor temp was about 45°, attic might've been 65°, house was 70°
I can see daylight near the soffit baffles but not through the ridge vent, however I suppose I wouldn't necessarily see it up there with the shingle overhang.
Just for giggles, here's a picture of what it looked like last winter
>The roofer I just had out that recommended the power vent went into the attic...
What means "...the power vent went into the attic..."?
The power vent needs to exit the attic, but preferably at the gable, not through the roof.
>"Outdoor temp was about 45°, attic might've been 65°, house was 70°"
During the shoulder seasons on sunny days it's not unusual for attics to hit north of 60F (even when it's below 40F on calm sunny days). An asphalt shingled roof is essentially an unglazed solar collector. That's also how it can get to be north of 110F in the attic when it's only 80F outside in summer.
Walter is correct to suggest you verify the venting details of the attic, but that's really more about moisture purging than temperature.
Thanks Dana - I meant I had a roofer out and he recommended a power vent that would exit from the roof. I'm not certain how near to the gable it would be
By the way, thank you all for the input. Really appreciate it
Please post a close up photos of the ridge vent from inside the attic.
I wonder if the plywood or shingles are blocking the air flow.
Walta
Deleted
Pics
Deleted
Deleted
Trying the image load again
As expected the furnaces is practically tight to the rafters, which is like putting a radiator next to that section of roof deck.
It looks like 2x10 rafters(?), which are 9.25" deep, and would be deep enough to install R23 (5.5" ) or R30 rock wool (7.25" ) in the bays near the barely insulated furnace & AC coil. They will friction-fit in place, but installing a strip of aluminized fabric radiant barrier to the rafter edges to hold them up for the longer term wouldn't be insane, and would add a bit to the overall heat blocking the roof deck in that area. Install the batts down to 6-12" or so above the floor, and to at least above the collar ties- maybe even to the ridge, without actually blocking the ridge vent. Looks like you have 5-6 bays 16" on center (?), with something like 8-12' of bay to be insulated. Rock wool is preferred here- unlike fiberglass rock wool won't shed friable fiber into the attic air, and it's completely fireproof. There would still be convective heating of the roof deck via the channel, but it would even-out the overall roof deck temperatures by eliminating the big hot spot.
The smaller hot spot, the exhaust venting it still important. That furnace can handle 105' of vent length + 7 ells- it should be possible (and not very expensive) to route it out to a gable wall.
Dana - interesting suggestion with rockwool and radiant barrier. Nobody I've spoken to so far has brought that up... It sounds preferrable to completely sealing the soffits and ridge vents and using spray foam... now just need to figure out if that is a more cost effective approach than spray foam
My take is that the melting is fairly well distributed along the whole attic, meaning that you need whole attic solutions.
Insulating with R20-R25 foam above the roof deck, R30 in the rafter bays, and re-roofing is going to run north of $15K probably north of $20K, but isn't terrible compared to a roof all cut up with dormers and a gazillion roof penetrations. Even then the furnace is best vented out a gable, not the roof.
Jon, yes the trend we've seen is that the dams are fairly evenly distributed across the entire roof line
Are they they the same on both pitches?
My house had ice damming problems on all sides when I first moved in, but now are predominantly on less-insulated sections of the roof deck, particularly at a valley below a skylight. When re-roofing it will be getting a foam-over, and aerogel insulated skylights to replace the pre-existing units.
I would say that they are generally worse on the back of the house which faces south and which also happens to be the side that the furnace sits on
To summarize, I'd:
1) get the air leak tests done and seal as indicated
2) use an IR thermometer or camera to find hot spots and insulate them as possible
3) add R15 to all the rafter bays (to greatly reduce attic heat getting to the snow). Don't block vents.
4) if the above isn't enough, add a pair of matched supply/exhaust fans to remove the heat (during ice dam conditions).
This shouldn't be that expensive.
Thanks Jon. When you say add a pair of matched supply/exhaust fans, do you have thoughts as to whether those would be best installed in the gable wall ends, one near the peak of the roof on one end and another on the opposite end, or other?
Certainly up high at opposite ends of the attic, but either through the roof or in the gable ends. Then test/adjust such that the attic is at the same pressure as the interior (just below the ceiling) when the fans are on (a secondary goal is to stop, not increase in or ex filtration).
Got it, that makes sense
Quick update for what it's worth - Just had a spray foam insulation contractor over to provide an estimate. $4,500 (ballpark) to spray foam the entire attic with 5" of closed-cell, creating a conditioned space, sealing off exterior ventilation. I was told that fire retardant "paint" was not required by code in this area for closed-cell spray foam. I was also given an option to use open-cell spray foam insulation which would be about half to 3/4 of the price for ~10". Fire retardant paint would be required. I haven't read great things about using open cell so it was an... interesting suggestion.
At a buck a board-foot that's only 900 square feet of roof area. At $1.30/b-f (HFO blown foam) that's only 700 square feet of roof area.
The house/roof somehow looked bigger than that to me in the pics. Maybe it's the camera angles, or not getting the whole house in the picture in most of the exterior shots. Is the footprint really on the order of 600 square feet per floor? I was thinking it was at least 2x that big.
If yes that would make the 70K furnace & 3 tons of AC a bit gonzo-oversized, not that it would be unusual to see that type of oversizing. I guess that makes you prepared for when hell freezes over, as well as BEFORE it freezes over, eh? :-)
So, are those 2x12s (11.25" deep) or are they 2x10s (9.25" deep)?
The 5" thermal bridge through 5" of foam is only about R6, which is enough to be an ice-dam starter though it won't be nearly the current rolling disaster. The path through 9" of rafter it's about R11 which is measurably better. With ~3" of foam and ~5.5-6" of mid-density fluff in a 2x10 rafter bay you'd get a high performance at a comparable possibly lower price. At 3" anybody's brand-X foam is plenty of dew point control for R23 rock wool (5.5") in a zone 5 location. That's probably the way to go if 2x10s.
With 4" of foam + R30 rock wool in a 2 x 12 bay it would beat code min with plenty of dew point margin. (If it's R7.5/inch HFO blown foam 3" foam + R30 rock wool would get you there with an inch to spare and still have at least some dew point margin at the foam fiber boundary.)
You'll still need to air seal and insulate the gable ends. How deep are the studs at the gables?
Dana, the house is 31' long x 25' wide so about 775 sq ft per floor. The rafters are 1.5"x9". I've attached the two quotes I received here for spray foam application (open cell and closed cell) for additional reference
It is a bit chilly tonight so I kicked the heat on downstairs to 70°. Thermostat upstairs is reading 74°, thermometer in the rafters (closer to the soffits than the ridge vent - about 1/3 inward from vent) reads 51°. This is the side opposite the furnace. Outdoor temp is 43°. My laser temp gun reads 55° inside the ridge vent and 45° at the soffit baffels.
Some researchers have found that vented attic-outdoor temperature difference needs to be < 8F when it's 15F to 22F outside.
What is the temperature of the roof deck next to the furnace compared to the equivalent height locations elsewhere in the attic?
The closed cell foam quote indicates R6.9/inch foam at either ~2" or ~5.5", which is likely to be an HFO blown product. At 3" that would be R21-ish, which is adequate dew point control for R30 rock wool batts which would bring it to a bit ahead of code min at a total foam+ fiber depth of 10.25". That's what I'd do. YMMV.
3 tons of AC for a ~1500' house that isn't over-glazed on the west side is definitely on the ridiculous side of oversizing. (MY house has it about that bad! :-) ) That's about 3x oversized. Similarly the 70K furnace is going to be about 3x oversized too. Welcome to the "industry standard", as it currently exists... :-(
If enough is good, more is better, and even MORE is even better right? Ha SMH.
I'll have a look this evening at the temps near the furnace.
Just so I make sure I'm understanding your approach:
1) leave soffits and ridge vents open? Or create air channel against roof deck (how far up if so?)
2) Spray foam applied 5.5" on roof deck, ~2" on gable walls
3) Rockwool batts up against that? (Gable walls too?)
4) Aluminized relective fabric (I forget the actual name) up against Rockwool
>"If enough is good, more is better, and even MORE is even better right? "
Code minimum is better. LESS thermal bridging than 5.5" of rafter depth is necessary if you really want to nail the ice damming problem. (Ask Joe Lstiburek up in Westford! :-) : https://buildingscience.com/documents/insights/bsi-063-over-roofing )
So, are they 2x10 or are they 2x12 rafters?
Responses to the questions in order:
1) Foam over the vents- even 1" of foam is sufficiently low permeance to protect the roof deck with is the primary point for roof venting.
Just 3" is enough for dew point control to hit code-min using much cheaper batts for the rest . Going for 5.5" foam is extravagant and has an environmental footprint. Since only R20=3" of foam is needed for dew point control on R30 fiber insulation (~R50 total, meeting the R49 code min) the expense of the additional 2.5" to hit 5.5" of foam isn't necessary. In fact the cost savings of that 2.5" will pretty much pay for the R30 rock wool AND the perforated radiant barrier to hold it in place with money to spare.
Even at box store pricing R30 rock wool costs less than $1.50 per square foot. The marginal cost of that additional 2.5" couple of foam is going to run about 2 grand for this roof, R30 rock wool would be about a grand, and 1000 square foot rolls of perforated RB run less than $150 (~$175 at box store pricing, including sales taxes.)
2) If the gable walls are 2x6 framing there is no need to use foam on the gables- R20-R23 batts (no foam) will get you there. (A flash'n'batt of 1" foam + R15 batts also gets you there in a 2x6.) If 2x4 it takes 3" to hit the MA code min R20.
3)Rock wool batts are air retardent enough to function without an interior side air barrier, and much cheaper R than closed cell polyurethane.
4) Only PERFORATED aluminized fabric types of radiant barrier can be used for holding up the batts. (Many manufacturers make that stuff- it's typically on the order of 15 -17 cents per square foot at box store pricing, in 4' x 250' rolls, less than 15 cents per square foot at internet stores.) This type has a ~1/4" grid of tiny needle-prick perforations that allows moisture to leave (slowly) via vapor diffusion. DON'T use unperforated radiant barrier. Unperforated goods would created a moisture trap, bubble-pack stuff would have that problem too, and doesn't always have a sufficient fire rating to be left exposed.